In a recent statement that has caught the attention of the tech community, Demis Hassabis, CEO of Google DeepMind, expressed significant scepticism over Chinese AI startup DeepSeek’s claim of developing its artificial intelligence system for under $6 million. Hassabis described the figure as “exaggerated and a little bit misleading,” arguing that the reported cost likely reflects only the final training stage of the model—a stage that represents just a small slice of the overall development expenses.
DeepSeek, which has been in the spotlight after unveiling a popular chatbot and accompanying AI model, asserts that its breakthrough was achieved using a fraction of the expenditure incurred by major U.S. competitors such as DeepMind and OpenAI. According to the startup, it spent approximately $5.6 million on computing resources, leveraging older Nvidia chips during the training process. This low-cost approach has ignited debates over whether the startup’s economic model could disrupt the prevailing financial paradigms in AI research and development.
Hassabis, however, remains unconvinced that DeepSeek’s achievements mark a transformative shift in AI technology. “We don’t see any new silver bullet technologies,” he noted, implying that the startup’s approach might have benefited from established methodologies, including potentially adapting techniques pioneered by Western firms. Though he did not delve into specifics, his remarks underscore the view that DeepSeek’s cost claims do not necessarily herald a new era of cost-efficient AI development.
The discussion comes at a time when leading U.S. tech companies are committing to massive investments in the artificial intelligence sector. Alphabet Inc., the parent company of Google, has announced plans to allocate $75 billion in capital expenditures in 2025—funds that will bolster its cloud computing division and support advanced AI services like its Gemini project. In a similar vein, Microsoft has outlined a strategy to invest around $80 billion in building AI-enabled data centres across the globe.
Industry analysts have raised questions regarding the comprehensiveness of DeepSeek’s financial disclosure. Many point out that the nearly $6 million figure appears to cover only the computing costs, leaving other significant development expenses unaccounted for. This partial transparency has led some experts to speculate that the startup’s pricing model might trigger a competitive price war in the AI services market, potentially challenging companies like OpenAI that operate with much higher overhead costs.
Further complicating the scenario, U.S. authorities have launched an investigation into DeepSeek’s supply chain practices. Officials are examining whether the startup bypassed existing chip export bans by channelling purchases through Singapore. Additionally, probes are underway to determine if affiliates of DeepSeek employed a technique known as distillation—a process where one AI model is trained using the outputs of another—to obtain data from OpenAI’s systems.
As the inquiry unfolds and the debate over development costs continues, industry leaders emphasize the need for a comprehensive view that encompasses all facets of AI production, including ethical considerations and full-scale investment analysis. While DeepSeek’s claims have undoubtedly spurred lively discussion, experts agree that understanding the complete financial picture is essential to gauge the true impact of its innovations on the rapidly evolving AI landscape.
—
For further updates on this developing story, stay tuned to Techuncode.com
and then