The AI landscape is in turmoil. A little-known Chinese company has developed an AI chatbot that rivals market leader OpenAI’s technology but at a fraction of the cost. This comes at a time when US tech giants have committed billions to AI investments—an unwelcome development for them, but a potential boon for global productivity.
DeepSeek, as the company is called, challenges the prevailing AI development model favoured by Big Tech. Instead of relying on vast amounts of data and computing power to refine generalist models, DeepSeek uses a more targeted approach, significantly reducing training time, costs, and the need for high-end processors.
This shift explains why Nvidia, a leader in AI chip production, saw its stock dip. If AI can be built with lower computing power, fewer advanced chips are required. However, it remains unclear exactly how much DeepSeek has spent on developing its model, or whether China has found ways to bypass US restrictions on Nvidia’s most powerful chips.
A new AI paradigm?
Initial reactions from US tech leaders suggest that DeepSeek has made a major breakthrough. Nvidia itself praised it as an “excellent AI advance,” while reports suggest that Mark Zuckerberg has assembled a “war room” of engineers to understand how DeepSeek achieved its results.
But stepping back, the real question is whether widely accessible, affordable AI models are actually good news. In an era of rapid technological progress, it is unsurprising that breakthroughs occur at an accelerating pace.
If more people can soon leverage AI to boost productivity at low cost, the implications could be transformative—especially for economies struggling with long-term structural growth deficits. To illustrate, a look at US productivity growth trends provides some context.
The productivity puzzle
Despite widespread claims about the limitless potential of technology, productivity growth has generally been slowing. Occasional surges occur, but they tend to fade quickly. In Europe, where working hours have declined, the downward trend is even more pronounced.
This issue affects nearly all major economies (with the exception of India). Ageing populations and weakening productivity growth constrain potential GDP expansion, forcing economies to rely increasingly on debt to sustain even modest growth.
A productivity boost from AI?
If AI development becomes more accessible and affordable, it could be a game-changer. Personally, I find myself relying on AI tools like ChatGPT more and more—whether for research, writing, or generating ideas. That’s just a small fraction of what these models can offer.
That said, I remain cautious about DeepSeek, particularly given concerns over data security and information flows to China. Still, if AI democratization helps revive productivity growth, it might be a price worth paying.
And if, in the end, everyone needs to upgrade their PCs and tablets with a powerful Nvidia processor, that might not be the worst outcome—especially if it leads to stronger economic growth.
Jeroen Blokland analyses key financial and macroeconomic trends in his newsletter The Market Routine. He also manages the Blokland Smart Multi-Asset Fund, investing in equities, gold, and bitcoin.