The head of one of the largest US endowments says OpenAI and Anthropic are in big trouble (www.businessinsider.com)

Scott Wilson, head of one of the largest US endowments, has raised alarm bells about the financial sustainability of AI giants OpenAI and Anthropic. At a recent tech gathering, he expressed skepticism about their hefty investments, suggesting that they are becoming vulnerable to cheaper, high-performance alternatives from Chinese companies like DeepSeek and Alibaba. Wilson believes that the unsustainable spending by these frontier labs could lead to a significant downfall, as many companies are opting for cost-effective, open-source models over pricier offerings from traditional U.S. firms. Wilson's concerns are backed by emerging data, showing that DeepSeek has taken a larger share of text-model requests on platforms like OpenRouter compared to OpenAI. In contrast, prominent investor Vinod Khosla vocally disagrees, contending that the true competitive advantage lies in the complex infrastructure behind these models rather than the models themselves. Khosla argues that companies like OpenAI, which develop custom chips tailored to their models, can achieve lower operational costs compared to open-weight models depending on external cloud services. This debate underscores a pivotal moment in the AI/ML landscape, where balancing innovation and profitability is becoming increasingly critical as competitors rush to fill market gaps.