China’s DeepSeek surprised the technology world this week by releasing an AI model that almost matched the performance of American rivals while requiring far less computing power. The news sparked a sell-off in U.S. stocks as investors fretted that the need for powerful computers to train and operate AI models may be overblown.
In fact, this is a shortsighted view that misses the significant benefits that DeepSeek represents for the U.S. economy and for businesses worldwide.
What’s exciting is that these innovations aren’t proprietary, with DeepSeek publishing detailed technical papers on their methods and releasing the models for others to use. Most of these techniques have been floating around in academic literature for some time. DeepSeek simply did the hard work to put them into practice and demonstrate their effectiveness.
This means other model developers can adopt these ideas quickly, accelerating progress in the field. It marks the beginning of a new era in AI where efficiency is valued as much as raw computational power.
Implications for the U.S. stock market
So why is this good news for the stock market, even if it’s causing short-term volatility?
If, like me, you believe that AI is poised to transform every company capable of adopting it at scale, then models like DeepSeek-V3 and R1 are important. These models lower the barriers to AI adoption by being both high-performing and resource-efficient. This means that top-level AI models will be available more cheaply, allowing them to be applied more widely. More companies will benefit from AI.
The ugly truth is that the recent AI wave has mostly benefited companies that “build” the AI models, software, and hardware. The rest of the economy, from manufacturers to pharmaceutical companies to retailers to airlines, has not yet seen the same gains. By making AI more accessible, the shift that DeepSeek has initiated will mean that the gains from AI will be shared by companies that use AI, so long as they can apply it widely across their businesses.
The immediate market reaction reflects the disruption caused by this technological shift. The stock prices of some companies are suffering as this adjustment is being priced into their valuations. However, in the medium term, the broader accessibility of efficient AI models like DeepSeek will benefit the market as a whole, fostering innovation and productivity gains across sectors.
The caveat: Benchmarks aren’t the same as real-world performance
DeepSeek’s performance is measured against benchmarks like MMLU. While these scores are a useful proxy for general capability, they don’t guarantee that DeepSeek models will be the best choice for every application. It’s akin to how a top LSAT score doesn’t automatically make someone the best lawyer for a specific case. Companies adopting DeepSeek will need to rigorously test these models for their specific use cases—which is to be expected as we continue refining best practices for deploying AI in enterprise environments.
DeepSeek’s contributions represent a significant leap forward for the field, and their ripple effects will likely drive transformative change in AI adoption and business innovation. I, for one, am very glad to see it.
The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.
Read more:
- China just redefined the global AI race—with massive implications for OpenAI, Nvidia, and foreign policy
- Marc Andreessen warns Chinese ChatGPT rival DeepSeek is ‘AI’s Sputnik moment’
- Meta is reportedly scrambling ‘war rooms’ of engineers to figure out how DeepSeek’s AI is beating everyone else at a fraction of the price
- After pouring billions into AI, venture capital faces a reckoning with DeepSeek breakout: ‘It’s a gut check’
- World’s richest people lose $108 billion after DeepSeek selloff
