Alphabet Q2: 2026 capex rises to $205B and free cash flow turns negative

In plain terms

Alphabet has just done what Wall Street both fears and anticipates: raising its 2026 investment bill for AI again. New capex guidance: around $205 billion. And this time, the counterpart is visible in the cash flow statement - the free cash flow for the quarter is negative, at -$5.9 billion.

Context

Since the beginning of 2026, the question is no longer "how much is Alphabet spending on AI?" but "to what extent can its cash flow mechanism absorb these expenses without the market turning around?" The Q2 2026 report shifts the dial. The capex guidance, regularly revised upwards since the initial announcement, now reaches an order of magnitude of $205B for the year.

The data

• Capex 2026 (new guidance): ~$205B (Tech in Asia).
• Q2 2026 revenue: ~177 trillion wons (~$119.8B) reported by ETNews.
• Quarterly free cash flow: -$5.9B (Tech in Asia) - a notable reversal for a usually massive cash-generating profile.

Analysis

The move is read at three levels. Compute: Alphabet is doubling down on its own TPUs and third-party GPUs to keep up with Gemini and meet cloud demand. Cloud: the commercial pipeline requires delivered capacity, not promised. Financial signal: a quarter of negative FCF at -$5.9B is not a disaster for a company with massive cash reserves - but it's a narrative pivot. Alphabet's historical position, "we can afford it," changes status: it's no longer a given, it's a choice backed by internal credit.

Scenarios

• Base (most likely): capex remains above $200B in 2027, with quarterly FCF inflection according to delivery schedule.
• High: new upward revision if Gemini enterprise takes off and Google Cloud wins heavy inference workloads.
• Low: forced rationalization if capex returns are delayed.

Risks

Dependence on TSMC/HBM availability, exposure to a tightening of exports to China, and above all: investor validation that volatile free cash flow is an acceptable mode.

Under the hood

Unlike Meta (solid advertising margins), Alphabet is paying for its AI transition on two fronts: consumer products (Search + Gemini), which consumes compute without obvious marginal revenue, and commercial cloud, which generates it but with delays.

So what

For executives building their 2027 AI budget: Alphabet's message is simple. AI infrastructure is expensive, for a long time, and even hyperscalers no longer seem to digest it without a wince. Expect a price impact on managed services - and a tightening of multi-year pre-payment/commitment conditions.