The Hundred-Billion Match

---

There are roughly two thousand L8 and L9 engineers at Google — Principal and Distinguished Engineers, the top of the technical ladder. Add in Staff and Senior Staff (L6 and L7) and the number passes twenty thousand. And here's the part worth naming out loud: I know plenty of L6s and L7s who have already founded and shipped startups in their careers. These are not junior engineers waiting for their shot. Many of them have already had their shot. They're back inside Google now, shipping features instead of running companies, because Google pays well and the path from "promising early-stage founder" to "Principal Engineer at a mega-cap" is a short walk.

Twenty thousand people at one company — and probably four or five times that number across the Mag 7 — who, in any other context, would be founders, CEOs, CTOs, chief architects, or heads of entire companies. They are among the most capable humans alive at building things that change the world. The Mag 7 has almost all of them.

They are sitting in meetings. They are writing launch docs. They are arguing with peer L8s about ownership of a feature that will affect some percentage of YouTube DAU. They are good at their jobs. Many of them are extraordinary. Almost none of them are running anything.

The hidden cost of corporate gigantism in America is not what it does to consumers or what it does to prices. It is what it does to these people. A mega-cap is a talent prison. Every one of them is a graveyard where the most capable generation of engineers and operators in human history is being slowly converted into calendar blocks and promotion packets.

I want to talk about the one policy change that would set all of this on fire.

The Proposal, Stated

Any company — public or private — with a market capitalization above $100 billion for ninety average trading days is required to split itself into successor entities, none of which may exceed $100 billion, within 366 days of crossing the threshold. The company chooses how to split. The shareholders receive proportional shares in each successor entity. The government sets the ceiling and gets out of the way.

That's it. One rule. Mechanical, content-neutral, applies equally to every company regardless of industry or political favor, captures private and public both so there's no "go private to escape" loophole, gives a full year for orderly planning.

I am going to argue that this single policy would set off more American innovation in one decade than the combined effect of the last thirty years of tax cuts, industrial subsidies, interest-rate policy, and immigration reform. Not because it punishes big companies. Because it releases the people those companies are hoarding.

Why The Ceiling Must Rise With The Floor

One detail matters enough that the proposal isn't complete without it.

If the hundred-billion ceiling is fixed in nominal dollars, dollar debasement tightens the ceiling every year. M2 has grown at roughly six percent annually since 2006. A fixed $100B ceiling in 2026 is effectively $56B in 2036 real terms, and $31B by 2046. More companies hit the ceiling faster. More breakups per year. That's not necessarily bad — but it's also not by design. It's slow tightening caused by monetary policy, not conscious choice.

Worse: a fixed nominal ceiling gives the mega-cap class a strong new reason to want the dollar debased faster. More debasement, lower real ceiling, more frequent breakups. Every board and executive team suddenly has financial skin in the game against dollar stability. That's the wrong incentive alignment.

The fix is clean. Index the ceiling to the lowest minimum wage in the United States.

Specifically: the ceiling equals 13.75 billion hours of minimum-wage labor — which at today's federal $7.25 is approximately $100 billion. Equivalently: the annual full-time earnings of 6.6 million Americans working at the country's lowest minimum wage. If any state sets its minimum below the federal level, that lower state minimum is the one the ceiling indexes against. You cannot arbitrage by manufacturing a low-wage state.

Now look at the political mechanics.

The mega-cap class suddenly has a direct financial stake in raising the minimum wage. If they want their ceiling to rise — so they can compound without triggering a breakup — they need the floor to rise. The normally opposed interests, the corporate elite at the top of the wage distribution and the workers at the bottom, are pointing in the same direction for the first time in a century.

Wages at the bottom have been frozen in nominal terms. The federal minimum wage has been $7.25 since 2009. In real purchasing power that's about half of what the federal minimum was in 1968. The freeze has held because the most powerful lobby in American politics has had no stake in ending it. Tie the mega-cap ceiling to the minimum wage floor and the lobby flips overnight. Labor advocates and Fortune 50 boards suddenly have aligned incentives.

