The CFO spoke and the air in the call felt different — not tense, exactly, but unmistakably shifted. I listened as the numbers landed: negative free cash flow, billions in AI spending, and an even larger bet for next year. If you own tech stocks, that was the moment you either leaned forward or quietly checked your positions.
I’ll walk you through what happened, why it matters, and where the risk actually lives.
On the earnings call, Alphabet disclosed a first: negative free cash flow — How a cash surplus turned into a hole
Alphabet closed Q2 2026 with negative free cash flow of $5.9 billion (€5.4 billion), a first since the company went public. That number arrived alongside a shock to guidance: 2026 spending was raised to as much as $205 billion (€189 billion), up from the prior range of $180 billion to $190 billion (€166 billion–€175 billion). The company also warned capex will “increase significantly in 2027,” and some analysts now model at least $262 billion (€241 billion) of spending over the next period.
Listen to the language: CFO Anat Ashkenazi framed the cash loss as a deliberate bet on AI infrastructure. I’ve heard executives call big bets “investments” before; this felt different because the size of it chewed through cash flow in real time. It’s one thing to promise future returns; it’s another to flip your cash statement negative while you build.
Why did Google’s free cash flow turn negative?
Because Google is pouring capital into data centers, chips, networking and custom servers to power large models — and doing so at a pace that outstrips current operational cash generation. The company told investors it expects free cash flow to remain pressured by “investments in technical infrastructure” that are meant to capture the AI opportunity.
On the balance sheet, capital spending now dwarfs routine needs — What that says about priorities
Slides on the call showed cloud growth and a line item that towers over the rest: capex. Google updated its 2026 plan and flagged meaningful increases into 2027. You can feel the strategic trade-off: growth for the future at the expense of present liquidity.
That trade-off is why some investors have turned cautious. Apple, which has largely avoided the same level of AI capex, outperformed parts of Big Tech and briefly reclaimed the title of the most valuable company. Meanwhile, banks of analysts — and outlets like Bloomberg and Reuters — are asking whether hyperscalers are simply spending ahead of validated demand.
How much is Google spending on AI?
Between the updated guidance and analyst estimates, we’re talking in the low hundreds of billions for 2026 and an even larger commitment thereafter: $200 billion (€184 billion) is the shorthand some circles use, but the explicit number on the call was up to $205 billion (€189 billion). That’s not a line-item tweak. It’s a capital program that can rewire an industry’s supply chain.
On the product stage, Gemini’s debut was loud — But competitors kept firing back
I remember the online buzz when Gemini 3 launched last year; it felt like a reset. Since then, the cadence from other players has picked up, and Google’s pace has slowed in public-facing headlines.
Google released Gemini 3.6 Flash but faces benchmark gaps against OpenAI, Anthropic and even Elon Musk’s Grok in some tests, while the flagship Gemini 3.5 Pro reportedly runs months behind schedule. CEO Sundar Pichai said Gemini 3.5 Pro is in testing and that Gemini 4 is being trained with “ambitious” goals. Jensen Huang at Nvidia has spent months telling investors he’s confident hyperscalers will see cash-flow inflection because agentic AI is proving its value — a public pep talk that reads like reassurance to anyone who sells AI compute.
All of that matters because the hyperscalers are Nvidia’s largest customers. If the returns on this buildout falter, Nvidia’s demand could wobble; if the returns materialize, the entire supply chain benefits. I see this as a high-stakes poker table: bets are huge, information is asymmetric, and players are sizing up each other’s hands.
On the revenue lines, cloud is strong even as Search cools — Where the growth is actually coming from
On paper, Google’s cloud is the bright spot. Cloud revenue hit $24.77 billion (€23 billion) for the quarter — an 82% year-over-year increase — and executives pointed to strong demand from both external customers and internal usage.
Yet core Search missed some expectations even as usage reportedly climbed during the World Cup. That split matters: cloud can justify heavy capex if margins and customer traction scale, but Search weakness reminds you that new AI products have to carry commercial weight — not just hype.
Will AI spending hurt Google’s stock?
Short answer: it can, if investors start to question the payback window. Negative free cash flow is a clear signal that investor patience is being tested. The market will price not just current spending, but the probability that those dollars produce durable, monetizable products. If AI tools like Gemini fail to convert usage into revenue quickly enough, the stock will react — and not kindly.
On the competitive map, hyperscalers are all-in while others are cautious — The strategic split you should watch
Across the industry, Microsoft, Amazon, Meta and Google have all committed colossal capex to AI. Apple has taken a different route, and that conservatism has been rewarded by some investors. You can read those moves as two philosophies: spend to own the infrastructure, or conserve cash while integrating AI more slowly.
If you ask me, both positions carry risk. The spenders risk overbuilding before demand matures; the cautious risk being outpaced in AI-enabled services. This is where the market’s nervousness comes from — and why a single quarter of negative free cash flow can feel seismic.
So where does this leave you as a reader, investor or product person? Google’s numbers show ambition: cloud momentum, aggressive capex, and a public roadmap for Gemini 4. They also show a company willing to trade near-term cash for long-term capacity. That gamble could reshape computing — or it could leave balance sheets strained and competition picking over returns.
The question is no longer whether Google can build the largest AI stack; it’s whether that stack will pay for itself fast enough to matter — and who benefits if it doesn’t?