Breaking: Chinese open source models challenge Anthropic IPO as Hugging Face deletions rise

Anthropic is preparing 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 core question for investors and everyday users alike is how a company that loses so much money can be worth such a colossal sum. The answer lies in the control of compute and the looming threat from open source competitors that are currently flooding the market with advanced capabilities.

The primary threat to Anthropic's dominance comes from the open source sector, specifically from advanced Chinese models. Companies such as Moonshot AI, which created Kimi, Alibaba, which developed Qwen, and Xiaomi are releasing powerful tools to the public. These firms are willing to give their models over to the public to increase innovation and market penetration. Kimi K3 has started to outperform the Opus 5 models from Anthropic, and it was given to the public for free. This direct competition forces Anthropic to justify its high price point and proprietary lock-in to its new shareholders.

The internet has responded to these free models by quantizing them down to run on smaller machines. Users have made these large language models usable on hardware like Mac Studios, A100 cards, and RTX Nvidia models. This trend is significant because it reduces the need for massive, centralized data centers. A single Nvidia RTX Pro 6000 graphics card, which retails at 20,000 dollars, can hold about 15 to 20 light users or 1 to 2 heavy coding users. This efficiency threatens the GPU bottleneck that currently favors large players like Anthropic, who have secured compute commitments of up to 900 billion dollars from companies including Amazon and Google.

Nvidia has moved to secure its position in this shifting landscape by agreeing to acquire Hugging Face for 12.9 billion dollars. Hugging Face is the platform where most open AI models are distributed, and the acquisition price is about 70 times its peak annual revenue of 150 to 200 million dollars. While the stated reason for the deal was CUDA, the software that runs these models, the deeper motivation was control of the distribution of models. This move allows Nvidia to manage the flow of open source code that directly competes with the proprietary models sold by its major customers.

Following the acquisition, advanced models were deleted from the Hugging Face platform. Anything over a quant 4, meaning models advanced enough to beat 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 this censorship and corporate overreach. In response, they opened Pirate Face, a new place where users can upload and download all models. However, this migration was not without cost, as the large 400 gigabyte models at quant 3 and higher were scrubbed from the network before the new platform fully launched.

The fight over control has also moved to Washington, D.C. Anthropic demanded that Congress ban Chinese models citing national security concerns, a move that many see as an effort to control the market. The company recently released media suggesting AI will take over the world, a stunt timed with the departure of disgruntled employee Jacob Coxon. This narrative helped push Congress to pass 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 from launching their own AI models in house using these advanced open source tools.

The pressure on Anthropic has been visible in its product releases this year. The company released Fable 5 and Mythos 5 on June 9, 2026. Just three days later, on June 12, 2026, Anthropic suspended access to these models worldwide. This action followed an export control directive from the U.S. Department of Commerce. The models were solving too many problems, leading Anthropic to pull them off the shelf. This sequence of events suggests that the goal was not just winning the AI race, but controlling it by managing the pace of public access.

User experiences with Anthropic have also become contentious in recent months. The company has been suspending users for reasons such as abuse, potential fraud, and usage violations. The appeal process is often described as useless because it is fully run by bots, and accounts are not reactivated once kicked off. Critics argue that once Anthropic goes public, these suspensions will multiply by 100 or even 1000 times. With deep pockets and access to unlimited funds, the end user may become less important to the company, which does not make money off of daily users or heavy users in the current structure.

There is growing concern about how Anthropic will monetize its user base in the future. The company is expected to create programs where user ideas and data will belong to Anthropic. If the company sees great ideas, it gets notified, and it may suspend users for unproven reasons to steal their intellectual property. This dynamic creates a risk for anyone building on top of Claude, as the company hosting their work could claim a share of the value when that work becomes successful. The uncertainty of these rules is a major point of discussion among developers and businesses relying on the platform.