nuclear-news

Screwed again: small investors to bail out billionaires from SpaceX, OpenAI, and Anthropic

In the past, companies had to wait, and have a track record of revenues and profits.

But that rule is gone. Now, a newly public company can qualify, simply based on size—how big it is.

This is work the Wall Street Journal is supposed to be doing. That’s what the Securities and Exchange Commission is supposed to be asking questions about. Instead it’s left to “Where’s your Ed at?” and a handful of podcasts who are showing the world that the investment thesis for the entire American AI industry is blowing up. Costs are rising, not falling.

Inside China / Business, Kevin Walmsley, Jun 07, 2026

SpaceX will soon go public, in an offering that will value the company at over a trillion dollars.

Anthropic and OpenAI are Artificial Intelligence companies, who also plan IPO’s for later in the year.

Recent changes to indexing rules will compel massive share buys into these companies by retirement and pension plans, and by passive ETF’s and mutual funds.

In the past, new companies were required to wait until insiders sold most of the shares after the lockup periods before being added to investment indices. Companies also needed to show a strong history of growth and sound financial practices.

The heads-up on this story comes from one of our favorite Substacks, Gold and Geopolitics. Our concern is that the public, normal people, are at least unaware, and maybe even indifferent, to how really screwed they are at the highest level, and by people at the highest level. It’s not the kind of story that is felt tangibly, at least not at first, like a big spike in gas or food prices. And it’s also difficult to follow, and that’s exactly what the architects of our financial and political system are counting on.

It involves the qualification requirements for a new company to be included in the NASDAQ-100. The regulators changed those rules, so that new public companies can be part of the index. In the past, companies had to wait, and have a track record of revenues and profits. Because trillions of dollars’ worth of pension investments—which is money invested on behalf of workers, millions of private retirement plans, plus exchange-traded funds and other mutual funds, are invested in those indices. And they do so, assuming that the top 100 tech companies, in this case, are well-managed and profitable businesses.

But that rule is gone. Now, a newly public company can qualify, simply based on size—how big it is. And they changed the definition of “public”. Companies can qualify as a megacap public company simply by being large at the time of its IPO, then limiting how many shares they sell to the public. The index now accepts a weighting multiplier. That is what “small float” means—insiders of the company still own almost all of shares, and so completely control the company, even though it’s “public”.

They also threw out the rule for four consecutive quarters—one full year—of profits.

With these rule changes, the guardrails are down and regular Americans will be forced to buy shares of bad companies, that don’t make money.

The NASDAQ, in this case, gets big fees from the Initial Public Offerings of companies that want to avoid the rules that used to govern the industry, and protect small investors, somewhat, and SpaceX, in this case, will have tens of millions of passive investors buying their shares. That massive, passive buying will put a floor under the stock price no matter what happens, at the same time that insiders are allowed to dump their stock. Insiders and early investors are restricted from selling their shares during the IPO; they must wait to do so, until later.

That was previously a major risk to company insiders, and early investors: if the IPO price is set too high, or if the company does poorly after the IPO, their shares will be worth far less than they had hoped, just as they and all the other insiders are selling. But with these rule changes, passive investors will be buying shares, every single month, because they’re buying the index.

This is the source document. Paragraph 2 explains the new “fast entry” rules for new companies that list on the NASDAQ exchange. If the company’s market cap is in the top 40 of companies already there, it’s a Fast Entry addition, and will be put into the index after 15 trading days. So 15 days after the IPO, SpaceX will go into the index. The company will be exempt from “seasoning and liquidity requirements”—seasoning is how much experience the company has, earning money, and liquidity is how much money it has in the bank. Exempt………………………………………………………………………………………………………………………………………………………………………………………………………….

This is work the Wall Street Journal is supposed to be doing. That’s what the Securities and Exchange Commission is supposed to be asking questions about. Instead it’s left to “Where’s your Ed at?” and a handful of podcasts who are showing the world that the investment thesis for the entire American AI industry is blowing up. Costs are rising, not falling. AI data centers today are budgeted to cost $50 billion per gigawatt to build—and they’re not getting built anyway. Soon they’ll cost $80 to $100 billion per.

And the costs are rising just as companies who use the AI are realizing they’re not getting their money’s worth. Uber is a client of Anthropic, and has already spent its entire 2026 budget. Will they load up on some more tokens, to get through the rest of the year? Doubt it—it was a “head exploding moment” to learn how much Uber spent on tokens, which did NOT result in useful consumer features.

Alibaba is a Chinese company, and their Qwen large language model is the world’s most popular AI tool for business owners outside the United States. Airbnb tried to use ChatGPT to design a new reservation feature on their app, and even though the CEO of Airbnb is good friends with Sam Altman at OpenAI, his company switched over to Qwen instead. It works faster and costs less.

And that is catching on. Other companies are quietly making the switch to Chinese large-language models because they cost far less, and they’re open source and easier for their teams to use in their companies. The performance of Chinese models is similar to Silicon Valley’s best products, and are easier to use, cost less, and are more efficient. Companies are enterprise users—they pay for tokens, and executives lose their jobs if other executives heads blow up when they see their AI bill and ask what they’re getting for it. Companies are looking for alternatives to OpenAI and Anthropic, and signing up for DeepSeek instead. That also means that data is not going through US data centers, it’s coming to China instead, where electricity also happens to cost a lot less.

That could be the biggest challenge of all. For active investors – not the passive ones — business models matter. Revenues and profits – they matter. And the cost of compute is what is driving these corporate users of AI. They pay for the AI. Their engineering teams use it, every day, to develop new tools and applications, and they’re switching over to Chinese LLM’s.

Anthropic and OpenAI are fundamentally bad companies, with bad valuations, and produce financial reports that not even their own top executives trust. Their customers are moving away. And that used to mean that the insiders cannot cash out and make billions of dollars. They might even go to jail. But that was before they changed the rules, and so they’ll make you buy them instead.

Be Good.

Resources and links:……………………………………………………………………………………….. https://kdwalmsley.substack.com/p/screwed-again-small-investors-to?publication_id=3320368&post_id=200908108&isFreemail=true&r=3alev&triedRedirect=true&utm_source=substack&utm_medium=email

June 11, 2026 - Posted by Christina Macpherson | business and costs