If you only read one AI headline today, don't just stare at Nvidia and don't just stare at Tesla Robotaxi.
Because what may truly change the U.S. AI landscape in 2026 is not one company beating earnings or one model scoring higher on a benchmark, but three super-giants that have not yet fully entered the public market and are all heading for Wall Street at the same time.
They are SpaceX, OpenAI, and Anthropic.
And the most interesting part is that even Gavin Baker, one of the investors who best understands semiconductors and tech, sees this wave of megacap IPOs as a very unusual capital test.
Who is Gavin Baker?
He is not a hype-driven influencer investor. He was early at Fidelity in Nvidia, Meta, and Uber, and later made big bets on SpaceX before rocket recovery was fully proven.
In other words, he is not the type to call Musk great after the rockets already landed; he was backing the story with real money while the rockets were still crashing.
So when someone like that comments on the IPOs of SpaceX, OpenAI, and Anthropic, he is clearly looking beyond the surface excitement.
Who is truly creating value?
Who is burning cash efficiently?
Who is merely wrapping a sexy story around a trillion-dollar valuation?
Baker's key point is that Musk has helped investors make money for twenty years, and he sees that as a kind of sacred contract: don't be too greedy on valuation. That matters, because it explains why SpaceX could keep rising in the private market and why investors were willing to trust Musk.
But that is also where the problem lies.
If SpaceX really lists at close to $1.75 trillion, can that still be called not greedy?
That is the real question this episode is trying to answer.
Before the video starts, please like and subscribe. The content here is worth revisiting.
On the surface, the IPOs of these three giants look like a super feast of the AI era. SpaceX tells a story about rockets, Starlink, xAI, space internet, and even future space-based AI infrastructure. OpenAI tells a story about ChatGPT, the AI application gateway, enterprise software replacement, and a new platform for human-machine interaction. Anthropic tells a story about Claude, enterprise AI, code generation, safety-focused large models, and a business model that is closer to profitability.
Each story is sexier than the last.
But that is exactly where capital markets become dangerous.
The sexier the story, the easier it is for retail investors to lose judgment.
When you look at SpaceX, it feels like missing it means missing humanity's trip to Mars. When you look at OpenAI, it feels like not buying it means missing the next operating system. When you look at Anthropic, Claude feels like the new Microsoft of the enterprise AI era.
But investing is not about whose story is bigger. It is about whose story can turn into cash flow.
Let's start with SpaceX.
The attention on this IPO is no longer just at the level of a normal tech stock. If it raises $75 billion at a rumored $1.75 trillion valuation, it could become one of the largest IPOs in human history.
When many people hear SpaceX, they first think of rockets, Starlink, Musk, and Mars colonization.
But what the market may really be buying is not the rocket itself, but the infrastructure imagination behind SpaceX.
The rocket is just the entry point. Starlink is the cash-flow story, and xAI is the AI narrative. If low-earth-orbit satellite networks, global communications, government contracts, military communications, AI compute leasing, and even orbital data centers are added later, SpaceX is no longer just a space company. It becomes a super-platform that blends space infrastructure and AI infrastructure together.
That is why the market is willing to give it such an absurd valuation.
But we must stay calm.
According to the article's breakdown, SpaceX's 2025 revenue looks to be around $18.7 billion, which is solid growth. But there is a crucial issue: some of the data comes after consolidation. In other words, once SpaceX, xAI, and X-related businesses are combined, the financial statements become more complicated.
Looking separately, Starlink is indeed a strong business. In 2025, revenue is about $11.4 billion and operating profit about $4.4 billion, with tens of millions of subscribers across more than 100 countries. This business is powerful and genuinely profitable.
But the launch business is not as easy as people imagine. In 2025, revenue is about $4.1 billion, yet it still runs an operating loss and requires continued heavy capex for Starship.
The most painful part is the AI segment, meaning xAI plus X. In 2025, revenue is around $3.2 billion, but the operating loss is as high as $6.4 billion, and capex is shocking. In other words, much of the money Starlink works so hard to earn is directly swallowed by the AI business.
This is SpaceX's biggest contradiction.
On one side, it has the world's strongest space-infrastructure story; on the other, it has an AI black hole that burns cash at high speed.
On one side, it makes you feel like the next ten years will be full of stars and oceans; on the other, the financials remind you that starry skies are expensive and very cash intensive.
Even more importantly, after listing, SpaceX will likely remain a highly Musk-controlled company. Ordinary shareholders will have very limited say, and short-term dividends should not be expected. What you are buying is not traditional shareholder rights, but a long-term vote of confidence in Musk's execution.
It is like handing your money to a genius driver.
The problem is that this car may drive to Mars, or it may keep flooring the accelerator and shake you so hard that you question life itself.
So is SpaceX strong?
Absolutely.
But the real question for investors is whether the market has already priced in many years of that strength.
Next is OpenAI.
Everyone knows its story. ChatGPT is no longer just a chat tool; it represents the AI application gateway and a new human-machine interface platform.
Over the past two years, most of the U.S. AI rally has been about Nvidia, GPUs, data centers, power, and cloud computing. In plain English, the market was betting on the AI picks-and-shovels sellers.
But if OpenAI lists, the market can finally price the application layer directly.
That is hugely important.
Once OpenAI becomes public, investors will look at it through a microscope.
How fast is revenue growing?
