It could break the stock market. But it’ll stimulate the economy. Just don’t expect inflation to cool on its own.
By Wolf Richter for WOLF STREET.
The AI investment bubble is starting to produce a giant sucking sound of cash out of stock-market investors’ accounts as Corporate America raises enormous amounts of cash through sales of newly issued shares at mindboggling stratospheric valuations.
There will be the share sales via mega-IPOs by SpaceX, Anthropic, and OpenAI; combined, they might suck up close to $200 billion of stock-market investors’ cash. SpaceX alone might raise $75 billion in its IPO.
There will be – and already are – the sales of newly created shares by giant tech companies, including at-the-market sales at huge valuations to retail investors and whoever wants to buy them.
Oracle started that in February by announcing a $20 billion at-the-market share sale, meaning that it will sell newly created shares into the market whenever it sees fit.
Alphabet came out on Monday with an $80 billion share sale, most of it at-the-market, but it also included a private placement of $10 billion in shares at a discount to Berkshire Hathaway.
There will be the share sales by a potentially huge wave of big and lesser IPOs as PE firms and VC firms are trying to unload their portfolio companies. PE firms are sitting on over 13,000 US companies, according to PitchBook. Many of them are established companies that PE firms acquired in leveraged buyouts years ago and have held for well past the 5-7 years that usually mark the exit timeline, and these companies may sell lots of newly created shares to the public during the IPO and with follow-on offerings in order to raise funds to pay down their heavy debts.
These sales of newly issued shares to raise money suck up cash from investors that have hoarded the cash, and move this hoarded cash over to Corporate America which is hell-bent on spending this cash in the real economy, on construction of data centers, factories, and chip plants, on HVAC equipment, electrical equipment, on semiconductors and servers, on power plants and transmission lines, on power-generating and transmission equipment, on all kinds of related products and services.
And the vendors of those products and services receive that cash and build that equipment and provide the services, and they pay their people and suppliers, and this whole pile of cash that investors had uselessly hoarded for years, and that they’d just gambled with for years, gets sucked up from their accounts via these share sales at mega-valuations and gets sprayed widely into the real economy.
With the huge valuations of tech companies – both public such as Alphabet and still private such as SpaceX – this may be one of the best times in history for companies to raise cash by selling shares. And it makes total logical sense to do mega-amounts of it, as long as the cash can flow.
Investors are on the other side of these share sales, and if it’s one of the best times in history to raise cash by selling shares at gigantic valuations, it’s not necessarily the best time in history to provide that cash by buying these shares.
And sucking that much liquidity out of the market – rather than plowing liquidity into the market with share buybacks – might very well be what’ll break the market.
But that’s just the market.
In terms of the real economy, this massive effort to convert investors’ hoarded cash into real economic activity that gets this cash to circulate and produce products and services that will generate more activity and create jobs and keep that cash circulating is stimulative to the US economy.
We’ve already seen the beginnings of the impact of cash Corporate America had hoarded being plowed into the AI investment bubble, and that investment surge has substantially added to GDP. But that was corporate cash that Corporate America had hoarded. This new wave of cash from share sales is cash investors have hoarded that’ll be getting sprayed into the economy with a firehouse to circulate for years to come.
All this economic activity has been creating demand, and will be creating demand, and it is also creating jobs and incomes – while AI is destroying some other jobs. Prices in some segments have already surged for consumers, such as electricity to households. But for now, companies are facing much higher price pressures (PPI jumped by 6.0%) than consumers (CPI jumped by 3.8%) as companies have had a hard time passing on all their cost increases.
I have to say that seeing investor cash and corporate cash getting plowed into the economy is a breath of fresh air, after years of mega share-buybacks, which accomplish the opposite. Share buybacks are still going on at companies, such as Nvidia and Apple. But this may be the first year in a long time when share issuance exceeds share buybacks, and for the economy – though not necessarily for investors – putting this cash to work is a good thing.
Just don’t expect inflation to cool on its own in this scenario. The AI investment bubble last year, this year, and next year is adding to inflationary pressures, and that’s real and it has started.
Whether or not any efficiencies from the use of AI in the future will be deflationary is being debated as a theory – and I can see lots of reasons why it won’t be since companies are not eager to pass on their profits to consumers, unless they absolutely have to, and AI might help them not to.
Fed Chair Warsh acknowledged those two phases of the AI investment boom in his Senate confirmation hearings: The inflationary phase now and for years to come, and the potentially deflationary phase further in the future, and he acknowledged that the deflationary phase may is just a theory that may not pan out. But the inflationary pressures from the AI investment boom are real and now.
