Switch, one of the largest privately held data center developers in the U.S., is in talks to raise billions of dollars at a valuation north of $50 billion, according to The Information. The report signals just how hungry investors are for the physical infrastructure powering the AI boom. If the deal closes at that price, Switch joins the small club of data center operators now valued like major tech companies.
What’s Happening
Switch builds and runs massive data center campuses, the kind of facilities that house the servers and chips behind modern AI workloads. The Information reports the company is negotiating a raise that would value it at more than $50 billion. That’s a striking number for a business that, until recently, lived in the unglamorous world of real estate and colocation.
A few quick facts for context:
- Switch was taken private in 2022 by DigitalBridge and IFM Investors in a deal worth around $11 billion.
- A $50 billion-plus valuation would mean its worth has roughly quadrupled in a few years.
- The company operates large campuses in Nevada, Michigan, Georgia, and Texas.
Why It Matters
The money chasing AI has shifted. For the past two years, the headlines went to model makers like OpenAI and Anthropic and to chip designer Nvidia. Now the attention is moving down the stack to the buildings, power, and cooling that make AI possible at all.
What stands out here is the valuation math. Data centers used to trade as steady, boring infrastructure assets. A $50 billion price tag puts Switch in growth-stock territory, which tells you investors expect AI demand to keep climbing for years, not quarters. They’re betting that whoever controls power and space controls a real chokepoint.
This also fits a clear pattern. Capital is flooding into anything that helps train and serve large models:
- CoreWeave went public and saw its value swing into the tens of billions.
- Crusoe, Vantage, and other developers have pulled in huge rounds.
- Hyperscalers like Microsoft, Google, Amazon, and Meta are spending hundreds of billions combined on AI buildout.
Switch raising at this level is another data point in the same story. The bottleneck for AI isn’t just smarter models anymore. It’s electricity, land, and the speed at which you can pour concrete.
The Power Problem Underneath
Here’s the part that makes these numbers make sense. AI data centers are enormous consumers of electricity. A single large campus can draw as much power as a small city. Utilities are struggling to connect new sites fast enough, and grid capacity has become the real constraint.
That scarcity is exactly why a developer like Switch can command a premium. Companies that already hold land, power contracts, and permits are sitting on assets that are hard to replicate. You can’t spin up a gigawatt of capacity overnight, no matter how much money you raise.
What to Watch Next
The deal isn’t done, and terms could shift. The Information frames this as ongoing talks, so treat the $50 billion figure as the target, not the final word. Still, a few things are worth tracking if you work in or around AI:
- Who writes the check. Sovereign wealth funds, private equity, and infrastructure investors have been the biggest backers of this category. The investor list will say a lot about where smart money sees the cycle going.
- Whether the valuation holds. A raise at this level sets a benchmark for rival developers. Expect more data center operators to test the market at higher prices.
- Power commitments. Any new capital almost certainly funds expansion, which means more pressure on grids and more deals with utilities and even nuclear providers.
For practitioners, the takeaway is simple. The cost and availability of compute over the next few years will be shaped as much by deals like this as by any model release. If you’re planning AI products or budgeting for inference at scale, the infrastructure layer is where the constraints, and the prices, get set.
More details are available in the original report from The Information.