Can a Chinese Startup Outsmart America's $500 Billion AI Plan?

Recently, a Chinese startup called DeepSeek has stirred up the AI world. It claims to have trained its models using older, less expensive hardware than what big U. S. companies like Nvidia use. This has some people wondering if DeepSeek could shake up the AI game and make America's big $500 billion AI project, Stargate, look less impressive. DeepSeek's approach is interesting. They've used older Nvidia chips, like the H800, instead of the latest and greatest models. This could mean that high-quality AI doesn't always need the most expensive tools. If true, investors might think twice about spending so much on AI infrastructure. But let's not forget about the big players in the AI world. Companies like Microsoft, Meta, Alphabet, and Amazon are still pouring billions into AI. Microsoft, for example, spent over $55 billion last year and plans to spend around $80 billion this year. Meta's spending jumped by 40% last year and is set to increase by 67% this year. Alphabet is planning to spend around $75 billion, and Amazon could spend about $105 billion. These companies are not backing down from AI. In fact, they're doubling down. Even after DeepSeek's arrival, the biggest tech companies are still investing heavily in AI infrastructure. This shows that DeepSeek might not be the game-changer some people think it is. Palantir Technologies CEO Alex Karp even said that DeepSeek's advancements should inspire U. S. tech companies to keep innovating. With all the big tech companies planning to spend over $320 billion on AI this year alone, it seems like DeepSeek might actually be a boost for domestic AI investment. So, while DeepSeek is an interesting player, it doesn't look like it's going to crash America's AI plans. Instead, it might just be another reason for big tech to keep investing in AI.

Can a Chinese Startup Outsmart America's $500 Billion AI Plan?

Continue Reading...

Actions