The DeepSeek AI cost claims are “exaggerated and misleading,” according to Demis Hassabis, CEO of Google DeepMind. Speaking at the Artificial Intelligence Action Summit in Paris, he explained that the reported $5.6 million cost only covers the final training round of DeepSeek’s AI model. This is just a small part of the total development cost.
What DeepSeek Reported
DeepSeek claimed it developed its popular chatbot and AI model for only $5.6 million. This announcement sparked attention as it was significantly less than the budgets of U.S. rivals like OpenAI and Google DeepMind. However, Hassabis clarified that this figure does not include all phases of AI development, such as research, data collection, or earlier training rounds.
“They reported only the cost of the final training round,” he said. “That’s a fraction of the total cost.”
Industry Scrutiny on DeepSeek AI Cost
DeepSeek’s claims have raised questions among industry experts. The startup revealed it trained its AI using older Nvidia chips. U.S. authorities are investigating if DeepSeek bypassed a chip ban by purchasing semiconductors through Singapore.
Additionally, OpenAI and Microsoft are exploring whether DeepSeek used “distillation” to train its models. Distillation allows one AI model to extract data from another. If true, this would explain how DeepSeek kept costs low.
No Silver Bullet in AI Development
Hassabis dismissed the idea that DeepSeek has disrupted the economics of AI development. He explained that Google’s AI model, Gemini, is far more efficient in cost-to-performance ratios than DeepSeek.
“There’s no silver bullet here,” Hassabis said. “DeepSeek is not an outlier on the efficiency curve.”
Silicon Valley’s Continued AI Investment
Despite DeepSeek’s low-cost claims, companies like Alphabet and Microsoft are still investing heavily in AI. Alphabet plans to spend $75 billion on capital expenditures in 2025, focusing on cloud computing and AI services like Gemini.
Microsoft has also announced an $80 billion investment in AI-enabled data centers. These moves reflect a broader strategy to compete with China’s growing influence in AI technology.
Final Thoughts
The DeepSeek AI cost discussion highlights the complexity of AI development. While the startup’s claims are intriguing, industry leaders like Google and Microsoft continue to focus on efficiency and innovation. As the AI race heats up, transparency and innovation will play key roles in determining long-term success.
