Alphabet jumps as Anthropic expands multi-gigawatt Google TPU deal for 2027
Alphabet shares rose after news that Anthropic expanded its Google TPU compute commitment to about 3.5 gigawatts starting in 2027, deepening demand for Google Cloud’s AI infrastructure. The deal, alongside a long-term Broadcom TPU co-design agreement extending through 2031, reinforced expectations for accelerating AI-driven cloud revenue growth.
1. What’s driving the move
Alphabet (GOOGL) moved higher as investors reacted to signs of rising demand for Google’s in-house AI accelerators (TPUs) and related cloud infrastructure. The key catalyst was Anthropic expanding its partnership to secure roughly 3.5 gigawatts of next-generation Google TPU capacity expected to come online from 2027, a scale-up that signals materially higher long-duration AI compute consumption routed through Google’s ecosystem.
2. Why the market cares
The announcement strengthens the narrative that Alphabet is not only competing in AI software, but also building a differentiated AI infrastructure stack—custom silicon plus cloud capacity—capable of supporting frontier-model customers at very large scale. It also adds visibility to Google Cloud’s AI workload pipeline, which is increasingly viewed as a primary driver of incremental growth and operating leverage versus the more mature advertising business.
3. The supply chain angle (Broadcom) and what it implies
The TPU expansion also highlighted Alphabet’s silicon roadmap and supply assurance, with Broadcom disclosed as a long-term partner for future generations of Google TPUs through 2031. For markets, that combination—multi-year silicon development plus a mega-customer compute commitment—reduces uncertainty around Alphabet’s ability to deliver AI capacity at the pace needed and supports the thesis that TPUs can be a durable cost/performance advantage as inference demand scales.
4. What to watch next
Investors will watch for any follow-on disclosures tying the expanded TPU capacity to Google Cloud backlog, AI services pricing, or margin trajectory, as well as any incremental capex signals required to support multi-gigawatt deployments. Near-term sentiment may also be influenced by upcoming quarterly results and management commentary on AI infrastructure utilization, cloud growth, and monetization progress.