Between October 29 and 31, 2026, Alphabet, Amazon, and Meta reported third-quarter earnings within days of each other — and the market reacted harder than it had all year. On Thursday, Alphabet's shares fell 12% over two days, its worst stretch on record, wiping out roughly $500 billion in market value. On Friday, the Nasdaq dropped 2.9%, Nvidia fell 4%, Broadcom lost more than 5%, and Qualcomm slid 11%. The trigger wasn't weak results — revenue grew at all three companies. The trigger was the price being paid for that growth.
Alphabet, Amazon, Meta, and Microsoft together guided to more than $740 billion in capital expenditure this year. Alphabet's free cash flow was negative for the second straight quarter (-$9.1 billion for the quarter); Amazon's trailing twelve-month free cash flow stood at -$42.4 billion, versus +$14.8 billion a year earlier. Five companies — Microsoft, Alphabet, Amazon, Meta, and Oracle — issued a combined $326 billion in debt in 2026, triple last year's total, according to S&P Global Market Intelligence. The one exception was Microsoft, whose capex came in $5 billion below guidance while free cash flow stayed positive; the company attributed the gap to delayed third-party capacity, not a deliberate pullback.
When growth stops being a convincing argument
Until this week, the investor logic was simple: more data-center spending meant brighter prospects. The question has now shifted to "when does this start paying for itself." Meta's ad revenue grew 26%; Amazon's AWS cloud segment grew 39% — and both stocks still fell (Meta 10% the prior week, Amazon 7% after earnings). For the first time at this scale, the market isn't doubting that demand for AI infrastructure exists — it's doubting whether that demand is being financed sustainably, largely through debt rather than through the companies' own cash flow. For Alphabet, there's an added wrinkle: AI Mode in Search is starting to eat into the very ad revenue that funds its own development — search ad growth slowed to 12%, its weakest quarter in five.
The bill that eventually lands on small businesses
For a small business owner, this isn't an abstract Wall Street story. Every current price point for ChatGPT Business, Gemini for Workspace, Claude, Copilot, and the dozens of AI features baked into CRMs and ad platforms was set under conditions where growth at any cost was rewarded, not penalized. If investors keep demanding a return on capital rather than just revenue growth, vendors are left with three levers: raise subscription and API prices, tighten free-tier limits, or push costs onto integrators and partners. Mark Zuckerberg himself admitted that Meta's Muse AI agent is "very compute-intensive to run" — and that's an agent the company is handing out nearly for free to build its user base.
Picture a concrete case: a small ad agency in Lisbon manages eight clients and relies on a stack of three AI subscriptions — one for copywriting, one for automated reporting, one for a support chatbot. If even one of the three vendors raises its price by 30–40% within six months, the agency's margin on those accounts could evaporate, because the tools' cost was never built into client pricing as a variable — it was treated as a fixed line item. The sensible move isn't to panic or abandon automation, but to build an annual price-review clause into client contracts and avoid making a core business process entirely dependent on one irreplaceable vendor.
It's worth holding on to the skeptical side too: a stock sell-off isn't proof that "the AI bubble has burst." Demand for compute is genuinely growing — Google, Amazon, and Microsoft all reported cloud-segment growth in the very same quarter. The gap between corporate optimism and market nervousness isn't about whether demand exists; it's about how that demand is being financed — with debt rather than profit — and how fast that debt will need servicing. This looks more like a risk repricing than an industry collapse, but that kind of repricing usually reaches end-customer prices faster than it reaches the headlines.
For a business building processes on top of external AI services, the next two or three earnings cycles from the big vendors are worth reading not as news from another world, but as a signal about future pricing. That's exactly why, at Dayava, we design automations from day one so the core logic doesn't hinge rigidly on a single AI provider, and so swapping the engine under the hood doesn't mean rebuilding the whole system.
If you want your processes to hold up against swings in the AI market — leave a request at dayava.pt/contactos/.
Source: TechCrunch, "Cracks are emerging in the AI spending boom"