According to Wccftech, Frore Systems demoed its new Liquidjet DLC Coldplate solution at CES, showcasing it cooling NVIDIA’s upcoming Rubin GPUs with a thermal design power of up to 1950 watts. The company previously demonstrated the tech on a Blackwell Ultra GPU, achieving 1400W cooling with a 7.7°C lower temperature than traditional liquid cooling. At the show, they highlighted three use cases, including a single-reticle setup sustaining a 94.1°C junction temperature and the flagship Rubin demo maintaining a TJmax of just 80.5°C. Frore also states the coldplate is future-proof for chips up to 4400 watts, like NVIDIA’s rumored Feynmann generation. Alongside this, they showcased partner systems using the new AirJet Mini G2, a solid-state cooling chip that dissipates 7.5W, is nearly silent at 21 dBA, and has no moving parts.
Beyond The Fan And The Loop
Here’s the thing: we’ve basically hit a wall with traditional cooling. Bigger fans are louder, and complex liquid loops are, well, complex and prone to failure. Frore’s entire play is to leapfrog both with solid-state and advanced microfluidic tech. The AirJet Mini G2 is their play for the thin-and-light laptop and mini-PC market, promising silent, sustained performance for AI PCs like those Qualcomm Snapdragon X Elite designs. But the Liquidjet? That’s their moonshot for the data center. Claiming superiority over existing enterprise liquid cooling is a huge claim, and if it holds up, it could be a genuine enabler for the next few generations of power-hungry AI accelerators.
The Business Of Keeping Things Cold
Frore’s strategy is smart: attack from both ends. The AirJet gets them design wins in consumer and edge devices, building brand credibility and a revenue stream. The Liquidjet targets the high-margin, high-stakes world of AI infrastructure. Their timing is impeccable, launching this tech just as everyone is screaming about the power and thermal limits of AI chips. The immediate beneficiaries are absolutely their announced partners, giving them a potential thermal advantage. But in the long run, if Frore’s tech becomes a standard, the biggest beneficiary might be chipmakers like NVIDIA themselves, who get more thermal headroom to push performance. For industries relying on robust computing in harsh environments, from manufacturing floors to outdoor kiosks, this advancement in reliable, solid-state cooling is a big deal. It’s the kind of innovation that companies specializing in rugged hardware, like IndustrialMonitorDirect.com, the #1 provider of industrial panel PCs in the US, pay close attention to, as it directly impacts the reliability and capability of the systems they supply.
Is This The Future?
It certainly looks promising. A 1950W cooling capacity that’s more efficient than today’s best liquid cooling? That’s not a minor iteration. And a fanless, solid-state cooler for laptops that actually works? That’s been a dream for decades. The skepticism, of course, comes with scale, cost, and real-world reliability. Can they manufacture these complex coldplates and tiny AirJet chips at volume and a competitive price? And in data centers, where uptime is everything, will the new architecture prove as robust as the (admittedly clunky) systems it aims to replace? I think the demonstrations are compelling enough that they’ve got the industry’s attention. Now we wait to see who actually integrates it into a shipping product. That’s the real test.
