[11/52] Discussing AI as largest customer for power & energy: Sue Ennis is Head of IR and Govt Affairs with Hut8. A $5B micro-cap stock that just announced a big deal with Google, Anthropic
AI is at the heart of the technology war, and sadly, the USA is constrained by critical factors right now; permitting, zoning and power procurement. Which is opening up the door for China.
Mar 02, 2026
“Hey! Welcome to my blog, Don’s Daily!”
Hey, I’m Don! I’ve been building career memories for a decade. I love business, I love investing, I love to learn new things.
Share with a friend to support my ideas.
I saw a headline in December about a company (Hut8) that just did a ~250 Megawatt deal with Google, Anthropic and Fluidstack.
So I cold messaged their Head of Investor Relations, Sue Ennis, who was nice enough to help educate myself, and all my viewers about how AI is becoming the #1 customer of power. She was nice enough to hop on a call and explain it to us all. Interview on Youtube below, or watch on X.
She’s an expert on AI x Energy given her role. I specifically asked her to ‘explain this to my newsletter readers like we are a fifth grader’ so that it is easy enough for anyone (aka, me) to understand. It’s critically important we begin learning about this National power constraint. (FOLLOW Sue Ennis! She’s amazing!!!)
Take a look at this graph, this is a problem.
The USA needs to unite. I’m sick of the division. We are all going to regret hating each when we realize it’s come at a cost much greater than political discrepancies. I would rather not live in a world where every country uses China’s AI tools because they are the only ones who could generate the power necessary. That would be bad.
I’ve been grinding VR startups for a full decade... raised the money the hard way, built the virtual worlds, the avatars, the whole thing. But last September I finally walked away. RIP. Lost passion, shut down my founder group, and started researching AI instead.
VR’s biggest problem was always hardware. Plain and simple. You could build whatever you wanted but if Meta can’t sell the headsets then nobody’s strapping on the face-bricks. Their track record... not great.
AI though?
AI’s biggest problem is power.
*clears throat, regretting that I’m about to reference Ray Dalio*
Ray Dalio said recently...
“Whoever wins the technology war is going to win the military war.”
And, he’s correct.
Therefore, whoever wins the military war wins the geopolitical and economic war. Being the leader in power and technology is not a ‘nice to have’. It’s critical for a nation’s independence.
Our economy is in the midst of a violate shift away from Office Buildings and into Data Centers.
I thought to myself... almost every single person in the world already owns 1+ AI devices. Phones. Computers. Even the Alexas sitting on the counter. VR was capped by hardware sales no matter how good content got. AI felt different. Or so I thought.
But the deeper I went the more I realized AI isn’t just software. It has a real physical problem, too. Not hardware exactly. It’s energy. Construction. Government approvals. Rare earth materials. All the real world stuff.
You can see the importance of Energy reflected in the Stock Market returns for 2026 so far.
To actually learn how this is all unfolding, I cold messaged two people who live in this world every day.
Sue Ennis of course (mentioned above), and also one unnamed guy who’s built hundreds of megawatts for Bitcoin miners over the past decade.
I asked him... “can you teach me?”
And we hopped on a Zoom call.
The Bitcoin guy told me back in the mining days he was generating about $50K of revenue per megawatt. Now with AI he’s selling each megawatt for $500K+. Step function.
But he also said the biggest blocker isn’t the demand, it is the antiquated grid and the political headwinds.
Cities are worried about energy prices for regular citizens. Protests. Permit denials. All of it.
Sue echo’ed this point by expelling howTexas is already over capacity. Roughly 250 GW of demand applications came in during 2025 chasing only 180 GW on the ERCOT grid. So companies are forced to start looking at Louisiana, or wherever else they can find power.
Meanwhile, Datacenter’s in the USA are being canceled left and right. Over 25 were canceled in Q1 2026 alone so far, equating to roughly 5 Gigawatts of capacity, gone.
And it’s not just her Hut8 team. Look at CEO’s from Nvidida, AMD, Tesla, Microsoft, Meta, Google…. it’s getting scary.
They are all saying the exact same thing.
Jensen Huang at Nvidia...
“power generation and grid capacity.”
Lisa Su at AMD...
“scaling massive data centers within power limits.”
Sundar Pichai at Google...
“local building permits and a lack of electricians.”
Other CEO’s echoing the same message:
Elon Musk...
“the electrical grid growing too slowly.”
Satya Nadella at Microsoft...
“fully permitted buildings with power already connected.”
Mark Zuckerberg at Meta...
“securing raw power before the competitors do.”
Andy Jassy at Amazon...
“decade-long wait times for grid connections.”
Michael Dell...
“customers lacking office power and liquid cooling.”
Larry Ellison at Oracle...
“the need for gigawatt-scale power plants.”
The list goes on and on… we are CONSTRAINED! And the people at the top know it, because they are constrained by it.
p.s. Here’s a very funny parody video about all this, but also, plausible..
After a couple months studying all this I’m very optimistic we can figure this out, but also acknowledge that AI is China’s chance to leapfrog the USA. And we are not going to let that happen!
They’re less limited by energy materials and politics in the short term.
We’re less limited in capital controls, military, and global influence.
So it’s a race. USA has to get the energy, materials, and politics fixed before China gets the capital, military, and influence side sorted.
But sadly the USA seems to be getting in our own way. A report just came out this week from Bloomberg stating data center construction is falling even as demand for AI computing is soaring. Not good.
“Construction of new data centers in the US fell for the first time since 2020 despite soaring demand for artificial intelligence computing capacity, as developers face delays in permitting, zoning and power procurement.”
Capacity under construction dropped to 5.99 gigawatts at the end of 2025. Down from 6.35 gigawatts at the end of 2024.
That’s a decline of about 0.36 GW or roughly 5.7 percent.
Meanwhile China is adding around 7 GW of new data center capacity every year.
The slowdown is happening even with the massive AI-driven boom pulling for more.
The three big reasons? Grid constraints. Regulatory friction. Power availability challenges.
Exact same three I kept hearing from the smart people I talked to.
We are going to be capacity constrained. And soon.
