Breaking: Nvidia Skews AI Supply Chain Ahead of Anthropic’s Massive IPO

Anthropic is preparing to take its Claude models public in an initial offering that projects a valuation between 2.5 trillion dollars and 3 trillion dollars. This massive figure stands in stark contrast to the company’s current financial reality, which includes a deficit of about 500 billion dollars and annual revenue of less than 20 billion dollars. The core question facing investors and the public is how a company losing 500 billion dollars can command a value in the trillions. The answer lies heavily in the compute commitments Anthropic has made to other companies, including Amazon and Google, which total up to 900 billion dollars. These commitments secure the infrastructure needed to run the models, but they also highlight the dependency on hardware providers.

Nvidia holds a dominant position in this hardware supply chain, and its recent moves have skewed the market heavily in favor of large players. The company drives demand up while simultaneously lowering the available supply for the general market. By giving its primary supply to large entities like Anthropic, Grok, and OpenAI, Nvidia leaves the consumer market to fit the bill. This strategy ensures that the major AI labs have the necessary resources to scale their operations, while individual users and smaller competitors face higher costs and longer wait times for access to high performance graphics cards.

The economics of this disparity are visible in the cost of a single Nvidia RTX Pro 6000 graphics card. This card retails at 20,000 dollars as of now, a price that does not include the servers, switches, CPUs, and processors needed to run the machines, nor the data center facility costs or electricity. A single card can hold about 15 to 20 light users running light compute at about 60 to 70 tokens per second. Alternatively, it can support 1 to 2 heavy coding users running Claude Code at about 40 to 50 tokens per second, assuming constant use 24 hours a day. Each heavy user on the max plan pays 200 dollars a month, while each light user pays about 20 dollars a month on average for a base plan.

When buying the cards from Nvidia direct at a wholesale price of 3,000 to 4,000 dollars, the GPU alone takes about 10 months of revenue from a heavy user to pay back. The same holds true for the 15 to 20 light users at 20 dollars a month. Most light users will not use the chat for more than a few hours a day, which means the bandwidth number can stretch 10 times, but the breakeven and acquisition cost remains huge. The compute and GPU bottleneck is a real problem for these companies, and Nvidia knows it. By controlling the flow of these cards, the chip giant effectively dictates the pace and scale of the AI industry.

Nvidia recently agreed to acquire Hugging Face, the platform where most open AI models are distributed, for 12.9 billion dollars. Hugging Face makes about 150 to 200 million dollars a year in its peak years, making the 12.9 billion dollar price about 70 times its revenue. While 12.9 billion dollars is pocket change for Nvidia and its CEO Jensen Huang, the acquisition has significant implications for the market. The stated reason for the deal was CUDA, the software that runs all these models, but the deeper reason was control of the distribution of models on the market, especially the open source market.

The open source market had been flooded by advanced Chinese models, including Kimi from Moonshot AI, DeepSeek, Qwen from Alibaba, and models from Xiaomi. These companies were willing to give the models over to the public to increase innovation. Kimi K3 started to outperform the Opus 5 models from Anthropic, and it was given to the public. The internet used these large language models, quantized them down, and made them usable on smaller machines like Mac Studios, A100 cards, and RTX Nvidia models. This trend threatened the proprietary models of major US companies by offering comparable performance at a lower cost.

After Nvidia acquired Hugging Face, advanced models were deleted from distribution. Anything over a quant 4, meaning advanced and capable of beating out Claude's models, was removed. The 4 bit, 6 bit, 8 bit, and 16 bit full sized models were banned from the Hugging Face platform. Developers on Discord rallied to stop the censorship and corporate overreach, and they opened Pirate Face, a place where you can upload and download all models. However, this was not before the large 400 gigabyte models at quant 3 and higher were scrubbed from the network. This move solidified Nvidia's grip on the supply of high performance AI tools.

Anthropic has also been involved in a broader effort to control the market through legislative means. The company demanded that Congress ban Chinese models because of national security, when the real goal was to control the market. Anthropic recently did a publicity stunt by releasing and circulating media that AI will take over the world when one of its disgruntled employees, Jacob Coxon, quit amidst his disappointment in AI. The goal was to get the government to pass bills banning models that were a huge threat to Anthropic's bottom line. The panic worked, and Congress passed a bill that now bans any government and its affiliates from using any Chinese models.

The recent suspension of models further illustrates this strategy of control. Fable 5 and Mythos 5 were released on June 9, 2026, but Anthropic suspended access to them worldwide on June 12, 2026, after the U.S. Department of Commerce issued an export control directive. The models were solving too many problems, and Anthropic pulled them off the shelf. This was not about winning the AI race; it was about controlling the AI race. By removing these powerful tools from the market, Anthropic maintained a level of advantage that justified its high valuation despite its massive deficits.