The Numbers Are Staggering — and Nobody Knows Where It Ends

Big Tech will spend nearly $700 billion on AI infrastructure in 2026 alone, according to Fortune. That's one year.

Over five years, Goldman Sachs Global Institute — co-authored by George Lee and Lucas Greenbaum — puts total AI capital expenditure estimates at $4 trillion to $8 trillion. The range alone reflects genuine uncertainty about where this spending will ultimately peak.

Blackstone just committed $5 billion in equity to a new AI infrastructure joint venture with Google, according to CNBC. The deal brings 500 megawatts of compute capacity online by 2027, powered by Google's in-house TPU chips — a direct shot at Nvidia's stranglehold on AI hardware. The venture will be run by Benjamin Treynor Sloss, former Google chief programs officer.

Solar Is Winning. Gas Is Still Cashing Checks.

The green energy transition and the fossil fuel boom are happening at the same time, not instead of each other.

BloombergNEF projects solar will become the largest single source of electricity by 2035, surpassing coal, oil, and natural gas combined. Matthias Kimmel, head of energy economics at BloombergNEF, told TechCrunch that this shift is happening on pure economics — solar is simply cheaper, and prices are expected to drop another 30% by 2035.

Simultaneously, data centers are driving 370 gigawatts of new natural gas capacity and 110 gigawatts of new coal capacity, per BloombergNEF. Because gas and coal can run 24 hours a day, 7 days a week — and solar can't — fossil fuels are projected to supply 51% of incremental electricity generation for data centers through 2050.

The AI revolution is accelerating solar deployment while extending fossil fuel dependency. Both trends are occurring in parallel.

Your Electricity Bill Is Going Up. Here's Why.

U.S. electricity demand was essentially flat for two decades. Then ChatGPT happened, according to 24/7 Wall Street analyst Don Lair.

Building new transmission lines takes years. Permitting a natural gas plant takes longer. The grid wasn't designed for this surge in demand.

Traders on Polymarket are currently putting a 93.5% implied probability on at least one qualifying AI data center moratorium passing into law by end of 2026. With WTI crude sitting at $99.89 per barrel as of April 27, 2026, fuel costs for marginal power supply are climbing.

Major data center clusters are concentrated in Virginia, Texas, Ohio, and Arizona. The strain is local. The bills are local. The profits flow elsewhere.

China Subsidized Solar's Rise — Now It's Eating Everyone's Lunch

Solar's dramatic cost collapse stemmed from two factors: China's industrial policy — direct subsidies to manufacturers, flooding the global market — and mass manufacturing economies of scale, according to BloombergNEF's Kimmel.

Pakistan added 25 gigawatts of solar in just two years after natural gas prices spiked post-Ukraine invasion. Compelling economics drive rapid deployment.

America is benefiting from cheap solar panels that Chinese state subsidies made possible. The dynamic rarely surfaces in Washington policy discussions.

The Smart Money Is Moving Fast

Armada, a San Francisco-based modular data center builder, just raised $230 million in a Series B round at a $2 billion valuation, per CNBC. Their partner: Johnson Controls, which is building a 400,000-square-foot factory in Arizona called Galleon Forge One, expected to create more than 500 jobs.

Armada's modular units — already deployed by the U.S. Navy and offshore oil rigs — can be set up in days instead of years. Customer bookings grew 540% between FY25 and FY26. Q1 FY27 alone posted roughly 2,000% year-over-year growth.

CEO Dan Wright framed the opportunity in national security terms: "The AI race will not be won by one-off projects. It will be won by the companies and countries that can manufacture, deploy, and continuously improve AI infrastructure, with speed, scale and sovereignty."

On the infrastructure hardware side, Vertiv Holdings posted Q1 2026 adjusted EPS of $1.17 against expectations of $1.01, on $2.65 billion in revenue — up 30% year-over-year, per 24/7 Wall Street. Americas organic sales were up 44%. The picks-and-shovels strategy is delivering results.

What Goldman Actually Said That Nobody's Reporting

Goldman Sachs' framework — published May 1, 2026 — emphasizes that the $4-8 trillion buildout estimate is NOT a fixed number. It's highly sensitive to four assumptions: how long AI chips last before replacement, how expensive next-gen data centers become, what chip architectures dominate, and whether power and labor bottlenecks slow deployment.

"Current estimates of the ultimate scale of the AI build-out are far more conditional than they appear," Goldman wrote.

Everyone citing trillion-dollar figures is working with provisional assumptions. The investment surge is real. The underlying math carries significant uncertainty.

What Comes Next

AI infrastructure represents the biggest capital investment cycle in a generation. Solar will eventually dominate the grid — that's driven by economics, not ideology. But through 2050, natural gas and coal will continue generating baseload power because the sun doesn't shine when servers need it most.

Regular Americans will see higher electricity bills. Tech companies will capture most of the economic upside. The grid wasn't built for this demand, and policymakers lack a comprehensive strategy to address it.

The AI race is accelerating. The energy math is unforgiving. And the costs are mounting.