When Chinese AI startup DeepSeek launched its R1 model requiring far less computational power than their American peers in January, it was only exposing the limitations Silicon Valley had already begun acknowledging. As millionaire investor Marc Andreessen noted, “we’re increasing GPUs at the same rate, but we’re not getting the intelligence improvements at all out of it.” That’s even before factoring in the supply chain challenge of providing enough chips to keep delivering the current rate of unsustainable growth.

The time of aggressively scaling through investing in more data appears to be drawing to a close. As the big players move to refocus on the composition and quality of specialized data sets, not sheer volume, the same principle needs to filter down to businesses and how they frame their AI strategy.