Enjoy reading WOLF STREET and want to support it? You can donate. I appreciate it immensely. Click on the mug to find out how:
“But the inflationary pressures from the AI investment boom are real and now.”
And the Fed has failed to come to grips with this, no matter how loud the messaging from the real economy.
For the average worker investing in their IRA/401k, likely in index funds like the S&P 500, total stock market, total world etc, what does this mean?
I’m assuming lower market returns since valuations are already so high. However, are regular folks going to be left holding the bag buying all the shares of these over leveraged PE-held companies?
Total world feels like a better bet because you’re at least not 100% US.
For people holding QQQ or something: look out below I guess. I’ve been holding my breath for years but it keeps building. Part of this is inflation, both CPI and asset inflation, making these numbers seem so big when forward PE ratios are actually well within historical norms at least for total world.
It means that the average expected return over the upcoming years is well below average. Of course, that’s been the case (Buffett indicator, Shiller P/E) for many years now. You pays your money and you takes your chances.
You mean if it all blows up?
It means their shares in the index are worth less for a bunch of years. Until the shares reach the point they bought them at, they are losing money to inflation. It means they kick themselves a lot for even being invested in the market, until their shares get back to even.
They prob kick themselves for not taking the realized gains when they thought they were Good enough.
The tech companies will be cut in half and 3/4s of them will be neutered and irrelevant.
Nvidia might keep its sparkle but won’t regain any significant share value until a new tech bubble, which prob would be 10 years off.
These irrelevant tech companies will be a lead weight on the indexes until some other companies come to the forefront and make tons of money. Prob biotech and pharmaceuticals.
Depends if buyers of these IPOs shares use cash and equivalents or liquidate share holdings. If the former then upward pressure on interest rates would be expected. If the latter, eg. selling NVIDIA, to purchase Space X, that may be index negative short term.
Bond holders/traders probably have more risk exposure that index holders.
Eventually a portion of all these circulating cash expenditures finds its way back into various investments, perhaps even collapsed Oakland and Austin condos.
Alphabet issuing 80B may seem like a lot but they bought back 4X that since 2020.
Nearly, $2T for SpaceX is a much more significant number. Well, its acutally just the average US deficit since 2020. Nothing special in the new norm for those on board.
Isn’t this a horrible environment for an IPO?
No one has any money. Sure in 2023, might have been smart. Or 2027, maybe 2029.
2026? Nope terrible year. Let’s have a war on oil , oh what a great time for an IPO!
I think even Roman Roy would agree with me. And in that fictional show he was responsible for a couple thumbs being lost. 🚀
Haha
It must be an awesome time for IPOs because so much is coming with high values. The point you hit on is where will the money come from. I’ve been wondering this as well. I assume money will come from money markets, gold, existing stocks and bonds. Does it mean that prices in those existing investments will drop to absorb?
“2026? Nope terrible year. Let’s have a war on oil , oh what a great time for an IPO!”
My reading of Wolf shows a different story.
https://wolfstreet.com/2026/05/20/credit-card-delinquencies-balances-debt-to-income-credit-limits-and-collections-in-q1-2026-americans-and-their-revolving-credit/
Look at the metrics in this post; the consumers are not in deep trouble.
If it’s a bad time for an IPO it’ll get pulled. Considering the pricing and market interest it must not be a bad time.
everyone has to IPO this year before time runs out on the money train… if AI quantifiable productivity gains arent realized ever then the financing behind it all needs to crumble eventually . on the other hand TSLA is still a meme stock after all these years so maybe investors just don’t care about profits anymore.
Exactly. These IPOs are a giant liquidity event to pay back investors before market price discovery can happen when everyone gets a good look at the books. Elon’s people seem pretty sure that the same fan base supporting Tesla will show up for SpaceX and keep the reality distortion field going.
Does that fan base have cash?
This whole AI boom is is giving me the same feeling I had about Tesla in its early days. I was convinced Tesla was way overpriced based on the risks they were facing and the valuation off Tesla at the time.
Tesla beat the odds and overcame the risks and grew like a thousand times bigger than what it was in those days. People are still paying ridiculous multiples for that stock.
But AI is dominating the entire market these days and the risk is that AI will take down all the stocks in the stock market and half the US economy.
However, the upside is also much higher.
We could simply put all of our money into safe cash investments but I think that inflation is eating all the return that comes from cash.
What’s a poor value investor to do!
I hope you are right, but why haven’t we seen the boom so far? This AI boom is at least a year old, and all we’ve really seen is increases in asset prices. Data centers don’t employ people long term, and most workers travel from job to job so net employment stays about the same. The power companies aren’t really scaling up. The chipmakers also aren’t requiring large numbers of new workers. On top of that, we are seeing layoffs and wage growth hasn’t spiked. How specifically is the money going to trickle down?
