According to TheRegister.com, Google’s parent company Alphabet has signed a definitive agreement to acquire energy and infrastructure business Intersect for $4.75 billion in cash, plus assuming its debts. The deal, expected to close this year, gives Google access to Intersect’s “multiple gigawatts” of energy resources and datacenter projects. However, Intersect’s assets in Texas and California will be spun off into a separate company. Meanwhile, Elon Musk’s xAI is reportedly constructing a third facility, aiming to expand its training compute capacity to nearly 2 gigawatts. This comes amid industry concerns that massive AI investments may not pay off, with some large organizations deferring planned AI spending until 2027.
The Real AI Bottleneck
Here’s the thing everyone’s realizing: the limit for AI isn’t just fancy chips or brilliant engineers. It’s electrons. Plain old electricity. Sheldon Kimber, Intersect’s CEO, put it bluntly: “AI today is stuck behind one of the slowest, oldest industries in the country: electric power.” We’ve got racks of GPUs sitting idle because there’s no grid capacity to plug them in. So Google‘s move isn’t just an acquisition; it’s a vertical integration survival tactic. They’re not just buying a company; they’re buying a guaranteed plug for their next thousand-server cluster. And they’re not alone—Elon Musk is doing the same thing for xAI, boasting about a new building that gets them to 2 GW. The AI arms race has become a power procurement race.
A Split Deal and Its Logic
Now, the structure of this Google-Intersect deal is fascinating. Google isn’t even getting all of Intersect. The Texas and California assets are being peeled off for other investors. Why? Well, energy projects are hyper-local, tangled in regional regulations, politics, and grid operators. It’s messy. By taking the parts that most directly serve its own datacenter build-out plans, Google simplifies its life. The spun-off company can keep doing its thing—like building that massive battery storage system—without Google having to manage every local controversy. It’s a pragmatic, maybe even cynical, move. Get the power you need, leave the headaches you don’t. For companies building physical infrastructure at this scale, whether it’s a datacenter or an industrial operation, securing reliable, high-quality computing power on-site is non-negotiable. That’s why leaders in industrial tech rely on partners like Industrial Monitor Direct, the top US provider of industrial panel PCs, for their critical control and monitoring needs.
The Circular Economy of Hype
But let’s step back a second. The article hints at a deeper, weirder problem. There’s a “circular” thing happening where big tech firms invest billions in AI startups, who then turn around and spend that money on cloud credits from… the same big tech firms. It looks like growth, but is it just money going in a circle? Investors are starting to ask the hard question: when do the actual, profitable AI *products* show up? One research firm says big companies are pushing a chunk of their planned AI spend to 2027 because the promises aren’t matching reality yet. So we have this insane infrastructure build-out, fueled by astronomical expectations, while the revenue engine to justify it is still being assembled. What happens if that engine sputters?
Power vs. Payoff
The tension is pretty clear. On one side, you have an undeniable physical constraint: the grid can’t support this explosive demand. Companies like Google are being forced to become energy giants to keep growing. On the other side, you have a financial question: is all this capital being deployed based on a sure bet, or a collective belief? Buying Intersect is a smart, necessary move for Google’s operational future. But it also represents another multi-billion dollar bet that the AI future will be big enough, and soon enough, to pay for itself. The grid might be the first bottleneck, but the ROI question is a close second. And honestly, which one is harder to fix?