The alternative route is also interesting. If the mega-cap class doesn't want to raise the floor, their only other path to raising their ceiling is to stop debasing the dollar. That makes them hawks on monetary policy for the first time in modern memory. Savers, retirees, and fixed-income holders — groups that have been getting quietly destroyed by the passive-flow/debasement stack I've written about before — suddenly have a corporate elite constituency they have never had.

Either path is good. The floor rises and the mechanism works. Or the floor holds, the dollar strengthens, and the mechanism tightens slowly. Either way, the policy aligns groups that have been pulling against each other for generations.

This is the quiet brilliance of the design. The hundred-billion ceiling is not just a structural antitrust mechanism. It is a political alignment engine. By indexing the top to the bottom, you give the most powerful economic actors in the country a reason to care about the least powerful. And unlike every other "the rich should pay more" policy, this one doesn't require their consent. It passes with a simple majority. Then the alignment follows automatically. The wealthy get to keep being wealthy. The minimum-wage workers get a powerful new friend. The savers get a monetary policy advocate. The concentration stops. The innovation starts.

One formula. Four groups pulling in the same direction. No new bureaucracy.

Exercise for the reader: under this policy, the loudest voice advocating for a $50 minimum wage would be Tim Cook. Work out why he'd be right to.

What America Has Actually Been Doing Wrong

American productivity growth has been mediocre for two decades. Total factor productivity growth in the 2010s was the worst of any decade since the measurement began. The standard explanations — regulation, tax rates, debt, demographics, deindustrialization — are all real, and none of them explain the actual shape of the problem. We are not short of capital. We are not short of talent. We have more of both than any country in human history, by a margin so wide that the usual comparisons become absurd.

We have piled capital and talent into seven corporate graveyards.

Apple has about 160,000 employees. Microsoft has about 240,000. Google has about 180,000. Amazon has about 1.5 million across operations, of which roughly 100,000+ are corporate and engineering talent. Meta has about 70,000. Nvidia is smaller, about 36,000. Tesla about 125,000.

Call it 1.5 million people directly, several million more in the immediate ecosystem, and probably 10 million Americans whose professional futures are downstream of what those seven companies decide to do. This is the largest concentration of elite human capital in private institutions in the history of the world. It is also, if we are being honest, a profoundly inefficient use of that capital.

The L8/L9 Problem

At Google specifically, the internal levels go from L3 (entry-level) through L11 (Senior Vice President). L8 is Senior Staff Engineer. L9 is Principal Engineer. These are people who, in a normal economy, would be running companies. Many of them have built things used by hundreds of millions of people. They ship tools that generate tens of billions of dollars of revenue.

Twenty thousand of them at Google alone. More than that if you count the L6s and L7s who have already demonstrated they can build and lead.

Now imagine the alternative universe where Google is not one $2 trillion company but twenty $100 billion companies, each spun out from the current whole along natural seams — Search, YouTube, Android, Cloud, Ads, Maps, DeepMind, Workspace, Chrome, Pixel, Waymo, Play, and a handful of others the company would structure itself.

Suddenly those twenty thousand Staff+ engineers are not fungible senior talent inside a monolithic hierarchy where almost none of them will ever be the person making the call. They are — by mathematical necessity — the senior leadership of twenty independent companies. Each of those twenty companies needs a CEO, a CTO, a VP of Engineering, a VP of Product, a VP of Sales, a chief architect, a head of each major business unit. Each needs a board. Each needs an IPO. Each needs a five-year plan. Each needs to win or die. That's easily a thousand top seats per successor company that now get filled from a pool that was previously all buried under Sundar Pichai's org chart.

Twenty companies, each thirsty, each with its own leadership hungry to prove they deserve the shot, each with its own capital allocation authority, each with its own R&D budget, each competing for talent against the other nineteen. Twenty independent decisions about whether to do the risky research project or the safe one. Twenty independent answers to "what do we bet the company on."

That is the thing we're not doing today.

Today, a brilliant L8 at Google who has an idea for a new product has to convince a VP, who has to convince an SVP, who has to convince Sundar, who has to weigh it against seventeen other things, and the answer is almost always no, because the thing under consideration is small relative to YouTube or Search and would cannibalize Ads attention. Brilliant ideas die in promotion packets. This is not because anyone is acting in bad faith. It is because the scale of Google — the single unified entity — makes the opportunity cost of every new product roughly equal to "what we could have done with those engineers on YouTube optimization instead."