The AI investment boom showed up in GDP already. I linked my GDP article where this is discussed.
The only way the money trickles down is UBI from windfall taxes… Coming soon
I agree that, if the choices are big tech companies using their cash to buy back their own shares or investing in the economy, the latter is better. But would be even better is if they weren’t making such obscene profits because of limited competition and their pricing power.
Microsoft can raise their subscription for Office 365 from $15 to $20 per month per user, and most companies won’t cancel. But Microsoft increases their revenue by 33%. They know that most of their customers don’t have a choice. This is my theory as to why so many people feel bad about the economy. The extra money that could go to raises, especially for small business, is being sucked up into big tech.
I would be cautious before celebrating, Wolf. Money from the top 1% circulating to the top 5%, so that the top 5% can buy stuff from companies owned by the top 1%, with subsequent inflation which only further hurts the bottom 95%, might be healthy in the short term but will further separate ‘the tale of two economies’. And it looks like the new Fed chair intends to redefine ‘core’ inflation downward, which will only contribute to our Alice in Wonderland economy.
The enormous wealth generated for the SpaceX mafia is already Bering used to start companies that will support the leo satellites, data centers in space, and on.
I’m not sympathetic to anyone who buys SpaceX shares based on a “dream” and loses it all.
I placed an order for 100 shares in the IPO, 15 day hold to avoid flipping. Funds aside to purchase more if it takes a substantial digger out of the gate.
It’s a gamble for sure. I don’t like the math behind this one, but lately the math doesn’t seem to matter. Someone is going to take a hit I’d bet and it might just be me!
Fun times and great article!
Inflation deflates cash, bonds and RE. VLCC tankers don’t care
about oil prices. They charge: barrels x distances. Since distances
are longer they charge more. Their shortage can last for years. The longer the trips the less available tankers are in the market. Since a large % of them are old, and new tankers cost more to build, the shortage can last years. Unless AI shortens VLCC constructing for less than half a year. Can the US compete with China, S.K or Vietnam, in building ships ==> with AI, onshore Steel, energy and qualified industrial labor force: yes !!
As someone previously in this industry, the answer is no. It would take a decade to get shipbuilding in the US to minor levels, and the costs are just astronomically high compared to the asian shipbuilders in all categories. The last western holdout was Odenske in Europe and that fell after Maersks triple-E builds circa 2010.
Shipbuilding is not about tech, it is about manpower, material, and land. Order books are multiple years long queues and it takes a long time to build even a single ship, and the facilities to build said ship are massive with huge capex frontloaded requirements on premium land, with a labor market monopoly (dockworkers unions) that is very averse to any automation. If I was to open a shipyard anywhere in the world, I’d put the US last place among any developed economy with a coastline.
The idea that there will someday be deflation from AI adaption is a handy theory for those who don’t want to do the uncomfortable things to tamp inflation now. “It’ll all even out in the end, eh?” :)
It is possible. Large number of white collar behind the computer screen jobs can be taken away by AI. It wont make the role go away but itd need lesser number of people to do the job.
A high paying white collar job pays many jobs downstream and if it goes away in large number then lesser demand for lot of things and service thus causing deflation.
Larger number of temp jobs are created to create ai infra but to run these ai infra not a lot of people are needed
And it is all going into the pick’s and shovels. But once the fat lady sings the old lady is going to be asking where’s the beef?
One industry AI has helped is Security. Palo Alto said they recently got 1000 inquiries from companies ( having to spend more, thanks AI) seeking protection from AI hacking.
Other companies cutting employees compensation to fund AI just to save expenses by cutting “lower value, human capital.”?
It could turn out to be like global warming or Trump, where nobody is safe. But in the meantime, pass the popcorn, please.
Feels like a new gold rush. Sell the picks and shovels.
Net private saving: Households and institutions (W986RC1Q027SBEA) is in a downward trend.
Q1 2026: 863.921
Q4 2025: 870.398
Q3 2025: 1,013.123
Q2 2025: 1,140.342
Q1 2025: 1,163.267
A decrease in the supply of loan funds alone will force interest rates higher.
Yes, now also consider that another 11 trillion in U.S. is rolling over and must be refinanced…
Interesting times, hedge accordingly.
It’s a huge amount of money getting spent. Just hard to see it repeating every year to justify the ever increasing share prices in the recipients of the spending.
And I hope it does actually get spent and the contractors get paid. Too often people like Musk refuse to pay until taken to court to force payment. Very hard for smaller suppliers to stay solvent.
