The Biggest Funding Story of the Week Has Nothing to Do With AI Software
A Houston energy startup no one was talking about last month just landed $1.75 billion to power Microsoft's data centers. Here's what that actually means.
Joulent launched in June 2026. By July 1st it had closed a $1.75 billion strategic investment from National Grid Ventures, giving the UK utility a 35% stake at a $5 billion implied valuation. The company’s first project is called Kilby: a 2.67-gigawatt gas-fired power facility in West Texas being built 50/50 with Chevron, supplying a Microsoft-operated data center campus under a 20-year power purchase agreement starting in 2028.
For context on why that number matters: demand from data centers powering generative AI services rose 17% in 2025, against 3% growth in overall global electricity demand. The gap between those two numbers is Joulent’s entire business case. #
This isn’t a startup in the traditional sense. There’s no app, no freemium tier, no growth hacking. Joulent is purpose-built energy infrastructure, and the customer is already signed for 20 years. Most founders spend years searching for product-market fit. Joulent’s CEO walked in with a Microsoft PPA and a Chevron co-development agreement before the company had a public presence.
I want to apply first principles thinking here, a model I cover in Mental Models: How to Think, Act, and Win. Strip away the AI hype and ask a simpler question: what does AI actually need that doesn’t exist yet at the required scale? The answer isn’t better models. It isn’t faster chips. It’s reliable, gigawatt-scale power delivered directly to data centers at speed, without straining existing grids. Joulent started from that constraint and built the company backward from it. That’s the discipline most energy startups skip because it’s easier to pitch software.
Here’s the contrarian read most people will miss. National Grid’s investment is expected to generate returns exceeding the 9 to 10% return on equity it achieves from regulated networks, reflecting a slightly higher relative risk profile, per J.P. Morgan analysts. That language matters. This isn’t altruistic infrastructure investment. It’s a utility betting that contracted AI power demand is more valuable than its core regulated business model. When incumbent utilities start repricing their own risk to chase AI energy contracts, you’re not watching a trend. You’re watching an industry restructure in real time.
——————————————————————————————————————
If this was useful, share it with someone who builds things. And if you want the full toolkit of 50 mental models, my book is coming soon.