AI IPOs 2026: Anthropic, OpenAI and SpaceX Reality Check
A great narrative and a great investment are not the same thing. In 2026, that distinction is worth billions — especially when it comes to the most anticipated AI IPOs in history.
We use narratives to explain or describe events because the truth is rarely simple — and the mental effort required to grasp it fully is high. So, for practical reasons, human beings revert to stories. According to the Cambridge Dictionary, a narrative is essentially a form of storytelling. Not lies — just simplified versions of reality that work most of the time, until they don’t.
Almost everything we believe to be true is, strictly speaking, not quite true. Take 1+1=2 as an example. That may sound definitive, but it is itself a story. Let me show you why.
If you have €1 and I give you another €1, you have €2. That seems true — not just a story. But if you have one apple and I give you another, you don’t have two identical apples. You have two apples, but the result is not the same as doubling the original. Truth is complicated. If something seems simple and obvious, it is almost always a narrative. That doesn’t mean it’s wrong — just that it’s a simplified version of the truth that doesn’t always apply.
The Investing Narrative Everyone Knows
One of the most common stories in investing, made popular by Peter Lynch, is that you should invest in companies whose products you love. Apple is a clear example. Most iPhone users have been — and still are — very satisfied customers, and early investors did very well. All of that is true. But because it’s simple, it’s still just a narrative — much like the optimism surrounding the AI IPOs of 2026, which often boils down to a compelling but simplified story.
The bigger picture is that customer satisfaction is a good place to start. But whether an investment is profitable depends far more on the business model than the product itself. A great product inside a weak or untested business model isn’t an investment plan — it’s a hope.
The AI IPOs 2026 Narrative
This brings us to 2026, which is shaping up to be the year of the big AI IPOs. Anthropic, OpenAI, and SpaceX are all planning to go public — marking a defining moment in the AI IPO story for 2026.
The narrative fueling excitement goes something like this: disruptive technology creates dominant companies; dominant companies deliver exceptional returns; therefore, buy. This isn’t wrong — it accurately described Google, Meta, and Spotify. But it quietly leaves out the variable that determines whether you actually make money: the price you pay relative to the maturity of the business model.
Google, Meta, and Spotify all went public after roughly a decade in business. They had proven revenue models, clear unit economics, and years of operational data for investors to scrutinize. Anthropic, OpenAI, and SpaceX are going public after just two to three years. Their business models remain largely unproven at scale — and yet they are coming to market at valuations of 20 to 25 times Price/Sales.
What the Numbers Actually Say
For context: Google at IPO traded at approximately 13 times sales — and it already had a dominant, proven advertising engine. OpenAI at 25x sales has neither the revenue history nor the margin visibility to justify even that multiple. The number is not just high. It is high without the foundation that has historically supported it.
Google: 10 years old, proven business, 13x sales. Spotify: 12 years old, established user base, 5x sales. OpenAI: 3-year-old, unproven model, 25x sales.
History offers a useful lesson. Cisco defined the Internet era. The technology won, and the company thrived. But investors who bought at the peak of the 2000 bubble waited over twenty years to break even. The narrative was right. The price was not.
Why We Are Sitting These Out
We are passing on all of them — and if SpaceX’s mid-June listing is on your radar, our dedicated analysis is worth reading first. Read our full SpaceX IPO analysis here.
This is not a statement about product quality, which is genuinely impressive. It is not a prediction that these companies will fail. We believe unproven business models and high valuations don’t align with the risk/reward profile we are willing to accept for our clients.
The current AI investment narrative is seductive precisely because it is simple: buy for growth now and worry about price later. That is unlikely to be true, not because AI is overhyped as a technology, but because price always matters and business models always matter. Two to three years is not enough time to know whether either holds up.
We may be wrong. But if we are, it will not be because the narrative was right — it will be because the market rewarded timing over fundamentals. That is not a trade-off we are willing to make on behalf of our clients.