TPG is not the only company whose announcements improve on its accounts. The rest are at Unaudited, where the filings get read so the press release doesn’t have to be.
TPG Telecom (ASX: TPG) – Investor Risk Register
The hard questions brokers can’t always write about. Unfiltered risks, disclosure gaps, and governance angles management would rather keep buried – but investors deserve to see. Analysis draws on public filings, broker commentary, regulatory submissions, blog posts #65-#101, and documented consumer evidence. Updated August 2026 with the HY26 half-year result (Post #100 – Beyond the Peak; Post #101 – The Wall and the Ladder): the “275 per cent” profit growth that was three million dollars, the interest saving consumed by the rent on assets sold, the arrival of the tax bill, the full $2.105bn spectrum schedule, the dividend raised into it, and the going-concern-flagged wholesale counterparty on which the growth story now depends. Prior updates: the 2020 Scheme Booklet vs 2025 accounts (Post #92 – Pawning the Silver), the 2026 AGM remuneration vote, and the post-Investor Day broker round.
Recent developments – 2025-2026
HY26 – the result beneath the headline August 2026
The half-year to 30 June 2026 was reported as 275 per cent profit growth. The statutory movement was $32m to $35m – three million dollars, nine and a half per cent. The 275 exists because the prior half’s statutory NPAT of positive $32m was re-presented on a pro forma basis as a $20m loss, adjusted as if the Vocus arrangements had run for the whole earlier period. Convert last year into a loss and this year becomes a resurrection. The same manoeuvre appears twice more on one slide: “underlying EPS up 2.7 cents” was in fact 3.7c down to 3.6c against a pro forma comparative; “free cash flow up 720 per cent” was $119m down to $93m on the same trick. Add back the discontinued operations that flattered last year and profit went $61m to $35m – down forty-three per cent. It is all reconciled in the company’s own appendix, slides 26 and 27, where the statutory column sits beside the pro forma one like a witness told not to volunteer anything.
“Twenty-nine per cent of the bottom line is a bookkeeping correction, and the deck says the half ‘benefitted from’ it – the same verb Swoop used about its own mystery millions.”
Post #100The bill for the silver – the saving that ate itself Credit where due
The one genuinely real, permanent benefit in the result deserves stating plainly, and this register states it: selling the fibre and paying down debt cut net bank interest from $122m to $37m – an eighty-five million dollar saving in a single half. Nobody should pretend otherwise. The trouble is what happened to it. In the same table, lease interest rose $27m – the company now rents the transmission network it used to own, a fifteen-year agreement signed on the way out the door – and an entirely new line appeared worth $36m, the cost of the handset-receivables financing programme. Sixty-three million dollars of an eighty-five million dollar saving, re-spent on the two transactions that produced it. Net benefit: twenty-two million – and profit before tax from continuing operations rose exactly $25m, of which 88 per cent is that financing line. EBITDA contributed eight. There is no operating improvement in this result; there is a balance sheet, and the balance sheet keeps most of what it gives. The guidance, characteristically, promises full-year financing costs “$100m+ lower” – pro forma for the new leases, which is to say, lower if you adjust out the thing that ate the saving.
“A pawnbroker offers better terms. Redeem the ticket and the silver comes home. Here the silver is gone and the weekly payment runs to 2040.”
Post #92 / #100The tax shield – gone, and converging with the spectrum wall Materialised
This register flagged the depleting tax shield in Post #81 and put it in the same window as the spectrum bill in Post #85. It has now arrived, ahead of schedule, and the CFO said it aloud: TPG is “now a net taxpayer, having utilised historic revenue losses against the gain on last year’s Vocus transaction,” with a go-forward rate of ~30%. The mechanism is the sting. The accumulated losses that sheltered the company since the merger were not exhausted by trading – they were consumed by the accounting gain on the fibre sale. The transaction that produced the headline profit in FY25 also burnt the shelter that made the profit tax-free. Cash tax this half: $21m, against nil a year earlier. And because the rate now applies to any recovery, every dollar the operating story improves is taxed thirty cents before it can accumulate toward the spectrum bill – the company has made its own recovery less useful to itself, precisely as the instalments come into view.
The $2.105 billion spectrum wall – front-loaded, and the cash isn’t there FY28-32
The full schedule, confirmed at Investor Day and unchanged: $840m in 2028 (850/1800MHz), $570m in 2029 (700MHz), $315m in 2030 (3.4GHz), $380m in 2032 (2GHz) – $2.105 billion in total, each instalment due roughly two months before the licence lapses, none of it negotiable with the Commonwealth. The shape is the danger: this is front-loaded, $1.41 billion in the first two years, two-thirds of the wall in 2028 and 2029, back to back. The biggest cheque is first and the second-biggest right behind it. There is no gentle ramp to build toward.
Set it against what the business actually makes. EBITDA of around $1.65bn for the year sounds ample – but EBITDA is the number a capital-hungry business reaches for when the capital is the problem. Out of it comes the cash cost of financing, the lease principal (the rent on the towers, the fibre and the Optus network it used to own), capital expenditure, and now tax. What is left – free cash flow before the dividend – lands near $300m a year. Against that sits a dividend near $392m. The dividend is larger than the free cash flow that funds it, so retained cash is roughly nil, and the single 2028 instalment is nearly three times a whole year of the cash that exists before the dividend is paid. The 2028 payment alone exceeds annual free cash flow before dividends. The market did the arithmetic and sold the stock ~8% on the day the schedule was confirmed.
The four doors – every funding source shut or shutting Capital risk
The wall cannot be paid from the business, so it must come from somewhere else. There are four somewheres, and the company has bricked up or narrowed each.
- Retained cash – nil, because the dividend consumes the free cash flow. The one lever fully in the company’s control, and it has just been raised in the wrong direction.
- New debt – facilities were cut $797m to $360m in the half; leverage is near 2.9x on the S&P measure, already through the company’s own 2.75x target. Funding even half the wall with debt moves the ratio toward 3.4x, closing on the 3.75x covenant. At the company’s own cost of debt, one $840m instalment costs ~$46-55m a year in interest – most of a continuing after-tax profit of perhaps $70m, permanently, and it lands below the EBITDA line so the EBITDA story survives while the earnings hollow out.
- Asset sales – the towers (OMERS, 2022), the fibre (Vocus, 2024), the receivables (securitised, twice) are gone. What remains is $2.3bn of property, plant and equipment (the network they still run) and $8.9bn of intangibles (goodwill and spectrum, unsaleable without selling the company). There is no fourth asset of the size that matters.
- Equity – demonstrably rationed. In November 2025 the company sought $550m and was given $300m, a shortfall AFR attributed to the reputational weather around the Triple Zero matter. When TPG last asked the market for equity, it got a little over half of what it wanted – at a share price (down ~38% over five years) that is itself flattered by the raise that lifted it. The equity is being held up by the very raises that would have to be repeated, at scale, to pay the government.
The good case is merely difficult: a dividend quietly cut, a raise ground out at a poor price, the ratio nursed just under the covenant, and the company survives it smaller and poorer, paying the government out of money it does not currently make. What there is not, on these numbers, is a version where the dividend raised in August and the spectrum wall that starts in 2028 both survive to 2032. Something gives, and spectrum cannot – you cannot run a mobile network on an underlying basis.
Net current asset deficiency – $677m, and every strut shortening Going Concern
Note 1(a) discloses a deficiency of net current assets of $677m – current liabilities exceeding current assets by that much – improved from $917m six months earlier, with the auditor’s going-concern paragraph beside it and no analyst asking a question about it. In fairness, in isolation this is not the alarm it looks like: read the composition and most of the current-liability side is trade payables of $765m and deferred revenue of $313m – money owed to suppliers that rolls perpetually, and customer prepayments discharged by providing service rather than cash. A telco with millions of monthly-paying customers funds a gap like this out of next month’s receipts, by design, and the directors are entitled to be satisfied it can. This register has no reason to doubt them, and says so. It notes only the direction of the struts holding it up. The deficiency improved $240m in the same half the company took $687m of securitised receivables through its cash flow. The facilities cited as support were cut $400m. The cash backing it is $55m. And the “history of generating profits” the note invokes sits against accumulated losses of $7.405 billion – a company that, across its entire existence, has never cumulatively made money. The deficiency is not the danger. The danger is that every strut is being shortened at once, and the note exists because the auditor thought it worth writing down.
Two ratios, one morning – the leases the headline omits Balance sheet
The Directors’ Report states net debt (excluding leases) to EBITDA of 0.8x, “material headroom relative to the bank covenant limit of 3.75x.” Some twenty minutes into the same webcast, the CFO told analysts the figure was about 2.9x on the S&P basis, down from 3.0x at December, targeting below 2.75x by year-end. Zero point eight on the page; two point nine on the call; same company, same morning, same six months. The difference is $2,747m of lease liabilities – nearly double the $1,435m of bank borrowings, and largely the creation of the very sale-and-leasebacks that produced the flattering ratio. Sell an asset and the debt used to buy it leaves the page; lease it back and an obligation of similar size returns under a different heading. The rating agency counts both; the headline counts one. And the deleveraging the presentation frames as transformation was one tenth of a turn in six months, still above the company’s own target.
MOCN economics – a cost creation dressed as a cost saving Escalating
The MOCN with Optus commits TPG to ~$1.59 billion over 11 years – roughly $72m a half – a fixed obligation that does not flex with subscriber volumes and escalates as Optus rolls out more 5G on the shared towers. Yet the statutory cost of provision of telecommunication services rose only $22m in the half. The gap is instructive, and worth a question to management: a ~$72m half-year cost that shows as a $22m rise implies large offsets – the operating costs of the 755 owned sites decommissioned to make way for the deal, and, on this register’s understanding, a portion of the arrangement being capitalised rather than expensed, with the right-to-use component recognised on the balance sheet and unwound below the line as depreciation and lease interest, where it never touches the gross margin the deck celebrates. If so, “mobile gross margin up 4.2 per cent” is struck on a cost base that never carried the full MOCN cost. The margin did not absorb MOCN; the demolition of TPG’s own regional network absorbed most of it, and the accounting relocated the rest. Eighteen months in, postpaid growth is flat, and Vodafone pays Optus weekly for access to a network on which Optus outcompetes it with better spectrum – complete with coverage carve-outs where Optus customers reach areas Vodafone customers cannot.
The dividend – raised above free cash flow, funded from the depreciation gap Escalating
The interim dividend was raised to 10 cents – roughly $196m against a half-year profit of $35m – and the CFO gave the funding arithmetic himself: EBITDA guidance, plus the capex cut from $750m to $650m, falling further. Depreciation and amortisation is guided at ~$1,280m against capex of $750m and falling; the gap between what wears out and what is replaced is where the dividend comes from, and management has now told the market it intends to widen it. Post #81 argued exactly this from the accounts as an inference; it is now a plan, described on a webcast, by the officer responsible for it. The franking was cut from 30 per cent to 25 – a listed company advising shareholders not to expect much Australian tax, in the same result that announced it had become a net taxpayer. A dividend paid from earnings is a company sharing success. A dividend paid from the space where reinvestment used to be, raised into a $2.1 billion wall, is a company consuming itself on schedule and inviting shareholders to admire the pace.
The FY25 base this was built on Context
For the record the register was assembled on, and against which HY26 should be read: stripping out the ~$45m non-recurring tax benefit, TPG generated approximately $7m in pre-tax profit from continuing operations in FY25 on $4.2bn of revenue – a 0.17% margin, with ROIC at 5.42%, below any reasonable estimate of WACC. In October 2024 it agreed to sell its fibre infrastructure and EGW fixed business to Vocus (completed 31 July 2025); by the company’s own disclosure the divested fixed business represented ~68% of the gross margin of the entire EGW segment – the high-margin infrastructure earnings, gone to a rival, for a $473m pre-tax gain, with a 15-year TAWFA leaseback recognising a $789m lease liability against a $509m right-of-use asset. Three assets – towers, fibre, receivables – one manoeuvre: cash today, a cost every year after. Selling the receivables also shrinks the capital base, so ROIC ticks up not because the business earns more but because there is less of it to measure.
“This is what selling the silver looks like when it is dressed as capital management.”
Post #92Provisions – $115 million with no breakdown Opacity risk
“Other provisions” jumped from ~$2 million to $115 million in FY25, with $118 million adjusted during the year – meaning ~$112 million in future cash outflows sit on the balance sheet, expensed but unpaid. Attribution is a single line: “Vocus separation obligations.” No sub-categorisation, no utilisation schedule, no independent verifiability. The contingent liabilities note simultaneously assesses no material effect on financial position – despite active ACMA investigations into incidents where customers died. The opacity is the issue, not just the number.
Disclosure rollback – churn and ARPU removed Investor concern
TPG discontinued mobile churn reporting and removed the ARPU sub-component breakdown (base/roaming/interconnect) at precisely the moment mix shift, tenure, and elasticity matter most. When a company needs four different profit definitions – statutory, pro forma, guidance basis, underlying NPATA – to tell its story, the simplest explanation is that none of them individually tells a good one.
