Google Burned Through More Cash Than It Made Last Quarter. Thanks, AI.
Google's Q2 2026 numbers are out, and they tell two very different stories simultaneously. On one hand, the company pulled in $119.8 billion in revenue, comfortably beating analyst forecasts. On the other, its stock dropped roughly 4.5 percent overnight and kept sliding. The reason: Google is now spending faster than it earns, and for the first time since going public, it has posted negative free cash flow.
The revenue breakdown is roughly what you'd expect. Search remains the biggest earner at $63.3 billion. Google Cloud brought in $24.8 billion, up nearly 24 percent from the previous quarter, which suggests genuine demand for its AI services rather than just hype. YouTube ads generated $11.1 billion, a 12-plus percent jump quarter-on-quarter, partly because Google has been making those ads longer, which is the kind of thing that works until users get annoyed enough to care. Subscriptions, platforms and devices added another $12.9 billion.
Strip out non-cash investment income and Google's operating cash flow for Q2 sits at around $39.1 billion. That's a healthy 40 percent improvement on Q2 2025. The problem is what's sitting on the other side of the ledger.
Earlier this year, Google told investors to expect capital expenditure somewhere between $180 billion and $190 billion for 2026, already roughly double the $91 billion it spent in 2025. That number has since been revised upward to $205 billion. In Q2 alone, the company spent $44.9 billion building out AI infrastructure. With $39.1 billion coming in and $44.9 billion going out, the resulting free cash flow was -$5.8 billion.
Free cash flow matters because it reflects what a business actually has available to operate with, without selling assets or borrowing. Google slipping into negative territory on this metric is not a crisis, the company still holds over $100 billion in reserves and remains extraordinarily profitable overall. But it is a signal worth paying attention to.
Google's leadership is framing all of this as intentional and temporary. The company notes its capex is now roughly six times what it spent in 2022, before the AI spending race properly kicked off, and has flagged that expenditure will likely climb further in 2027. The message to investors is essentially: trust the plan.
Investors are not entirely convinced. AI spending across the tech industry is expected to exceed $700 billion this year, and the returns remain difficult to quantify. Google is better positioned than most to absorb this kind of outlay, with its advertising business and cloud revenue providing a solid base. It also designs its own AI chips, with the latest Tensor 8i and 8t reportedly offering improved efficiency for data centre workloads. But efficiency gains only help so much when the overall scale of spending keeps expanding.
There are other pressures building. Google reportedly delayed the release of its Gemini 3.5 Pro model, which is still in limited testing with a small group of partners. Internally, there have been concerns that the model is not delivering the performance gains needed to stay competitive with OpenAI's GPT 5.6 or Anthropic's Claude Mythos. A wave of departures among senior AI researchers has not helped matters.
The next few months will be telling. Google has the financial cushion to sustain this level of spending, but it cannot sustain it indefinitely without something to show for it.