Two of the most hyped companies in America are heading for the stock market, and everybody wants a piece. OpenAI and Anthropic — the outfits behind ChatGPT and Claude, respectively — have both filed confidentially to go public.
I’ve been investing for 45 years, including a decade as a stockbroker on Wall Street. Here’s what that decade taught me: The stocks everybody’s dying to own are usually the ones you should think hardest about.
I’m not saying these aren’t great companies. They might be. But a great company and a good stock aren’t the same thing — especially when the price already assumes everything goes perfectly.
Here are five reasons to keep your hand off the buy button, at least for now.
1. You can’t buy them directly, and the back doors are pricey
Both companies are still private. Their shares don’t trade on any exchange, and you can’t buy them through a regular brokerage account. Anthropic filed confidentially with the SEC on June 1, and OpenAI announced its submission within days.
A confidential filing is a first step, not a promise. The timing hinges on SEC review and market conditions, so there’s no firm date for either one.
There are back doors to get pre-IPO shares, but they’re expensive. A couple of closed-end funds already hold slices of these companies. The Destiny Tech100 (DXYZ) trades on the NYSE, and the ARK Venture Fund (ARKVX) holds stakes in both. Regular investors can buy in.
But Destiny charges around 2.5% a year and ARK’s net fees run close to 3%, according to The Motley Fool. A plain index fund charges a few hundredths of a percent. Those fees eat into your returns every single year.
Here’s the second catch. When companies are this hyped, these funds can trade well above the value of what they actually hold — a fat premium to what’s called net asset value.
Then the IPO hits. Everyone dumps the fund to buy the real shares, the premium collapses, and latecomers are left holding the bag. That’s exactly what burned some SpaceX proxy funds before its debut.
Want to skip the funds and buy true pre-IPO shares instead? That’s mostly off-limits unless you’re an accredited investor — think a seven-figure net worth or a six-figure income.
2. The thing they sell is turning into a commodity
Here’s the part that should worry any potential investor. The core product — the AI model itself — is getting cheaper fast, thanks to pressure coming from China.
Open-source Chinese models now run 60% to 90% cheaper than the top models from OpenAI and Anthropic, according to CNBC. One of them, Zhipu’s GLM-5.2, landed within a percentage point of Anthropic’s Opus model on a closely watched benchmark at roughly a fifth of the cost, Reuters reported.
When a product does the same job as those from multiple other companies, it becomes a commodity. How do commodities compete with one another for market share? Price. That’s why commodities don’t typically command fat margins or trillion-dollar valuations.
True, one AI model may become demonstrably the best and best-selling, as Apple is with the iPhone. But can you tell today which company that will be? I can’t. I’ve been using Google’s Gemini, Anthropic’s Claude and OpenAI’s ChatGPT for more than a year now, and from what I can tell, there’s no obviously dominant choice.
Quick aside — most internet financial advice comes from people who weren’t alive during the last recession. I’ve been writing about money for more than 35 years. Want rock-solid advice? Sign up for the free Money Talks Newsletter. Takes 10 seconds. No fluff. No spam.
3. One of them is burning cash like there’s no tomorrow
OpenAI is losing money at a pace that’s hard to wrap your head around. Internal projections reported by Fortune point to a roughly $14 billion loss in 2026, with cumulative losses that could top $115 billion before the company turns a profit.
In 2025, OpenAI spent about $1.69 for every $1 it took in, per Fortune. It’s also committed to spending around $1.4 trillion in computing deals over the next eight years.
Anthropic’s picture looks tighter — it’s aiming to break even by 2028, and most of its money comes from enterprise customers. But “less scary than OpenAI” still isn’t the same as profitable.
4. There’s no cash cow underneath
This is the one people forget. Compare these two to other AI giants your money might touch.
Google’s parent, Alphabet, earned about $132 billion in profit in 2025. Search and ads print money, and that money pays for every AI bet Google makes. If Gemini never earns a dime, Alphabet is still just fine.
SpaceX, which went public this June, has Starlink subscriptions and government contracts humming underneath its AI ambitions. OpenAI and Anthropic have no such backstop. They’re close to pure plays — burning cash to build the future with no legacy business bankrolling the wait.
And if you own an S&P 500 index fund, you already own slices of Google, Nvidia and the rest — which means more of your money may already ride on a handful of tech giants than you think.
5. You’d be buying a story, not a balance sheet
A near-trillion-dollar price tag before the public has seen a full set of audited financials isn’t investing. It’s betting on a story.
And parts of the story carry an asterisk. Some of Anthropic’s headline revenue is reported on a gross basis, which can make the top-line number look bigger than a clean comparison with rivals would, as VentureBeat and others have noted.
Then there’s regulation. This June, the government forced Anthropic to pull two of its most advanced models offline for weeks over export controls. When Washington can flip a switch and freeze your product, that’s a risk no spreadsheet fully captures.
The bottom line
I’m not telling you these stocks will flop. In today’s speculative environment, it’s likely they’ll go up. SpaceX, another company with no earnings, went public and immediately soared from its IPO price of $135 to as high as $225. (About a month later, however, it’s changing hands at $134 — meaning everyone who was “lucky” enough to buy it at the peak price is now underwater.)
So it won’t be shocking if Anthropic and OpenAI do well when they first go public. But at these prices, the reward is mostly priced in — and the risk is mostly yours.
Here’s what I’d do: At the minimum, wait for the public S-1. That’s the document that turns a hype-filled story into audited numbers you can actually judge. Read it. Then decide.
One of the money rules I try never to break: If I can’t judge how an investment really works, I don’t buy it.
Add a Comment