The base TPG does not own Structural
The largest single addition to TPG’s wholesale channel – the ~135,000-subscriber Moose Mobile base it competed hard on price to win off Optus, the number meant to make the December subscriber line sing – belongs to Swoop Holdings (ASX: SWP). And Swoop’s own auditor has recorded a material uncertainty as to the company’s ability to continue as a going concern. Not this register’s characterisation: the auditor’s, in Swoop’s own filed half-year report, on the same day Swoop’s board issued a release headlined “Swoop Delivers Strong Organic Growth.” A wholesale base is not a subscriber base. It is capacity rented to an intermediary that holds the billing platform, the customer relationship, and the choice of network – held entirely at somebody else’s discretion, and in this case the somebody is a company whose discretion is currently constrained by its own accounts and an auditor’s warning.
Read Swoop’s accounts the way this register reads TPG’s, and the resemblance is uncomfortable. Operating cash flow for the half was negative $4.1m; free cash flow negative $14.9m, plugged by a ten-cent equity raise and the sale of a shareholding. The one quarter that looked positive – $2.75m of operating cash – contained a single line of $4.18m described only as “Other (Commercial agreement receipts),” a payment with no counterparty named, arriving in the very quarter Swoop signed its wholesale agreement with TPG. Strip that receipt out and the quarter is cash-negative, and so is the year. This register asserts nothing about who paid it – nothing is disclosed, and speculation is not analysis. It observes only that upfront incentives and migration payments are ordinary features of tier-one wholesale deals; that a receipt of that size is material to the result; and that shareholders have been told what the payment was called but not who made it.
Behind that sits a company selling everything not bolted down – Vonex gone, the voice business gone to Pivotel, the Melbourne fibre network contracted for sale at $11m against a murmured build cost several times that – with the going-concern note itself listing asset sales as a source of funds to survive the next twelve months, and a Westpac facility maturing in June 2027. The mobile base the whole edifice rests on has been flat at 127,000 for eighteen months, on an average margin of five to seven dollars a user. That is the counterparty. That is who TPG discounted its wholesale rate to supply.
Three doors out – and in two of them, the base goes back to Optus Scenario risk
The 135,000 can leave TPG’s network by three different doors, and in two of them the base TPG paid to win ends up somewhere else entirely. These are analytical possibilities, not predictions.
- Swoop survives, smaller. The likeliest and least bad: redundancies, price rises, cost discipline, the base stabilising below where it started. TPG keeps the relationship but carries a thinner book than it bid for, at the thin margin that made the contract winnable. The +180,000 becomes a number nobody mentions again.
- Swoop is absorbed. The pattern this sector has followed for a decade – sub-scale resellers do not fail so much as get bought. And the buyer of a distressed Optus-heritage MVNO group is, on any survey of who acquires these assets, at least as likely to be Optus-aligned as Vodafone-aligned. In that scenario the base migrates a second time – off Vodafone, back toward Optus – and TPG will have warehoused a customer base for eighteen months before handing it back to the network it came from. On the revealed preference documented across four Moose posts, the customers would go willingly, because they were trying to get back to Optus the entire time.
- Swoop does not make it. The least likely, but the one the auditor’s language obliges an analyst to contemplate. In an administration a customer base is an asset sold to whoever bids, and TPG would find itself simultaneously a trade creditor, the wholesale supplier, and a bidder against everyone else for a base it currently carries. If the winning hand is Optus-aligned, the largest wholesale “win” TPG has booked in years exits the network in a single transaction, and the growth story exits with it.
There is a question the contract would settle and the disclosures do not: does TPG’s wholesale agreement carry minimum-volume commitments? If it does, a shrinking Moose base is Swoop’s problem – a company under a going-concern cloud, contractually obliged to pay for capacity it is no longer using. If it does not, the shrinkage is TPG’s, and the revenue line thins every month the exodus runs. Both answers are interesting; neither has been disclosed.
The channel check – the base is walking, and it can be watched Live evidence
This is not forecast; it is happening. Across four posts this register has documented the Moose migration in real time: an ombudsman confirming a spike in complaints on the record; support queues past four hours; a company advising its own customers not to telephone it, on account of what it calls overwhelming demand; a nine-year review score in decline; and a roll-call of departures whose destinations are almost uniformly Optus MVNOs – amaysim, Dodo, Coles, Onepass, aldi, Spintel, Yomojo, Aussie Broadband. The customers are not merely leaving Moose; they are steering deliberately back onto the network they were marched off. The single most repeated sentiment in the archive reduces to five words: I need to stay on Optus.
And for the first time there is a number, supplied from inside the market rather than inferred from outside it. The chief executive of Yomojo, an Optus-network reseller and therefore an interested party, has told this masthead that customers arriving from Moose went from around 6.5 per cent of his inbound ports in 2025 to more than half across the two months since the migration began, and more recently in excess of nine in ten. It is a ratio, not a headcount; it measures the success of a campaign built for exactly this purpose; and it is offered with those qualifications firmly attached. But it describes a direction that does not arrive by accident – and one that points away from TPG. These are customers who were meant to land on Vodafone’s network and are instead landing back on Optus’s, under a rival’s badge.
The size of the leak is disputed but not trivial. This register’s own estimate is that fifteen to twenty thousand services fall away between active porting-out and the disengaged who let the disconnection clock expire. Industry sources put it higher – a SIM swap this messy, on their reckoning, is lucky to retain seventy-five per cent, which would be closer to thirty-five thousand gone. Take the low end or the high, and the point holds: the wholesale base TPG booked as growth is being sold a network its customers are actively fleeing, and the channel checks show them fleeing to the competitor. The migration lands in 2H26; the arrivals will push the reported wholesale SIO count up while the departures push the underlying customer count down, inside the same line, and the line will report the first. TPG competed hard on price to win 135,000 customers. What it appears to have bought is the right to supply a shrinking base, owned by a distressed intermediary, whose customers are porting back to Optus in real time.
“A number preserved on a slide, a customer base voting with its SIM card, and a network being filled faster than it is kept. It is the whole company in miniature.”
Post #101ACMA investigation – active, minimal disclosure Active
TPG disclosed an active ACMA investigation in its FY25 annual report with almost no detail. In a year when two customers died from Triple Zero failures, this is not a minor compliance footnote. The scope has not been disclosed. Based on publicly documented complaint patterns and regulatory precedent, plausible areas of inquiry include: systemic complaint patterns escalated through TIO Systemics; sales conduct and coverage representations through dealer channels, MOCN marketing vs actual coverage representations (see Dalton NSW); complaint-handling failures and non-fulfilment of agreed remedies; credit, collections and hardship practices involving vulnerable customers; and potentially emergency call failures directly.
For precedent on financial exposure: ACMA’s largest ever fine was $12 million (Optus, 2024, Triple Zero failure). Telstra paid $3 million for a 90-minute outage. Proposed legislative reforms would lift the maximum civil penalty to $10 million, with penalties up to 30% of adjusted turnover for the most serious breaches – and would remove the current two-step process, enabling direct enforcement. Against a $7 million pre-tax earnings base, even a mid-range penalty is material.
Two Triple Zero deaths – disclosed, bonus paid anyway Safety failures
TPG’s annual report discloses two customer deaths from Triple Zero failures. The company has since established a compliance uplift program, free handset replacements for affected customers, and ongoing monitoring – real costs, not separately quantified in the accounts. The Senate examined the matter. The Board assessed all risk gateways as passed for FY25 STI purposes and awarded the CEO a $250,000 discretionary bonus the same year.
“44 minutes of analyst Q&A, and the sell-side let management walk the entire call without a single governance question.”
Post #66TIO – complaints surged, baseline still elevated Q3 FY26 data
In Jan-Mar 2025, TPG complaints rose 32.4% QoQ (to 846) and 37.6% YoY. Vodafone mobile complaints rose 66.7% for poor coverage and 64.9% for service dropouts – against a market baseline of +0.6% QoQ. Telstra and Optus trended the opposite way. The TIO Systemics team is engaged, signalling potential ACMA referral, and cases are reportedly missing internal deadlines, some dragging 6-7 months.
Q3 2025-26 data confirms the elevated baseline has not gone away: Vodafone’s complaint volumes remain higher year-on-year than both Telstra and Optus while the broader industry moderates. A TIO metric was added to the FY26 executive scorecard after the FY25 surge – but a year of remediation commitments has not returned the complaint floor to pre-surge levels. Complaints are not just reputation – they are an earnings metric: every escalation is handling cost, TIO fees, remediation credits, and reacquisition spend, and they now feed directly into the cost-out ceiling documented on the Execution tab.
OAIC complaint – s13G privacy breach alleged Active
OAIC complaint CP25/03169 accepted, alleging breaches under APP 10, 12 and 13 (accuracy, access, and correction of personal information) and invoking s13G of the Privacy Act 1988 – serious and repeated interference with privacy. Maximum civil penalty for corporations: the greater of $50 million, three times the benefit obtained, or 30% of adjusted turnover. Potential outcomes include a published determination, mandatory compensation, and referral to the Federal Court.
ASX Compliance – case file opened Listing Rules 3.1/3.1B
A formal submission was lodged with ASX Compliance in October 2025. The case focuses on: removal of key KPIs (churn, ARPU breakdown) precisely when clarity is most needed; the handset-receivables accounting reversal; inconsistent commentary around churn causation and MOCN economics; and potential disclosure gaps regarding Felix Mobile’s underlying profitability and cannibalisation risk.
The ACCC domestic roaming inquiry – the hope, and what it reveals Active inquiry
On the HY26 call the CEO was warm – repeatedly, unmistakably – on the ACCC’s inquiry into mobile services, and on the prospect of mandated network sharing or domestic roaming as a remedy to the resilience problems the June outage exposed. Read against the balance sheet, the enthusiasm is its own disclosure. The third carrier, having spent ~$1.59 billion to rent Optus’s regional network and decommissioned 755 of its own sites to do so, is now warm on the prospect of regulated access to Telstra’s as well. There is a rational version – a smaller carrier seeking access to infrastructure it cannot economically replicate, which reasonable people will argue before the Commission. But a company whose two stated routes to a viable future are a satellite constellation it does not own and a regulatory outcome it cannot command has quietly changed the meaning of the word challenger. A challenger builds a thing that competes; what is being described is a tenant applying for a better lease. The outcome of the inquiry is a genuine swing factor for the investment case – and its centrality to the pitch is itself a measure of how few levers remain internal.
Collections risk and ACCC exposure – a Federal Court test case Escalating
Reports of collections activity on disputed balances, on deceased account holders, and potentially invalid default listings. Dealer-channel conduct – over-promising coverage, misrepresenting plans, commission-driven sales – remains a standing ACCC risk. Rising complaints point to possible non-compliance with the TCP Code and misleading or unconscionable conduct under the ACL. Post #87 sets out the ACCC’s Federal Court proceeding against the mercantile-agent structure now under scrutiny, in which TPG features as a bespoke Annexure B client – the “independent” solicitor, the agent he pays, and the creditor he acts for revealed as functionally one hand in three gloves.
National network outage – June 2026 Operational
A national Vodafone outage affected millions, generating over 8,200 DownDetector complaints before 10am – and the status page, hosted on the very systems that failed, went down with it. The disruption was brief, but outages of this scale typically drive elevated port-out and churn in subsequent weeks – relevant given Vodafone postpaid recorded zero net adds across FY25 and stayed flat through 1H26. Affected customers retain TIO complaint pathways for service loss and associated business income impact. The forum evidence of that churn is already visible (see Execution and Counterparty tabs).
Whistleblower investigation – closed, no findings shared Closed
Written correspondence confirms TPG’s CEO personally contacted a board member and shareholder of the complainant’s current employer during an active consumer dispute. TPG initially denied whistleblower protections before reversing, acknowledging coverage under the Corporations Act, and referring the matter to KPMG FairCall. TPG then shut down the KPMG whistleblower portal mid-investigation. A twelve-page independence challenge against the external investigator went unanswered. The investigator found nothing; the investigation closed with no findings shared. The CEO contact has not been denied at any point. Separately, Vodafone disclosed a complainant’s employment history to a journalist – a named email identifies Vodafone as the source. Both are documented.
The matter has since progressed into formal legal channels. If it reaches the Federal Court, Corporations Act s1317AE(2) cost-protection provisions can restrict a company’s ability to recover its own legal costs in whistleblower proceedings – even on a successful defence. Discovery, forensic analysis, and expert evidence are largely unrecoverable. A prolonged defence can become a high-cost, low-recovery contest – a dynamic that weighs heavier when underlying profitability is thin.
The independence question – three foundation seams Governance nexus
Post #86 documents a foundation that moved $3.3 million in a year and kept $840 of it – zero volunteers, one part-time employee, reserves that never escape four figures, and a board drawn entirely from one building that changes everything except who controls it. Post #88 follows a director who sat at a remarkable set of seams – shareholder of the employer, foundation director, near-quarter-century partner at the firm that acted for the telco in its merger – and who, after a whistle was blown and a CEO reached for a telephone, exited the law firm within a month and the foundation board six months later, each departure quiet and for nothing in particular. The philanthropy is real. The independence is not.
