Porter Stansberry published an essay on May 13, 2026 called “The Greatest Financial Market Bubble In History Is Forming A Top.” The essay ran on his Substack, Porter’s Journal, the Porter & Co. member site, and The Burning Platform. Fourteen minutes of reading that compresses a career’s worth of credit-cycle thinking into one argument.
The argument is specific. The financial press treats the AI rally as an earnings story. Stansberry says it is a credit story, and the credit cycle just turned.
That distinction matters. An earnings story means companies are making money and the market is pricing future profits. A credit story means companies are borrowing money to build infrastructure, and the borrowing is what is driving the market. The difference shows up when credit tightens. Earnings stories survive tightening because the underlying business keeps generating cash. Credit stories break because the financing structure depends on continued access to cheap borrowing.
The Numbers Stansberry Built the Thesis On
The essay is built on a stack of specific data points, each one a piece of the credit structure.
The CAPE ratio sits at 40. The cyclically adjusted price-to-earnings ratio smooths earnings over ten years to filter out short-term noise. The historical mean is 17. Three prior peaks exist in the modern record: 1929 at 21x, 1972 at 21x, and the dot-com peak in March 2000 at 35x. The 2026 reading is the highest sustained CAPE in the history of U.S. equity markets. A reversion from 40 to the historical mean of 17 implies a price decline of roughly 58 percent.
Stansberry’s point about CAPE is structural. He argues that a bubble is defined by the velocity of the deviation from trend, not the absolute level of valuation. When prices move two standard deviations above their normalized growth path, the asset has reverted to the trend line every time. The reversion is violent because the leverage that drove the move up must be unwound quickly. The selling is margin selling, which means it is forced, which means it accelerates.
The credit indicators support the valuation picture. The ICE BofA U.S. High Yield Option-Adjusted Spread closed May 11 at 2.79 percent, roughly half its post-1996 average. When the spread between safe debt and risky debt is that narrow, investors are assuming risk does not exist. Margin debt at FINRA member firms hit $1.22 trillion in March 2026, up 38.7 percent year-over-year. Money is virtually free, and free money gets borrowed and deployed into marginal projects.
The AI Infrastructure Build and the Debt Behind It
The four largest hyperscalers, Amazon, Alphabet, Meta, and Microsoft, collectively guided to between $610 billion and $725 billion of capital expenditure in 2026. Total data center spending is projected to reach $3 trillion by 2029.
The spending is real. The question is how it is financed. Data center debt issuance hit $625 billion in 2025, four times the $166 billion issued in 2023. Oracle alone accumulated roughly $100 billion of debt and closed a $16 billion financing on a single Michigan data center in April 2026. Meta was preparing as much as $25 billion in new investment-grade bonds.
Stansberry frames this through the Austrian economist Ludwig von Mises and the concept of malinvestment. Cheap credit creates the illusion that capital is infinite. When capital appears infinite, every project looks accretive at the margin. That includes the marginal data center that will never earn its cost of capital. The malinvestment is the buildout that only makes sense while money stays free.
CoreWeave: The Case Study
CoreWeave is where Stansberry’s argument gets concrete. The company has raised $28 billion in equity and debt in the twelve months preceding the essay. In March 2026, it closed an $8.5 billion delayed-draw term loan. That loan was the first investment-grade financing in history secured by GPU hardware and customer contracts. Moody’s rated it A3, three notches into investment grade.
Before March 2026, no rating agency had given an investment-grade rating to a loan backed only by GPU chips. The reasoning was straightforward. GPUs are volatile, short-lived, and easily made obsolete. Moody’s overrode that reasoning and gave the facility a rating that utilities and railroads typically receive.
The structural mismatch is where Stansberry focuses. The bond matures in March 2032, a six-year maturity. The assets backing it are Hopper-generation GPUs. Nvidia is already several development cycles past Hopper. The loan is secured by hardware with a two-to-four-year lifespan, financed over six years. The cash flows depend on payments from OpenAI, a company projected by its own backers to lose $35 billion in 2027.
The circular financing pattern is the deeper layer. Microsoft’s remaining performance obligations, the contracted future revenue commitments on which its 26 percent cloud growth narrative depends, total $625 billion. Forty-five percent of that is concentrated in a single customer: OpenAI. Microsoft has invested billions in OpenAI so that OpenAI can pay Microsoft for Azure. Nvidia committed to invest $100 billion in OpenAI so OpenAI can buy Nvidia products.
Total AI revenue is estimated at less than $50 billion annually. Total AI investment is above $1 trillion. The ratio is 50 to 1, financed in part by six-year debt against two-year assets.
The Historical Parallel Structure
Stansberry’s method in this essay is historical analogy, and the structure is deliberate. He identifies five credit-cycle peaks and traces the trigger in each one.
