As AI data centers drive up electricity prices, London-based startup Tem thinks AI would possibly be capable to assist clear up it, too.
Tem has constructed an power transaction engine that depends on AI to chop costs relative to different power merchants. The corporate has signed up greater than 2,600 enterprise clients all through the U.Okay. on the promise that purchasing power from its utility division can save them as much as 30% on their power payments.
The startup not too long ago closed an oversubscribed $75 million Collection B led by Lightspeed Enterprise Companions with participation from AlbionVC, Allianz, Atomico, Hitachi Ventures, Revent, Schroders Capital, and Voyager Ventures, TechCrunch has solely discovered.
The spherical values Tem at greater than $300 million, a supply acquainted with the deal informed TechCrunch. The startup plans to make use of the funding to assist broaden to Australia and the U.S., beginning with Texas.
“We’re in a pleasant place the place we form of have management over our personal profitability. So I may have chosen to not increase in any respect and had a stunning, good bootstrap enterprise in some methods,” Joe McDonald, co-founder and CEO of Tem, informed TechCrunch. “Properly, we’re not that form of enterprise. We all know what we need to obtain as somebody who desires to go public over time.”
Tem is a basic market play, matching electrical energy mills with customers. The corporate deliberately began by focusing nearly solely on renewable power mills and small companies to fill either side of the ledger. “The extra decentralized and the extra distributed, the higher it’s for the algorithms,” McDonald stated. “However this works all the best way as much as enterprise.”
The corporate’s clients embody fast-fashion retailer Boohoo Group, delicate drink firm Fever-Tree, and Newcastle United FC.
Techcrunch occasion
Boston, MA
|
June 23, 2026
Presently, Tem is working what quantities to 2 totally different companies. One, known as Rosso, is the transaction engine that matches suppliers with patrons. Right here, machine studying algorithms and LLMs assist predict provide and demand.
The aim with Rosso, McDonald stated, is to chop prices by eliminating a number of layers which might be current in present power markets. “In every of them, you’ve acquired totally different groups doing totally different jobs, taking totally different ranges of revenue from again workplace to buying and selling, buying and selling desks to different buying and selling desks, and possibly 5 to 6 intermediaries in complete that allow the movement of cash to maneuver from one aspect to the opposite,” he stated.
With AI, he stated, “you now have a chance to switch the people, the labor prices, and the disparate techniques into one single transaction infrastructure.” The aim is to make the worth that clients pay for electrical energy nearer to the wholesale value.
The opposite a part of Tem, known as RED, is a “neo-utility” constructed to show the worth of Rosso.
“Once we first began, we tried to promote our infrastructure to the power firms, and we acquired nowhere,” he stated. RED is at present the one utility utilizing Rosso, and McDonald stated its progress has pushed the corporate to prioritize it over opening Rosso to others.
Sooner or later, although, Tem plans to permit different utilities in.
“In actuality, it doesn’t matter how good [RED] is; it’s not going to get above a 40% market share. And it shouldn’t, as a result of that turns into a monopoly in itself. So, me, I’d a lot slightly go to get entry to all of the transaction movement,” McDonald stated.
“Long run, we actually don’t thoughts who owns the client, who owns the technology so long as our infrastructure is getting used,” he stated. “That is simply an infrastructure play in the identical means AWS was, or Stripe was.”