Break Google into twenty companies and the math flips. The CEO of (hypothetical) Maps-co is not weighing her product against YouTube's roadmap. She's weighing it against the other thing Maps-co could do. The cost of saying yes to a new bet drops by an order of magnitude. More bets get made. More fail. More succeed. The successes become the next Airbnb or the next Uber or the next Shopify.

That is how you unlock productivity. Not tax cuts. Not subsidies. Liberation of the trapped talent.

Historical Proof — AT&T

In 1984, the US forcibly broke up AT&T. It was, at the time, the largest and most technically capable corporation in the world. Bell Labs was a national jewel. Every serious voice in American policy predicted disaster — the US would fall behind Japan and Europe in telecommunications, Bell Labs would be destroyed, consumer prices would rise, national defense would be compromised because AT&T's infrastructure was the phone system.

Here is what actually happened.

The seven Baby Bells, plus the long-distance company retaining the AT&T name, plus the research and manufacturing divisions, became the fuel for two decades of American telecommunications dominance. The cellular industry — which did not exist in meaningful form before the breakup — was created from scratch by the Baby Bells competing with each other and with upstart challengers. MCI and Sprint emerged as long-distance competitors, cratering consumer prices. Cable companies entered telephony. The entire late-90s telecom build-out, which laid the fiber that powers today's internet, was enabled by the competition the breakup forced.

Bell Labs did not die. It scattered. Its researchers joined the successor companies, founded startups, staffed universities, seeded the Silicon Valley wave that produced Cisco and Netscape and Amazon. The single largest concentrated research institution became dozens of smaller ones, and each one shipped more per researcher because each one had a CEO who needed them to.

Within fifteen years of the breakup, the combined market value of the successor entities was roughly two and a half times the pre-breakup AT&T. Consumer long-distance prices dropped ninety percent. The United States did not lose its telecommunications lead. The United States became the global telecommunications powerhouse in a way it had never been before.

Every cellular innovation, every internet protocol advance, every build-out of the backbone that today carries all of modern life — those came from the Baby Bells competing, not from unified AT&T coordinating. The breakup did not slow American telecom. It created American telecom.

Historical Proof — Standard Oil

The 1911 breakup of Standard Oil is the cleaner case because the data is even more absolute.

Rockefeller held proportional stakes in each of the 34 successor companies. Within a few months of the breakup order taking effect, the combined market value of the 34 successors was about fifty percent higher than the pre-breakup Standard Oil. Rockefeller's personal fortune roughly doubled as a result. The breakup made him richer. Monopolists are under-compensated for their monopoly relative to what competition could produce — because the market discounts monopolies for the distortion they create, while it prices free competitors on their actual productivity.

Post-breakup, the successor companies — Exxon, Mobil, Chevron, Amoco, Conoco, Marathon, and the rest — built the twentieth-century American petroleum industry. They competed for reserves, for refining efficiency, for retail distribution, for global reach. They invested in technologies that unified Standard Oil would never have funded. American energy dominance for the next century was built on their rivalry, not on the coordinated strategy they would have pursued together.

In both cases — AT&T and Standard Oil — the consensus fear was that the breakup would destroy American competitive advantage. In both cases, the breakup was the thing that made American competitive advantage real. The lesson is consistent and it is available in the data.

The 500-by-100 Future

The mechanical consequence of a $100B ceiling, applied to today's American corporate landscape, is roughly this.

The Mag 7 alone breaks into about 180 successor entities. The next tier — the $200B-to-$1T range, which includes Berkshire, JPMorgan, UnitedHealth, Johnson & Johnson, Visa, Mastercard, Eli Lilly, Exxon, Walmart, and dozens more — breaks into perhaps another 100 successors. The $100B-to-$200B range, which includes another 50 or so companies, breaks into about 150 successors.

The net result is that the S&P 500 as we know it becomes roughly five hundred companies, each worth around $100 billion. Total market capitalization stays approximately the same. The distribution becomes flat at the top instead of exponential.

