July at Temperate: Novel tech, mechanical hurdles
What we did in July
Built our TRL5 prototype. Our first unit is cooling air - with a delta of 4 degrees C between air going in and air coming out. We’re delighted that the device is actually cooling, but we’re not yet seeing the volume that we’re looking for. Fergus is hard at work discovering what’s wrong and fixing it. More data on this in next month’s update. We’ve learned an awful lot getting this one built, and are already designing on the next version. Key improvements below.
Patent Search Results Are In. Our invention is novel! While there’s still work to do before our final submission in May 2026, our IP attorneys are confident we’ve created something unique, and that our rights to it will be fully protected. We all heaved a sigh of relief.
Demo Day at Google London. We pitched at a climate-focused demo day hosted at Google’s King’s Cross office. We had great feedback, with many hints to look into data centres. That event yielded 15+ follow-up chats with interested investors. (Don’t worry—we’re not raising just yet.)
First NED. We’re thrilled to welcome Michael Reid to our board as a Non-Executive Director. Michael was the first person who really “got it” and quickly became our earliest champion. He’s an engineer experienced in sustainable cooling, a technical founder in his own right, and brings experience launching home appliances in a number of our target markets: India; Japan; and the USA.
What we’re learning (the hard way)
Building something that’s never existed before means plenty of failures, and we’ve learned from each one. Here are a few of those hard-earned gems from the most recent prototype build.
Iteration time. Our most recent build took 3 months - this is obviously far too long, so we’re going to bring everything in house to shrink build time to 3 weeks. The long build time was due to the bespoke nature of the parts we’re working with (I’m looking at you, Fresnel mirrors), long shipping schedules and extra work on getting components ready for install like polishing scuffed parts. For future prototypes, we’re taking more control over the process by cutting parts in-house and keeping the build entirely in the lab.
Weight. Once the parts arrived and we started assembly, we realised that our unit was approaching 50kg - driven by the weight of our cooling panels. Adding more panels means more cooling, but makes the unit too heavy. In our next build, we’re going to re-engineer the panels to shed weight without losing cooling power.
Airflow. Our current build is not getting sufficient air flow to effect a strong heat transfer. When designing the next version, air-side performance is now our main technical focus. We’re working with fluid-dynamics specialists to model new fan, duct, and geometry options that increase flow without wrecking our noise or power budgets.
Modelling. We’ve cut analysis time for our coating design by 95% thanks to support from one of our technical consultants, who suggested a simplified model structure while keeping all the critical analysis. Compute was originally costing £40/hr+ and taking 2-3 days to get results - we can now get the same results for multiple scenarios in an hour.
Plans for August
Collect Cooling Data. After patching up the current TRL5 prototype, we’ll be collecting and analysing performance trends—expect to see the results in next month’s update.
Next Prototype: even more cooling. We’ve got two weeks of design and modelling planned to take all of these lessons into the next version of our prototype. We’re aiming for improved airflow for more efficient bulk heat transfer, i.e. more cold air, higher total cooling capacity.
Exploring Data Centre Cooling. We’re researching AI Infrastructure as a potential fast-follow second use case. Our hunch? Our cooling can improve the efficiency of data centre by augmenting their current setup.
Metrics (July)
Metric Number Status
Monthly burn. £6,000 🟢
Runway remaining 14 months 🟢
Divergence vs budget +£18,000 🟢
How You Can Help
We’re a tiny team doing a ridiculous thing. If you can help, we’d love to hear from you:
Non-Executive Director Search. Our dream NED would have a background in mechanical engineering, deep cooling sector experience, and knowledge of at least one major target market: India, Brazil, or the UAE.
Data Centre Cooling Expertise. Know how large-scale data centre cooling works? We’d love to pick your brain over a 20-minute coffee chat.
We’re moving fast, learning loads, and building what we believe will be the next generation of cooling technology. Thanks for being part of the crew.
—
The Temperate Team
(Less hot air, more cold air)