The $250,000 bonus – on $7 million profit Remuneration
In a year of two Triple Zero deaths, an active ACMA investigation, a whistleblower matter, surging complaints, zero postpaid growth, and $7 million in underlying pre-tax profit, the Board awarded CEO Inaki Berroeta a $250,000 discretionary bonus for “outstanding leadership in a period of significant transformation.” Two other executives received $100,000 and $85,000. Total discretionary bonuses: $435,000 – over 6% of the entire underlying earnings base, paid as discretionary top-ups to three people.
“The discretionary bonus pool alone is about three weeks of the company’s entire annual profit. What, precisely, would a bad year look like?”
Post #67Board independence – three of ten, still outnumbered Below ASX standard
TPG’s board now has three independent directors out of ten after a June 2026 appointment – up from two, still short of the ASX recommendation of a majority of independents and an independent chair. The Chairman (Canning Fok) is an Executive Director and Co-Managing Director of CK Hutchison – not independent by definition. The independents (including Helen Nugent AC and Paula Dwyer) carry the bulk of the governance workload and chair the critical committees, yet are structurally outnumbered on every contested vote. The remaining seats: shareholder nominees (CK Hutchison, Vodafone Group), a legacy company insider (former General Counsel), and the CEO. The merger that justified this structure was 2020; the register has since changed shape entirely – Soul Patts fully exited, Robert Millner’s seat vacated. AustralianSuper, one of TPG’s largest institutional holders, has no board representation at all.
CEO equity grant – December 2025 Alignment concern
The Board approved a CEO equity grant in December 2025 – while the ACMA investigation was active, TIO complaints were surging, the compliance uplift program was underway, and the whistleblower investigation had commenced. The critical question: if vesting hurdles are EBITDA-based rather than ROIC or TSR, the grant rewards a metric that can climb while the business destroys value. EBITDA growth via cost-cutting on a sub-WACC return base is not value creation – it is a more efficiently managed value-destroying business. The HY26 result sharpens the point: mobile gross margin rose 4.2 per cent while premium customers spun down and the cost of provision was flattered by MOCN accounting – margin as a cost story, not a customer one.
Risk gateways – form without substance Governance concern
The Board stated all risk gateways were passed for FY25 STI purposes. During the gating period: two customers died from Triple Zero failures; ACMA opened an investigation; Vodafone TIO complaints rose ~24% YoY; a compliance uplift program was initiated; and a whistleblower investigation commenced using a Band 1 external investigator. All gateways passed. If none of this trips a gateway, the question is not whether the gateways were technically satisfied – it is what they are there for. The Qantas parallel is instructive: gateways that existed on paper and never operated in practice. Decoration – the appearance of accountability without its substance.
The central governance question
The chronological alignment of disclosure rollback, safety incidents, rising complaints, negative analyst revisions, remuneration awards, and senior-executive conduct creates one investor question: is TPG managing through the challenges, or managing the optics of them? The Qantas, AMP, and Crown precedents share a pattern – a board that lacked genuine independence was structurally incapable of challenging management when challenge was most needed. The cost was borne by minority shareholders, customers, and employees.
The vote that moved fourteen-fold 2026 AGM
There is one number a board writes down once a year that it cannot blame on the weather: the vote against its remuneration report – the single resolution the law hands shareholders purely to say what they think of executive pay. At the 2025 AGM, that vote was 0.87% against. Nine hundred and ninety-nine shareholders in a thousand raised no objection. At the 2026 AGM, twelve months later, the same resolution drew 12.17% against – a fourteen-fold increase in a single year. It is not, technically, a “strike” (that needs 25%), and the board will cling to the gap between twelve and twenty-five. But a fourteen-fold surge in dissent on the one vote given to register displeasure is not noise. It is a specific accusation, from the owners, about a specific thing: the pay.
And here is the tell that turns a data point into a verdict. On the same poll, the resolutions to actually hand the CEO his equity – the short-term and long-term grants – sailed through at 99.73% and 99.55%. A year earlier those same grants drew 11.6% against; this year, next to nothing. The dissent did not disappear – it moved, off the grants (which shareholders must approve or watch the reward paid in cash instead) and onto the remuneration report, the one vote that changes nothing and is therefore the only one worth spending on a message. Shareholders waved the plumbing through and marked down the house. That is precisely what an owner does when he has stopped quarrelling with the mechanism and started quarrelling with the result.
87.64% of the maximum – on a $7 million business Pay vs performance
For the year ended 31 December 2025, the CEO was awarded a short-term incentive of $3,053,206 – split half cash, half deferred rights, so the word restraint might be applied to a payment of three million dollars. That award represented, in the report’s own words, 87.64% of the maximum. Not of target. Of the maximum. In a year the continuing business earned $52 million before tax – and $7 million before a $45 million R&D tax benefit did the lifting.
Read those figures in the same breath, because the board did not. A CEO collecting the better part of nine-tenths of his maximum possible bonus, from a business that converted $5 billion of revenue into $52 million of profit – and would have converted it into $7 million without the tax benefit. The bonus scheme, asked to grade the year, returned an A-minus. The accounts, asked the same question, returned a 1% margin and a return on capital below the cost of the capital itself – the technical definition of a business destroying value while it operates.
Nor did it stop at the framework. Having run the scorecard and arrived at its number, the board then exercised discretion to add a further $250,000 to the CEO “one-off,” plus $100,000 to one lieutenant and $85,000 to another. Discretion is the word that matters – the board’s own judgment, unmediated by any formula, applied after the machine had already delivered its verdict. Presented with a year the machine graded at nearly 90% of maximum, the board’s considered human addition was: more. The formula was generous; the humans were more generous still.
“A grading instrument that returns an A-minus for a year like this one is not measuring performance. It is laundering it.”
Post #91NPS at maximum, complaints at a multi-year high Scorecard contradiction
One studies the scorecard for where the generosity came from, and the answer is instructive. The company’s Net Promoter Score – its own house measure of whether customers are happy – was recorded, for the company’s brands, as an outcome at maximum. Maximum. This in the year the Telecommunications Industry Ombudsman’s tables show Vodafone complaints running some 40% adrift of Telstra and Optus, the two large rivals driving their numbers down while TPG’s stable drove its own up.
A satisfaction metric at maximum and a complaints metric at a multi-year high, in the same company, in the same year, purporting to describe the same customers – and only one of them was invited onto the scorecard. Which prompts the question the whole bonus turns on: whose picture reached the boardroom. The masthead asserts no mechanism; it observes only that the instrument the company builds and marks itself printed perfect, the instrument kept by the TIO printed worst in years, and the board wired its bonus to the first and left the second off the dashboard. A company is entitled to measure itself. It is not usually so fortunate as to mark its own exam and bank the result.
The complaints metric added only after the failure year closed Timing
The most quietly damning entry in the file is a matter of timing. TIO complaints – the measure on which TPG spent 2025 conspicuously losing to its rivals – were added to the executive scorecard for 2026. Prospectively. The metric on which management performed worst was admitted to the instrument that pays management only after the year in which the failure occurred had been safely closed, its bonuses counted and banked. Accountability arrived – with the immaculate timing of a smoke alarm installed the morning after the fire, and wired to next year’s kitchen.
A board does not ordinarily bolt a complaints metric onto executive pay in a year its own satisfaction score printed at maximum; the two moves sit oddly together, unless something had made the complaint figures suddenly, awkwardly visible. The masthead notes only the sequence: the gap widened, the gap became public, and the metric appeared – in that order.
The $2.8 billion “dividend” that was a partial liquidation Capital return
The pay was being paid for something. In November 2025, shareholders approved a capital distribution of $2.826 billion – $1.52 per share by capital reduction, plus a 9-cent unfranked dividend. The company is candid about the source, because the accounts leave it no choice: the distribution was, in TPG’s own words, “funded by proceeds from the Vocus transaction.” Funded, that is, by selling the fibre network. Nearly $3 billion returned – sourced not from anything the business earned but from a piece of the business the board sold, and unfranked, because there were no taxed profits behind it for credits to attach to. It was not a dividend. It was a partial liquidation, conducted with a smile and applauded as capital management.
The recurring dividend tells the same story in miniature. Stripped of the Vocus-funded special, it runs to some $334 million a year – two payments of 9 cents, both unfranked, the balance sheet’s own admission that there were no franking credits because there were no taxed Australian profits to generate them. The HY26 dividend was then raised to 10 cents, funded on the CFO’s own account by the capex cut – the pattern not merely continuing but accelerating into the spectrum wall.
“The shareholders were paid handsomely. They were simply paid out of the furniture.”
Post #91The death and the raise, in one announcement Disclosure
On 18 November 2025, TPG lodged a statement with the ASX confirming what it called a “tragic loss of life” following calls from a device that could not reach Triple Zero. The company added, in the same announcement, that its network had been “operational” at the time and that it had met its “regulatory obligations.” It then explained why it was telling the market at that moment: it was providing the information “for transparency because of heightened public awareness,” while “currently undertaking the institutional component of the Reinvestment Plan.” The bookbuild, the statement noted a paragraph later, “received strong demand.” The death and the capital raising were disclosed in the same announcement, under one continuous-disclosure header, with the raise named as the context. The masthead sets the fact down as the company recorded it, and leaves the reader to make of the pairing whatever the reader will.
What a bad year would have to look like The question
Total the year the board graded at nearly 90% of maximum: two Triple Zero-linked deaths disclosed; the premium brand added net nobody (2,846,000 to 2,846,000); the continuing business earned 1% on revenue; complaints ran 40% adrift of rivals who were falling; ACMA opened an investigation the company disclosed in one sentence and declined to describe; other provisions swelled from $2 million to $115 million; the cash flow was flattered by selling the fibre network and factoring customers’ own phone instalments at a ~$95 million derecognition cost. Against that catalogue, the machine returned 87.64% of maximum and the board added a quarter-million on top. The shareholders, reading the same words, wrote down 12.17%. It is the sound of a shareholder base that has stopped believing the adjective – “transformative” doing the work that “underlying” did at Qantas and “challenging” does at every company managing its own decline.
“What, precisely, would a bad year have to look like? A grading instrument that returns an A-minus for a year like this is not measuring performance. It is laundering it.”
Post #67 / #91The composition problem – growth in the cheap lines, decline in the profitable ones Core failure
Post #85 sketched this a year ago from an investor-day slide; the HY26 appendix proves it. Every line growing is a cheap line or somebody else’s: Digital First (felix, iiNet, TPG-brand) up 114,000 at a $26 ARPU; wholesale MVNOs up 104,000 at a thin wholesale rate. Every line shrinking is a valuable one: NBN down 84,000 at a $74 ARPU; postpaid down 15,000 year-on-year at $49; Fixed Wireless, the highest-margin product in the fixed book, down a thousand; Vision Network’s margin per user collapsing to $4.70. The net subscriber figure looks calm; the composition is a company hollowing out from the inside – the same body count, worth less per head. Jefferies gave it the correct name in June and the half confirmed it: spin-down. Every net addition came in at roughly half the postpaid rate or belonged to somebody else. The company describes the outcome as market-share leadership; its own broker describes it as customers spinning down. And each step toward wholesale, digital and prepaid lowers the revenue quality of the whole – a downgrade of the base disguised as growth of it.
Twenty-four cents – the price rise that bought nothing Elasticity ceiling
An analyst asked, twice, where a ~$3.60 (ex-GST) postpaid price rise had gone; postpaid ARPU moved twenty-four cents ($48.51 to $48.75). Against inflation of 3.6% that is a real decline of ~3%, after raising the price. The concession, extracted on the second ask: promotional activity – a 100-day, 50%-off run on Vodafone from March to June, right up to the balance date. Postpaid began the half at 2,846,000 and ended at 2,846,000, fifteen thousand fewer than a year earlier. Raise the back book, half-price the front book, net twenty-four cents, add nobody. The deck calls it “ARPU growth across all products accelerating.” When deep discounting cannot move the postpaid base, the organic demand picture is weak and price-rise churn is quietly offsetting the promotional gains. And the promised second-half improvement comes from repricing felix and the low-end plans – the discount brands, shortest tenure and highest churn, the only cohort still turning up. They are raising the price on the only customers who are still arriving.
Fixed Wireless went backwards – the valve leaking inward Thesis broken
For two years Fixed Wireless was the answer to the NBN question – the higher-margin product that would catch the fixed-line base as it fell. It is the single most profitable line in the fixed book: an ARPU of ~$51 that is almost entirely margin, because unlike NBN (where ~$74 of ARPU yields ~$22 after the NBN Co access charge) it runs on TPG’s own network and keeps nearly all of it. This half it fell from 285,000 to 284,000 – and it did so with every tailwind management could name: 5G standalone live, addressable market up 15%, more than 70% of the base on 5G, the MOCN footprint fully deployed, 2100MHz refarmed for 5G in Sydney and Melbourne, a new Wi-Fi 7 modem shipped. The deck asserts it “returned to growth in 2Q26,” which, given the half closed at minus one thousand, means the first quarter was worse. This register named a year ago the cease-sale dynamic in congested metro pockets, where the network is full and a customer at an affected address cannot sign up – the overcapacity coming home in a company that cut capex. But that explains the congested suburbs; it does not explain the absence of growth across the new MOCN footprint, the freshly available addresses the deal was sold to open up, advertised and attached to the better modem. A product handed more coverage and better equipment that still does not grow has a problem that is not coverage or equipment. It is demand, or the brand, or both.
