2Y 4.800 =5Y 4.890 ▼10Y 5.070 ▼30Y 5.410 ▼VIX 24.56 ▼MOVE 98.15 ▼DXY 103.82 ▲USD/JPY 159.41 ▼GOLD 3,755 ▼COPPER 5.70 ▼WTI 82.70 ▼BTC 62,795 ▼ETH 1,777 ▼S&P 7,324 ▼NDX 27,116 ▲MSTR 96.50 ▼
Rates · Liquidity · Scarcity · Credit · Power · 2026-11-18

The Biggest Share Sale Ever Is Money Leaving, Not Arriving

The full market version

Anthropic sold 801.4 million new shares at 83 dollars each on 18 November, plus 23.6 million shares from existing owners. That raises at least 68 billion dollars, and a further 26 billion came in a private sale at the same price. The total is more than 94 billion dollars.

For scale: SpaceX raised 75 billion in its base deal and 86.2 billion including the extra shares its banks could sell. So this is the largest amount of money ever raised by selling shares in a company. It starts trading on Thursday.

The obvious way to read it is as a debut: the first of the big artificial intelligence labs to go public, ahead of a rival that says it will wait until after 2027. The more useful way to read it is as a withdrawal.

Why a share sale takes money out of the system

When a company sells new shares, no new money is created. An investor hands over cash and receives a share. The company now holds the cash and the investor holds a claim on future profits rather than something they can spend today.

The cash does not disappear, but it changes hands from people who were willing to hold it as money to a company that will spend it slowly. And 94 billion dollars moving in one week is a large amount of money to find at short notice.

The question is where it comes from. For the past four years, spare cash in the financial system had a home: the Federal Reserve's overnight reverse repo facility, where banks and funds could park money and earn interest. On 17 November that facility held 0.400 billion dollars. It is empty.

So the buyer of this deal is not spending cash that was sitting idle. That cash has already been used elsewhere. They are either selling something to raise the money, which pushes down the price of whatever they sell, or drawing on deposits that would otherwise have been lent or invested.

One precise point, because it is easy to get wrong: a share sale does not drain the banking system of reserves. The money the company receives lands in a bank account somewhere in the same system, so the total is unchanged. What changes is who holds the money and in what form. The thing that actually drains reserves in December is the tax date and the government's borrowing, not this deal.

The tank it draws on is large, and this is 3.2% of it

Banks held 2,960.6 billion dollars of reserves at the Federal Reserve in the week to 11 November. Against that, 94 billion is 3.2%.

That is not a crisis, and it would be wrong to describe it as one. But it is a withdrawal big enough to be felt at the edges, in a market that is already charging extra for cash. The overnight lending rate, SOFR, printed 3.93% on 18 November against the 3.90% the Fed pays on reserves. Three hundredths of a percentage point above, and the first time it has been above since 9 November.

Three hundredths is a small number. What it tells you is that the buffer which used to absorb events like this has become thin enough that the event now shows up in the price of overnight money.

It arrives in the busiest month of the year

The Treasury says it will borrow 800 billion dollars in the January to March quarter. December has a corporate tax payment date on the 15th, which pulls cash out of the banking system, and year-end accounting rules on the 31st, which make banks reluctant to lend. The government bond market already reflects the pressure: the 10-year yield closed at 5.10% on 17 November, and the extra return investors demand for holding a long bond rather than rolling short ones — the term premium — is 1.01%, the highest since 2014.

A tax date, a wall of government borrowing and the largest share sale ever, all inside one month. None of them alone breaks anything. Together they explain why the site's tightening gauge read +0.545 on 18 November, with all five of its components positive, and why the assets that pay no cash at all fell hardest while the dollar rose.

There is an honest counter-argument. Ninety-four billion dollars of demand for a company that is not yet profitable is not what a shortage of money looks like. That is why the first week of trading, not the price it was sold at, is the real evidence.

What being in an index does, and when

The Nasdaq allows a newly listed company to be added to the Nasdaq-100 quickly, and the committee has done it before. That buying is real, but it is automatic rather than considered, it arrives after the fact, and it is measured in a few billion dollars rather than in the size of the deal.

The S&P 500 is different. To be included, a company needs a record of actual profit. A lab at this stage of building does not have one, so the biggest pool of automatic money in the world is closed to this stock for now. The 94 billion is therefore people making a judgement about artificial intelligence, not a rule following a list. People change their minds, and they do it quickly.

The other date that matters is the lock-up. Six months from Thursday lands in the second half of May 2027, when the early owners are first allowed to sell. Between now and then, the price is being set on a small number of shares relative to the size of the company.

One bet, three names

This is where the piece stops being about Anthropic. On 18 November the assets in this book moved like one instrument with three labels. Strategy's shares fell 20.29% over the week. ASST fell 24.5% in a single session. Bitcoin fell 5.47% over five sessions. Gold fell 5.22% over the same five, and the S&P 500 fell 4.09%. The dollar index rose 0.74%.

