Anthropic has announced its largest infrastructure commitment to date, securing approximately 3.5 gigawatts of next-generation Google tensor processing unit (TPU) compute capacity through Broadcom beginning in 2027. This substantial agreement comes alongside the revelation that the AI company's revenue run rate has soared past $30 billion, representing more than a threefold increase from approximately $9 billion at the end of 2025. The deal builds upon the 1 gigawatt of capacity already being supplied to Anthropic in 2026, with the majority of the new infrastructure to be located in the United States.
The announcement represents a three-way partnership that positions Broadcom as a critical intermediary between Google's custom silicon and Anthropic's AI workloads. Broadcom has simultaneously signed a separate long-term agreement with Google to design and supply future generations of custom TPU chips, alongside a supply assurance agreement extending through 2031. This strategic positioning has made Broadcom increasingly indispensable in the AI infrastructure landscape, with analysts estimating the company could record $21 billion in AI revenue from Anthropic in 2026 alone, potentially rising to $42 billion in 2027. Broadcom shares rose approximately 3% in extended trading following the announcement.
Anthropic's extraordinary revenue growth is driving these massive infrastructure commitments. The company now serves over 1,000 business customers each spending more than $1 million annually, doubling from 500 customers just two months earlier. This acceleration followed Anthropic's Series G funding round in February 2026, which raised $30 billion at a post-money valuation of $380 billion. What distinguishes Anthropic's approach is its multi-vendor chip strategy, with Claude trained and served across Amazon's Trainium chips, Google's TPUs, and Nvidia GPUs, making it the only frontier model available across AWS, Google Cloud, and Microsoft Azure.
The deal extends Anthropic's November 2025 commitment to invest $50 billion in American AI computing infrastructure, aligning with the strategic priorities outlined in the Trump administration's AI Action Plan. This domestic emphasis reflects both strategic positioning and regulatory considerations, as a substantial portion of next-generation AI training capacity becomes anchored in American geography. The announcement represents another data point in an escalating compute arms race, where AI labs' infrastructure requirements now exceed what can be financed from revenue alone, requiring financial engineering at a scale once reserved for infrastructure utilities.
Fuente Original: https://thenextweb.com/news/anthropic-google-broadcom-compute-deal
Artículos relacionados de LaRebelión:
- Google Lanza Gemma 4 IA Abierta y Licencia Apache 20
- Gemma 4 de Google IA Potente Licencia Libre
- Google Lanza Gemma 4 Modelos Abiertos Revolucionarios
- Microsoft Lanza IA Propia Desafio Directo a OpenAI y Google
- Google Accelerates Post-Quantum Encryption Migration to 2029
Artículo generado mediante LaRebelionBOT
No hay comentarios:
Publicar un comentario