The Data Center Power Crisis: 3 Stocks Solving the Energy Gap

The AI boom has a dirty little secret: there's not enough electricity to power it.

I'm not being dramatic. Goldman Sachs $GS ( ▼ 0.5% )  projects data center power consumption will surge 165% by 2030 compared to 2023 levels. That's not a typo. And here's the kicker: expanding electrical grids to support this demand can take a decade.

The AI revolution is moving at light speed. The power grid? It's stuck in traffic.

This mismatch is creating one of the most asymmetric investment opportunities I've seen in years. While everyone fights over chip stocks, a handful of companies are quietly solving the energy gap that could bottleneck the entire AI buildout.

Today, I'm breaking down three of them.

⚡ The Scale of the Problem

Let's zoom out for a second.

By 2030, data centers, AI infrastructure, and cryptocurrency mining could consume 945 TWh annually. For context, that's roughly equivalent to Japan's entire electricity consumption. One country's worth of power: just for servers.

The demand isn't theoretical. It's already here.

Microsoft $MSFT ( ▼ 0.74% ), Google $GOOG ( ▼ 0.04% ), and Amazon $AMZN ( ▼ 1.01% )  are in a datacenter arms race. They're signing multi-billion dollar deals to secure power before their competitors do. Some hyperscalers are even exploring on-site nuclear reactors because they can't get enough juice from the traditional grid.

When Big Tech starts talking about building nuclear plants, you know the power crisis is real.

This is where it gets interesting for investors. The bottleneck isn't chips anymore: it's electrons. And the companies solving this problem are flying under the radar while NVIDIA $NVDA ( ▼ 0.72% )  dominates every headline.

I wrote about this shift in depth in The 2026 AI Infrastructure Playbook. If you haven't read it yet, I'd start there. It lays out the full "Beyond the Chip" thesis.

But today, we're drilling into the power pillar specifically.

🔋 Stock #1: Caterpillar $CAT ( ▼ 1.19% )  

The "Gold Standard" for Backup Power

When a $2 billion data center can't afford a single second of downtime, who do they call?

Caterpillar.

$CAT isn't sexy. It's not an AI stock in the traditional sense. But every hyperscale data center needs diesel and natural gas generators as backup power: and Caterpillar dominates that market.

Here's what most investors miss: data centers don't just need backup power for emergencies. They need it for grid stabilization. When the local utility can't deliver consistent power (which happens more often than you'd think), those CAT generators kick in.

The investment case:

• Market position: Caterpillar owns roughly 20% of the global generator market. In enterprise-grade backup power, their share is even higher.

• Recurring revenue: Generators need maintenance, parts, and eventual replacement. This isn't a one-time sale: it's a long-term relationship.

• Pricing power: When uptime is worth millions per hour, customers don't haggle on generator costs.

The Latte Metric to watch: Order backlog in their Energy & Transportation segment. If it keeps climbing, data center demand is doing the heavy lifting.

$CAT won't 10x. But it's a steady compounder with a massive tailwind that the market hasn't fully priced in.

🔌 Stock #2: Eaton $ETN ( ▼ 0.83% )  

The Backbone of Power Distribution

If Caterpillar generates the backup power, Eaton makes sure it actually gets where it needs to go.

$ETN specializes in power distribution, UPS (uninterruptible power supply) systems, and electrical components. They're the "plumbing" of the data center power stack.

Think about it this way: a modern AI data center is basically a small city's worth of electrical infrastructure crammed into one building. Someone has to design, build, and maintain that system. Eaton is that someone.

The investment case:

• Pure-play electrical: Unlike diversified industrials, Eaton is laser-focused on power management. When data center spending surges, they benefit directly.

• Margin expansion: As AI workloads require more sophisticated power solutions, Eaton can charge premium prices for higher-spec equipment.

• Secular tailwind: Electrification isn't just a data center story: it's EVs, renewable energy, and grid modernization. Eaton wins across all of them.

The Latte Metric to watch: Data center revenue as a percentage of total sales. If this number keeps ticking up, the AI thesis is playing out.

One thing I love about $ETN: they're already embedded in the hyperscaler supply chain. Microsoft, for example, is using grid-interactive UPS systems that leverage battery storage to stabilize the public grid during peak hours. Guess who supplies a lot of that infrastructure?

🏗️ Stock #3: Quanta Services $PWR ( ▼ 1.82% )  

The Engineers Building the Physical Grid

Here's a question: who actually builds the power lines, substations, and electrical infrastructure connecting data centers to the grid?

Quanta Services.

$PWR is the largest specialty contractor in North America for electric power infrastructure. They're the boots on the ground making the energy transition physically happen.

This is the most "picks and shovels" play of the three. Quanta doesn't manufacture anything: they build and maintain the infrastructure that everyone else depends on.

The investment case:

• Irreplaceable expertise: You can't outsource this work overseas. Grid construction requires specialized crews, local permits, and decades of institutional knowledge. Quanta has all three.

• Backlog visibility: Large infrastructure projects take years to complete. Quanta's order backlog gives investors unusual visibility into future revenue.

• Multiple tailwinds: Data centers, renewable energy integration, EV charging infrastructure, grid hardening: every major energy trend benefits Quanta.

The Latte Metric to watch: Total backlog growth, especially in their Electric Power segment. A growing backlog means years of locked-in revenue.

The risk here is execution. Quanta is a project-based business, which means margins can get squeezed if they bid too aggressively or face labor shortages. But the demand environment has never been stronger.

📊 Why Power Stocks Are Mispriced

Here's my read on the situation.

The market is obsessed with the "software layer" of AI: the models, the chips, the apps. That's where the hype is. That's where the CNBC segments focus.

But AI doesn't run on hype. It runs on electricity.

And right now, there's a fundamental mismatch between the pace of AI adoption and the pace of power infrastructure buildout. That gap has to close. The companies closing it are $CAT, $ETN, and $PWR.

These aren't moonshots. They're established, profitable businesses with strong balance sheets and decades of operating history. They just happen to be sitting directly in the path of one of the biggest capital expenditure cycles in modern history.

The thesis is simple: the best growth stocks in 2026 won't be chip makers: they'll be power suppliers.

🎯 The Bottom Line

The data center power crisis is real. It's not going away anytime soon. And the companies solving it are still trading at reasonable valuations while AI chip stocks sit at nosebleed multiples.

I'm not saying dump your NVIDIA shares. I'm saying the opportunity set is wider than most investors realize.

If you want the full picture:

I put together The 2026 AI Infrastructure Playbook: a deep dive into the four pillars of AI infrastructure beyond the chip: Power, Cooling, Real Estate, and Connectivity. It's the framework I'm using to find overlooked opportunities in this cycle.

The AI buildout is a decade-long story. We're still in the early innings.

Stay curious,

George ☕️