BREAKING: The AI Buildout Just Got Too Big For Private Money

$450 billion in noncancelable obligations against $16.3 billion of half-year revenue. It’s going public because nothing else can carry it.

JUST IN: THE BUILDOUT OUTGREW ITS LENDERS

Flatbed pulled in ahead of me carrying two transformers strapped down like artillery. Oversize placards, pilot car, the whole production. You see these every day out here now and almost nobody writes about them, which is strange, because the transformer is the actual bottleneck in the AI buildout. Not chips. Copper and steel and an 18-month lead time.

Anthropic filed a draft prospectus. The numbers in it explain the flatbed.

Four hundred and fifty billion dollars of obligations you cannot walk away from, against sixteen billion of half-year revenue. That is the entire trade. Everything else is commentary.

Man Gets into Cybercab…You Won’t Believe What Happens Next (AD)

This man just got inside a Cybercab, Elon Musk’s new car that has no steering wheel or pedals.

What happens next will shock everyone.

Click here to see this exclusive footage.

THE BREAKDOWN: START WITH THE PEOPLE SAYING NO

1. A research firm says the equity is worth $144 billion, not $2 trillion. New Constructs ran a reverse discounted cash flow against the leaked numbers and called it “the most ridiculous IPO of 2026” and an “unprecedented test of investor gullibility.” To justify $2 trillion, by their math, you need a 10% after-tax operating margin starting immediately and held for a decade, revenue up 1,000% this year, 100% next, then 56% a year through 2035. That lands you at $3.6 trillion in revenue. The 2025 operating margin was negative 175%.

2. The $80.7 billion loss is not what the headline says, in both directions. About $72 billion of it is a non-cash remeasurement of convertible notes and financing instruments. Actual burn is closer to $8.7 billion. So the scary number is mostly accounting. But the bulls do not get to wave it away either, because those instruments convert at listing, and conversion is real dilution landing on whoever buys the stock.

3. The multiple depends entirely on which revenue number you pick, and all three are going to circulate. Against 2025’s full-year $4.6 billion, $2 trillion is about 435 times sales. Against the $16.3 billion half, about 123 times. Against that half annualised, about 61 times. The low number is only honest if the growth continues, which is precisely the thing in dispute. Anyone quoting one of those three without saying which is selling you something.

4. The bull case is not stupid, and here it is at full strength. Revenue went from $1.1 billion to $16.3 billion in twelve months. That is not a company struggling to find demand. SpaceX posted enormous losses and then ran the largest IPO ever, and investors looked straight past them. If the growth holds for even two more years, every multiple above collapses to something ordinary. The bet is that revenue grows into the obligation before the obligation comes due.

$45M raised and the price is still $3.30 (AD)

This is the first time in history the FAA has given a personal flying car a path to market.

Doroni started in a garage.

Today it has FAA clearance.

70+ manned test flights completed.

600 customers already reserved one.

That is $240 million in potential revenue before a single delivery.

16,500+ investors are already in.

$45M+ raised. Nasdaq $DRNI reserved.

Invest at $3.30 →

Morgan Stanley says this market hits $1 trillion by 2040.

The average person loses 9 days a year to traffic.

Doroni is building the alternative.

120 MPH. 100-mile range. Charges at home in 25 minutes.

First deliveries targeted for 2028.

The pre-IPO share price is $3.30.

Own a Piece of the Flying Car at $3.30 →

This is a paid advertisement for Doroni Regulation A offering. Please read the offering circular at https://invest.doroni.io/

THE FRICTION: ONE COMPANY, TWO PRICES

The equity side is being pitched a two-trillion-dollar growth story. The credit side, looking at the identical company, wants more than ten percent to lend against it, up from eight and a half on the last deal.

Lenders get paid back before shareholders and they get paid a fixed amount, so they spend their lives thinking about what goes wrong. When equity and credit disagree this violently about one balance sheet, the historical record is not kind to equity.

THE MACRO EXTRACTION

Here is why this lands on a road newsletter instead of a tech one.

The reason this company is going public is not ambition. It is arithmetic. OpenAI’s last round took months to pull together more than $120 billion and was the largest private raise in history. Nobody thinks that gets repeated this quarter. Private markets have hit their ceiling, and $750 billion of commitments has to come from somewhere deeper.

The deepest pool of capital in the world is the public market. The public market is pension funds and index funds. Index funds are the default option in your 401(k).

So the sequence is: company lists, company enters an index, and a retired machinist in Beaumont who has never typed a prompt owns a slice of $450 billion in noncancelable compute obligations. He will not be consulted. He will not be told. It will simply appear inside the fund he was defaulted into in 1998.

I have written this story six times this month without the name attached. The data centres bidding for Texas permits and river water. The chips posted as collateral. The Fed hiking into an oil shock. The ten-year at 5.36% repricing every asset with a future in it. This filing is where all of it has to clear.

Transformer on that flatbed is going somewhere with a substation and a fence. Somebody already signed for it. The question the prospectus is really asking is whether the revenue shows up before the invoice does.

— Brook

The Bare Economy. From the road. For the road.