For twenty-three years, America has been running out of shares.
Not out of companies. Out of the stock those companies have on offer. Buybacks retired it, cash acquisitions absorbed it, and a generation of businesses stayed private rather than listing, so the supply of public equity shrank in almost every year since 2003. That shrinking float quietly did some of the work that earnings are supposed to do.
Barclays expects 2026 to be the year it turns, with roughly 1.5 trillion dollars of net new stock arriving over the next two years. The largest companies in the world are now selling the same thing every private company is selling.
Key Takeaways:
A two-decade support is being removed. Net US equity issuance turns positive in 2026 for the first time since 2003, on Barclays' numbers.
The cash ran out. Alphabet's free cash flow went negative last quarter for the first time since it went public. Meta's fell to $784 million.
Equity became cheaper than debt. Earnings yields near five percent against an all-in cost of borrowing closer to six flipped the financing decision.
Supply is expanding at both ends. 5.62 million new business applications last year, and only three in ten expect to ever hire anyone.
Private rounds now clear a higher bar. Longer diligence, tighter terms, more proof before the wire.
Twenty-three years of a shrinking float
The black line is the whole argument. In every year it sits below zero, the market ended up holding less stock than it started with, because buybacks and cash acquisitions retired more shares than new issuance created. That arithmetic supported prices without anyone deciding it should. It is the quietest bull-market tailwind of the last two decades, and almost nobody outside equity capital markets desks thinks about it.
In 2026 it crosses. Not because sentiment changed, and not because anyone rang a bell. It crosses because a handful of companies made a financing decision.
They are not selling stock because they want to
Alphabet, Amazon, Meta and Microsoft have announced something in the region of 725 billion dollars of combined spending on AI infrastructure. That money has to come from somewhere, and for most of the last decade the answer would have been obvious. These are the most cash-generative businesses in the history of commerce. They funded everything out of operations and still had enough left over to retire their own shares by the hundreds of billions.
That is no longer true. Alphabet's free cash flow went negative in the second quarter of 2026, the first negative quarter since the company went public. Meta's came in at 784 million dollars, its lowest in nearly four years. Street forecasts have the four of them combined dropping toward roughly four billion dollars in the third quarter, which for companies of this size rounds to nothing.
So the money has to be raised, and the only question is in what form. Their earnings yields sat near five percent this summer while their all-in cost of borrowing sat closer to six, which made selling equity cheaper than issuing debt for the first time in this cycle. They are not selling stock because they are optimistic about the price. They are selling stock because it is the cheaper of two expensive options.
The bond market is not making it easier
The thirty-year Treasury has traded above five percent for twenty-seven days this year, about one session in five, and the last time it spent that long up there was 2007. The difference is that the Fed's own policy rate is more than a percentage point lower now than it was then, which means investors are demanding more to lend long than they did going into the financial crisis.
Credit noticed before equity did. Five-year protection on Oracle trades near 215 basis points against 144 at the start of the year, one notch above junk after the company committed seventy billion dollars to its own buildout. Protection on Nvidia set a record near seventy-nine. Spreads move before multiples do, which is worth remembering when the equity market looks calm.
A note on what this chart does not prove
The Nasdaq's recent peak sits close to the largest listing ever priced, and the chart above is the version of that observation making the rounds. We are including it because it is being discussed, and we are labelling it honestly. One offering does not move a thirty-trillion-dollar index. Two lines falling together is not a mechanism, and treating it as one is how confident people get things wrong.
The supply argument does not need it. Net issuance turning positive for the first time in twenty-three years is a structural change measured over a year, not a story about any single week or any single company.
The other end of the market is doing the same thing
While the capital markets were re-equitizing at the top, Americans filed 5.62 million applications to start new businesses last year, a record, with monthly filings running near half a million. Almost none of it is venture-shaped. The Census Bureau tags an application as high propensity when the business looks likely to employ somebody other than the founder, and that share has fallen to about thirty percent from thirty-eight percent in 2019.
Now put that next to venture. Funding topped 200 billion dollars for a second straight quarter while deal count fell to a decade low, and 263 mega-rounds absorbed 81 percent of everything deployed. Record dollars, fewer doors.
So there is more equity arriving at the top of the market, more companies forming at the bottom of it, and the machinery in the middle did not get any wider. Companies with real revenue and no interest in a mega-round still need a legal way to sell equity to people who are allowed to buy it. That is the entire reason online capital formation exists, and this is the first month the data argued for it out loud.
What it means if you are the one raising
For two decades the supply of equity available to a public investor got smaller every year, which pushed capital outward, into private rounds, into venture, into anything that would sell them a claim on growth. That pressure is what made a competitive seed market feel normal.
Now the largest companies in the world are offering that same investor a liquid claim on the AI buildout, at a price they can check every second, while the thirty-year Treasury pays above five percent to do nothing at all. A private round has to clear a higher bar than it did eighteen months ago, and it has to clear it against alternatives that are easier to buy.
That shows up as longer diligence, tighter terms, and more proof asked for before the wire. It is not a bad market to raise into. It is a market that pays the founders who came prepared, and charges everyone else for the difference.
FAQ
What does net equity issuance turning positive mean?
It means the market is being handed more stock than it retires. For roughly two decades, buybacks and cash acquisitions removed more shares than new issuance added, so the supply of public equity shrank almost every year. Barclays expects 2026 to be the first year since 2003 that the arrow points the other way, with roughly $1.5 trillion of net new stock arriving over two years.
Why are the largest tech companies issuing equity instead of borrowing?
Because the cash is gone and debt got expensive. Alphabet, Amazon, Meta and Microsoft have announced roughly $725 billion of combined AI infrastructure spending. Alphabet's free cash flow went negative in the second quarter of 2026 for the first time since it went public, and Meta's fell to $784 million, its lowest in nearly four years. With earnings yields near five percent against an all-in cost of debt closer to six, selling stock became the cheaper option.
Does more equity supply mean stock prices fall?
Not mechanically, and anyone claiming certainty is overreaching. What changes is that a support the market enjoyed for two decades is being removed. A shrinking float did some of the work that earnings are supposed to do. When supply expands instead, prices have to be carried by fundamentals alone.
How does this affect a private company raising capital?
It changes what your round competes against. For twenty years a shrinking supply of public equity pushed capital outward, into private rounds and venture, because investors needed somewhere to buy growth. Now the largest companies in the world offer that same investor a liquid claim on the AI buildout, priced every second, while the thirty-year Treasury pays above five percent to do nothing. A private round has to clear a higher bar than it did eighteen months ago.
Why does record business formation matter here?
Because supply is expanding at both ends at once. Americans filed 5.62 million applications to start new businesses last year, and only about three in ten expect to ever employ anyone besides the founder. More equity is arriving at the top of the market and more companies are forming at the bottom, while the machinery in the middle that connects companies to investors did not get any wider.
Educational only. Not legal, tax, accounting, or investment advice, and not a recommendation to buy or sell any security. Deal Box is not a broker-dealer, placement agent, investment adviser, or custodian. All offerings are issuer-direct and issuer-approved. Market data and third-party figures are as reported at the time of writing and are not projections of any outcome.