Those are not six separate events. They are one repricing of the rate at which future money is discounted, sorted by how far in the future each asset's value sits. The asset that pays nothing, ever, fell most. The shares of a company whose balance sheet is that asset fell next. The hedge against the currency fell alongside them, because the thing it protects against became cheaper. The index fell least, because its cash flows arrive soonest.

The practical point is about Thursday, not about last week. Adding a fourth position does not make the book more diversified. It makes it more concentrated in the same bet, wearing a fourth name. Diversification that only works while money is plentiful is not diversification — it is borrowing with extra steps.

The outlook: what the bulls need, and what the calendar actually delivers

The bullish argument going around is specific: bond yields fall, oil falls and the dollar falls, and that sets off technology, growth and every asset whose value sits far in the future.

The first problem is that all three went the other way this week. The 10-year yield rose twelve hundredths of a percentage point to 5.10%. The dollar index added 0.74%. Brent crude rose 0.46% even as American crude fell 5.46% — a repricing of one location, not a collapse in the oil price.

The second problem is the calendar. Thanksgiving week has a reputation as the best week of the year. Across forty years from 1986, the S&P 500 rose in that week 65% of the time, with a typical gain of 0.36%. That is positive, and it is reliable, and it is not the best week of the year: a typical gain of a third of one percent is an ordinary week with a good hit rate.

So the honest framing is conditional. The fuse the bulls describe needs the real yield — the return on a government bond after inflation is taken out — to fall first, because that is the leg that dragged everything down. Nothing in the calendar provides it. What would change the picture is specific: a 10-year back below 4.98% with the inflation estimate still near 2.21%, which would mean the bond was giving back what it took.

What would prove this wrong

The claim is narrow. Ninety-four billion dollars raised in new shares is money leaving the same pool that funds everything else in this book, and it lands in the month that pool is thinnest. It is not a claim that the sale causes a fall, and it is not a forecast of Thursday's opening price.

The first test is the price. If Anthropic holds above 83 dollars through its first week while Strategy stabilises above 90 and bitcoin holds 63,000, then the cash came from outside the risk markets, the withdrawal was smaller than the arithmetic suggests, and this reading is wrong. If the shares trade below what they were sold at while the collateral keeps falling, the marginal buyer has gone and the sale was the signal rather than the cause.

The second test is the overnight market. Three hundredths of a percentage point above the administered rate is small. If that gap widens while reserves stay flat, the binding constraint stops being the interest rate and becomes the cash itself, and every asset in the book gets a second, unrelated reason to fall.

THE CALL
CallThe largest share sale ever is a withdrawal, not a deposit. Anthropic raised more than 94 billion dollars at 83 a share on 18 November — private cash moved into a corporate treasury in a week when nobody had spare cash to lend — against an overnight facility holding 0.400 billion and bank reserves of 2,960.6 billion, into a month that already carries a corporate tax date on the 15th and 800 billion dollars of government borrowing the quarter after. The 10-year closed at 5.10%; the extra return demanded for holding long bonds is the highest since 2014; the overnight rate printed three hundredths of a percentage point above what the Fed pays on reserves. The same week, every asset that pays no cash fell hardest: Strategy -20.29%, ASST -24.5% in a session, bitcoin -5.47%, gold -5.22%. That is one bet in four names, and Thursday adds a fifth.
Evidence83 dollars a share against a 70 to 78 range; 801.4 million new shares plus 23.6 million existing; at least 68 billion raised plus a 26 billion dollar private sale, more than 94 billion in total, against SpaceX at 86.2 billion including the extra shares. Overnight cash facility 0.400 billion (17 Nov); reserves 2,960.6 billion (11 Nov); overnight rate 3.93% against 3.90% paid on reserves (18 Nov); 10-year 5.10% and real yield 2.893% on a 2.21% inflation estimate (17 Nov); term premium 1.01%; Treasury borrowing for the January to March quarter 800 billion. Marks: Strategy 92.65 (-20.29% on the week), ASST 17.525 (-24.5% on the day), bitcoin 63,053 (-5.47%), gold 3,780.40 (-5.22%), S&P 500 7,307.99 (-4.09%), dollar index 103.79 (+0.74%). Thanksgiving week across forty years: typical gain 0.36%, up 65% of the time.
InvalidationAnthropic holding above 83 dollars through its first week while Strategy stabilises above 90 and bitcoin holds 63,000 would say the cash came from outside the risk markets, that the withdrawal was smaller than the arithmetic suggests, and that this reading is wrong. The second test is the overnight market: if the three-hundredths premium over what the Fed pays on reserves widens while reserves stay flat at 2,960.6 billion, the binding constraint stops being the interest rate and becomes the cash itself.
Levels2Y 4.83 / 10Y 5.12 / 30Y 5.46 / 2s10s 29
DISCLOSURE — NOT INVESTMENT ADVICE
NoticeMacroTape publishes macroeconomic analysis and market commentary for information and education only. Nothing published here is financial, investment, legal or tax advice, or a personal recommendation, and no author is a registered investment adviser. Calls, Levels and invalidation points are analytical views about how a market mechanism works, not instructions to trade. Readers act on their own responsibility and carry their own losses.