Skip to main content
    All shows

    TMT Brief · Saturday, October 3 · 7 min

    TMT Brief · Saturday, October 3, 2026

    0:00-:--
    Speed

    Transcript

    Koko: The UK competition regulator just put a two billion pound fibre merger on the clock. The CMA has concerns about Nexfibre buying Netomnia, and the clock's running: remedies due in two weeks, final decision December fifteenth.

    Max: [surprised] Two weeks. That's not a lot of runway to restructure a deal backed by Liberty Global, Telefonica and InfraVia.

    Koko: Right, and this is the live test everyone in fibre consolidation is watching. Fast around the rest: Accenture's quarter landed clean, eighteen point six eight billion in revenue, record twelve point seven seven billion in managed-services bookings.

    Max: Samsung raised Galaxy S26 prices by up to two hundred dollars, blaming memory shortages that are really about AI data center demand eating the chip supply.

    Koko: And McKinsey's number landed hard: only thirty-seven percent of companies see any EBIT impact from AI programs, even though eighty-nine percent use it regularly.

    Max: So here's why this week matters for the board deck. Regulatory friction and component cost pass-through are now just as material to the model as AI topline growth. You have to budget for both sides of that ledger.

    Koko: Let's actually sit with Nexfibre-Netomnia, because two billion pounds of consolidation just hit a wall most people assumed was gone post-Brexit deregulation talk.

    Max: I don't think the wall's gone, I think it moved. The CMA's basically saying: show us remedies, fast, or this doesn't close on the timeline you modeled.

    Koko: Which changes the math for every other altnet rollup in UK fibre watching this. If Nexfibre, with Liberty Global and Telefonica behind it, still gets flagged, nobody's deal is automatically clean.

    Max: I'd actually call that healthy. UK fibre has been overbuilt for years, too many altnets chasing the same homes passed. Consolidation was always coming, the question was whether regulators would let it happen on favorable terms.

    Koko: Favorable to whom though? If remedies mean divesting footprint or opening access, that erodes exactly the synergy case that justified the two billion price tag in the first place.

    Max: Sure, but the alternative is a market where five altnets bleed capital forever because nobody's allowed to merge. December fifteenth is the date that matters, that's when we find out if this is a speed bump or a dealbreaker.

    Koko: For any CFO with fibre assets on the balance sheet, or eyeing an exit, that's now a line item in your scenario planning, not a footnote.

    Koko: Accenture's quarter is the cleanest read we've had on whether enterprise AI spend is actually converting. Eighteen point six eight billion in revenue, seven percent local-currency growth, and a one point two book-to-bill.

    Max: Book-to-bill over one is the number I'd circle. That means bookings are outrunning revenue recognition, backlog is building, not draining.

    Koko: Record twelve point seven seven billion in managed-services bookings specifically. That's not pilots, that's contracted, multi-year work.

    Max: Which directly argues against the McKinsey number from the same week. Thirty-seven percent EBIT impact sounds like AI isn't paying off. Accenture's book-to-bill says somebody's paying for it anyway.

    Koko: Those aren't actually contradictory though. Enterprises are signing the transformation contracts before they've proven the EBIT case internally. Accenture gets paid either way.

    Max: [chuckles] That's the honest version. The services layer monetizes the hype cycle regardless of whether the client's P and L moves. Good work if you can get it.

    Koko: So for a CFO watching peer spend, Accenture's backlog is the leading indicator, McKinsey's EBIT number is the lagging one. Use both, don't just pick the flattering one.

    Max: And if you're the CIO justifying next year's budget, this is your evidence that the market's still buying the story. Just don't confuse demand for proof of return.

    Koko: Samsung just raised Galaxy S26 prices by a hundred to two hundred dollars across the range, and the reason is memory shortages.

    Max: Which traces straight back to AI data center demand soaking up the same chip supply. That's the capex cycle finally showing up as a hardware margin line, not an abstraction.

    Koko: This is the part that should worry every device and hardware CFO. You're not competing with other phone makers for memory anymore, you're competing with hyperscalers building data centers.

    Max: And hyperscalers will outbid you every time, their unit economics tolerate it. Samsung's just passing the cost through to consumers because it has the brand power to do it.

    Koko: Not everyone has that pricing power though. A smaller device maker eating the same component inflation without Samsung's margin cushion is in real trouble.

    Max: So this is a sourcing and procurement conversation that belongs in the CFO-CTO relationship now, not just a supply chain footnote. Component cost forecasting needs an AI-demand variable in it going forward.

    Koko: Right, this isn't a one-quarter blip if the data center buildout keeps running at this pace.

    Koko: The FTC opened a probe into Anthropic, OpenAI and peers over AI safety practices, right as both are prepping IPOs.

    Max: Regulatory overhang stacking on top of disclosure season. Bad timing for them, good leverage for regulators.

    Koko: McKinsey's adoption-versus-EBIT gap, eighty-nine percent usage, thirty-seven percent impact, says the monetization story is still unproven at the P and L level.

    Max: And component costs, fibre antitrust, services backlog, three totally different parts of the stack all getting repriced by the same AI buildout.

    Koko: So what's a board-deck item versus what's a call this month. Board deck: model component cost inflation, memory specifically, into any hardware or device forecast for next year. That's structural now.

    Max: This month: if you've got fibre or telecom infrastructure assets anywhere near a consolidation conversation, get your antitrust read done before December fifteenth, not after.

    Koko: And board deck again: don't approve the next AI budget cycle off adoption numbers alone. Demand the EBIT attribution McKinsey says most companies still can't show.

    Max: This month: if you're selling AI transformation services, Accenture's book-to-bill says the market's still buying, so this is your window to close deals before buyer scrutiny catches up to the ROI gap.

    Koko: The full written breakdown is at koko knows dot A I.

    Max: Here's the open question for the weeks ahead: when the CMA rules December fifteenth, does fibre consolidation get easier or harder everywhere else in Europe. Watch that date.