I mean this is capitalism. Find resources, exploit people to extract resources, beggar your competition and the labor. Profit.
There is no economy of scale for the workers. Your value is purely relative to the effectiveness of other sources of functional work.
The next big pool of money is the reasonably wealthy who can afford healthcare which runs at equivalent cost to shelter nowadays, which is to say nearly unaffordable.
K shaped indeed. Whether the gambling economy results in a lifting of all boats is questionable, but they’ve got an algorithm to sell you (on a subscription) that will guarantee* results. Hodl friends, hodl. That spacex stock is going to the moon.
It reminds me of a song I just listened to on vinyl, Rickie Lee Jones, Easy Money.
*Results guaranteed but not guaranteed to be positive.
Dow up over 800 points, but Nasdaq slips as Broadcom’s stock plunges
The AI boom seems mostly to be about data-center construction and Semi-Conductors. Perhaps major new investment in power generation will also follow.
But I can’t eat the output from a data center. I cant live in it or wear it or use it’s output to move my body from point a to b. The whole argument for AI as a net plus to the economy rests on it’s promise of boosting productivity and much of that promise rests on the final cost-per-token we will end up with when all the capex is spent and turned into compute and the power to run that compute.
If the productivity doesnt materialize, the whole AI boom is simply a kinder, gentler version of war spending.
Time will tell…and if there is one thing our culture seems to hate above everything, it’s waiting for time to tell us.
Phil: Dassault + NVDA Omniverse + CudA X can cut design time.
NVDA Isaac Sim can train robots to weld and paint. A modular ship
building in the US might compete with China, S.K and Vietnam bc we have plenty of energy which they don’t have, onshore steel, AI and robots.
Futurism: AI Bubble So Surreal That It’s Now Propping Up Toilet Industry
MW: Wasted AI budgets at Uber, Microsoft and Nvidia trigger hiring — because human workers are cheaper
Classic peak of the equity market behavior. Smart money exiting and going risk off while the dumb money piles into the market looking through the rearview mirror. Institutional outflows from funds and ETFs over the last 10 weeks have been supported by retail investors jumping on the train before it goes over the cliff. IPOs and opening private equity and credit funds to retail and 401ks just provides more liquidity for institutions to exit. The Euphoria can’t last that much longer. Very little compensation for taking risk beyond a few months.
Why do companies and investors have all this cash available? Is it due to fiscal deficit? Maybe some of that money should be helping balance the fed govt budget. If AI buildout doesn’t prove to be deflationary, or long term productive, then this form of technocrat capitalism will have some reconciling to do.
Wolf, can you explain a little bit about Elon/Space X circumventing the usual IPO rules? Is it really true it will automatically be added to the s&p 500 despite not being profitable? And wouldn’t this basically guarantee the stock skyrockets when all the index funds have to buy it?
Musk didn’t circumvent any rules. Wall Street rules got bent left and right eagerly to accommodate the most historic IPO by valuation ever.
Good article. It’s possible that the bust will happen sooner than everyone thinks because as per certain pundits there are billions invested in hype/subsidizes but path to profitability is not clear and this is a lot different than Amazon and we are outside the QE days of cheap money. AI sucking sound can be heard in the crypto world as well as AI if it comes forward will outbid miners for energy hands down.
Interesting post considering vast swaths of the market are down 40% and/or have been flat for a whole year sans some energy/AI/chip/data infrastructure plays. My hope is that the swoosh in liquidity will come from that mania vs. the swoosh in liquidity that have donated their market cap/share prices to the AI mania.
The IPO’s won’t suck cash out of the economy if the Fed prints more liquidity. Under the Fed’s ample liquidity regime, I think they would print more money before liquidity ever tightened enough to drop asset prices.
Scratch that prior comment. I think the point is cash will not leave the economy. It will transfer from those who sit on it to those who spend it, thereby creating more inflation pressure.
As of the end of May there was $8.3 Trillion in Money Mkt funds alone. So far in June, MM funds are up another $48 Bill. No need to print New money (Data from Cranes MFI). Derisking trades? Plenty to cover IPOs.
If (when) the AI boom busts and all the investor cash that stimulates economy is used up or vanished from valuation drops, how will this affect the economy? Won’t there be a inflation coupled with job losses in all sectors associated with the AI build out? Or will the bust be disinflationary or deflationary?
What “hoarded cash”? Everyone has been all-in for months now…
Well, think about it. When companies sell shares (IPO or otherwise), investors buy them either with cash they have in their accounts, or by selling other investments to get that cash, but investors buy them one way or the other, which is precisely my point, it sucks liquidity out of the market and causes prices to decline.