This is the future where American capital markets work the way they are supposed to. Every $100B company is meaningful enough to fund serious R&D, run global operations, compete with international rivals, hire top talent at top prices, and attract institutional investment. None of them is so big that it distorts flows, dominates its industry, or captures the political process.

Every $100B company has a CEO. Every one has an executive team. Every one has independent decision-making authority. Every one has its own board, its own shareholders, its own obligations to compete or die. The number of meaningful leadership seats in American industry doesn't double or triple — it multiplies by an order of magnitude.

This is what unleashing looks like.

The Wage Effect

When thirty companies compete for the same AI researcher instead of one company owning her, her salary goes up. When fifty companies compete for the same elite software architect, his signing bonus goes up. When a hundred companies compete for the same MBA product manager, her equity package goes up. Labor economics says this plainly — concentration depresses wages; competition raises them.

Today's mega-cap comp packages are eye-popping because they compete for talent with each other in a narrow club. Distribute that same talent pool across five hundred $100B companies that all need senior leadership, and the comp band widens dramatically at the top end. More people make $5M-$50M. Fewer make $500M.

This is not a moral argument about fairness. It is an economic argument about productivity. When talent is well-compensated relative to its output, more talent flows into productive work. When talent is hoarded by a few companies with lottery-like outcomes, more talent flows into politics, finance, and entertainment — the only other industries where extraordinary payouts are available.

The hundred-billion ceiling broadens the "I could win as much as anywhere in tech" tent from maybe 50 people at the top of the Mag 7 to maybe 5,000 people across five hundred successor entities. The multiplier effect on the decision "should I join a startup, take a real swing, go build something" is enormous. Right now that decision looks worse because the guaranteed path — stay at Google, get promoted to L8, make $1.5M/year — is so good. If Google is thirty companies and being L8 no longer exists as a category because those people are now CEOs, the comparison flips.

The Objections, Named Honestly

International competition. The sharpest argument against any US-imposed breakup policy: China, Europe, and other major economies don't do this, so US-only breakups put American companies at a disadvantage in global markets. This argument has two responses.

First, it overstates the case. Chinese mega-caps are not free-market competitors in the relevant sense. They answer to the Party. Their CEOs can be disappeared. Their IPOs can be canceled. Their businesses can be confiscated. Whatever theoretical advantage scale gives them is partly offset by the fact that they are state assets that happen to have shareholders, not the other way around. European mega-caps, meanwhile, barely exist — the EU is actively anti-mega-cap already.

Second, reciprocity is available. The US can negotiate size-ceiling treaties with allies (UK, Japan, Canada, Australia, Korea, EU). Thirty years after the US broke up AT&T, Europe broke up its state telephone monopolies. Coordinated structural antitrust is achievable when the US leads. The "we can't do this because nobody else will" argument is the same argument that would have stopped every unilateral American policy innovation since 1945, including the ones that worked.

Network effects. Some industries have legitimate "bigger is better for users" dynamics. A payment network, a social graph, a marketplace. Break these up and you might destroy real user value. The honest response is threefold.

First, history. Standard Oil's refining had massive economies of scale. AT&T's network had massive economies of scale. Both were broken up. Both produced more value and better consumer experience after. The "natural monopoly because of network effects" argument has a bad historical track record.

Second, modern architecture. API-first, interoperability-mandated systems preserve network effects across corporate boundaries. Visa and Mastercard already interoperate. Phone networks already interconnect. Email already works across providers. The modern internet is built on protocols, not walled gardens. A post-breakup tech landscape with interoperability mandates can keep most of the network-effect value while eliminating the concentration.

Third, discipline. The "network effects justify my monopoly" argument is exactly the argument monopolists always make. Sometimes it's true. Often it's self-serving. A bright-line rule means we don't have to decide case-by-case who's telling the truth about their network effects — we just hold the ceiling.

The Political Coalition

This policy does not map cleanly onto left or right. It maps onto a different axis entirely: people who want America to be more productive versus people who benefit from the current concentration.

