OpenAI and Anthropic, once hailed as the future of AI, are now facing significant market doubts. Some experts propose nationalization to ensure public benefit.
TL;DR
- OpenAI and Anthropic's trillion-dollar valuations are under scrutiny as market doubts grow.
- Experts propose nationalization to ensure these AI labs serve the public interest.
- The economic viability of these companies is questioned due to high costs and competition.
What happened
OpenAI and Anthropic, founded by AI developers concerned about corporate AI development, have faced market doubts recently. Their IPO filings in June sparked buzz about trillion-dollar valuations, but recent headlines highlight public backlash to AI datacenters and Nvidia's slumping stock. SpaceX's stock price also tanked shortly after its IPO, raising questions about the profitability of leading AI labs.
The market is reassessing the financial value of these companies. Evidence suggests that the economics of big AI labs may not guarantee a booming return on investment. Frontier AI models are expensive to train and depreciate quickly, and enterprise clients are minimizing AI token usage. Additionally, open-source and Chinese competitors are offering similar models for free, further depressing prices.
Why it matters
If the market deems these companies as not financially viable, they may collapse. However, their products and innovations are valuable to society. Nationalization could ensure these AI labs serve the public interest, similar to other national labs that have produced world-shanging innovations.
Transitioning these companies into public agencies would require restructuring. The innovation function could be publicly managed with congressional oversight, and AI operations could be managed as a commodity resource, similar to public utilities. This could reduce the overall resource and environmental cost associated with AI and align the models with democratic values.
Key facts
- OpenAI and Anthropic filed for IPOs in June with buzz about trillion-dollar valuations.
- Nvidia's stock has been slumping, and SpaceX's stock price tanked shortly after its IPO.
- Frontier AI models are expensive to train and depreciate within months.
- Enterprise clients are minimizing AI token usage to reduce costs.
- Open-source and Chinese competitors are offering similar models for free.
- The US has a history of managing national, regional, and state supercomputing centers.
- Other countries, including Switzerland, Spain, and Singapore, are already operating public AI labs.
- Congress manages a $200bn R&D portfolio, which could include frontier AI development.
Context
OpenAI and Anthropic were founded by AI developers who feared unrestrained corporate AI development. Their founders proclaimed that their new labs could be trusted to develop the technology in humanity's best interest. However, both companies have since become corporate behemoths, raising questions about their commitment to the public interest.
The economic viability of these companies is questioned due to high costs and competition. Frontier AI models are expensive to train and depreciate quickly, and enterprise clients are minimizing AI token usage. Additionally, open-source and Chinese competitors are offering similar models for free, further depressing prices.
Nationalization could ensure these AI labs serve the public interest. The US has a history of managing national, regional, and state supercomputing centers, and other countries are already operating public AI labs. Transitioning these companies into public agencies would require restructuring, but the benefits to the public are clear.
