It's a truly incredible time to be alive. Who would have thought fifty years ago that we would be talking about launching humans into space as if they were catching a taxi to go to the mall? What would Issac Asimov say if he walked into a house in a local neighborhood and saw that the door had been opened by a robot? That's the reality that we're living in right now. Sitting on my porch and reading Peter's post, it's all just so surreal. It's happening so fast that most people don't see it. The dualities of fear and wonder fight for our attention at the same time, it's difficult to know where we'll be in just a few months, let alone a year or two. Great post, Peter, loved the recent podcast too!
Sharing this news with family, Peter realize only God=God and we need the 1 & only Savior Jesus Christ. The greatest truth never changed, sealed in God’s Word, the #1 Bible. Take heed friends.
The convergence at the infrastructure level is underappreciated. AI’s exponential gets discussed constantly, but electricity demand projections for AI data centers could reshape geopolitics of energy in 10 years just as much as model capabilities reshape labor markets. The physical constraints are the real ceiling.
Maybe not going to be a problem after all. Data center innovation is happening all the time. The problem is that the MSM all parrot the same story and few do original research on hot trends and new ideas. So everyone thinks the end is nigh. Here is a new idea. Let's get Mythos working on this!
---
This new chip could slash data center energy waste
April 10, 2026
University of California - San Diego
Summary:
A new chip design from UC San Diego could make data centers far more energy-efficient by rethinking how power is converted for GPUs. By combining vibrating piezoelectric components with a clever circuit layout, the system overcomes limitations of traditional designs. The prototype achieved impressive efficiency and delivered much more power than previous attempts. Though not ready for widespread use yet, it points to a promising future for high-performance computing.
The energy section is the one that doesn't get enough attention in these convergence pieces. 410,000 megawatts of filed demand against a grid that did about 85GW peak last summer. Even if half those filings never materialise, the gap between what's been requested and what can actually be delivered is enormous.
The fusion timeline is the real variable though. If it delivers partial commercial output by 2032, the constraint disappears. If it doesn't, we're looking at a decade where AI capability outruns the physical infrastructure to deploy it. Two very different worlds depending on which side of that fork we land on.
Was at BYD’s visitor center in Shenzhen in February. Group of European industrial executives, dark suits, pens hovering.
The technician drove a six-inch nail through a conventional nickel-cobalt cell. Whoomp. Thermal runaway, 500°C, thick smoke. Then the Blade cell, same nail. Nothing. Room went quiet.
Kept thinking afterward about everything upstream of that demo. The suppliers. The decade of capex. The engineers who learned to make lithium iron phosphate cells at that scale.
Look at the humanoid bill of materials now. Actuators, motors, rare earth magnets, battery packs. A million EVs already paid for all of it. The price collapse is the EV supply chain spilling into a new category.
So your 36-month call feels right. The hardware keeps getting cheaper.
What I can’t figure out yet is who gets to train on a million real kitchens.
Like Jack Dorsey, this past year I awake each morning filled with equal measures of existential dread and hope/optimism. I'm never quite sure how I feel after each Moonshots pod, but I gotta get that geeky fix!
If fusion delivers even partial commercial output by 2032 (which multiple credible paths now suggest is plausible) the energy constraint on AI, robotics, and every other exponential technology evaporates. This is the unlock that changes everything.
----
2032? Why does everyone keep ignoring Helion's contract with Microsoft to deliver fusion powered electricity to a MS data center by 2028?
Startup begins work on US fusion power plant. Yes, fusion.
By Peter Behr, Christa Marshall | 07/31/2025 06:18 AM EDT
Helion Energy intends to deliver electricity to Microsoft within three years.
In another sign of fusion energy development’s quickening pace, Helion Energy, a startup with $1 billion in private funding, announced Wednesday it has commenced construction of its first planned power production reactor.
The Washington state-based company said the site work keeps it on track to deliver electricity within three years to Microsoft under a 2023 purchase agreement, and “one step closer” to realizing the vision behind its unique fusion energy technology, said David Kirtley, Helion co-founder and CEO, in a statement.
Microsoft Chief Sustainability Officer Melanie Nakagawa, noting that “the path to commercial fusion is still unfolding,” said the company was proud to support Helion’s pioneering development.
Helion is one of three developers to amass $1 billion in investments, according to the 2025 annual report by the Fusion Industry Association. The company is aiming to build the first U.S. power plant to harness fusion energy for the grid at a site in Malaga, Washington, which is roughly halfway between Seattle and Spokane.
AI gets smarter, robots get cheaper, and suddenly the real constraint shows up in a place nobody was watching closely enough. Energy. You can feel it in that Texas number alone. Demand running ahead of infrastructure by that margin doesn’t resolve neatly. It forces tradeoffs somewhere in the system.
The optimism around fusion is doing a lot of heavy lifting here. It’s the release valve everyone hopes arrives on time. Maybe it does. But until it’s real at scale, the system has to operate under pressure.
What’s interesting is how fast the narrative jumps to abundance. That tends to happen right before friction shows up. Every major shift promises a world with fewer constraints. In practice, the constraints just move.
The opportunity is obvious. So is the tension. The people who do well in this kind of transition aren’t the ones chasing the headline. They’re the ones watching where the bottlenecks are forming before everyone else notices.
It's a truly incredible time to be alive. Who would have thought fifty years ago that we would be talking about launching humans into space as if they were catching a taxi to go to the mall? What would Issac Asimov say if he walked into a house in a local neighborhood and saw that the door had been opened by a robot? That's the reality that we're living in right now. Sitting on my porch and reading Peter's post, it's all just so surreal. It's happening so fast that most people don't see it. The dualities of fear and wonder fight for our attention at the same time, it's difficult to know where we'll be in just a few months, let alone a year or two. Great post, Peter, loved the recent podcast too!