Vision Network – the own-fibre margin, quietly collapsing Unremarked
Vision Network – TPG’s own fibre-to-the-premises network, the supposed high-margin alternative to reselling NBN – added four thousand subscribers this half while its margin per user collapsed: $16.93, then $11.17, then $9.85, then $4.70 across four disclosed halves. A 72% fall on the product meant to be the premium end of fixed, its ARPU falling too – the only fixed line where the price is going backwards, not up. The one asset TPG built to own the customer economics in fixed is earning less per user every six months, heading toward earning nothing. Nobody asked about it on the call.
Wholesale – a rounding error consuming the capacity the profitable product needs Dilutive
Wholesale is presented as proof of the network’s appeal – the MVNO line up to 214,000. On an incremental basis it is close to worthless. The entire wholesale mobile revenue line is around $45 million, less than 2% of group revenue; Consumer Mobile alone is roughly 24 times its size. The per-service economics collapse as the base grows: blended wholesale revenue per service roughly halved as the MVNO base doubled, the arithmetic of a tiered rate card and bigger, cheaper cohorts. TPG built the channel from ~7,000 services in 2023 to 214,000 today and it still generates less than the handset-resale line. And the wholesale customer is the thinnest there is – no retail relationship, no brand loyalty, revenue at a wholesale discount, and traffic that consumes the same congested capacity the high-margin Fixed Wireless product needs and cannot get. Every discounted wholesale gigabyte sold into a full cell is a Fixed Wireless customer who could not sign up. It is filler poured into the hole where the value used to be – and the largest piece of it, the Moose base, is the going-concern counterparty documented on the Counterparty tab.
The limit of cutting – margin held up by savings that run out Floor in sight
Management’s answer to all of this is that what matters is mobile gross margin, up 4.2%, and that costs are being taken out to defend guidance. The margin did rise, for three reasons, none repeatable and two finite: the MOCN step-down in the cost of carrying regional traffic (a one-time benefit that annualises out); operating costs pinned flat against 3.6% inflation ($46m of a $100m cost-out delivered, but a program with an end); and felix being cheap to run as well as cheap to buy, so it lifts the margin percentage as it grows even as it lowers revenue per head. That is not operating strength – it is a company defending a margin line with the cost ledger because the revenue ledger will not cooperate. You can cut your way to a margin; you cannot cut your way to a customer. Once MOCN annualises and opex cannot be cut against inflation again, the mix damage has nowhere left to hide. Management’s own guidance – margin growth to continue exceeding service-revenue growth – is the tell: they expect to keep growing margin through costs, and the cost line has a floor. In a service business that floor is already visible in the complaint data.
Churn visibility – management flying blind Board-level concern
Management has publicly admitted they “don’t know” why customers churn – despite running post-exit surveys and Next Best Action modelling in Siebel built to surface churn risk, and despite citing exit-survey data to the ACCC when lobbying for MOCN approval. Churn disclosure has since been removed entirely. Without robust churn attribution – and without disclosing it – sustained net-add underperformance continues unmonitored by the market. Not an operational gap; a board-level governance question, and it lands hardest precisely as the June outage and the Moose migration drive the churn nobody can now see.
Felix – prepaid economics dressed as SaaS Narrative risk
TPG presents Felix as the “Netflix of telco.” The analogy works in headlines and collapses under scrutiny. Netflix retains subscribers a median ~31 months at falling marginal cost. Felix’s estimated tenure is under 15 months, churn runs 6-10x higher, and marginal cost rises with scale on unlimited plans. Felix ARPU is ~$26 – roughly half Vodafone postpaid’s ~$49. Every migration from postpaid to Felix is a revenue downgrade for the Group, and the HY26 mix shows the substitution accelerating: Digital First up 114,000 while the premium base stands still. Felix bills like SaaS. It does not earn like SaaS.
Pricing above the stronger rival – the tariff of a cornered brand Post #92
The convention is that the challenger undercuts the incumbent – price is the one lever a weaker brand can always pull. On the plans advertised as at mid-2026, Vodafone has misplaced it: its plans sit above Optus’s for comparable data, the third-placed brand, fresh from a national outage, pricing above the network it partly rides under the sharing deal and cannot beat. And the premium buys a conspicuously un-premium experience: a disjointed app, and the moment a customer drives off the shared footprint onto Vodafone’s native network, coverage thins and data slows. Pricing dates fast and varies by promotion – readers should confirm current rates directly with the carriers.
Fixed broadband – structural decline, not a transition ~20% of EBITDA at risk
NBN lost ~116,000 subscribers net in FY25 (down 6.9%), and a further 84,000 in the half to June 2026. Fixed Wireless – the supposed offset – went backwards. Jefferies flags the NBN speed-boost program as a potential “churn event” benefiting challengers; Aussie Broadband and Superloop keep recording SIO wins half after half. Fixed contributes a fifth-plus of Group EBITDA. Management calls the market “structurally challenged” – an admission there is no credible path back to growth – and now faces, per Morgan Stanley, a Starlink direct-to-device threat to the fixed-line base on top of it.
The sponsored content playbook – Post #68 December 2025
In December 2025, TPG CEO Inaki Berroeta published a paid opinion piece in The Australian promoting the MOCN deal, regional coverage, and spectrum cost concerns. The same day, a second piece – labelled “Sponsored Content” – promoted the MOCN through TPG’s General Manager of Strategy. Neither mentioned: the $7 million pre-tax earnings base, zero postpaid growth, the $115 million provisions jump, the already-active ACMA investigation, the two Triple Zero deaths, or the whistleblower investigation just commenced. The CEO equity grant was approved that same month, while the investigation was live.
Results-day coverage – “bolsters profits” then “275 per cent” Feb & Aug 2026
On 27 February 2026, AFR ran results-day coverage headlined “Demand for Netflix-style mobile plans bolsters TPG Telecom profits.” The stock dropped 2.72% on the day – the market read the results and sold. Zero postpaid growth was buried in paragraph six. The August HY26 result repeated the pattern at higher amplitude: a “275 per cent profit growth” headline for a three-million-dollar movement built on a restated prior-year loss, reconciled only in the appendix. In both cases the wire number and the market reaction pointed in opposite directions, and the reconciliation sat where a reader had to reach for it. Independent analysis of both results is on this masthead; 115,000+ investors found it here.
“The company disclosed the pro forma basis and published the percentage, and it knew precisely which of those two things would appear on the wires by lunchtime. It was not the basis.”
Post #100Investor Day and HY26 – the market stops forecasting and prices it ~8% fall, then the cuts
The June 2026 Investor Day confirmed the spectrum bill, the 1H26 composition problem, and the pivot to cost-cutting; the stock fell ~8% on the day. The bear case that had been building analytically since FY25 stopped being a forecast and became a price move. The HY26 result and the broker round that followed – UBS Neutral at $3.97, Morgan Stanley cutting to $3.40 on the Starlink thesis – confirmed the direction: every broker that has updated since the Investor Day has cut its target or flagged softening postpaid, regardless of headline rating.
The survey vs the SIM card – Post #90 July 2026
A UBS consumer survey reported Vodafone loyalty firming and Optus losing share. The same fortnight, the Moose Mobile migration moved ~135k customers off Optus onto Vodafone’s network – and the Whirlpool forums filled with those customers porting straight back out to Optus resellers, one describing Sydney CBD coverage as “deader than disco.” Perception is a survey of people who haven’t tried it; reception is the verdict of people who have. The UBS fieldwork also closed before the 19 June outage it could not have captured. The only poll that counts settles when the trading detail becomes audited fact.
The narrative gap – slides vs. reality
- “Customer and network experience is thriving.” – Complaints trend the wrong way; Vodafone still elevated against peers as of Q3 2025-26 TIO data, and the Moose migration is churning back to Optus in real time.
- “Customers responding positively to simpler brands.” – Vodafone postpaid net adds: zero in FY25, flat through the half despite a 100-day 50% promotion.
- “ARPU growth across all products accelerating.” – Postpaid ARPU moved 24 cents after a ~$3.60 rise; the average is carried by tiny increases on the cheap brands. Not acceleration – dilution masked.
- “Mobile gross margin up 4.2 per cent.” – Struck on a cost base flattered by MOCN accounting and held-flat opex, while the premium base stood still. Margin as a cost story, not a customer one.
- “Spectrum renewals later in FY27 are manageable.” – Now confirmed at $2.105 billion FY28-32, the 2028 payment alone exceeding annual FCF, the dividend raised into it. “Manageable” is marketing language.
- “We are winning a growing share of mobile customers at higher ARPU.” – Growth is digital-first brands at ~$26 ARPU, half of postpaid, and wholesale at a thinning rate. Not higher-ARPU share – mix compression.
Current ratings and targets – to August 2026 (post-HY26 result)
Broker notes following TPG’s HY26 half-year result (August 2026), the June 2026 Investor Day, and the ~8% share price fall that met the spectrum schedule. Despite headline ratings holding in several cases, target prices have been cut, EPS forecasts revised down, and the underlying commentary flags the same structural concerns raised on this page: softening postpaid, rising spectrum costs, reliance on the digital/value segment to carry growth, and now a Starlink disruption thesis. Most critical and most recent notes first. Red = negative/underweight, amber = neutral, green = positive/outperform.
Earlier ratings – to March 2026 (pre-Investor Day)
Retained below for context on how targets have moved. The trajectory is one-directional: Morgan Stanley Underweight continuously since November 2023; Ord Minnett’s target cut from $6.60 (April 2024) to $4.14 (December 2025), a 37% reduction over twenty months; UBS from $5.40 (January 2026) to $3.80 then $3.97.
The broker silence – what was never asked
In 44 minutes of analyst Q&A with eight brokerages on the FY25 results call, not one question was asked about: Triple Zero deaths, ACMA investigation, governance, executive remuneration, or the whistleblower matter on foot. On the HY26 call, six analysts took the microphone; not one asked how net profit grew 275 per cent, nor about the $677m net current asset deficiency, nor the going-concern paragraph attached to it. The annual report and half-year report disclose all of these. That silence is not the absence of a problem. It is the absence of accountability.
What has converged is the direction: every broker that has updated since the Investor Day has either cut its target or flagged softening postpaid, regardless of headline rating. Targets now sit in a band from $3.40 (Morgan Stanley) to $4.30 (Jarden), the bulls betting on cost-out and digital-brand mix to do the work subscriber growth is not. The HY26 result did not resolve the disclosure-visibility problem documented on this page. It confirmed the numbers the disclosure rollback had been obscuring.
The schedule is now confirmed: $840m in 2028, $570m in 2029, $315m in 2030, $380m in 2032 – front-loaded, $1.41bn in the first two years, each instalment due roughly two months before the licence lapses. On the company’s own numbers, free cash flow after the dividend is approximately nil, and the 2028 instalment alone is nearly three times a full year of free cash flow before the dividend. Undrawn facilities were cut from $797m to $360m in the half. What is the explicit, itemised funding plan for the 2028 payment specifically – what combination of retained cash, new debt, equity, capex deferral, or asset sale – and when will shareholders be given firm figures rather than “cash flows and borrowing headroom”? If the plan is debt, what is the pro forma effect on net debt to EBITDA against the 2.75x target and the 3.75x covenant?
The interim dividend was raised to 10 cents – roughly $196m against a half-year profit of $35m – and the CFO stated the increase is funded by the reduction in capital expenditure from $750m to $650m, a gap between depreciation (~$1,280m) and capex he intends to widen. Given a $2.105 billion spectrum bill beginning in 2028, a free-cash-flow-after-dividend position of approximately nil, a depleted tax shield, and undrawn facilities cut to $360m, on what basis does the Board consider raising the dividend – rather than retaining the cash toward the spectrum instalments – to be in shareholders’ interests? At what underlying earnings level would the dividend be covered by profit rather than by the depreciation-capex gap, and at what point does the Board reduce or defer it rather than re-lever a balance sheet it spent two years deleveraging?
The MOCN arrangement is worth approximately $1.59 billion over 11 years – roughly $72m per half – yet the statutory cost of provision of telecommunication services rose only $22m in the half. Please reconcile the difference. Specifically: how much of the ~$72m half-year MOCN cost is expensed within cost of provision versus capitalised as a right-of-use asset and recognised below the gross-margin line as depreciation and lease interest? What was the operating cost of the 755 decommissioned owned sites that the arrangement replaced, and how much of it fell away in the period? If a material portion of the MOCN cost sits below the gross-margin line, does the reported “mobile gross margin up 4.2 per cent” reflect a cost base that never carried the full MOCN cost – and what is mobile gross margin with the full arrangement expensed?
The largest single addition to the wholesale channel is the ~135,000-subscriber Moose base, owned by Swoop Holdings (ASX: SWP), whose auditor has recorded a material uncertainty as to its ability to continue as a going concern. Does the wholesale agreement carry minimum-volume or minimum-spend commitments – and therefore, does a shrinking Moose base fall on Swoop (obliged to pay for unused capacity) or on TPG (a thinning revenue line)? What is TPG’s total credit and revenue exposure to Swoop across the wholesale agreement and any related receivables? What contingency governs the Moose base in the event of a Swoop acquisition or administration, and has the Board assessed the risk that the base migrates off TPG’s network entirely – a risk the documented porting evidence suggests is already materialising?
