Breaking: Anthropic IPO Valuation Hints at Massive Gaps in AI Business Model

Anthropic is preparing to go public in an initial public offering that projects a valuation between 2.5 trillion dollars and 3 trillion dollars. This financial leap places the maker of the Claude models at the center of a contentious debate regarding who controls the future of artificial intelligence. The company faces a significant challenge in justifying such a high price tag given its current financial performance and the massive obligations it carries to its partners.

The core of the financial question lies in the stark gap between the projected valuation and the company's actual earnings. Anthropic makes less than 20 billion dollars a year in revenue while running a deficit of about 500 billion dollars. Investors are being asked to believe that a company losing 500 billion dollars is worth 2.5 to 3 trillion dollars. This discrepancy raises urgent questions about the sustainability of the business model and the expectations of the market for future growth.

A major driver of these costs is the company's commitment to computing power. Anthropic carries compute commitments to other companies, including Amazon and Google, of up to 900 billion dollars. These agreements lock the company into massive expenses that far outpace its current income. For everyday users and businesses relying on Claude, this means the infrastructure behind their subscriptions is supported by a web of financial promises that are difficult to reconcile with the current bottom line.

The economics of running these models are complex and heavily dependent on hardware efficiency. A single Nvidia RTX Pro 6000 graphics 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. These estimates assume constant use 24 hours a day, a scenario that few individual users actually achieve in their daily workflows.

Pricing for these services varies significantly based on usage intensity. 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. The hardware required to support these users is expensive, with the RTX Pro 6000 retailing at 20,000 dollars as of now. This price does not include the servers, switches, CPUs and processors needed to run the machines, the data center facility costs, or the electricity required to keep them running.

When buying cards from Nvidia direct at a wholesale price of 3,000 to 4,000 dollars, the cost recovery timeline remains long. The GPU alone takes about 10 months of a heavy 200 dollar a month user to pay back, and about the same for the 15 to 20 light users at 20 dollars a month. While most light users will not use the chat for more than a few hours a day, which can stretch the bandwidth number by 10 times, the breakeven and acquisition cost stays huge for the provider.

This compute and GPU bottleneck is a real problem for these companies, and Nvidia knows it. Nvidia drives demand up and lowers supply, giving all its supply to the large players like Anthropic, Grok and OpenAI. The consumer market is left fitting the bill for these constraints. To solidify its position, 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, so the 12.9 billion dollar price is about 70 times its revenue. The stated reason for the acquisition was CUDA, the software that runs all these models, but the deeper reason was control of the distribution of models on the market. This move was particularly aimed at the open source market, which has been flooded by advanced Chinese models like Kimi from Moonshot AI, DeepSeek, and Qwen from Alibaba.

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 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. This happened 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, when the real goal was to control the market. The panic worked, and Congress passed a bill that now bans any government and its affiliates from using any Chinese models. This legislative action followed a period where Anthropic released Fable 5 and Mythos 5 on June 9, 2026. The U.S. Department of Commerce issued an export control directive, and Anthropic suspended access to them worldwide on June 12, 2026.

In recent months Anthropic has been suspending users with reasons like abuse, potential fraud, and usage violations. The appeal process is fully run by bots, and as soon as they kick you off they will not reactivate the account. Once Anthropic goes public, the suspensions will multiply because it will have deep pockets and access to unlimited funds. Dario Amodei, the CEO of Anthropic, said in an interview that some ideas are worth billions, and that Anthropic will likely have some sort of royalty or partnership program.

Under this proposed plan, when a person or company finds the cure for cancer, some of that accreditation and funding will go to Anthropic. This should scare anyone who builds on top of Claude, as the company that hosts your work may claim a share of the prize when your work becomes worth billions. The company is trying to pass bills that would stop other companies like Verizon from launching their own AI models in house by using these advanced open source models.