Breaking: Nvidia’s Supply Grip Shapes Anthropic’s IPO and AI Control

Anthropic is preparing to file for an initial public offering that projects a valuation between 2.5 trillion dollars and 3 trillion dollars. This valuation 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 central question for investors and observers is how a company losing hundreds of billions of dollars annually can justify a valuation in the trillions. The answer lies heavily in the control of compute resources, a market dominated by Nvidia.

Nvidia holds a decisive grip on the supply of the graphics processing units required to run large language models. The chip maker drives demand up while lowering supply, effectively giving the cards to major players like Anthropic, Grok, and OpenAI. This strategy leaves the consumer market to fit the bill for the infrastructure. As of now, a single Nvidia RTX Pro 6000 graphics card retails at 20,000 dollars. This price tag does not include the servers, switches, CPUs, processors, data center facility costs, or the electricity needed to keep the machines running.

For the everyday user, the economics of accessing these models are becoming increasingly difficult to understand. A single RTX Pro 6000 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. Heavy users on the max plan pay 200 dollars a month, while light users pay 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 a heavy user’s payments to pay back. The same timeframe applies for the group of light users.

The bottleneck in compute and GPU availability is a significant problem for AI companies, and Nvidia is aware of this leverage. By prioritizing large institutional clients, Nvidia ensures that the major players have the resources to scale their models. The consumer market is left with higher costs and fewer options. This dynamic supports Anthropic’s valuation narrative, as the company secures its position at the top of the supply chain. The compute commitments Anthropic carries to other companies, including Amazon and Google, amount to up to 900 billion dollars.

The fight over who controls AI is not just about hardware. It is also about the distribution of models. Nvidia 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. The stated reason for the acquisition was CUDA, the software that runs these models, but the deeper reason was control over the distribution of models, especially in the open source market.

The open source market has been flooded by advanced Chinese models, including Kimi from Moonshot AI, DeepSeek, Qwen from Alibaba, and models from Xiaomi. These companies are willing to give their models over to the public to increase innovation. Kimi K3 started to outperform the Opus 5 models from Anthropic and was given to the public. The internet community 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 premium pricing structure of closed models.

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 this perceived censorship and corporate overreach. They opened Pirate Face, a place where you can upload and download all models, but not before the large 400 gigabyte models at quant 3 and higher were scrubbed from the network.

Anthropic demanded that Congress ban Chinese models because of national security, although the real goal was to control the market. The company recently executed a publicity stunt, releasing and circulating media that AI will take over the world. This coincided with the departure of disgruntled employee Jacob Coxon, who quit amidst his disappointment in AI. The goal was to get the government to pass bills banning models that posed a 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.

Anthropic is also trying to pass bills that would stop other companies like Verizon and other networks from launching their own AI models in house by using these advanced open source models. In June 2026, Anthropic released Fable 5. The government forced them to take it down. Fable 5 and Mythos 5 were released on June 9, 2026, and 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 action was less about winning the AI race and more about controlling it.

In recent months, Anthropic has been suspending users with reasons such as abuse, potential fraud, and usage violations. The appeal process is often viewed as useless because it is fully run by bots, and accounts are not reactivated once users are kicked off. Once Anthropic goes public, these suspensions are expected to multiply significantly because the company will have deep pockets and access to unlimited funds. The end user will not be as important to the company as the shareholders. Anthropic does not make money off of daily users, and it certainly does not make money off of heavy users.