The Panic of 1857 started with the failure of the Ohio Life Insurance and Trust Company. The trigger was a single institution that declined to roll a marginal piece of credit. The panic spread through the banking system because the underlying credit structure was overextended. The railroad boom of the 1840s and 1850s had been financed almost entirely by debt, and the debt had been priced as if railroads could never lose money.
The 1929 crash was powered by call money. Investors borrowed against their stock portfolios to buy more stock. When the call money market tightened, the selling was forced. Margin calls triggered more margin calls. The leverage that drove the market up unwound it in days.
The 1973-74 bear market followed the OPEC oil shock. Stansberry’s framing is that the oil shock was the catalyst, but the vulnerability was already built into the credit structure through the Nifty Fifty valuation bubble. Companies like Polaroid and Avon traded at 50 to 90 times earnings. The oil shock tightened credit, and the overvalued stocks reverted.
The 2000 dot-com bust involved vendor financing. Cisco and other equipment makers financed their own customers’ purchases. The revenue growth was real on paper, but it was funded by the sellers themselves. When the financing dried up, the revenue disappeared. Cisco took 25 years to recover its 2000 high. The internet was real; the capital allocated through Cisco’s stock was speculative.
The 2026 AI bubble, in Stansberry’s framework, is the fifth instance of the same pattern: genuine technological transformation financed by mispriced credit, where the technology is real and the capital structure built on top of it is unsound.
What Stansberry Sees in Bitcoin and Gold
The essay includes a market signal that ties the credit-cycle thesis to the monetary thesis Stansberry has been building for years.
Bitcoin hit an all-time high of $126,198 on October 6, 2025. On May 11, 2026, it traded at $81,224, down 36 percent from the high and 22 percent year-over-year. Gold was making new highs over the same period. Stansberry’s read: Bitcoin tends to lead the credit cycle on the way up and on the way down. Gold tends to lag credit on the way up and lead defaults on the way down. The divergence, gold rising while Bitcoin sells off and equities ripped vertically into a reversal, is what the leading edge of a credit contraction looks like.
This connects to his broader framework. The 2029 monetary reset thesis, documented in his book published earlier in 2026, argues that Social Security trust fund depletion will force a government default before the end of the decade. The Permanent Portfolio framework, adapted from Harry Browne, allocates 25 percent each to Forever Stocks, long-dated Treasuries, gold and Bitcoin, and cash. The credit-cycle essay adds a deflationary chapter between now and then, complementing the inflation endgame. The credit contraction comes first. The inflationary policy response comes second. The Permanent Portfolio is designed to survive both.
The Track Record on Credit
Stansberry’s credit calls have a documented history. His research published in the mid-2010s warned that the post-2008 credit cycle would produce the worst corporate bond default cycle in history. The argument was that the 2009 government intervention cut the default cycle short, leaving zombie debt that should have defaulted but was instead rolled forward and refinanced in 2011 through 2013. High-yield bonds traded at yields below 5 percent during that period. Stansberry called it insane and positioned subscribers in distressed debt strategies.
The May 2026 essay is the evolution of that framework applied to a different sector. In 2015, the malinvestment was in corporate bonds broadly. In 2026, it is in AI infrastructure specifically. The Austrian credit-cycle indicators he references are the same. The pattern of genuine innovation funded by mispriced credit, followed by a forced unwinding, is the pattern he has been tracking for over a decade.
The difference in 2026 is the scale. Data center debt issuance in a single year, $625 billion, is larger than the entire corporate bond market was in the periods he was warning about a decade ago. The six-year loan against two-year assets is a structural mismatch that did not exist in prior cycles because the underlying assets, railroad bonds or internet equipment, at least had longer functional lives.
Where This Leaves the Reader
Stansberry’s essay ends with a list he says to put on a wall:
- An eighteen-day vertical run in semiconductors
- Thirty percent concentration of the S&P 500 in seven stocks
- $725 billion of capex from four companies in one year
- $625 billion of data center debt in a single year
- Six-year A-rated loans against two-year assets
- A 22 percent year-over-year decline in Bitcoin into new all-time highs in gold
- A CAPE of 40
- A high-yield spread of 2.79 percent
The trigger, in his framework, is always the marginal piece of credit that someone declines to roll. In 1857 it was a single Ohio trust company. In 1929 it was call money. In 2000 it was vendor financing. The 2026 equivalent has not been identified yet. The structure that makes the trigger matter is already built.
The essay is Stansberry’s most differentiated contribution to the AI bubble conversation. Other publishers frame it as a valuation problem. Stansberry frames it as a credit problem, and the distinction determines what happens when the cycle turns. A valuation correction is a drawdown. A credit unwinding is a cascade. The Permanent Portfolio is his answer to both. The Porter Stansberry dossier traces the career arc behind that framework, and the Gods of Gas Rice brothers story shows the conviction-led positioning the credit thesis sits alongside. The gold deflation warning is where Stansberry names the deflationary chapter he expects to precede the inflationary endgame. More Porter Stansberry files are collected in the guru dossier hub.