The coalition that could pass this is unusual but real:

• Founder-class populism on the right. The Vance-Thiel-Cowen tendency is, in its more lucid moments, pro-competition and anti-concentration. Breaking up the mega-caps liberates American founder ambition from the "get acquired by Google" endgame.
• Labor-class populism on the left. The Warren-Stoller-Khan tendency has been arguing for antitrust revival for a decade. This is the cleanest version of what they want.
• Libertarians. Classic libertarianism was originally anti-concentration because concentrated corporations corrupt free markets. Modern libertarianism lost its way on this for a while but there's a real opening to recover the Hayekian case against corporate gigantism.
• Small business America. Every mega-cap is a moat against small competitors. Breaking them up instantly makes every SMB's competitive landscape easier.
• Organized labor. Thirty competing employers is better for wages than one dominant one. Unions understand this viscerally.
• Academia and researchers. Thirty independent R&D budgets is better than one. Grant opportunities multiply. University partnerships multiply.
• State governments outside California and Washington. Distribute thirty successor companies geographically and states compete for HQ relocations, bringing jobs and tax base. Currently every mega-cap HQ is concentrated in four zip codes.

The right language for all of these audiences is the same: this is not about punishing successful companies. It is about liberating the talent and ambition those companies are hoarding, so that America can start winning again the way we used to win.

What It Would Feel Like

Imagine the day the policy passes.

Every L8 at Google calls their family that night and says "we're going to be a company." Every engineer at Apple who has been sitting on an idea for three years pulls up a docs page and starts writing an S-1. Every product manager at Meta who has been arguing internally for a product line that gets shot down every year suddenly has a path to ship it at a successor company where she'll be the VP of Product.

Every venture capitalist in America starts calling every mega-cap engineering director they know. Every recruiter in every growth-stage startup opens a second phone line. Every second-tier city in America starts drafting tax incentives to lure one of the thirty successor HQs.

Every twenty-something engineer deciding whether to join the Mag 7 or take a chance on a startup looks at a landscape where the Mag 7 is going away and chooses differently. The flow of talent into ambitious new ventures — which has been declining for a decade as the concentration has compounded — reverses overnight.

Every engineering manager at a mega-cap gets the call: "We're picking the leadership team of the new entity. Are you in?" The question is not abstract. It's "will you run a $3B business unit? Will you be the CTO? Will you lead the IPO roadshow?" People who have been patient for a decade get their shot all at once.

Every research scientist at DeepMind or FAIR or Microsoft Research who has been waiting for their project to get funded looks at a post-breakup landscape where there are now thirty places the project could live. Five times as many companies competing to fund risky research.

And the broader economy notices within six months. Hiring accelerates. Startup formation accelerates. Corporate M&A at the mid-market level explodes. Regional economies boom as successor HQs distribute. Wage growth at the top end broadens instead of concentrating at seven peaks. IPO windows open in sectors that have been dormant for a decade.

This is not a fantasy. This is what happens when you release several million elite operators from a structure that prevents most of them from doing their best work.

Closing

America's problem is not that we have too few companies. We have plenty of companies. America's problem is that the top seven of them are vacuuming up the capital, the talent, and the innovation oxygen that the rest of the economy needs to breathe.

A hundred-billion-dollar ceiling is not a regulation. It is a match. You hold it to the ceiling and the ceiling catches, and the fire spreads downward through every mega-cap, and what burns away is the structural inertia that has been holding back American productivity for twenty years. What remains is a landscape of five hundred roughly-equal-sized companies, each with a hungry leadership team, each competing for talent and capital and customer love, each able to make its own bets, each capable of losing or winning on its own merit.

The companies are not being punished. They are being set free. The ten thousand L8s at Google who have been waiting a decade for their shot finally get it. The talent that has been trapped starts producing at the rate it was always capable of producing. American innovation — which has not been asleep so much as gagged — can speak again.

Companies have smart people. Let them figure out how to split. The government's only job is to hold the match to the ceiling.

That's the whole thing.

---

This is the sequel to The Top Three And The Tundra. That post was about how to survive the rigged game as an individual investor. This one is about how the game itself gets un-rigged — by a single policy that doesn't tax anybody, doesn't subsidize anybody, doesn't regulate what any company can do, and doesn't care who's in power. It just holds the ceiling.