China does more with less

The sudden rise of DeepSeek has been fascinating and instructive on so many levels. Just like that, it forced US-China headlines away from Trump’s tariffs, triggered a trillion-dollar bloodbath on Nasdaq, sent the volatility index (VIX) skyrocketing, rattled the US technology establishment and poked big holes in American tech giants’ justifications for sky-high valuations and closed ecosystems.

Already a shocked American financial press is wondering if this is just “AI’s Sputnik moment”, as venture capitalist Marc Andreessen put it, or a Pearl Harbor-like hit on financial markets that robbed Silicon Valley’s Mag-7 of a trillion dollars in one trading session.

And even though the broader market began its recovery shortly after the initial shockwave, America’s AI leaders, as well as the Trump administration, will wonder for a long time how China managed to pull off such a feat so quickly – and what else might be on the horizon.

Much attention has been given to the implications of open-sourcing AI models – rightly so – but the bigger question is how a relatively unknown Chinese AI firm, in a single stroke, made America’s top tech firms, and its government, look unprepared.

For years, US policymakers and Silicon Valley leaders assured the world that China’s AI development was being throttled by strict sanctions on advanced semiconductors.

The Biden administration’s export controls were supposed to slow down Beijing’s AI ambitions by cutting off access to high-end Nvidia chips.

Yet DeepSeek’s success has exposed the ineffectiveness of Washington’s broader tech containment strategy. If China can produce cutting-edge AI without the latest chips, what other technological barriers will it overcome, and how easily?

The impact on the US tech giants is equally embarrassing. OpenAI, Google DeepMind, and Anthropic have poured billions into training their models, relying on vast resources and highly exclusive hardware. Their argument has been that AI development is an expensive, capital-intensive process, justifying their sky-high valuations and closed-source business models.

DeepSeek’s emergence undermines this premise. It suggests that China, without access to America’s best chips, has been able to develop a rival system at a fraction of the cost. Investors are now asking if DeepSeek can do this, why are American firms spending so much more?

Then there’s the open-source dilemma. US AI companies, with the exception of Meta, have been fiercely protective of their models, keeping them proprietary and commercialised.

DeepSeek’s release challenges this status quo by offering an alternative that is both high-quality and freely available. If this trend continues, US firms may be forced to rethink their approach or risk being outmaneuvered by more agile competitors. Meanwhile, the US government, which has been debating regulations on AI openness, faces a new strategic concern. Does restricting open-source AI at home simply push the advantage to China?

Financial markets certainly took note as Nvidia alone lost $600 billion in market value on Monday — the largest single-day loss for any company in history. And it’s no surprise that investors are beginning to question whether the AI boom, as hyped by Wall Street, is truly sustainable when disruptive competition can emerge from unexpected corners of the world; especially players that the US has been bending over backwards to keep out of the race altogether.

All of this makes the Trump administration’s position more precarious. The idea that Washington can maintain control over key technological advancements is now in doubt.

If DeepSeek is just the first major shockwave from China’s AI ecosystem, does the US need to prepare for more surprises? The real question is whether the American response will be one of adaptation or another round of sanctions that fail to slow the inevitable.

With a paranoid, pro-tariff, protectionist president in the White House, it’s not hard to venture a guess.

Copyright Business Recorder, 2025