The Story Everyone Is Repeating
Ian King’s pitch opens with a story that has been retold across the financial internet: Elon Musk built xAI’s Colossus data center in Memphis in 122 days, then doubled its size in 92 days. The construction speed was impressive. The power problem it exposed was the actual lesson.
Colossus needs roughly 2 gigawatts of electricity. For context, that is more than Hoover Dam produces in a year. The local utility, Memphis Light Gas and Water, could spare eight megawatts of the 150 Musk wanted for phase one. When the grid says no, you have two choices: wait a decade for transmission upgrades, or build your own power plant.
Why He Built His Own Plant
Musk chose to build, and the reason is simple economics of time. A new transmission line can take a decade of permitting, land disputes, and construction. A data center operator does not have a decade. The chips are ordered, the customers are paying, and the power has to arrive on the same schedule as the servers.
The workaround is on-site generation. Instead of waiting for utility interconnection queues, a data center can generate power behind the meter, right next to the servers, on its own timeline. That is what happened at Colossus, and it has become the template the rest of the industry is now studying.
King’s framing is memorable for a reason. “For the first time in his life, Elon found something his billions couldn’t buy.” Strip away the marketing and the point holds: money can buy chips, land, and talent, but it cannot buy a faster grid.
The Company Behind the Plant
The on-site power at Colossus comes through a joint venture called Stateline Power. Solaris Energy Infrastructure owns 50.1 percent of that venture, and SpaceX owns the other 49.9 percent. Solaris is the operating partner, and it is the company King calls the “Musk Master Key.”
Solaris did not appear out of nowhere. It spent roughly a decade as Solaris Oilfield Infrastructure, providing mobile power to remote drilling sites where the grid did not reach. That oilfield work taught the company how to drop turbines and generators in rough, grid-free conditions, and the skill transferred directly to data centers. Our Solaris Energy Infrastructure explainer covers the company and its contracts in detail.
Why the Deal Matters
The significance of Colossus is not one power plant. It is that a hyperscaler, desperate for electricity, chose to bypass the grid entirely and sign a long-term contract with an on-site provider. That validates a business model the wider market had not fully priced in.
It also reframes who the bottleneck is. For years the assumption was that compute was the scarce resource. Colossus suggests the scarce resource is delivered electricity, and the companies that can deliver it in weeks, not years, are the new chokepoint. The demand shift behind that idea is the subject of our piece on data center power demand.
The Part Worth Skepticism
The story is clean, which is exactly why investors should slow down. The Musk connection is emotionally potent, but the $86 million King’s headline references is capital committed to the joint venture, not a personal bet on Solaris shares.
It is also worth remembering that the deal has not reached full power yet. The joint venture ramps toward full output in 2027, and there is a two-to-three-quarter window where debt and capital spending build before the contracts throw off meaningful cash. The story is real; the timeline is longer than the headline suggests.
The broader point for any investor: the fact that a problem is real does not mean every company adjacent to it wins. On-site power is a competitive business, and winning the Colossus deal does not guarantee Solaris wins every deal that follows. For a wider look at the sector, see our guide to AI data center power stocks.
The Other Customers
The Colossus deal gets the headlines, but the more important fact for investors is that Solaris did not stop there. The company has signed long-term, gigawatt-scale power deals with two other hyperscalers, entirely separate from the Musk relationship.
That is the part that separates a real business from a one-off Musk story. A single joint venture with SpaceX could be a curiosity, a contract that exists only because Musk needed power in a hurry. Three separate hyperscaler relationships suggest the on-site model travels beyond one customer and one city.
It also means the bull case does not depend entirely on Musk. The fleet more than tripling by 2029 and the $1 billion in annual cash earnings are driven by the whole book of contracts, not just Colossus. The Musk connection is what gets the pitch attention; the other two customers are what give the business its staying power.
The honest way to weigh this is to ask whether the non-Musk contracts are real and durable. If they are, the story is broader than the headline. If they are still in the early, non-binding phase, then the Musk deal is doing more of the work than the pitch admits.
Ready to see the research? Click here to access Ian King’s report.
NewsletterVetter is an independent publication. We receive compensation from some of the services we review through affiliate links. Nothing on this site is investment advice. Always do your own research.