Alphabet bumps annual capex to $205B to bridge AI supply gap
Alphabet reported Q2 revenue of $119.8 billion and raised its annual capital expenditure forecast to $195B-$205B to fund AI and cloud infrastructure growth. The company also announced plans to shift more computing demand to third-party cloud capacity while internal capacity is constrained.
Key Takeaways
- Google Cloud revenue jumped 64% to $22.4 billion, reaching a total order backlog of $514 billion
- Annual capital expenditure guidance increased to a range of $195 billion to $205 billion
- Alphabet is renting GPU capacity from SpaceX for $920 million per month to bridge immediate supply shortfalls
- Gemini app monthly active users reached 950 million with a processing rate of 22 billion tokens per minute
Why It Matters
Alphabet's pivot to renting third-party cloud capacity signals that infrastructure timing, not just capital, has become the primary bottleneck for AI market leaders. By utilizing external providers like SpaceX and CoreWeave, Google is choosing short-term margin pressure over the risk of losing enterprise cloud market share during its internal build-out. For the streaming and video ecosystem, this massive infrastructure spend will eventually lower the cost of high-density compute, but in the near term, it confirms a supply crunch that may limit the roll-out of compute-intensive features. Watch for whether Google Cloud's 64% growth rate can outpace the rising cost of capital in a higher-rate environment.
Additional Context
The $920 million monthly agreement between Google and SpaceX, first reported in June 2026, involves the lease of approximately 110,000 Nvidia GPUs housed in SpaceX’s Memphis data center complex. Per TechCrunch, the deal includes specific termination rights if SpaceX fails to deliver full GPU allotments by September 30, 2026, underscoring the urgency of Google’s capacity needs. This shift toward 'neocloud' providers follows a similar pattern at Microsoft, which has secured bridge capacity from specialized firms like CoreWeave to sustain its own AI services during peak demand. While hardware spend is peaking, Alphabet is also moving to commercialize its internal efficiency gains. Per IG.com, Alphabet began selling capacity on its custom Tensor Processing Units (TPUs) to external cloud providers in July 2026. This move is part of a broader strategy to transform internal cost-saving technology into a secondary revenue stream. Analysts from Bloomberg noted in July 2026 that these custom chips could significantly offset infrastructure costs once hardware sales begin contributing more meaningfully to the bottom line in late 2026 and 2027. Simultaneously, Google is expanding its AI distribution through strategic partnerships. Per a July 2026 report from The Information, Apple has finalized a deal to use Gemini as the backend for its next-generation Siri features, potentially granting Google default AI presence on over 1.4 billion iPhones. These distribution wins contrast with the current financial pressure; Alphabet's free cash flow dived to negative $5.9 billion in Q2 2026, down from a $10.1 billion surplus in the previous quarter, as the company prioritizes infrastructure build-outs over short-term liquidity, according to data from Fierce Network.
Read full article at siliconangle.com
Get this in your inbox → Subscribe
Enjoy our coverage?
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source