I was on an earnings call the moment the numbers landed: executives promising a brighter year while the spreadsheet bled red. You could feel the room shrink — optimism up front, anxiety behind the figures. I’ve seen this before, and I’ll tell you what to watch next.
At the open of the call the CFO stated exact figures: Google’s free cash flow plunged again, and the company raised its spending plan for the year.
Google reported free cash flow of negative $9.07 billion (€8.44 billion) for the quarter, deeper than its prior negative $5.9 billion (€5.49 billion). That’s about a $3.17 billion drop in one quarter—roughly $3.17 billion (€2.95 billion)—and a signal investors don’t like: growth in machine power is outpacing immediate returns.
Quarterly revenue of $127.58 billion (€118.68 billion) missed expectations, long-term debt rose to $125.1 billion (€116.34 billion), and management disclosed future purchase commitments approaching $1 trillion (€930 billion). On the call, CFO Anat Ashkenazi said Google expects to spend between $195 billion and $205 billion this fiscal year ($195B = €181.35B; $205B = €190.65B), most of it on AI infrastructure and TPU inventory.
Why did Google’s free cash flow drop?
Because capex and inventory for AI systems are massive and front-loaded. You’re paying for racks, chips, and data-center expansion today so models can run tomorrow. Ashkenazi warned that scaling TPU systems and stocking components will continue to pressure free cash flow even as leadership promises attractive returns on invested capital later.
At the podium Sundar Pichai introduced a new frontier model right before the market closed.
Google debuted Gemini 4 Carbon — branded as its strongest frontier model — and said it’s already used internally to boost productivity. Pichai called it a major step and said external access will be tightly controlled, initially opening to a subset of trusted cyber defenders and reportedly being tested by the Trump administration.
Philipp Schindler admitted AI Mode and AI Overviews drive user engagement but have “lower monetization” today; the company plans to increase ad saturation in those experiences. If you’re trying to parse what happens next: adoption exists, but monetization lags.
What is Gemini 4 Carbon and who gets access?
Carbon is positioned as Google’s most capable frontier model yet; internally it’s meant to improve productivity across Search and Workspace, and externally it’s being limited to select defenders at first. That guarded release aims to balance control and reputation — and to keep the model from becoming a public relations liability while the company figures out commercial paths.
Watching market reaction: shares fell sharply and skeptics resurfaced within hours.
Shares dropped about 11% after hours. Investors seemed unconvinced that AI productivity gains will translate to near-term cash flow recovery — especially given the scale of capital being deployed. Michael Burry and Torsten Slok have both raised alarms: Burry warned about an AI bubble potentially popping, and Slok argued AI has delivered little financial payoff outside big tech.
The balance sheet was a leaking bucket, and the AI pitch became a siren song — seductive but risky if the economics don’t match the hype.
On the ground at Google products you can see both the promise and the problem.
Search queries driven by AI are the strongest in years, Ashkenazi said, and there’s clear demand for AI subscription offerings. But higher query growth with “lower monetization” creates a mismatch: more user activity, less ad revenue per query. That’s what’s suppressing free cash flow even as usage climbs.
Will Google’s AI investments pay off?
Management insists yes — they expect an ROIC above cost of capital and eventual strong operating cash flow. I believe the wager is defensible: Google has the data, cloud, TPUs, and Search distribution few firms match. But the timeline matters: you’re buying a promise that may take years to repay, and the market is impatient.
In the broader market you can see the same strain: big bets, thin near-term returns.
Apollo’s chief economist Torsten Slok wrote that AI has delivered little financial payoff across much of corporate America. That criticism matters: if adopters outside tech aren’t seeing revenue bumps, the narrative of AI as an immediate growth engine weakens, and capital markets may punish companies front-loading investment.
I’m not telling you to panic, but I am asking you to widen your view: look at cash flow trends, capex cadence, and monetization strategies rather than headlines about model rankings and internal productivity gains. Google still controls powerful assets — Search, Android, Cloud, TPUs, and the Gemini family — but control is not the same as instant profit.
You can track the signal in three places: capex guidance, Search monetization metrics, and TPU/inventory disclosures. Watch how Google tightens ad density in AI Mode and whether Gemini Carbon is used to lift per-query revenue. If those levers don’t move soon, the market may demand clearer evidence of payback — and faster.
If Google’s AI promises a golden year internally, can it convince the market before cash flow becomes the headline?