There is a compelling case to be made that memory chips are becoming as strategically important to the modern economy as oil once was. This is not merely a statement about demand for a hot product; it is a claim about the role these components now play in the foundation of economic power. But for that thesis to hold — even if the torrent of AI spending eventually slows — several things have to break the right way. The central one is energy.
The Energy Equation Beneath the Silicon
It is easy to talk about chips as if they exist in isolation, but they are inseparable from the infrastructure required to power them. Running these chips at scale consumes enormous amounts of energy, and with oil prices currently very high, companies are being forced to look elsewhere to fund and feed this infrastructure. That search has opened the door to a range of alternatives: renewable energy, nuclear power — particularly the small modular reactor programs that have drawn so much recent attention — hydrogen power, and solar.
The underlying reality is simple and stark: energy is expensive, and these chips demand a great deal of it. For the broader thesis to survive a slowdown in spending, the cost and consumption of energy must come down in step with the cost of the chips themselves. Everything has to stay aligned. If chips get cheaper but the power to run them does not, the economics break.
Assuming, however, that we are not heading into a sudden collapse in demand — given the supply bottlenecks and the sheer scale of appetite for compute — the question becomes one of where to invest. And here a crucial distinction emerges between the future and the present.
Solve the Bottleneck That Exists Now
A common mistake is fixation on the future: on what the energy landscape might look like four, five, six, or ten years out. But the energy bottleneck is not a future problem. It is a problem that needs solving today. That reframing changes where the smart money should go.
The most reliable bet is plugging into existing infrastructure — current electrical grids and existing electricity companies, whether state-owned or private. There are real advantages to this approach: certain tax credits, and the ability to use wiring and cables already in the ground. It is extraordinarily difficult to compete with an ever-present, ever-advancing technology if you are relying on energy that does not yet exist. Bet on energy that is still on the drawing board, and you will run out far faster than the competition. The infrastructure that is already there is the foundation; the speculative buildout is a luxury layered on top.
The Quiet Value in Asian Chipmakers
Turning to the semiconductor market itself, one of the more interesting observations is that Korean chipmakers trade at a relatively steep discount to the broader semiconductor space. The instinct is to ask whether the market is getting this right or missing something — and the answer leans toward the latter.
Part of the explanation has nothing to do with technology and everything to do with culture and communication. The American economy is intensely media-heavy. The constant stream of headlines that gets pumped out can correlate directly with rises and falls in the market valuations of public companies. Asian markets — in South Korea, Japan, and China — are simply not as oriented toward this media-heavy presence. So part of the lower valuation of these companies is not a reflection of inferior technology, but of the fact that they are far less public about broadcasting every development. The United States does the storytelling well, but that same tendency can get in its own way.
Consider the depth of capability at firms like Samsung, with its Advanced Institute of Technology and its enormous research-and-development apparatus. There is a great deal of runway left in these companies. South Korea is a hard-working, stable-minded country, and its trajectory tends to be linear. As it progresses, it may not take steps as aggressive as those favored in the United States, but it can grow just as fast — if not faster — precisely because it stays appropriately situated. Steady positioning, rather than dramatic leaps, can win the long race.
Will the Money Run Out?
The primary spenders driving this buildout right now are the big hyperscaler megacap technology firms. When one of them — Alphabet, for instance — recently went to the capital markets to raise money, it amounted to only about 2% of its float, with the possibility of placing some of it with a partner like Berkshire. It was not, in itself, a dramatic event. But if these moves are only the early days, and the pace is accelerating, the natural worry is whether the money runs out or the system becomes too levered.
That particular fear is probably overstated. There is not much real chance of the money drying up, because middle-market spending is set to increase. Many middle-market private equity firms have averaged solid returns since roughly 2010 to 2012 and have substantial capital to deploy. The large players — firms like Blackstone and KKR — are sitting on enormous reserves of dry powder. And the megacap firms themselves are multi-trillion-dollar enterprises; if they need to issue new shares or sell bonds, they can do so readily, placing large volumes of corporate debt at favorable rates.
The deeper risk lies elsewhere.
Geopolitics, Compliance, and the Real Constraint
The harder challenge is the constant shifting of geopolitical tensions and policy. The big technology companies operate in — or touch — nearly every country on earth. That global footprint means their compliance burdens, enterprise risk management demands, and privacy obligations are formidable, because they are fighting an enormous range of regulations across many jurisdictions at once.
For these firms, continued growth becomes a matter of playing the cards at hand: appeasing a multitude of stakeholders, each with different and often conflicting interests. That is a far more difficult task than raising capital. You can sell a great deal of corporate debt at good rates, but if five different countries impose five different policies on importing and exporting physical chips, competing on pricing becomes dramatically harder. The constraint is not the balance sheet — it is the patchwork of rules the goods must pass through.
The Volatility of Policy Itself
A defining and underappreciated feature of the current environment is the sheer volatility of policy. The previous administration tended to move in large-scale strokes — sweeping infrastructure spending bills that, once passed, were largely left to play out. The current administration operates differently: policy changes constantly. One day there are tariffs on something; the next day they are gone.
The most important thing investors and companies can internalize is that policy will change, and that consistency cannot be relied upon in these markets. The best posture is preparation for anything. For companies, the most valuable policy is flexibility and adaptability. If you remain ready for change while continuing steadily down your path, then when the supply chain shock or the policy shock arrives, you are not caught off guard.
A Cold War Echo and the Pull Toward Unity
One of the bottlenecks slowing the buildout is that policy itself delays bringing new capacity online in the United States. As the race for AI research and development increasingly becomes a geopolitical contest, there is a real possibility that some of the red tape gets cut.
There is a useful parallel here to the period from the 1970s to the 1990s and the dynamics of the Cold War. Among the world's governing bodies, the European Union currently faces the heaviest regulatory burden, being the most regulated government body in the world. For the United States, unity and a certain American dynamism will be critical forces shaping the remainder of this AI cycle in its current form.
If other countries pull ahead, American businesses are likely to band together more tightly. The pressure of competition — the imperative to be the first or at least the best country to win the AI infrastructure race — could well lead to a removal of some of that red tape. There is something genuinely interesting in this: the same psychological and geopolitical forces that push nations against one another may end up binding domestic players together. Rivalry abroad can become cohesion at home.
Conclusion
The story of memory chips as a new strategic commodity is ultimately a story about systems aligning — silicon with energy, valuation with reality, capital with capability, and policy with ambition. The technology is not the bottleneck; the power to run it, the rules that govern it, and the willingness to build it now rather than someday are. The investors and companies who understand that the binding constraint is energy and policy in the present — not capital in the future — are the ones positioned to navigate what comes next.