No One Holds It All
AI, space, and energy have become one race — and no country holds everything it takes to win. Only one place on Earth holds all the pieces and stays free.
In February 2026, a rocket company bought an artificial-intelligence company for the largest sum in the history of private business. SpaceX absorbed xAI in a deal that valued the combined firm at about $1.25 trillion.¹ Within months it folded the AI in as its own division and filed with the U.S. government for permission to put up to a million satellites in orbit — not to carry the internet, but to run data centers in space.²
You do not have to admire the people involved to read the signal. That is real money — more than the yearly output of most countries — all moving in one direction. Capital that size does not chase a slogan. It chases a certainty. And the certainty is this: artificial intelligence has become a physical build so vast that the most vertically integrated company on Earth is trying to leave the planet to keep building it.
Here is the tell, though. Even that company cannot do it alone — not the compute, not the energy, not the machines, not the minerals. If a giant with a trillion dollars and its own rockets cannot assemble the whole thing by itself, that is the first clue to the real shape of this race. It is bigger than any company. And, as it turns out, bigger than any country.
Start by seeing the race for what it has become. We talk about the AI race, and separately about a new space race, and separately again about an energy crunch. They are not three races. They are one. Artificial intelligence runs on computing power, computing power runs on electricity, and electricity has become the wall everyone hits first. Space is both its own frontier and the place that computing may have to go to find the power and cooling the Earth can no longer spare. Energy sits underneath all of it. You do not lead in one of these anymore without leading in all three. Win the whole stack, or win nothing.
Every one of those frontiers runs on the same short list. Call it the stack: capital to fund it, the best AI models, the chips to run them, the compute to house them, the energy to power it, the critical minerals inside every machine, the factories to build the hardware, the rockets to reach orbit, and — beneath all of it — the people to design, build, and operate the whole thing.
No nation on Earth holds the entire list. Not one. Not the United States. Not China.
The question is not who is ahead this morning. It is who can assemble the complete stack — and do it without risking the freedom of the human being. Too much has been paid, over too long, to win freedom in this world to now hand all the power of our data to a country that does not believe in it.
Most of the world’s great powers are in this race. But two are structurally built to reach the end of it: China and the United States. And they are built in opposite ways.
China assembles the stack by command, inside one border. One government, one plan that outlasts any single leader, holding the world’s manufacturing floor, most of its mineral refining, an enormous buildout of power, and a state-run constellation of AI satellites already going up — 2,800 planned, the first dozen already in orbit, built to process data in space itself, edge computing in orbit.³ It is the most complete stack any single country has ever tried to build alone. And it is built the way a controlled system builds everything: from the top, by order.
The United States holds the other half of the world’s advantages — the frontier AI models, the capital, the software, the rockets, and more raw computing power than anyone else on Earth.⁴ What it does not hold is the physical base. Its industrial spine is thin. Its workforce is aging. It does not fabricate its own most advanced chips; almost all of them come from a single company in Taiwan.⁵ And it does not mine or refine its own critical minerals — China refines more than 70% of them and controls roughly 90% of the rare-earth magnets that go into everything from fighter jets to electric motors.⁶ The brains are American. The body is somewhere else.
This is not a new discovery in Washington. The United States has spent years trying to pull its supply chains out of China, ever since it learned how dependent it had let itself become — and it is one of the few things both American parties agree on. Tariffs, the CHIPS Act, executive orders, a scramble for nearshoring and for bringing factories home. The direction is set, and it is not reversing.
But you cannot decouple into thin air. When China restricted rare-earth exports in 2025, it took only weeks for a major American automaker to shut a plant for lack of magnets.⁷ A stack built entirely at home is a fantasy — the United States does not have the minerals in the ground or the chip factories in the desert to do it, and building them will take a decade this race will not wait through.⁸ So the honest question was never whether to decouple. It is: decouple to where?
There is only one answer that is near, allied, already integrated, and physically complete. Not across an ocean — on the same landmass. The pieces the United States is missing are, most of them, next door. North and south.
Look at what the neighbors actually hold.
Canada holds the parts the American stack is short on: energy at continental scale; critical minerals in the ground from a stable, allied supplier — exactly the diversification the whole West is now scrambling to find;⁹ and the continent’s largest supply of fresh water, a resource that in the age of data centers has stopped being a footnote and become strategic.¹⁴ It holds something less obvious, too. The deep-learning breakthrough that made this entire moment possible was kept alive, through the years most of the world ignored it, largely inside Canadian universities — by the researchers who went on to win computing’s highest honor.¹⁰ Canada has been in the brains of this race from the start.
Mexico holds the build. Its factories already make the hardware of North America; its aerospace clusters in Querétaro and Sonora turn out engines and precision components for the world’s largest manufacturers, anchored by a dedicated aerospace university and more than twenty-five thousand aerospace-related graduates every year.¹¹ These are not cheap hands. They are engineers and builders — the youngest large workforce on the continent, at a median age near thirty while the United States and China both age past forty.¹² A country that graduates that many engineers is not built to sell its labor. It is built to own a share of what it makes.