Statutory NPAT rose from $32m to $35m – three million dollars – yet the result was presented as 275 per cent profit growth, a figure available only by restating the prior half to a $20m pro forma loss. Of the $35m reported, approximately $10m is a one-off, non-cash amortisation correction (Note 4) that the investor deck states the half “benefitted from.” Excluding that correction and including discontinued operations, total NPAT fell from $61m to $35m. Which figure does the Board regard as the fairest representation of the half’s performance, and will it commit to leading future results with the statutory movement rather than a pro forma percentage that appears on the wires before the reconciliation on slide 27 is reached?
The Directors’ Report states net debt (excluding leases) to EBITDA of 0.8x against the 3.75x covenant; the CFO cited 2.9x on the S&P basis on the same webcast. The difference is $2,747m of lease liabilities beside $1,435m of bank debt, substantially created by the sale-and-leaseback transactions that produced the lower figure. Which ratio does the Board consider the truer measure of the group’s leverage, and why is the covenant-facing 0.8x the one that leads the Directors’ Report? Given the deleveraging over the half was one tenth of a turn (3.0x to 2.9x on the S&P basis), on what basis is the period characterised as balance-sheet transformation?
Fixed Wireless – the highest-margin product in the fixed book, running on TPG’s own network – fell from 285,000 to 284,000 in the half, in which 5G standalone launched, the addressable market rose 15%, more than 70% of the base moved to 5G, the MOCN footprint went fully live, 2100MHz was refarmed for 5G in Sydney and Melbourne, and a new modem shipped. Cease-sale in congested metro pockets explains the constrained suburbs; it does not explain the absence of growth across the newly available MOCN-footprint addresses. What were gross additions and disconnections for Fixed Wireless across the new footprint specifically, and what does management attribute the absence of net growth to, if not coverage or equipment?
Note 1(a) discloses a deficiency of net current assets of $677m and attaches a going-concern paragraph, against cash at bank of $55m and undrawn facilities cut to $360m. The improvement from the $917m December deficiency coincided with $687m of securitised receivables taken through operating cash flow. How much of the $240m improvement is operating versus the receivables sale? What proportion of the current liabilities is the current portion of borrowings and lease liabilities (cash-settled maturities) versus trade payables and deferred revenue (self-refilling)? And will the Board disclose the maturity profile of borrowings and facilities against the 2028 spectrum date?
Mobile churn and the ARPU sub-component breakdown were both removed at the moment mix shift, tenure, elasticity, a national outage, and a major wholesale migration make them most material. What are current churn and ARPU-component metrics versus prior corresponding periods? Given the June 2026 outage and the Moose migration, what were mobile port-out and churn volumes in the following weeks, by brand? Will the Board reinstate churn and ARPU disclosure so that “intention to try,” margin, and subscriber quality can be tested against outcome?
The 2020 Scheme Booklet promised “Australia’s leading challenger full-service telecommunications company” on a “more diverse earnings base across fixed broadband and mobile,” and expressly warned a shareholder wishing to keep “exposure predominantly to fixed broadband” might vote against. In October 2024 the Board sold that fixed exposure to Vocus and leased the fibre back for fifteen years (a $789m lease liability against a $509m right-of-use asset, ~$280m to the P&L on completion). Morgans called it a “reversal” of the merger argument. How does the Board reconcile the 2020 convergence rationale with the 2025 divestment of the convergence, and what is the current strategic thesis the Vocus sale advances? What is the total expected cash cost of the TAWFA over fifteen years against the ownership economics it replaced?
On the HY26 call, management’s two stated routes to a viable coverage future were low-earth-orbit satellite (“complementary technology”) and a favourable outcome from the ACCC mobile services inquiry, having spent ~$1.59 billion renting Optus’s regional network and decommissioned 755 of its own sites. Both routes rely on assets or outcomes the company does not own or control. What is management’s own base case for the business if the ACCC inquiry produces no mandated roaming remedy and LEO direct-to-device competition intensifies – the scenario Morgan Stanley’s Underweight is built around? What is being built, as distinct from rented or requested?
The vote against the remuneration report rose from 0.87% at the 2025 AGM to 12.17% at the 2026 AGM, while the resolutions to grant the CEO equity passed at 99.73% and 99.55% – shareholders directing their entire objection at the pay framework, not the grants. What specific feedback did the Board receive from shareholders and proxy advisers behind that shift, and what concrete changes to the remuneration structure will it make – given a “strike” at the next AGM would trigger a potential board spill? Will financial-performance and ROIC or TSR weightings be materially increased relative to internal satisfaction measures?
The FY25 scorecard recorded the company’s Net Promoter Score at maximum, in the same year TIO data showed Vodafone running ~40% adrift of Telstra and Optus – both of which reduced their volumes while TPG’s rose. How can an internal satisfaction measure sit at maximum while the independent complaints measure sits at a multi-year high for the same customer base? What is the methodology, sample, and channel for the internal NPS, and will the Board disclose it so shareholders can reconcile the two readings? Why was the TIO complaints metric added to the scorecard only prospectively, for 2026, rather than to the FY25 outcomes it would most obviously have captured?
A third independent director was appointed in June 2026, bringing the total to three of ten against an ASX recommendation of majority independence. The merger that justified the original structure occurred in 2020; the register has since changed fundamentally – Soul Patts fully exited, Robert Millner’s seat vacant, and AustralianSuper, a major holder, without representation. Does the new appointment represent a genuine path to majority independence, or a like-for-like replacement that leaves the structure unchanged? When will the board reflect the company – and its register – as they currently stand?
The TPG Telecom Foundation moved approximately $3.3 million in a year while retaining reserves of around $840, with no volunteers, one part-time employee, and a board drawn substantially from within the company and its associated firms. To what extent do grant recipients overlap with entities connected to directors of the Foundation or the parent, and what recusal processes govern grant approvals where a director has an association with the beneficiary? Will the Board commission an independent review of the Foundation’s grant-making and director-independence arrangements?
The annual report discloses an ACMA investigation and two customer deaths from Triple Zero failures. What is the scope of the investigation? What are the expected financial consequences – investigation costs, potential penalties, compliance-program costs – and why does the contingent liabilities note assess no material impact on financial position despite active investigations into incidents where customers died?
The KPMG whistleblower portal was shut down mid-investigation. A twelve-page independence challenge against the external investigator went unanswered before the investigation closed with no findings shared. What governance process allowed the portal closure while the matter was active? Was the independence challenge formally considered by the Board, and if so, what was the outcome?
Other provisions jumped from ~$2 million to $115 million with attribution only to “Vocus separation obligations.” What is the full breakdown? What is the expected utilisation timeline? Does any component relate to compliance remediation, Triple Zero response, regulatory investigation costs, or the whistleblower matter?
Felix ARPU is ~$26, roughly half of Vodafone postpaid. The HY26 mix shows Digital First up 114,000 while the premium base stood still. How much of Felix and digital-brand growth is incremental versus cannibalisation from Vodafone postpaid or prepaid? What proportion of net adds came from customers who previously held a higher-ARPU product within the Group – and what is the resulting drag on blended ARPU and Group margin dollars?
A 120-year-old investor that held through mergers, regulators, and a pandemic sold over $650 million of TPG shares and exited entirely between March and April 2026, taking its board seat with it; AFR has reported Vodafone Group PLC tracking the share price for a potential exit; and Morgan Stanley reports SingTel exploring a minority partner for Optus. What is management’s view of the register transition underway, and how does it affect strategic options – capital allocation, M&A, and the spectrum funding dilemma – particularly if CK Hutchison or Vodafone Group reduces?
TPG pays an estimated $25-35 million annually to license the Vodafone brand, with Vodafone nominees on the board – raising the question of whether their employer’s interest in licence revenue is fully aligned with TPG shareholders’ interest in minimising costs. Given the brand’s long-running perception challenges, an elevated TIO baseline, zero premium postpaid growth, and a national outage, should management consider retiring the Vodafone brand for a unified TPG identity – eliminating the licence cost and providing an opportunity to rebuild brand equity? (Across the Tasman, Infratil retired the Vodafone name for One NZ and reported metrics ahead of the old brand – see Post #92.)
This page synthesises publicly available company reports and calls, broker and analyst commentary, consumer forum posts and social media content, media articles, and independent analysis published in Posts #65-#101 on vodafail.com.au. It draws on TPG Telecom Limited’s Appendix 4D and Half-Year Financial Report for the half-year ended 30 June 2026 (including the Directors’ Report, consolidated financial statements, statement of financial position, and Notes – in particular Note 1(a) going concern and Note 2 segment reporting), the HY26 Investor Presentation and results webcast, the FY25 Annual Report and Remuneration Report, the 2026 Investor Day materials and the ACMA spectrum determination as reported therein, prior ASX announcements, and published broker notes. It also draws on Swoop Holdings Limited’s (ASX: SWP) filed half-year report, Appendix 4C quarterly reports, investor presentation and ASX announcements. Insider and industry comments are treated as anecdotal and unverified. Published for educational, informational, and public-interest purposes only. Not financial, investment, or legal advice. All views are opinion, commentary, or interpretation of public material. Market-capitalisation and plan-pricing figures are approximate, drawn from public data at the relevant dates, and subject to change.
Where figures have been derived by arithmetic, annualisation, or estimation from filed figures – including the illustrative free-cash-flow build, the estimated interest cost of debt-funding a spectrum instalment, the pro forma leverage impact, the per-service wholesale revenue estimate, and the composition of the net current asset deficiency – those derivations are described as illustrative or estimated, are not forecasts, and may prove materially inaccurate; the underlying figures are the company’s own. Subscriber, ARPU and financial figures are reproduced from the companies’ own disclosures.
The recording by an auditor of a material uncertainty as to going concern, in respect of either TPG Telecom Limited or Swoop Holdings Limited, is a standard audit disclosure and is expressly not a determination of insolvency. Nothing on this page asserts, implies, or should be read as asserting that TPG Telecom Limited or Swoop Holdings Limited is insolvent, is unable to pay its debts, is being wound up, or will fail; the directors of each company have stated their satisfaction as to the ability to meet obligations as they fall due, and this page records that. The scenarios described in respect of Swoop (survival at reduced scale, acquisition, or administration) are expressly hypothetical analytical possibilities, not predictions or statements of fact. The $4.18 million “commercial agreement receipt” is disclosed in Swoop’s own Appendix 4C; its counterparty is not disclosed, and this page expressly does not assert, imply, or invite the inference that TPG Telecom or any other identified party made that payment. Estimates of subscriber losses from the Moose Mobile migration are disputed analytical estimates. Porting figures attributed to a competing reseller are reproduced as provided by that company’s chief executive, whose interest is disclosed; they are unverified, describe one reseller’s inbound porting mix only, are a ratio and not a headcount, and are not a measure of total market movement. References to the November 2025 equity raising and its attributed cause reflect information available to this masthead and are presented as attribution, not established fact. The description of TPG as a “terminal business” is expressly an opinion held by observers and shared by the author on the basis of the disclosed facts, not a statement of fact as to solvency.
Nothing on this page alleges or implies that TPG Telecom Limited, Swoop Holdings Limited, Telstra, Optus, Macquarie, or any related entity or individual has engaged in misleading conduct, breached any accounting standard, or contravened any law, regulation, code or standard. The relevant financial statements were reviewed or audited and the reports issued; nothing here suggests otherwise. References to the Triple Zero matter, the ACMA investigation, and the ACCC inquiry reflect the company’s own disclosures and the existence of public processes; nothing here speculates on the outcome of any regulatory process, and the author extends condolences to the families involved. All views are the author’s honest opinions, formed on reasonable grounds from the material described, and published in reliance on the protections afforded to honest opinion, fair comment, and publication in the public interest under the Defamation Act 2005 (NSW) and its equivalents. Readers must conduct independent research and seek professional advice before making any investment, legal, or financial decisions. If TPG, Swoop, or any stakeholder believes any statement is inaccurate, please contact us – we will review and update where appropriate. The author has an active dispute with TPG Telecom Limited and has made protected disclosures under the Corporations Act 2001. The author holds a very immaterial shareholding in TPG Telecom Limited (ASX: TPG) and no interest in Swoop Holdings Limited or any competing entity named.
The Series
Thirty-two posts. One thesis. TPG Telecom is not a turnaround story. It is a repricing story running out of road, wrapped in transformation language, and governed by a board that appears to have confused the slides with the accounts.
Posts #65 through #96 document the financials, the governance failures, the regulatory exposure, the shareholder exodus, the ad agencies, the philanthropy, the disappearing acts, the complaints the company reports across three lines so nobody totals them, the coverage map redrawn by press release, the assets sold and rented back, the commissioned report that convicts itself, the fortnight the price changed four times, the base that fled a reseller rather than be moved onto the network, and the morning the market finally did the arithmetic. The EBITDA was $1,660 million. Pre-tax profit was $7 million. The ratio between them is 237 times. Everything else follows from that.
