Everyone’s worried AI will take over the world. But right now, AI can barely keep the lights on.

Welcome to the great AI crunch.

AI isn’t hitting one wall—it’s hitting five:

An energy crisis.

An AI compute crisis.

An AI storage crisis.

An AI skills crisis.

An AI security crisis.

If crisis is the fire, then opportunity is the forge. That's precisely where we're zeroing in.

And we’re not the only ones.

Right now, every genius hedge fund is diving deep down the rabbit hole of AI infrastructure.

Steve Cohen's Point72 Asset Management plans to raise $1 billion specifically for investing in AI companies

Bridgewater Associates, Man Group Plc, Highbridge Capital Management, and Simplex Asset Management are actively developing or investing in machine learning technologies.

Meanwhile, Meta’s planning a massive $60-65 billion investment in AI infrastructure for 2025. And Amazon Web Services (AWS) is investing at least $11 billion in Georgia to expand AI and cloud computing infrastructure.

Why? Because the demand is going nuclear, and our current solutions can’t keep up.

Before we get into the numbers—and solutions—there’s one thing you should know:

FREE Live “Turning Point” Event

Next week, Team Altucher is headed to Austin, Texas to talk opportunities and portfolio-swelling solutions—LIVE.

This one-of-a-kind event will be streamed to you absolutely FREE of charge on Tuesday, March 11th at 2pm ET. And it’s all taking place online, so you can watch the panel in action from the comfort of your own home.

This won’t be anything like our normal weekly calls… this is a HUGE event with the entire Altucher team.

And we’re going deep on the biggest topics in tech right now, including…

→ Are artificial intelligence stocks in a bubble that’s ready to pop?

→ How will Robert F. Kennedy Jr. shake up drugmakers and biotechs?

→ Can quantum computers become the next trillion dollar-industry, or are they just sci-fi hype?

We’ll even share our BIGGEST predictions for the rest of the year and beyond.

We’ll be in touch soon with all the details on how to access the livestream. But for now, mark this time off on your calendar and prep your notepad.

AI’s Five-Alarm Fire

Let’s start with energy:

U.S. data centers alone consumed 176 TWh in 2023 (4.4% of all U.S. electricity), and projections show this soaring to 325-580 TWh by 2028.

Texas alone currently hosts over 340 data centers, consuming nearly 8 gigawatts of power, which accounts for 9% of the state's electricity demand.

It’s estimated Texas will need to add electricity equivalent to 30 nuclear reactors by 2030 to meet soaring demand from data centers.

U.S. power consumption as a whole is expected to reach record levels in 2025 (4,179 billion kWh) and 2026 (4,239 billion kWh), surpassing the 2024 record of 4,082 billion kWh.

Compute:

Beyond energy, compute hardware supply is also becoming another pencil-thin bottleneck.

The most advanced AI chips are in short supply worldwide.

This is creating compute crunch: even with abundant capital, cloud providers can’t get enough high-end chips quickly. Independent AI developers face long waits or high prices for GPU instances.

Demand for AI compute is far outpacing the ability of traditional centralized infrastructure to scale.

Storage:

Over the past decade, the amount of data worldwide has skyrocketed, jumping from 2 zettabytes in 2010 to an estimated 147 zettabytes in 2024.

That is 402.89 million terabytes created every day.

To put this into perspective, one HD movie is about 5 gigabytes. Meaning, we’re producing 82,330 HD movies worth of data per day. 

You'd have to watch over 57 movies every minute—nearly one movie per second—just to keep pace with the daily data creation.

If you tried downloading all the data created last year using a typical home internet connection, you would need billions of years—even longer than the universe has existed.

And here’s the thing:

Data storage capacity requirements are expected to surge by 150% in the next couple of years.

That’s assuming everything stays as-is (a risky assumption) and there isn’t some big breakthrough in AI or quantum computing that requires even more data.

Traditional storage systems are already struggling to cope with the volume, velocity, and variety of data generated by AI applications.

Skills:

IBM's 2025 annual report revealed that 38% of their planned AI projects were delayed due to a shortage of qualified AI specialists, despite offering salaries 30% above the industry average.

They’re not the only ones.

Job postings surged 2,000% in 2024.

And, according to a survey conducted by Pluralsight, only 23% of executives say they have the AI skills they need to stay ahead.

By the end of 2026, almost all companies (more than 90%) are expected to experience significant challenges due to a shortage of AI skills among their workforce.

This skills gap is not just a minor inconvenience - it's projected to have a massive economic impact, potentially costing the global economy up to $5.5 trillion.

Finally, the big kahuna…

Security:

In the U.S., many data centers are clustered in a few geographic hot spots, such as northern Virginia’s “Data Center Alley.”

This creates what security experts call "juicy targets" and what I call "putting all your digital eggs in very expensive, very hackable baskets."

In 2023, the FBI warned that Chinese hackers are already "pre-positioning" in American infrastructure, essentially leaving digital time bombs.

A successful cyber attack on one of the major cloud regions (for example, an AWS us-east data center cluster) could cripple AI services nationwide.

Physical incidents—whether sabotage, natural disasters, or accidents—at a handful of key data centers could also have outsized impacts.

YET…

Retrofitting or “hardening” these sites is extremely costly and slow. (Current regulations only partially address these risks.)

The Solution

In short, today’s AI infrastructure has a “too big to fail” problem.

Historically, when faced with "too big to fail" architectures, elegant solutions have often emerged not through incremental fixes…

But through fundamentally rethinking the architecture itself.

More on that, though, tomorrow.

And much more on that on Tuesday’s LIVE event in Austin.

Stay tuned for all the details.