News
Nov 18, 2025
News
Enterprise
Scaleups
Artificial Intelligence
Global
NewDecoded
4 min read
Microsoft and NVIDIA are investing up to $15 billion combined in Anthropic under a landmark partnership announced Tuesday. NVIDIA will commit up to $10 billion and Microsoft up to $5 billion, while Anthropic has committed to purchase $30 billion of Azure compute capacity. The Claude AI developer will also contract up to one gigawatt of compute capacity, which industry executives estimate could cost between $20 billion and $25 billion. The partnerships make Claude the only frontier model available on all three major cloud platforms (Microsoft Azure, Amazon Web Services, and Google Cloud). Azure customers will gain access to Anthropic's latest models including Claude Sonnet 4.5, Claude Opus 4.1, and Claude Haiku 4.5 through Microsoft Foundry. Additionally, Claude will remain integrated across Microsoft's Copilot family, including GitHub Copilot, Microsoft 365 Copilot, and Copilot Studio.
For the first time, NVIDIA and Anthropic are establishing a deep technology partnership focused on optimization. The companies will collaborate on design and engineering to optimize Anthropic models for peak performance, efficiency, and total cost of ownership. Simultaneously, NVIDIA will optimize future architectures specifically for Anthropic workloads. Anthropic's initial compute commitment will utilize NVIDIA's Grace Blackwell and Vera Rubin systems.
The announcement has intensified scrutiny around circular investment patterns in AI. Investors are increasingly uneasy that the AI boom has outrun fundamentals, with business leaders noting that circular deals where one partner props up another's revenue add to bubble risks. This follows similar concerns raised after NVIDIA's $100 billion investment in OpenAI, where the chipmaker invested in a customer that then committed to purchasing its products. Circular investments may give investors an inflated perception of true demand for AI infrastructure. Recent analysis shows that while tech companies have invested approximately $560 billion in AI infrastructure over two years, they've generated just $35 billion in combined AI-related revenue. An MIT report found that 95% of organizations are getting zero return despite enterprise investment reaching $30-40 billion.
The deal reflects how the AI industry is consolidating around a few key players. D.A. Davidson analyst Gil Luria noted that the partnership reduces the AI economy's reliance on OpenAI, with both Microsoft and NVIDIA now diversifying their frontier model dependencies. Anthropic, founded in 2021 by former OpenAI staff, was recently valued at $183 billion and is projecting to potentially triple its annualized revenue run rate to around $26 billion next year.
This announcement represents more than just another AI partnership. It exemplifies a troubling pattern where chip manufacturers invest billions in AI companies, which then commit to spending even more on those same chips and cloud services. The structure creates a self-reinforcing cycle that may be artificially inflating valuations without corresponding end-user demand.
With valuations stretched to levels not seen since the dot-com bubble and 30% of the S&P 500 concentrated in just five companies, investors face a critical question: Are these deals building sustainable AI infrastructure, or are they financial engineering masking fundamental weaknesses in the business models? The coming quarters will reveal whether the massive spending translates into real revenue, or whether AI infrastructure follows the path of 1990s fiber-optic cables, sitting underutilized while markets await demand to catch up with supply.