The Year of the Trillion-Dollar IPOs
Just like the internet, AI will change the world. And just like the internet, AI seems to be starting with a bubble.
With SpaceX’s successful IPO, hundreds of people became millionaires, a few became billionaires, and one became a trillionaire. The truth is that they were already there based on the private market values assigned to the company—this was just a mark-to-market event. And given the company only floated ~5% of its total shares, there hasn’t been much real liquidity for insiders.
The private market has been incredibly good to SpaceX. They likely could have raised the ~$75 billion privately without any issue (they raised roughly $10–15 billion over the last two years alone), so it’s not entirely clear why they chose to go public. The typical explanation is VC pressure—funds need to return capital and push for liquidity—but that doesn’t seem to be the case here. SpaceX’s shareholder base is unusually long-duration, with insiders, sovereigns, and crossover capital that have been willing to roll forward rather than exit. There have also been structured secondary programs offering periodic liquidity to employees and early investors, which reduces the urgency to IPO.
So, if not liquidity, then what? A few possibilities:
Currency: Public equity can be used for acquisitions, compensation, and broader strategic optionality
Access to retail demand: Expands the buyer base meaningfully
Narrative timing: Peak enthusiasm + scarcity = premium valuation
Benchmark inclusion: Forces passive flows over time
Still, none of those feel strictly necessary. It may simply be that at a certain scale, going public is less about need and more about inevitability, or maybe it’s just fun to IPO.
Anthropic and OpenAI have also set the stage to IPO, and considering how the SpaceX issuance performed, I’d expect both to at least accelerating timelines to capture momentum.
First off, I agree AI will change the world. The productivity gains could be enormous. But that doesn’t mean any price is justified.
To illustrate how stretched valuations are, it helps to go back to basics. A trillion-dollar annuity, discounted at today’s risk-free rate (~4.5%), pays roughly $45 billion per year in perpetuity. That’s guaranteed cash flow. Equities are the opposite, cash flows are uncertain, outcomes are path-dependent, and competition, execution, and leverage matter. So, you start with that $45 billion baseline and then demand a risk premium on top. You should be earning meaningfully more than that to justify owning something with real downside.
Now layer in reality, these companies are not producing stable cash flow, they are currently deeply loss-making, and the capital intensity is massive. (See: The Tech Giants Are No Longer Asset-Light Businesses, and What If the Massive AI Spend Doesn’t Pay Off?)
Yet investors are underwriting outcomes where these businesses become some of the most profitable companies ever created. Last week, SpaceX traded at something like ~2,800x 2028E earnings. That projects a future where revenue scales massively, margins expand, competition is limited, and capital intensity declines, all the while the three companies all are in an intense competition for what is arguably a commoditized product (See: CEO of Microsoft on AI Agents & Quantum | Satya Nadella).
The bigger issue is that they’re losing money today, and not small amounts. Based on leaked and disclosed figures, OpenAI and Anthropic are burning tens of billions. The losses for these three companies were upwards of ~$40 billion in 2025, and capex (compute, infrastructure, energy) is still accelerating.
This isn’t SaaS with negative working capital and 80% margins. This is industrial-scale AI infrastructure, with economics that are still very much unproven. You could argue that costs will collapse with scale, models will monetize better over time, and winners will emerge with dominant positioning, but those are assumptions, not facts.
Separately, I would really like to understand how SpaceX received an investment-grade rating from S&P. Seriously, how?
You have:
Enormous capex requirements
Lumpy and long-dated project economics (Starship, Starlink)
Heavy reinvestment needs
Limited visibility into sustainable free cash flow
The rating only makes sense if you underwrite:
Durable, quasi-monopolistic positioning in launch and satellite internet
Long-term contracted revenue streams from governments and enterprises
Significant operating leverage as Starlink scales
A very high tolerance for execution risk
In other words, the rating appears to reflect less what SpaceX is today and more what it is expected to become. That may ultimately prove correct. I understand credit ratings are based on expected debt repayment capacity, not current profitability alone. Still, the decision is striking. Companies such as Amazon, Tesla, and Netflix spent years demonstrating profitability, cash-flow generation, and balance-sheet strength before achieving investment-grade status. SpaceX appears to have crossed that threshold largely on the basis of its market position, strategic importance, and expected future economics rather than a long history of proven free cash flow generation.
Whether that proves prescient or premature remains to be seen.
But it's a useful reminder of where we are in the cycle: we're not valuing these businesses based on what they are today; we're valuing them based on what they could become under highly optimistic assumptions.
The internet changed the world too. It also went through a period where capital was effectively free, expectations overshot reality, and valuations disconnected from fundamentals.
History didn’t say the technology was wrong. It just repriced the timeline.
—
Michele Chiacchio
June 23rd, 2026
Important Disclosures
All publications are provided for informational and educational purposes only and do not constitute financial, investment, or economic advice. Investing involves risk, including the possible loss of principal. The views expressed are solely those of the author and do not necessarily reflect those of any affiliated organizations.