Sharing this news with family, Peter realize only God=God and we need the 1 & only Savior Jesus Christ. The greatest truth never changed, sealed in God’s Word, the #1 Bible. Take heed friends.
The convergence at the infrastructure level is underappreciated. AI’s exponential gets discussed constantly, but electricity demand projections for AI data centers could reshape geopolitics of energy in 10 years just as much as model capabilities reshape labor markets. The physical constraints are the real ceiling.
Re: Energy Red Alert: 7x Demand Energy Surge
Maybe not going to be a problem after all. Data center innovation is happening all the time. The problem is that the MSM all parrot the same story and few do original research on hot trends and new ideas. So everyone thinks the end is nigh. Here is a new idea. Let's get Mythos working on this!
---
This new chip could slash data center energy waste
April 10, 2026
University of California - San Diego
Summary:
A new chip design from UC San Diego could make data centers far more energy-efficient by rethinking how power is converted for GPUs. By combining vibrating piezoelectric components with a clever circuit layout, the system overcomes limitations of traditional designs. The prototype achieved impressive efficiency and delivered much more power than previous attempts. Though not ready for widespread use yet, it points to a promising future for high-performance computing.
...
https://www.sciencedaily.com/releases/2026/04/260409101103.htm
Will robots need to participate in developing homes that accomodate both humans and robots?
Seems only fair. And like lightbulbs, will standards need to be implemented for categories of humanoid robots?
Robots have to fit into human designed houses. The only thing a robot needs is a charging cradle, like 7 of 9 on Star Trek.
Thank you for your work.
👍 concise
The energy section is the one that doesn't get enough attention in these convergence pieces. 410,000 megawatts of filed demand against a grid that did about 85GW peak last summer. Even if half those filings never materialise, the gap between what's been requested and what can actually be delivered is enormous.
The fusion timeline is the real variable though. If it delivers partial commercial output by 2032, the constraint disappears. If it doesn't, we're looking at a decade where AI capability outruns the physical infrastructure to deploy it. Two very different worlds depending on which side of that fork we land on.
$4,900 for Unitree. Peter, that’s the number.
Was at BYD’s visitor center in Shenzhen in February. Group of European industrial executives, dark suits, pens hovering.
The technician drove a six-inch nail through a conventional nickel-cobalt cell. Whoomp. Thermal runaway, 500°C, thick smoke. Then the Blade cell, same nail. Nothing. Room went quiet.
Kept thinking afterward about everything upstream of that demo. The suppliers. The decade of capex. The engineers who learned to make lithium iron phosphate cells at that scale.
Look at the humanoid bill of materials now. Actuators, motors, rare earth magnets, battery packs. A million EVs already paid for all of it. The price collapse is the EV supply chain spilling into a new category.
So your 36-month call feels right. The hardware keeps getting cheaper.
What I can’t figure out yet is who gets to train on a million real kitchens.
Good piece, Peter. Sending to my Shenzhen cohort.
Like Jack Dorsey, this past year I awake each morning filled with equal measures of existential dread and hope/optimism. I'm never quite sure how I feel after each Moonshots pod, but I gotta get that geeky fix!
Will humanity and singularity be decided enough to restrict abuse of these powerful tools?
If fusion delivers even partial commercial output by 2032 (which multiple credible paths now suggest is plausible) the energy constraint on AI, robotics, and every other exponential technology evaporates. This is the unlock that changes everything.
----
2032? Why does everyone keep ignoring Helion's contract with Microsoft to deliver fusion powered electricity to a MS data center by 2028?
Startup begins work on US fusion power plant. Yes, fusion.
By Peter Behr, Christa Marshall | 07/31/2025 06:18 AM EDT
Helion Energy intends to deliver electricity to Microsoft within three years.
In another sign of fusion energy development’s quickening pace, Helion Energy, a startup with $1 billion in private funding, announced Wednesday it has commenced construction of its first planned power production reactor.
The Washington state-based company said the site work keeps it on track to deliver electricity within three years to Microsoft under a 2023 purchase agreement, and “one step closer” to realizing the vision behind its unique fusion energy technology, said David Kirtley, Helion co-founder and CEO, in a statement.
Microsoft Chief Sustainability Officer Melanie Nakagawa, noting that “the path to commercial fusion is still unfolding,” said the company was proud to support Helion’s pioneering development.
Helion is one of three developers to amass $1 billion in investments, according to the 2025 annual report by the Fusion Industry Association. The company is aiming to build the first U.S. power plant to harness fusion energy for the grid at a site in Malaga, Washington, which is roughly halfway between Seattle and Spokane.
...
https://www.eenews.net/articles/startup-begins-work-on-major-us-fusion-power-plant-yes-fusion/
AI gets smarter, robots get cheaper, and suddenly the real constraint shows up in a place nobody was watching closely enough. Energy. You can feel it in that Texas number alone. Demand running ahead of infrastructure by that margin doesn’t resolve neatly. It forces tradeoffs somewhere in the system.
The optimism around fusion is doing a lot of heavy lifting here. It’s the release valve everyone hopes arrives on time. Maybe it does. But until it’s real at scale, the system has to operate under pressure.
What’s interesting is how fast the narrative jumps to abundance. That tends to happen right before friction shows up. Every major shift promises a world with fewer constraints. In practice, the constraints just move.
The opportunity is obvious. So is the tension. The people who do well in this kind of transition aren’t the ones chasing the headline. They’re the ones watching where the bottlenecks are forming before everyone else notices.
Awesome article Peter...
Say goodbye to Capitalism and say hello to Intelligenism!
Thanks. Excellent!
Great work Peter and your team from a participant 😉