Read them in order if you want the full picture. Or start with whichever number matches the question you came here to ask.
Post #65 – When the Music Stops: The $7 Million Telco, Shrinking to Greatness – $4.2 billion in revenue. ~$7 million in pre-tax profit. Vodafone postpaid flat in 2H25. Digital-first momentum diluting mix. NBN still shrinking. Marketing spend up without premium conversion. This is what “shrinking to greatness” looks like in telco – cut the cost base, blend the reporting, raise prices on cohorts, call it discipline, and hope the complaints don’t hit earnings before the next narrative refresh.
Post #66 – The $2B Problem TPG Can’t Afford – UBS doubles its spectrum renewal estimate to ~$2 billion for FY27-30 and forecasts negative free cash flow in FY27. The CEO confirmed costs exceed $1.5 billion on the earnings call then refused to quantify further. Against an underlying earnings base of ~$7 million, the company returned $3.3 billion to shareholders and may now need to go back cap in hand for capital it just gave away.
Post #67 – The Bonus Year: Thin Earnings, Thick Optics – Two Triple Zero deaths. An ACMA investigation. A whistleblower matter. Zero postpaid growth. ~$7 million in underlying earnings. The Board looked at all of that and awarded $250,000 in discretionary bonus to the CEO and $435,000 across three executives. All risk gateways were assessed as passed. The Qantas parallels write themselves.
Post #68 – Buying the Narrative: Sponsored Spin, Friendly Headlines, and the TPG Story Nobody Checked – In December 2025, TPG paid for two pieces in The Australian promoting the MOCN and the transformation narrative. Two months later, the AFR ran results-day coverage with the headline “bolsters profits” – unpaid, and substantially identical in framing. The stock dropped 2.72% on results day. 115,000 investors found independent analysis elsewhere.
Post #69 – The Smart Money Just Left the Building – Washington H. Soul Pattinson – a 120-year-old investment house that held TPG through mergers, regulators, and a pandemic – sold $200 million of shares at a discount. The AFR hints the remaining ~10% stake may follow. When the shareholders closest to the register are reducing, restructuring, or quietly preparing exit pathways, the gap between the slides and reality becomes impossible to ignore.
Post #70 – Who’s Watching the Watchers? Board Independence, Shareholder Overhang, and the Governance Vacuum Ten directors. Two independent. Twenty percent independence on a company with $4.2 billion in revenue and a $7 million pre-tax underlying profit. The Chairman represents CK Hutchison in Hong Kong. Two Vodafone nominees represent London. The former General Counsel represents institutional history. AustralianSuper’s 3.93% stake – retirement savings for ordinary Australians – has no representative in the room at all.
Post #71 – Nine Lives: The Ad Agencies Vodafone Burned Through on the Way to Zero Growth – Nine creative agencies in seventeen years. Clemenger. Host. Ogilvy. Cummins & Partners. JWT. MullenLowe. VMLY&R. Saatchi & Saatchi. Now Howatson+Company. The advertising industry called Vodafone “Australia’s most toxic client.” $40 million in go-to-market spend. Zero net postpaid growth. At some point you stop blaming the agency.
Post #72 – Marked Safe from the Whistleblower Policy – Australia’s whistleblower laws promise protection, independence, and accountability. In practice, the company you reported selects the investigator, the General Counsel managing the response sits across key oversight functions, and the board’s structure raises obvious questions about independent oversight. The law gives you a shield. The system hands it to the people you’re shielding against.
Post #73 – The Story Nobody Will Publish – Eleven journalists engaged with the story. Every one of them went silent. What the media won’t publish, the platform already has – and nearly three million people have read it. There is a particular silence in Australian media that sounds nothing like the absence of a story.
Post #74 – Nothing Out Here – Vodafone told Australia there’s nothing out here. Australia replied in the comments. A national campaign featuring Ali Wong and an emu, launched after the half closed, for a product the market had already returned its verdict on. The coverage claim and the coverage map did not agree.
Post #75 – The Gift That Keeps Giving – Twenty-six million dollars in charitable donations since 2002. A foundation co-directed by the CEO and the General Counsel. A twelve-page independence challenge filed against the external investigator – submitted through the KPMG whistleblower portal TPG subsequently shut down. The investigator found nothing. The philanthropy, on closer inspection, looks like infrastructure.
Post #76 – The Seat Nobody Wants – $304 million out in a fifteen-day window. Soul Patts’ board representative gone. Two independent directors holding up an $8 billion company like load-bearing walls in a house that keeps adding floors. An AGM in four weeks where the questions that matter most are the ones nobody is asking.
Post #77 – Houdini Never Filed a Form 605 – Robert Millner AO has impeccable timing. Soul Patts built a 120-year reputation on knowing when to leave. The Form 605 was filed. The seat was vacated. The questions that follow an orderly exit are sometimes more revealing than the ones asked while you’re still in the room.
Post #78 – Fifteen Years and a Footnote – Section 191 of the Corporations Act requires a director with a material personal interest to disclose it. Every boardroom has a chair nobody looks at until someone has to leave it. Not because the meeting ended. Because the Act said so.
Post #79 – Read Receipts – A consumer advocacy platform is supposed to attract consumers. What is not supposed to happen is traffic from inside the company itself. The audience nobody invited showed up anyway. The receipts were already published.
Post #80 – Two Companies in a Purple Coat – Forty million dollars buys a lot of billboards. It does not, apparently, buy customers. The Ali Wong campaign arrived after the half closed. The most expensive identity crisis in Australian telecommunications is playing out in real time, and the two companies inside the purple coat are pulling in opposite directions.
Post #81 – Transformational: A $7 Million Result With a $1.6 Billion Costume – The CEO called it transformational. The EBITDA was $1,660 million. Pre-tax profit was $7 million. The ratio between them is 237 times. Operating free cash flow “nearly doubled” – once you strip out the $687 million one-off securitisation, it actually declined. Four hundred and two people liked the LinkedIn post. The statutory accounts tell a different story.
Post #82 – The Cartographer’s Apology – Vodafone spent the first half of 2026 insisting the coverage was there: a television campaign, a coverage map redrawn by press release, an apology that somehow blamed the competitor. Then, in June, the network went down nationwide and the territory answered back. The map was not the territory. Confusing the two is the beginning of most disasters.
Post #83 – Bagged a Moose – Vodafone “won” 135,000 Moose and Swoop customers – except “won” flatters it, since their provider simply switched networks and nobody asked the customers. They don’t count until each swaps a SIM, which some are refusing to do. The sharp deals that made Moose worth having have already vanished. A migration is not a victory, however the slide is captioned.
Post #84 – A Record Quarter – Vodafone had a record quarter – of complaints. The spike that hit late last year never receded; its rivals’ numbers are falling while its own won’t, and the brand that once led this scoreboard now trails it per customer. It reports the family across three separate lines so nobody totals them. We totalled them.
Post #85 – The Uninvited Guest – Two years of admiring the slide; one morning with a calculator. TPG fell about 8% on Investor Day, the market having finally totalled a $2.1 billion spectrum bill, a flat premium base, and a return on capital below its cost. The deeper story is a company quietly rewriting its own scoreboard – merging premium and digital-first into a single line so the dilution stops showing. The arithmetic arrived anyway. It does not leave early.
Post #86 – $840 and a Rolodex – A charity that moved $3.3 million in a year and kept $840 of it. Zero volunteers. One part-time employee. Contributions that lurch on the donor’s whim while the reserves never escape four figures – and a board, drawn entirely from one building, that changes everything except who controls it. The philanthropy is real. The independence is not.
Post #87 – Acting On Instructions – What does it mean when the independent solicitor chasing your debt, the mercantile agent he pays, and the creditor he acts for are revealed by an ACCC filing to be, functionally, the same hand wearing three gloves? TPG isn’t named in the proceeding – it’s merely a bespoke Annexure B client of the machine now in Federal Court. The question the letter never answered was who, exactly, you were being pursued by. This one answers it.
Post #88 – Sequins and Silence – Meet the man who held every seam at once – shareholder of the employer, director of the telco’s foundation, near-quarter-century partner at its merger firm – until a whistle blew and a CEO reached for a telephone. What followed was less a career and more an escape act: the law firm partnership gone within a month, the foundation board seat six months after that, each departure quiet, orderly, and for nothing in particular. And then – only then, safely out the other side – the reinvention: a shimmering jacket, a Woody Allen tribute act, a second life as a stand-up comedian for a man who’d just spent half a year proving he could vanish from anything. He left the boardrooms in silence and found his voice under stage lights. The exits didn’t make a sound. The encore is still going.
Post #89 – A Run of Unfortunate Weather – How does a man go from “best month” praise to a Letter of Concern in the span of a single spring week? This is an anatomy of workplace weather turning on a dime – the sudden barometric plunge that arrives, with remarkable convenience, precisely when it’s needed. The performance was never the problem. The timing is the entire subject, and the piece invites you to time it yourself.
Post #90 – Net of the Sale – Three documents, one company, filed weeks apart. The survey said the customers were staying; the guidance said the premium base wouldn’t grow; the accounts said the half-billion-dollar profit was a fibre network sold and gone by next year. Two of the three were marketing. The third was audited – and it did not agree with the other two. An analyst’s eyesight, it turns out, depends entirely on whose stock he is holding: he watches the postpaid spin-down at Telstra with real concern, then turns to TPG and pronounces the same customers, halving their spend from fifty dollars to twenty-five, a strategy. It is the only condition in medicine that improves as the patient gets sicker.
Post #91 – Eighty-Eight Per Cent – Last year, under one per cent of TPG‘s shareholders objected to executive pay. This year, twelve per cent did – a fourteen-fold surge on the one vote shareholders are handed purely to register displeasure. In between, the company earned one per cent on its continuing business, handed back $2.8 billion it got from selling the fibre network, and paid its chief executive 87.64 per cent of the maximum bonus – in a year the premium brand added nobody, complaints ran forty per cent adrift of its rivals, its own satisfaction score printed at maximum, and a regulator opened an investigation it declined to describe. A scorecard that returns an A-minus for a year like that is not measuring performance. It is laundering it.
Post #92 – Pawning the Silver – Every great estate in decline reaches the same quiet moment: the day the family starts selling the good things to pay for the rest, and tells itself the sale is a plan. In 2020, TPG promised fixed and mobile under one roof. Six years on the roof has been sold, the fibre rented back for fifteen years, the towers sold and leased in again, and the customers’ own handset debts sold forward at a loss – $687 million of receivables converted into a headline that operating cash flow had “nearly doubled.” It had not doubled. It had been pawned. Strip out the sale and the number went backwards. A company that must sell its assets to report cash from operations is not generating cash. It is holding a garage sale and calling it a salary.
Post #93 – The Moose That Wandered Home – Moose Mobile spent eight years as a quietly beloved Optus reseller. Then it switched its wholesale network to Vodafone, told almost none of its customers, and let them find out from their bills. What followed, tower by tower and post by post, was a public exodus – customers mapping their own coverage loss, porting back to Optus resellers by name, one declaring he would sooner “coat my entire body in pollen and nail myself to a beehive” than go back to Vodafone. Nothing about the base changed but the network. When a customer will accept a beehive over your product, the product is the finding. The variable is Vodafone.
Post #94 – Zugzwang – Open Vodafone‘s website and you find two companies arguing. One asks fifty-eight dollars; the other, beside it, asks forty-five and offers more. Both are Vodafone. In a single fortnight the company changed its price four times – premium on the first, a discount on the fourth, a student cut on the twelfth, promotional plans on the fifteenth – on a list it had spent months preparing and held, at full freight, for the length of a long weekend. This is the chess position where every legal move loses and passing is not allowed. Raise the price and the customers leave; cut it and the margin does. They are no longer arguing about how to win. There is no way to win. They are arguing about which way to lose.
Post #95 – Statistically Insignificant – There is a number on page three of Vodafone‘s new commissioned report the company wants the market to see: $1.1 billion of customer benefit. There is a second, on page fifty-six, it would rather you missed: eighty – the number of new customers the whole thing rests on, bolted to an admission that those who left Vodafone “are not captured in the sample.” A billion-dollar coverage story built by surveying the people who stayed, hand-picking a happy customer, and formally excluding everyone who walked. The report even carries, on page sixty-two, the line that undoes it: “should not be relied upon by anyone else.” TPG commissioned proof it could not generate, and Deloitte, to its credit, printed every caveat needed to see through it. The customers who weren’t surveyed have already answered, in the only currency that clears. They ported out.
Post #96 – Herd Immunity – On 16 June, Swoop told the ASX it would grow the Moose base from 135,000 subscribers to 180,000. That same week, on the forums, its customers began writing their resignation letters. To lure them onto the Vodafone network they didn’t choose, Moose is raffling a car – fifty thousand dollars in prizes to persuade people to do a thing they apparently will not do for free – while quietly switching off the discounts its growth forecast depends on. Australians have watched this film before: documents filed with the corporate regulator show Circles.Life spent $2.87 of marketing for every dollar of revenue chasing scale it never reached, and was carried out feet-first. Nine agencies could not sell this network; fifty thousand dollars in prizes cannot give it away. TPG did not win 135,000 customers. It won a live demonstration that its own network is the one its new customers are queuing to escape.
