A little over a year ago a small Chinese artificial-intelligence lab shocked the world.
DeepSeek released a pair of models which performed almost as well as the best Western ones,
but were built for a fraction of the cost.
The market value of Nvidia and other providers of AI infrastructure briefly tumbled
as investors fretted (wrongly) that demand for their wares would slow
in the face of such a leap in the efficiency of model-making.
Yet the release on April 24th of the lab's new model, called v4, has been greeted with a shrug.
Why?
DeepSeek's latest release hits many of the same heights its predecessor did.
According to tests run by the company, the performance of its most powerful "Pro" system
falls only marginally short of the models put out by leading American competitors three to six months ago.
DeepSeek's v4 is cheap for customers, too.
An introductory offer makes it a thousandth of the price of the best American models for some uses.
Even after that rate expires on May 7th,
v4 will cost between a tenth and a quarter of American equivalents.
But it seems that, unlike DeepSeek's previous blockbuster, v4 was not cheap to build.
In 2025 the lab eagerly pointed out that the cost of training its AI was about $6m,
far below the going rate in the West.
The lab's technical white paper on v4 omits any estimate of this measure.
The fact that 16 months elapsed between v4 and its predecessor