Imagination Technologies has installed its seventh chief executive in a decade, with Markus Mosen taking the helm, and this time the British IP firm says the strategy is staying put: GPUs only. The company has shelved its CPU and NPU ambitions to pour resources into its core graphics business.
Chief revenue officer Jake Kochnowicz told EE Times that RISC-V CPU startups are starting to struggle and the market is ripe for consolidation. “We’re a GPU company, let’s focus on that,” he said, adding that the direction is largely the same as before and the leadership change is about moving faster.
The bet rests on a next-generation GPU architecture called F-series, a “significant refresh” financed by a $100M loan taken out two years ago. It targets automotive and data center workloads and scales to larger die sizes than the E-series, which tops out around 300 mm2 at TSMC N5. Two big customers have already signed on ahead of a launch around the end of 2026. The F-series will not hit 1 PFLOPS, but the generation after it could, Kochnowicz said.
China plays a central role. One customer is a large public EV maker using Imagination GPUs in ADAS systems, and the company can license IP for Chinese “big GPUs” as long as customers are not on the US entity list. Licensing terms prohibit military drone applications.
Imagination frames AI as a “first-class citizen” alongside graphics, but insists it remains a graphics company first. The bet is that consolidation will reward a focused GPU IP player over the next decade.