Post #97 – No Finite Term – The phrase sits in the accounts like a splinter: TPG’s handset receivables were sold “for no finite term.” Read it twice. The company has factored its customers’ own phone instalments to a Macquarie-arranged trust – the fourth such regime in a decade, HSBC to Greensill to a party unnamed to Macquarie, the same asset pawned, redeemed and pawned again – and the price of the latest sale was a $95 million derecognition cost, thirteen times the $7 million the continuing business earned before tax. A company that sells the same receivables four times is not managing capital; it is running out of it, and telling you so in a footnote. The “for no finite term” is the tell. There is no end date because there is no plan for the cash to come back – only the arrangement, renewed each time the previous one is spent, converting a stock the company owns into a flow that looks, from a distance, like earning.
Post #98 – Overwhelming Demand – There is a sentence a company writes only when the wheels have come off, and Moose wrote it to its own customers: please do not telephone us, on account of “overwhelming demand.” Demand is the word doing the concealing. A migration the market was told would add forty-five thousand customers has instead produced support queues past four hours, an ombudsman confirming a spike in complaints on the record, and a nine-year review score sliding in real time – the traffic is not people arriving to buy, it is people trying to leave and unable to get through the door to do it. The forums fill with the same five words, over and over, like a hostage note: I need to stay on Optus. TPG competed hard on price to win this base off Optus and onto Vodafone; the base is now besieging a call centre to get back. “Overwhelming demand” is what a company calls a stampede when it cannot admit which direction everyone is running.
Post #99 – Discretionary Spending – Two boards, two CEOs, one instructive contrast. TPG pays Iñaki Berroeta $8.2 million; Telstra pays Vicki Brady $7.5 million. TPG is worth about $7 billion; Telstra about $52 billion. When its network failed, Telstra‘s board docked its CEO $607,000 – for an outage that happened eight days after the year had even closed, before its own expert had reported. TPG‘s board, surveying a completed year that held two Triple Zero deaths, an ACMA investigation, a Senate appearance and a whistleblower matter, added $250,000 to its CEO on top of what the formula produced. The word that separates them is discretion – the judgement a board applies after the machine has spoken. One board used it to punish a failure that wasn’t yet its CEO’s fault; the other used it to reward a year of failures that were. The difference between the two companies is not the pay. It is what the pay is for.
Post #100 – Beyond the Peak – An executive supplied the phrase himself, in an answer about capital expenditure: TPG is “beyond the peak of the investments in 5G.” He meant the radio gear. Read against the rest of the half-year, it is the truest three-word summary of the company anyone has offered in two years. The premium subscriber base peaked in June 2025. The tax shield peaked and has been spent – the company is now a net taxpayer, having burnt its losses on the very fibre sale that produced last year’s headline profit. The $85 million interest saving from that sale peaked on arrival and was immediately eaten, $63 million of it, by the rent on the fibre and the fees on the receivables. Fixed Wireless peaked without ever climbing. And the “275 per cent profit growth” was three million dollars, manufactured by restating the prior year into a loss. Every source of improvement in this result was borrowed against something that will not be there next time. There is no such thing as a finished network – only a network you have stopped paying for, and a bill that arrives later under a different name. The bill is $2.1 billion, and it starts in 2028.
Unless otherwise stated, references to complaints, investigations, systemic issues, disclosure concerns, and governance failures are allegations, opinions, or analytical inferences based on public material, not findings of unlawful conduct.
Timeline of Events
April 2024 – Optus signs an MOCN network-sharing deal with TPG, an almost identical arrangement to the Telstra deal the ACCC had rejected the year prior. Ord Minnett, notably unmoved, records the net economic value as “not material enough at this stage to get overly excited, or otherwise.”
Mid 2024 – TPG removes mobile churn reporting and the ARPU sub-component breakdown from investor materials. The disclosure disappears at precisely the moment mix shift, tenure pressure, and postpaid stagnation are most acute – which is to say, at the moment it would have been most worth reading.
October 2024 – TPG sells its Enterprise, Government and Wholesale assets to Vocus for $4.7 billion. Morgan Stanley flags the loss of structural growth assets, and the company becomes, in its words, “predominantly a consumer mobile operator and NBN reseller.” The growth engine is sold; the consumer treadmill remains.
August 2025 – Macquarie returns from research restriction with an Outperform. 1H25 results land in-line, mobile net adds come in at +100k, and postpaid returns to marginal growth. The transformation narrative gains traction.
August–September 2025 – Jefferies cuts its price target after modelling churn, ARPU dilution, fixed-line headwinds, and MOCN execution risk – the first sign that direct analyst engagement with the operational red flags has begun.
October 2025 – A formal submission is lodged with ASX Compliance under Listing Rules 3.1/3.1B, and a case file is opened, covering the KPI removal, the Felix sub-brand disclosures, the handset-receivables flip-flop, and the inconsistent MOCN commentary.
November 2025 – TPG attempts to raise $550 million through a reinvestment plan at $3.61 per share. It raises $300 million, falling $250 million short. The AFR attributes the shortfall to the reputational consequences of a customer dying from a Triple Zero failure. The market, it turns out, was reading the parts of the annual report the analysts weren’t asking about.
December 2025 – CEO Inaki Berroeta publishes a paid opinion piece in The Australian promoting the MOCN and spectrum-cost narrative; a second sponsored piece appears the same day. The Board approves a CEO equity grant that same month – while the ACMA investigation is active and the whistleblower investigation has commenced. The ACMA investigation itself is disclosed in the FY25 annual report with the minimum of detail, covering matters the company declines to specify.
February 2026 – TPG reports its FY25 results. Underlying pre-tax profit: approximately $7 million. ROIC: 5.42%. Vodafone postpaid net adds: zero. Dividends paid: $335 million. The AFR runs “bolsters profits.” The stock falls 2.72%, and Macquarie cuts its FY26 EPS forecast by 73%. The headline and the tape disagreed, and the tape won.
April 2026 – The whistleblower matter progresses to formal legal correspondence, both sides now represented. In the same month, TPG shuts down the KPMG whistleblower portal mid-investigation. A twelve-page independence challenge filed against the external investigator goes unanswered. The investigator finds nothing. The investigation closes with no findings shared – the portal closed, the challenge unanswered, the conclusion unseen.
11 June 2026 – TIO complaint data for Q3 2025–26 confirms Vodafone’s complaint volumes remain elevated year-on-year against both Telstra and Optus. While the broader industry moderates, Vodafone’s baseline sits structurally higher than prior periods – a floor that a full year of remediation commitments, and a TIO metric added to the FY26 executive scorecard, has not brought down. A year of promises; the same floor.
18 June 2026 – Vodafone’s network crashes nationally. Over 8,200 complaints hit DownDetector before 10am – and the status page, hosted on the very systems that failed, went down with it. The matter continues progressing through legal channels.
Late June 2026 – Moose Mobile’s ~135,000 customers are migrated off the Optus network onto Vodafone’s, unasked – and the Whirlpool forums fill with those customers porting straight back out to Optus resellers, one describing Sydney CBD coverage as “deader than disco.” A UBS consumer survey the same fortnight reports Vodafone loyalty firming. The survey and the SIM card disagree, and the SIM card pays the bill.
Late March 2026 – Soul Patts files a further Form 605. Another $100 million out. $304 million in fifteen days. Robert Millner announces he will not seek re-election at the May AGM. The closest thing corporate Australia has to a Warren Buffett disciple looked at the boardroom table and decided the exit was worth more than the seat.
April 2026 – The matter progresses to formal legal correspondence. Both sides are now represented.
8 April 2026 – Soul Patts completes its full exit. 61.525 million shares at $3.98. $244.9 million. Total exit: ~$304 million across fifteen days. After 40 years on the register, they’re gone (as a substantial shareholder).
April 2026 – The KPMG whistleblower portal is shut down by TPG mid-investigation. A twelve-page independence challenge filed against the external investigator goes unanswered. The investigator finds nothing. The investigation closes with no findings shared.
May 2026 – Robert Millner retires from the TPG board following the AGM. Two independent directors now hold every committee that matters, outnumbered four to one on every vote.
11 June 2026 – TIO complaint data for Q3 2025–26 confirms Vodafone’s complaint volumes remain elevated year-on-year against both Telstra and Optus. While the broader industry trend shows complaints moderating, Vodafone’s baseline appears structurally higher than prior periods – a floor that has not returned to pre-complaint-surge levels despite a full year of remediation commitments and a TIO metric added to the FY26 executive scorecard.
Executive Summary
TPG Telecom (ASX: TPG) faces a compound risk set that has been building for two years and is only now beginning to surface in reported numbers.
The FY25 result was presented as transformational. Strip out a non-recurring $45 million tax benefit and the underlying earnings base was approximately $7 million on $4.2 billion in revenue. The company paid $335 million in dividends, awarded $435,000 in discretionary executive bonuses, and returned $3.3 billion to shareholders – all in the same year it generated a net margin of 0.17%.
The premium mobile brand added zero net postpaid customers. The fixed broadband base shed 116,000 NBN subscribers. Fixed Wireless – positioned as the growth offset – added 17,000 net. The MOCN arrangement with Optus, which cost approximately $122 million in its first year of operation, produced identifiable benefits of approximately $50 million. The gap is not narrowing.
UBS has since doubled its spectrum renewal estimate to approximately $2 billion for FY27-30 and is forecasting negative free cash flow in FY27. The CEO confirmed costs exceed $1.5 billion on the earnings call and declined to say more. The company that just returned $3.3 billion to shareholders may need to re-lever to fund a spectrum obligation it still won’t quantify.
The governance picture is equally concerning. Two independent directors out of ten. A non-independent Chairman. A CEO who received a $250,000 discretionary bonus in a year when two customers died from Triple Zero failures and an ACMA investigation commenced. A whistleblower investigation using a Band 1 external investigator that, to date, has not been the subject of a standalone market announcement.
Washington H. Soul Pattinson – a 120-year-old investment house that held TPG through everything – sold $200 million of shares at a discount in March 2026. The AFR signals the remainder may follow.
The bull case requires everything to go right. The evidence suggests very little is.
Regulatory and Compliance Exposure
OAIC and ACMA. The Office of the Australian Information Commissioner has accepted a complaint alleging breaches of APP 10, 12 and 13 – accuracy, access, and correction of personal information – and potentially invoking s13G of the Privacy Act 1988, which covers serious and repeated interference with privacy. Maximum civil penalty: the greater of $50 million, three times the benefit obtained, or 30% of adjusted turnover.
ACMA has commenced an investigation, disclosed in the FY25 annual report with no specification of scope. For benchmarking: ACMA’s largest ever penalty was $12 million, against Optus in 2024, for a Triple Zero failure. Proposed reforms would lift the maximum civil penalty to $10 million per breach – with penalties up to 30% of adjusted turnover for the most serious conduct – and would remove the current two-step enforcement process, enabling direct action. Against a $7 million pre-tax earnings base, even a mid-range outcome is material.
TIO complaint trajectory. In Jan-Mar 2025, TPG complaints rose 32.4% QoQ and 37.6% YoY. Vodafone mobile complaints rose 66.7% for poor coverage and 64.9% for service dropouts, against a market baseline of 0.6% QoQ; Telstra and Optus trended down. Engagement by the TIO’s Systemics team signals potential ACMA referral and systemic-process risk – not isolated case handling. And the surge did not subside: Q3 2025-26 data confirms Vodafone’s complaint volumes remain elevated year-on-year against both Telstra and Optus, its baseline sitting structurally higher than pre-surge levels while the wider industry moderates – a floor that a full year of remediation commitments, and a TIO metric added to the FY26 executive scorecard, has not brought down.
Each escalation carries direct costs: TIO lodgement fees, internal case management, formal written responses, remediation credits, and management oversight. At elevated complaint volumes, the combined cost stack across a year is material relative to the earnings base.
ASX Compliance A formal submission under Listing Rules 3.1 and 3.1B was lodged in October 2025. ASX Compliance has opened a case file. The submission covers: removal of key performance indicators (churn, ARPU breakdown); the handset receivables accounting reversal; inconsistent commentary around churn causation and MOCN economics; and potential disclosure gaps regarding Felix Mobile’s underlying economics, cannibalisation risk, and cohort-level breakeven.
Collections and conduct Reports of collections activity on disputed balances, consumer reports of collections activity involving deceased account holders, and potentially invalid default listings. Dealer channel conduct – including coverage misrepresentation, SIO cycling, and representations made during sales and retention calls that do not survive billing – remains a standing ACCC risk vector. These issues are not merely theoretical. Similar allegations and complaint themes appear repeatedly across publicly documented consumer complaints and dispute narratives.
Financial Implications
The Spectrum Wall
Management flagged funding from “cash flows and borrowing headroom.” The market did the arithmetic instead, and the stock fell ~8% on the day. Every option degrades the balance-sheet narrative management spent FY25 building: re-lever, dilute, cut CAPEX, or sell spectrum. Having returned $3.3 billion to shareholders, the company may now find the spectrum bill larger than what remains on the balance sheet to absorb it – and the $1.7 billion debt paydown it spent two years celebrating will need to be partially unwound before the decade is out. The capital was returned with a flourish. It will be borrowed back in silence.