And the United States holds what it has always held: the capital, the frontier science, the chip design, the software, and the launch.
Set them side by side and the arithmetic is almost embarrassingly simple: the piece each one lacks, another one has. Together, the three nations close the gaps each of them carries alone — water, energy, minerals, access to two oceans, people, and innovation. Apart, three incomplete stacks. Together, the only complete one on Earth that can be built by free people, under three flags, without asking Beijing’s permission for anything.
Seen this way, what North America can secure together has a name: continental security, continental energy, continental water, and the continental talent — the skilled hands and the trained minds — to compete with the entire world. None of the three holds it complete on its own. The three together do. That is what this moment is really about: for the first time in a long while, there is a clear, cold case in which together beats apart. Not out of sentiment — out of self-interest.
And here is what should end the argument: we already run this machine. North America is already one of the largest integrated economies on the planet — a single automotive part can cross these borders as many as eight times before the car is finished. We built the machine decades ago. We just operate it as three separate shifts instead of one company, and we have never once decided to aim it at the frontier on purpose.
Which brings up the thing we have to say plainly. The story of the United States, Mexico, and Canada cannot keep being only a story about drugs, cartels, addiction, migration, and trade deficits. Those are real. But while the three capitals argue about the border, the border itself is quietly carrying the supply chain that will decide whether a free world or a controlled one writes the rules of this century. The small story is starving the large one of air.
And there is a contradiction at the center of it that no one should pretend away. The United States is trying to decouple from China for its own security — and at the same time picking fights with, and laying tariffs on, the very neighbors who hold the pieces that would complete that decoupling.¹³ You cannot regionalize away from a rival while treating your own region as a threat. That is not strategy. It is the short-sighted politics that could hand the century to the one bloc that plans in decades.
And this is where the USMCA renegotiation — the one that looks like it is going from bad to worse — could serve something larger than it appears. It is the table where the three could shield, as a shared security concern for all of them at once, the pieces of the stack that decide the race: energy, minerals, compute, critical manufacturing. Not each one invoking “national security” against its neighbor — which is exactly what is happening now — but the three defining a single continental security. That takes setting the egos and the short-term political strategies aside, and putting first what is actually at stake: the AI and space race.
None of this rests on space arriving on schedule. Maybe orbit will carry the world’s computing and power one day; maybe it takes far longer than the announcements suggest, or never fully pays off. That validation is years away, and honest people should say so.
But that is exactly why the continent matters now. While we wait to learn whether the answer is above us, the answer is already beside us. North America is the near-term, provable base for leadership in AI, space, and energy — whether or not the orbital bet ever fully comes in.
Nobody knows what happens next. Not who wins, not what the finish line even looks like, not which technology turns out to matter most. That uncertainty is the point. In a race this long and this unclear, you do not bet on a single genius or a lucky break. You bet on the most complete and resilient base — the one place that holds every piece and can still build with free hands.
That place is North America. The stack is here.
Setting the egos aside, putting the AI and space race first, and running the machine as one team — that is what would bring closer the day this race is led by our hemisphere, and not another. Whether we secure the stack, or not, is the choice each of our nations still has to make. And that is where this series goes next.
Sources
CNBC — SpaceX–xAI merger valued at ~$1.25 trillion (Feb 2026).
FCC (Doc DA-26-113) / SpaceNews — SpaceX “Orbital Data Center System,” application for up to one million satellites (filed 30 Jan 2026).
SpaceNews; South China Morning Post; Live Science — China’s Three-Body Computing Constellation, 2,800 satellites planned, first 12 launched 14 May 2025.
War on the Rocks — US holds ~5× the aggregate AI-training compute of China and leads in high-end GPUs and design software.
Forbes; Bruegel — TSMC produces ~97% of the world’s most advanced semiconductors.
U.S.-China Economic and Security Review Commission (2025) — China refines >70% of critical minerals and controls ~90% of rare-earth permanent magnets.
GQG Partners; Chicago Council on Global Affairs — China’s 2025 rare-earth export restrictions forced a U.S. automaker to idle a plant within weeks.
Chicago Council; Andersen Institute / CFR — meaningful processing independence requires sustained investment “not yet mobilized”; a decade-scale effort.
FP Analytics; Canada’s federal Critical Minerals Strategy — substantial reserves; a stable, allied supplier.
CIFAR / Turing Award — the deep-learning revival was sustained largely in Canadian universities; its leaders won the 2019 A.M. Turing Award.
Prodensa; American Industries Group — Mexico’s Querétaro/Sonora aerospace clusters, dedicated aerospace university (UNAQ), 25,000+ aerospace-related graduates per year.
UN Population data / national statistics — median age ~30 (Mexico) vs. ~40 (US and China).
U.S. Section 232 tariffs affecting Mexican and Canadian inputs; concurrent trade disputes with both neighbors.
Environment Canada; UNU-INWEH — Canada holds a large share of the world’s fresh water; the U.S. Southwest (Colorado River, Lake Mead) and northern Mexico face water stress, worsened by data-center water use.