DeepSeek: The Start of The AI War
©Ici Beyrouth

“Artificial intelligence will have a more profound impact on humanity than fire, electricity, and the internet” according to Sundar Pichai, the CEO of Alphabet, Google’s mother company. It is indeed safe to say that AI will fundamentally transform every major industry on the planet, meaning that nations that master AI will also master the world. Until now, a small number of American companies have dominated the field of artificial intelligence. Firms like OpenAI, Nvidia, and Oracle have attracted massive investment, becoming some of the highest-valued companies in the world. But suddenly, that dominance appears less certain, with a new competitor on the block: DeepSeek.

What Happened?

Over the weekend, a chatbot released by an obscure Chinese AI startup called DeepSeek became the most downloaded app on iPhones, causing American tech stocks to plummet, wiping $1 trillion from US tech giants' valuations. This marked the single biggest market loss in US history. President Trump described the situation as a “wake-up call” for US companies, urging them to stay on top of the competition. But the biggest news was Nvidia, the rockstar of the AI world, which lost around $589 billion in market valuation in one day. To put things in perspective, Coca-Cola is valued at around $270 billion.

What Is DeepSeek?

DeepSeek is a Chinese startup. Just a few days ago, it released an impressive AI model called R-1, which triggered all this chaos. In terms of performance, the Chinese model currently seems to perform quite competitively with Western rivals like OpenAI’s ChatGPT. It also appears to have attracted a significant number of users, with DeepSeek topping the App Store rankings. But what’s really astonishing is the company’s claim that its new model cost a fraction of what American competitors have spent, both in terms of resources and time. DeepSeek says it spent only $6 million and just two months to develop the model, using Nvidia’s less-advanced H800 chip. In contrast, American tech firms have spent enormous amounts on computing power for their AI models. Just last year, four of the biggest tech firms spent about $280 billion on data centers. DeepSeek’s model implies that such high levels of spending and infrastructure may not be as necessary as we once thought.

Indeed, Nvidia, the American titan that has benefitted from AI firms' demand for advanced and expensive chips, becoming the world’s most valuable company, may no longer be able to justify its current production levels, potentially leading to decreased sales. If companies can now train their own models with fewer chips, this could be a big problem for Nvidia, which is why its stock price fell so sharply. Additionally, there could be significant consequences for the cloud computing giants like Google, Microsoft, and Amazon, all of whom have heavily invested in AI. Still, for OpenAI’s CEO Sam Altman, the picture is not so bleak. In a tweet on X, he said his company will deliver much better models, insisting that more chips are still necessary since “the world is going to want to use a LOT of AI”.

We don’t know yet how this will all play out. What is certain, however, is that the AI war has officially begun. DeepSeek is China’s direct challenge to US AI dominance. With AI poised to revolutionize every aspect of our lives, we can expect this battle to be intense. What’s unfolding feels similar to the space race between the US and the USSR. Competitors will spend large resources and use any tactics at their disposal to achieve dominance. The US cannot allow Chinese models such as DeepSeek “to risk our national security and leverage our technology to advance their AI ambitions,” US Representative John Moolenaar, who co-chairs a select committee on competition with China, said. He added, “We must work to swiftly place stronger export controls on technologies critical to DeepSeek’s AI infrastructure.” Moolenaar speaks with a bipartisan voice, as both parties are cognizant that whoever wins the AI race will own the future.

 

Comments
  • No comment yet