MOCN unit economics Fixed commitment of ~$143 million annually once fully ramped. Not volumetric – does not adjust with subscriber outcomes. Year-1 identified benefit: ~$50 million. Year-1 identified cost: ~$122 million. Net: approximately -$72 million in the first year. The original market logic – 100,000 to 200,000 incremental premium postpaid adds – delivered zero. The CFO’s response to questions about the breakeven hurdle was to say “break even is definitely not our aspiration” while declining to provide a revised target. The goalposts have been quietly moved to a field where the economics are worse.
Disclosure Churn removed. ARPU sub-components removed. Four different profit definitions in active use simultaneously. Broker EPS estimates for FY26 range from 6 cents to 19 cents across four analysts covering the same result on the same day. That dispersion is not analytical disagreement. It is the direct consequence of insufficient disclosure.
Dividend ~$335 million paid on statutory NPAT of $52 million – itself dependent on a non-recurring $45 million tax benefit. Payout ratio: 644%. All four covering brokers forecast 19-20 cents in dividends against EPS estimates of 6-19 cents for FY26-27, implying payout ratios of 100%-300%+. The dividend is a signalling device. Whether the signal survives the spectrum wall is the question the market has not yet priced.
Network, Product and Execution
Vodafone postpaid 2,846k at FY24. 2,846k at FY25. Zero net adds. The MOCN doubled the coverage footprint, cost $40 million in GTM spend, and was handed the Optus outage as a switching opportunity. Telstra grew postpaid. Optus grew postpaid. TPG did not. The CFO declined twice on the earnings call to provide the subscriber and ARPU assumptions underpinning FY26 EBITDA guidance. One analyst said it plainly: “Sounds like you don’t really want to go into specific assumptions around subs and ARPU.”
Felix and the “Netflix of telco” illusion – Felix ARPU: $25.75. Vodafone postpaid ARPU: ~$50. Felix has no fixed-term contract, no binding tenure, and churn running an estimated 6-10x higher than the Netflix benchmark it is measured against. It is automated prepaid billing dressed in subscription language. Every customer who migrates from Vodafone postpaid to Felix is a revenue downgrade. The 228,000 mobile net adds headline includes a large proportion of lower-ARPU digital-first customers. The premium engine is not growing. Felix is camouflaging it.
Fixed broadband NBN: -116,000 net subscribers, down 6.9%. Fixed Wireless: +17,000 net, despite the new modem launch. Metro suburbs face congestion-driven capacity constraints acting as effective cease-sale dynamics. CAPEX is being reduced. Challenger brands including Aussie Broadband and Superloop continue recording SIO wins half after half. Fixed contributes approximately 23% of Group EBITDA and has no credible path back to growth. Management calls the market “structurally challenged.” That is accurate. It is also an admission.
Coverage gaps – Over 140 fringe towns have been identified lacking network continuity between TPG’s native coverage and the MOCN footprint. Approximately 755 regional sites are being decommissioned with heavier reliance on Optus. Post-agreement, bargaining leverage weakens. Coverage gaps remain across regional NSW, coastal QLD, NT, southern TAS, WA, and SA – including areas where Optus maintains exclusive presence. “Double the Network” described a theoretical maximum. The practical reality involves a material rump of underserved geographies.
The app iOS App Store reviews for the Vodafone app frequently cite login failures, poor functionality, and weak usability. Management has noted the front-end rebuild is on a 12-15 month horizon. If the app is contributing to postpaid churn – and there is reasonable basis to think it is – deferring front-end remediation while CAPEX is cut raises the question of which problem is being prioritised, and why.
Governance Considerations
The governance picture at TPG is best understood not as a series of individual failures but as a structural condition that makes individual failures predictable.
Consider the composition. A non-independent Chairman representing CK Hutchison. Two Vodafone Group nominees carrying structural conflicts on the very brand-licensing decisions that pay their employer. A former General Counsel whose institutional knowledge insulates rather than illuminates. A CEO director who is not independent by definition. And, until recently, Soul Pattinson’s nominee – a seat whose mandate dissolved the moment the stake did. Robert Millner retired following the May 2026 AGM, and the seat left with him.
Which leaves the arithmetic starker than it was. Three independent directors now carry the full governance workload of a $5 billion-revenue company, and remain outnumbered on every vote that matters. A board built for a company that no longer exists, overseeing one that increasingly does – and structurally incapable, by its own composition, of mustering a majority the moment a majority is required.
The two independent directors – Helen Nugent AC and Paula Dwyer – chair and sit on every critical committee. Between them, they carry the full independent governance workload of a $4.2 billion revenue company. They are outnumbered on every vote. When the Governance, Remuneration and Nomination Committee recommends against a discretionary bonus, the full board – where eight of ten seats belong to shareholder nominees, legacy insiders, and the CEO – makes the final decision.
What passed through this structure in FY25: two Triple Zero deaths. An ACMA investigation. A whistleblower investigation run by a Band 1 external investigator. Surging complaint volumes. Zero premium postpaid growth. And $435,000 in discretionary executive bonuses.
All risk gateways passed.
The Qantas, AMP, and Crown parallels are not rhetorical flourishes. They are documented precedents for what happens when a board structure prioritises shareholder bloc representation over independent accountability. In each case, the pattern was the same: the failure was visible long before the crisis forced board renewal. The cost was borne by minority shareholders, customers, and employees. The directors who presided over the failure departed with their reputations intact and their fees collected.
AustralianSuper holds 3.93% of TPG – approximately $300 million of ordinary Australians’ retirement savings. It has no board representative. It relies entirely on two independent directors who are structurally unable to carry a vote against the combined bloc. That is not a governance framework. That is a quorum for rubber-stamping.
Brand and Reputational Risk
Vodafone Australia pays its parent company an estimated $25-35 million annually to use the Vodafone name. The Vodafone nominees on the TPG board represent a brand licensor with a direct financial interest in licence revenue continuity – an interest that may not be aligned with TPG shareholders’ interest in minimising costs or resetting brand positioning.
The brand itself has problems the licence fee compounds. “Vodafail” is resurfacing in mainstream coverage. Hundreds of documented consumer complaints describe billing errors, false credit flags, and complaint-handling failures that match the same systemic patterns. TIO complaint volumes are trending the wrong direction while competitors decline. The “Double the Network” campaign featured drone footage of Dalton NSW – a location with weak or no indoor coverage per Vodafone’s own maps. A trust-building campaign that contradicts verifiable network data does not build trust.
The question several analysts and commentators have raised – and which the Vodafone nominees on the board are structurally conflicted from answering objectively – is whether the brand is worth what it costs. A unified “TPG Mobile” identity would eliminate the licensing fee, remove the conflict of interest, and provide a clean reset opportunity. Whether the Board is capable of asking that question with the current composition is a different matter.
Media Coverage
December 2025 – Sponsored content, The Australian TPG CEO Inaki Berroeta publishes paid opinion piece promoting the MOCN and calling for “fair and predictable” spectrum pricing. A second sponsored piece the same day promotes MOCN regional coverage through TPG’s General Manager of Strategy. Neither mentions: zero postpaid growth, the $115 million provisions jump, the ACMA investigation, the two Triple Zero deaths, or the whistleblower matter that had just commenced. Both are labelled Sponsored Content.
February 2026 – Results day, Australian Financial Review Headline: “Demand for Netflix-style mobile plans bolsters TPG Telecom profits.” The CEO quoted extensively. Zero postpaid growth buried in paragraph six. No mention of ACMA, Triple Zero, the $115 million provisions jump, the $435,000 discretionary bonuses, or the $2 billion spectrum estimate that UBS would publish days later. The framing, in my view, closely resembles the paid content that appeared in The Australian two months earlier. The stock falls 2.72%. The market disagreed with the headline.
March 2026 – AFR Street Talk Washington H. Soul Pattinson sells $200 million in TPG shares at a discount. AFR Street Talk explicitly flags that the remaining ~10% stake could follow, with TPG’s free float climbing to 41% if the full exit proceeds. The block trade clears within minutes of market open.
June 2026 – Investor Day TPG’s 2026 Investor Day confirmed $2.1 billion in spectrum renewal costs between 2028 and 2032, with the first payment of $840 million due in 2028 – larger than annual free cash flow before dividends. Vodafone postpaid remained flat despite a 100-day 50% promotional discount running through the period. NBN lost another 35,000 subscribers. Digital brand growth continued at roughly half the postpaid ARPU. The stock fell approximately 8% on the day. The bear case stopped being analytical. It showed up in the price.
115,000 investors Independent analysis of TPG’s FY25 results published on Reddit reached 115,000 views. It covered the $7 million earnings base, zero postpaid growth, MOCN economics, and the governance issues that 44 minutes of professional analyst Q&A did not raise. The gap between what was published in mastheads and what investors actually needed to know was large enough to drive traffic of that scale.
Prepaid, Wholesale and Sub-Brand Strategy
The prepaid problem TPG’s prepaid base is structurally low-margin and transient by nature. A significant proportion is tourists on short-stay SIMs, promotional port-ins who chase introductory offers, and serial SIM-hoppers cycling between providers. These customers cover their variable costs. They do not fund MOCN economics.
Once you strip out transient volume, churn recycling, and promotional cohorts with poor retention, what remains is a segment that sustains the base – it does not shift the needle on a $1.57 billion network commitment premised on durable, high-ARPU postpaid additions.
Lebara, Kogan and Lyca TPG’s wholesale and MVNO portfolio delivers volume at lower ARPU and thinner margins. Lyca Mobile – acquired in 2025 – added over 95% of the wholesale base. The key question is whether these brands are genuinely incremental or effectively SIO filler that competes with Vodafone’s own base at lower price points. High-data wholesale and MVNO plans also consume Fixed Wireless capacity, limiting TPG’s ability to grow its higher-AMPU FHW product in the very markets where it should be winning.
Felix cannibalisation Felix ARPU is approximately $25.75. Vodafone postpaid ARPU is approximately $50. When Felix grows, the question is not how many subscribers – it is from where. Anecdotal market feedback and the flat postpaid print suggest Felix is taking share from within the Group rather than from Telstra and Optus. Growth that substitutes premium ARPU with discount ARPU compresses blended economics. It does not constitute market share gain. It constitutes a blender.
Methodology and Sourcing
This page synthesises: publicly released company results and investor call transcripts; broker and analyst commentary including Jefferies, Morgan Stanley, UBS, Macquarie, Morgans, and Ord Minnett; TIO quarterly complaint statistics; ACMA consultation documents and preference positions; OAIC published information; ASX announcements and Listing Rule submissions; consumer forum posts, social media, and ProductReview data; media coverage from the AFR, The Australian, and ABC; and independent analysis published in Posts #65-#81 on this site.
Industry and insider comments referenced are treated as anecdotal and unverified unless independently corroborated. All financial analysis is based on publicly disclosed figures and is subject to the limitations of external modelling.
All entities and individuals referenced retain the presumption of lawful conduct unless determined otherwise by a competent regulator or court.
Right of Reply
TPG Telecom Limited, Vodafone Australia, and any executives, directors, or representatives referenced on this page are invited to provide clarification, correction, or additional context on any matter raised.
If TPG believes any factual statement is inaccurate, incomplete, or requires qualification, it is encouraged to provide specific corrections with supporting documentation and any relevant contextual explanation.
Verified responses will be published in full and in context, without editorial distortion.
Silence will not prevent continued analysis based on publicly available information, regulatory disclosures, and documented evidence.
Disclaimer
This page is published for educational, informational, and public-interest purposes only. It does not constitute financial, investment, or legal advice and should not be relied upon for any decision-making purpose.
All views expressed are opinion, commentary, or interpretation of public material and anecdotal reports. While reasonable care has been taken, no guarantee of completeness or accuracy is made. Financial figures referenced are drawn from publicly released materials and are subject to revision, restatement, or differing interpretation.
References to regulatory scrutiny, governance risk, systemic complaint patterns, and potential areas of investigation are framed as matters of public interest and analytical commentary – not findings of misconduct.
Nothing in this publication constitutes a finding that TPG Telecom Limited, Vodafone Australia, or any individual has breached any law, regulation, or code unless such breach has been formally determined by a competent authority.
The author has an active dispute with TPG Telecom and has made protected disclosures under the Corporations Act 2001. The author holds a very immaterial shareholding in TPG Telecom Limited (ASX: TPG). These matters should be considered when evaluating the analysis presented.
Readers must conduct independent research and seek professional advice before making any investment, legal, or financial decisions.
Also by the author:
The rest of the register lives at Unaudited: Australian companies, read from the filings rather than the press release.
Telcos:
- Swoop Holdings: the 100,000 migrated services and the going concern doubt
- Superloop: the Sri Lanka staff and the rupee salary peg
- Circles.Life Australia: the accounts behind its exit
Everyone else:
- Retinue: $38.9 million lost without a gross profit and who gets paid first
- Lawpath: the shareholders, the valuation and the Westpac deal
- Emperor’s Garden: why the cream puff window closed