11.2 Billion: The First Chip Unicorn Will Arrive in 2026

The deal and the product

Frore Systems announced a $143 million D round led by MVP Ventures, with follow‑on participation from Fidelity, Mayfield, Addition, Qualcomm Ventures and others, taking its valuation to $1.64 billion (about RMB 11.2 billion) and lifting total capital raised to roughly $340 million. Frore does not design chips — it makes liquid‑cooling systems that keep high‑power AI accelerators from overheating, and its products are reportedly compatible today with NVIDIA (英伟达), Qualcomm (高通) and AMD chips and developer boards.

A timely pivot

The sharp angle here is timing. It has been reported that about two years ago Jensen Huang saw Frore’s demo and suggested the company move into liquid cooling. Frore did exactly that, shifting from mobile and fanless device thermal designs to data‑center liquid cooling. Why does that matter? Because AI model scale has driven GPU power and rack‑level heat to new heights — single cards now can exceed 700W TDP and cabinet power easily passes 100 kW — and air cooling can no longer keep pace. Liquid cooling is becoming a physical necessity for modern AI deployments, not just a nice‑to‑have.

Strategy, investors and geopolitical context

Frore’s strategy is deliberately non‑partisan: support multiple chip platforms rather than betting on a single winner. That makes its addressable market grow as competition among chip vendors intensifies. The investor mix — financial heavyweights alongside industry players such as Qualcomm Ventures — is a common signal that a hardware company has moved from technical validation toward commercial scale. It has been reported that US export controls and broader tech geopolitics are also reshaping procurement and deployment patterns for high‑end accelerators, which could influence where and how quickly liquid‑cooling installations roll out.

What this means for the market

Frore’s rise to unicorn status is notable because the cooling subsector has so far produced few scaled leaders; the space is still fragmented and early. If liquid cooling becomes standard across new AI data centers, companies that have secured cross‑platform compatibility and early customer integrations will enjoy sticky demand — replacing a deployed cooling solution is complex and costly. The bigger point: while chip design remains the headline battle, the infrastructure that enables those chips to run at scale is now a clearer investment frontier. Who wins the compute wars may still be uncertain, but the need to keep compute cool is not.