The race to build AI data centers is heating up. Switch, a Las Vegas-based data-center developer, is in talks to raise billions at a valuation over $50 billion. This would be a huge jump for a company that builds the physical spaces powering AI workloads.

Investors like Brookfield Asset Management and KKR are circling, with banks such as Goldman Sachs and JPMorgan advising on the deal. This funding round could set Switch up for an IPO next year.

Why the rush? AI models need massive computing power, which means thousands of GPUs running non-stop. Companies like OpenAI recently raised $122 billion, signaling a tidal wave of infrastructure spending is coming. That demand flows straight into data center construction and upgrades.

Big tech giants are also ramping up spending. Meta is reportedly seeking $13 billion to build a new AI data center in Texas. They plan to finance this with a mix of debt and equity, led by Morgan Stanley and JPMorgan. Meanwhile, Microsoft, Amazon, Google, and Meta are expected to spend over $5 trillion on cloud infrastructure through 2030.

Why Data Centers Are the New Gold Mines

Data centers are no longer just warehouses full of servers. They are the backbone of AI innovation. Companies are scrambling to secure land, power, and building space to meet AI’s heavy demands.

Switch’s expansion plans include a 382-acre campus near Pittsburgh, a rare land-and-power package that’s hard to replicate quickly. Investors see these physical assets as solid bets in a market where compute capacity is scarce.

Goldman Sachs says private infrastructure and real estate funds will play a bigger role financing this boom. These investors like the steady income and inflation protection data centers offer. The private infrastructure market has grown at over 11% annually recently and could accelerate even more.

The AI Infrastructure Spending Wave

The scale of AI’s compute needs is staggering. Training a top AI model requires tens of thousands of GPUs, consuming massive electricity. US utilities plan to spend $1.4 trillion by 2030 to power data centers.

Hyperscalers like AWS, Azure, and Google Cloud are key players. They continuously expand their GPU clusters and networking to keep up with AI progress. Their capex plans alone run into trillions over the next five years.

This surge in demand is reshaping the market. Data center REITs and developers are seeing their valuations climb as investors price in sustained growth.

As AI models improve faster than expected, infrastructure needs are accelerating. Industry experts predict AI progress will compress a decade’s worth of gains into just six months, pushing compute demand curves even higher.

This means financing rounds like Switch’s and Meta’s will become more common. Investors want to get ahead of the massive capital expenditures that will ripple through the tech and real estate markets.

In short, the AI boom isn’t just about software breakthroughs. It’s about building the physical spaces where the future’s smartest machines live and work. And that race is just starting.