Tem raises $75 million to reshape power markets using AI

As AI data centers drive up electricity bills, London-based startup Tem thinks AI could help solve this problem too.

Tem has built an energy trading engine that leverages AI to lower prices compared to other energy traders. The company has signed deals with more than 2,600 business customers across the UK, promising them savings of up to 30% on their energy bills by purchasing energy from its utilities division.

The startup recently closed an oversubscribed $75 million Series B led by Lightspeed Venture Partners with participation from AlbionVC, Allianz, Aomico, Hitachi Ventures, Revent, Schroders Capital, and Voyager Ventures, TechCrunch has learned exclusively.

The round values ​​Tem at more than $300 million, a person familiar with the deal told TechCrunch. The startup plans to use the funding to support its expansion into Australia and the United States, starting in Texas.

“We’re in a good position where we have some control over our own profitability, so we could have chosen not to raise at all and have a nice, nice bootstrapped business in a way,” Tem co-founder and CEO Joe MacDonald told TechCrunch. “Well, we’re not that kind of business. As someone who’s been trying to go public for years, we know what we want to achieve.”

Tem is a classic market play that matches generators and consumers. The company intentionally started by focusing almost exclusively on renewable energy generators and small businesses to fill both sides of the ledger. “The more decentralized it is, the better it is for the algorithms,” MacDonald said. “But this works all the way down to the enterprise.”

The company’s customers include fast fashion retailer Boohoo Group, soft drinks company Fever Tree and Newcastle United FC.

Currently, Tem runs two different businesses. One is a transaction engine called Rosso that matches suppliers and buyers. Here, machine learning algorithms and LLMs help predict supply and demand.

Rosso’s goal is to reduce costs by eliminating some of the layers that exist in the current energy market, McDonald said. “Each organization has different teams doing different jobs, different levels of benefit from back office to trading, trading desks to other trading desks, and there are probably five or six intermediaries in total that allow the flow of funds to move from one to the other,” he said.

Using AI, he said, “we have the opportunity to replace humans, labor costs, and disparate systems with one transactional infrastructure.” The goal is to bring the price customers pay for electricity closer to wholesale costs.

Another part of Tem is called RED, a “neo-utility” built to prove Rosso’s worth.

“When we first started, we tried to sell infrastructure to energy companies and it didn’t work out,” he said. RED is currently the only utility using Rosso, and McDonald said the company’s growth has led it to prioritize Rosso over opening it up to other companies.

But Mr. Thame plans to allow other power companies to enter at some point.

“In reality, it doesn’t matter how good you are. [RED] The market share will not exceed 40%. It shouldn’t do that because it would become a monopoly in itself. So I would rather have access to all the transaction flows,” McDonald said.

“In the long run, as long as the infrastructure is used, we really don’t care who the customer owner is or the generational owner is,” he said. “This is just an infrastructure effort, just like AWS or Stripe.”