AI's Next Phase: Infrastructure Control Drives Anthropic, OpenAI, and Nvidia

With Anthropic's valuation nearing the "high hundreds of billions" and Nvidia climbing "beyond a $5 trillion market cap", Wall Street analysts have "revised expectations upward" for the next phase of AI adoption. This era is less about model breakthroughs and more about who controls the underlying infrastructure, even as apps like ChatGPT, Claude, Gemini, Perplexity, Microsoft Copilot, Meta AI, Grok, and DeepSeek become everyday digital tools.

The shift in focus means that command over computing power is now paramount, rather than solely the innovation of AI models. Anthropic and OpenAI are at the forefront of building leading AI models, while Nvidia dictates the pace for the world's AI hardware.

Three key leaders—Dario Amodei at Anthropic, Sam Altman at OpenAI, and Jensen Huang at Nvidia—are increasingly shaping the sector. Their differing stances on computing power, supply chains, and global partnerships create strategic tension, notably the public clash between Amodei and Huang regarding China chip policy.

"If you're trying to figure out where the AI money is actually going, this is a good place to start," Markman writes.

Under CEO Dario Amodei, Anthropic has positioned Claude as a stability- and safety-focused system, an approach that has propelled its valuation into the "high hundreds of billions". Surging demand for its enterprise AI tools makes a "potential IPO" as soon as October "increasingly likely," according to Forbes markets contributor Peter Cohan.

Anthropic's financial disclosures reveal annualized revenue at "$1.4 billion", with "more than 500 customers" spending at least "$1 million" a year, Forbes contributor Jon Markman reports. This growth fueled a "$30 billion Series G round", which stands as the second-largest private raise in tech, only behind OpenAI's "$40 billion" last year.

OpenAI remains the benchmark against which many AI competitors are measured. Under Sam Altman, the company pursues a two-track strategy: rapid release of new GPT versions and deep integration of its technology into Microsoft’s cloud. This partnership grants OpenAI "privileged access to Azure’s large AI data centers", linking its expansion to Microsoft’s spending and regulatory exposure.

However, OpenAI faces structural pressures similar to Anthropic. Access to high-performance hardware remains uneven, training costs continue to climb, and governments impose stricter rules on advanced chip deployment. Markman notes that Microsoft ended Claude Code licenses after internal usage costs "ran past the annual AI budget months ahead of schedule", highlighting economic constraints even for top-tier models.

Jensen Huang, representing Nvidia, embodies the supply side of AI. He has "sharply criticized U.S. export controls", arguing they "gave Chinese companies the spirit, the energy, and the government support to accelerate their development", Forbes contributor Drew Bernstein writes. This tension significantly influences Nvidia’s global strategy.

While export rules have altered where Nvidia can sell its most advanced chips, they have also spurred a rush of domestic buying. Cloud providers, government-backed AI programs, and large enterprises are scrambling to secure capacity. Nvidia has accelerated new chip releases and strengthened joint development efforts with AI labs and cloud providers, solidifying its central role in the AI hardware stack.

Beyond these three leaders, other companies play vital roles in the AI economy. Google DeepMind advances its Gemini AI roadmap, positioning Google as both a rival and a major infrastructure provider. Meta continues to release powerful open-source AI models, increasing competitive pressure. AMD gains traction with its MI300 AI accelerators, while TSMC remains a critical manufacturer, and Broadcom provides essential networking technology.

Companies building advanced AI systems are increasingly reliant on chipmakers for computing power. Bernstein captures this dynamic in the clash between Amodei and Huang over export controls, calling it "the most important industrial policy question of the decade." Anthropic's growth fuels demand for Nvidia's chips, and Nvidia's architecture shapes how quickly model companies can scale; for instance, Anthropic secured "more than 220,000 Nvidia GPUs through SpaceX", as Markman highlights.

Wall Street analysts are closely monitoring several catalysts. Anthropic’s expected IPO filing will offer the market its first detailed look at the costs of building cutting-edge AI models. OpenAI’s next major model and its pricing strategy could shift the industry, and the spending by Microsoft, Amazon, Google, and Oracle on data center hardware, which fluctuates with Nvidia’s upgrades, serves as a key real-time indicator of AI demand.