George Gilder Golden Hour

Here is the idea.

The U.S. government is buying equity stakes in strategic American companies — not bailouts, not loans, but actual stock. They are taking ownership positions in industries deemed critical to national security.

George Gilder calls these “Golden Shares.” He has documented five of them in the last year. And he says a sixth is coming — a small American chipmaker with technology that solves a physics problem the AI industry has been running from for years.

The promo is called Convergence X, or the Golden Hour. It is published through Eagle Products, the same group behind Gilder’s Moonshots. The thesis combines a policy trend, a semiconductor bottleneck, and a technological breakthrough into a single investment argument.

Let me walk through each piece.

The Five Golden Shares

Between September 2025 and January 2026, the U.S. government took equity positions in five separate companies across different industries.

U.S. Steel. The government blocked Nippon Steel’s acquisition on national security grounds, then stepped in with an equity stake. The first Golden Share. Steel is the backbone of the defense industrial base, and the administration decided foreign ownership was unacceptable.

MP Materials. Rare earths are essential for defense electronics, magnets, and advanced manufacturing. China controls most of the supply chain. The government took a position in the only domestic rare-earth producer at scale.

Intel. $7.86 billion in CHIPS Act funding came with an equity component. Intel is the only American company that both designs and manufactures leading-edge chips. The government made sure it stays American.

Lithium Americas (Thacker Pass). Lithium is the mineral that powers the energy transition. Thacker Pass in Nevada is the largest known lithium deposit in the United States. The government bought in at the ground floor.

Trilogy Metals. A Defense Production Act investment in Alaska copper-zinc mining. The stock surged 230% overnight when the deal was announced.

Five deals in roughly five months. One per month, each in a different strategic industry. Steel, rare earths, semiconductors, lithium, base metals. The pattern is clear: the government is systematically building equity positions in the industrial base it considers essential.

What That Pattern Means

Gilder’s argument is that this is not random dealmaking. It is a new American playbook.

The government has always had tools to influence strategic industries — tariffs, subsidies, antitrust enforcement, the Defense Production Act. What changed is that it started buying stock. An equity stake is different from a grant or a loan — it gives the government a seat at the table, aligns their interests, and signals to private capital that these companies have sovereign backing.

The companies that get Golden Shares do not have to worry about hostile takeovers. They do not have to worry about being starved of capital. The largest investor in the world is on their side.

That is the macro argument. The micro argument is about the next one.

The Pick Clues

Gilder does not reveal the company in the free presentation. But he leaves enough breadcrumbs to narrow the field.

The target is under $20 per share.

It is a DoD Category 1A Trusted Supplier — the highest security designation a chipmaker can hold. That means the Department of Defense has certified it to fabricate sensitive circuitry for weapons systems, communications, and intelligence platforms. Only a handful of companies hold this designation.

The company has multi-beam lithography technology — an approach to semiconductor manufacturing that solves a physical bottleneck the entire AI industry is struggling with.

It has received approximately $400 million in Pentagon funding.

A retired four-star general sits on its board.

Add those clues together and the field narrows to a very small number of companies. One of them is Intel, which already has a Golden Share. The other is the pick.

The Reticle Limit Problem

The reason this pick matters has to do with physics.

Every AI chip today is made using photolithography — shining light through a mask (a reticle) onto a silicon wafer to etch circuits. The reticle has a maximum size. You cannot make a chip bigger than what the mask can expose in a single shot.

That maximum is called the reticle limit. And the AI industry has been bumping up against it for years.

The largest chips from Nvidia, AMD, and Cerebras are already near the reticle limit. Making them bigger is not an option unless you change how lithography works. That is where multi-beam lithography comes in.

Instead of exposing one reticle-sized block at a time, multi-beam technology writes patterns across an entire wafer in a single pass. It removes the size constraint entirely. A chipmaker with this technology can build computing systems across a full wafer rather than dividing them into hundreds of separate dies.

This is what Gilder calls “the end of chips” — the end of the era where progress meant smaller transistors, and the beginning of an era where progress means bigger wafers.

The Convergence Thesis

The Golden Hour sits inside a larger framework Gilder calls Convergence X. His argument is that eight separate technologies are converging simultaneously — AI, robotics, autonomous vehicles, edge computing, advanced semiconductors, quantum, next-generation wireless, and biotech.

Each one has been developing on its own timeline. What makes this moment different is that they are all starting to accelerate at the same time. When that happens, Gilder says, the compounding effect is larger than any single technology revolution.

The microchip revolution was one technology. The internet was one technology. Smartphones were one technology. Convergence X is eight at once.

If the thesis is right, the companies that sit at the intersection of these technologies — and that have the government’s backing — are positioned for a cycle of growth that looks like nothing the market has seen since the early days of the personal computer.

What to Make of It

The Golden Share trend is real — five documented deals in five months, across five strategic industries, with the government buying equity directly. The Category 1A designation is real, the multi-beam lithography technology is real, and the reticle limit is a real physics constraint the semiconductor industry has been managing around for years.

The question is whether the specific company Gilder has in mind is the one that breaks through. Multi-beam lithography is not commercially proven at scale. The company is small, which means it carries all the risks that come with small-cap investing — volatility, illiquidity, dependence on a single product cycle. The $400 million in Pentagon funding helps, but government contracts do not guarantee commercial success.

Gilder has been right about technology direction more often than not. He called the microchip revolution in the 1980s, the bandwidth revolution in the 1990s, and the blockchain shift in the 2010s. He also led subscribers into positions that lost money during the dotcom bust, because he was early and being early looks like being wrong.

The Golden Hour thesis is a bet that the government-industrial complex is building a new playbook and that one small semiconductor company with the right technology and the right connections is at the center of it. The thesis is coherent. The trend is real. What remains to be seen is whether the company can deliver.

Either way, the idea is worth sitting with.