What is enterprise customer retention like?
Can inference costs come down?
Can gross margins improve?
How exactly does it share revenue with Microsoft?
Will it eventually be a high-margin software company, or a model company that must always buy compute and always burn cash?
This is where OpenAI becomes awkward.
Its user base, brand mindshare, and developer ecosystem are arguably the strongest of the three. The article mentions a huge weekly active user base and many paid subscribers, with very fast revenue growth.
But it also burns cash fast.
The stronger the model, the higher the inference cost. The more users it has, the more calls it handles. The more calls, the bigger the compute bill.
It is like a restaurant with an ever-growing line outside, greater fame, and people calling it the future of Michelin three-star dining. But when the owner does the math, every bowl of noodles is burdened by surging gas, electricity, and labor costs in the kitchen.
So OpenAI's core question is not whether demand exists.
Of course demand exists.
The real question is whether that demand can be monetized at sufficiently high gross margins.
If OpenAI lists at a valuation near the trillion-dollar level, the entire AI application layer will be repriced. Microsoft will be viewed differently, Oracle will be viewed differently, and Nvidia and AMD will be viewed differently too, because as long as large-model companies keep expanding, compute demand will not disappear easily.
But that also creates a very real issue.
If OpenAI itself is listed, why would investors still take the detour to buy so-called AI shadow stocks?
In the past, many funds bought Microsoft because they could not buy OpenAI. They bought Amazon and Google because they could not buy Anthropic. They bought Tesla and satellite-themed stocks because they could not buy SpaceX.
But once those core assets themselves reach the public market, capital will make a fresh choice.
That is the scariest part of the three giant IPOs.
They are not just adding a few ticker symbols. They may pull money back from a pile of AI concept stocks toward the real core assets.
That is good for the strongest names, but for companies merely riding the AI theme, it could mark the start of valuation re-rating.
The third main character is Anthropic.
If OpenAI is the best-known AI application gateway, Anthropic is the most dangerous challenger in the large-model war.
Claude has built strong recognition in code generation, enterprise productivity, and long-context processing. More importantly, Anthropic's story is different from OpenAI's.
OpenAI sells the biggest traffic gateway, the strongest brand mindshare, and the largest developer ecosystem.
Anthropic sells enterprise customers, commercialization efficiency, a safety-first image, and a path that is closer to profitability.
That is why people who understand technology, semiconductors, and compute costs are especially interested in Anthropic.
The biggest problem in AI has always been that everyone knows it matters, but no one knows exactly how it will make money.
Why can Nvidia command such high valuations?
Because it is not selling a story; it is really making money, with very strong gross margin, revenue, and cash flow.
But large-model companies have often felt like this: many users, fast growth, huge stories, but money burns very quickly.
If Anthropic can prove first that large-model companies do not only burn money, but can also produce healthy enterprise revenue and profit, then the valuation logic for AI application companies will be rewritten.
The article mentions a key point: Anthropic appears much more capital efficient than OpenAI, may burn only a fraction as much cash, and its revenue growth and enterprise penetration are both very fast. More interestingly, it also faces a compute constraint.
Simply put, Claude is not unused; it is compute-limited.
It is like a Ferrari limited to 80 km/h, not because the engine is weak, but because there is not enough fuel.
If that compute bottleneck is removed in the future, its revenue and profit could have even more room to expand.
That is why the article describes Anthropic as a spring compressed to the limit.
Once released, the rebound could be astonishing.
But we should not only look at the bright side.
No one stays king forever in the AI model race. Google, Meta, OpenAI, and open-source models are all chasing it. Claude leading today does not mean it can still lead in two years. Silicon Valley never lacks talent mobility, nor does it lack new technical routes.
So Anthropic may have the best capital efficiency, the fastest growth, and the closest path to profitability, but it still faces competition risk, compute risk, and valuation risk.
Put all three companies together and the real underlying logic of this IPO wave becomes clear.
SpaceX represents the ultimate imagination of AI infrastructure.
OpenAI represents the ultimate imagination of the AI application gateway.
Anthropic represents the ultimate imagination of enterprise AI commercialization.
If they all enter the public market one after another, what they bring is not just three new stocks, but three new valuation anchors.
From now on, the market may not only ask whether Nvidia can sell more GPUs.
It will also ask whether SpaceX can connect space and AI infrastructure, whether OpenAI can turn AI applications into a new operating system, and whether Anthropic can turn enterprise AI into a genuinely stable money-making business.
That means the 2026 U.S. AI rally may be entering its third phase.
Phase one: the market trades chips.
Whoever has GPUs is king.
Phase two: the market trades cloud and power.
Whoever can build data centers and secure electricity has pricing power.
Phase three: the market starts trading real AI companies going public.
Whoever can prove AI is not just a cash burner but a cash-flow machine becomes the new core asset.
But the key point is this.
Megacap IPOs are not necessarily all good for the market.
Because an IPO is, by nature, a request for money from the market.
If SpaceX raises $75 billion, and OpenAI and Anthropic also hit the market at near-trillion-dollar valuations, then a lot of money will have to be reallocated out of secondary markets, institutional portfolios, index funds, and thematic funds.
In plain retail language, market money is not infinite.
If a big chunk of it goes to subscribe to SpaceX, OpenAI, and Anthropic, other AI stocks may temporarily lose blood flow.
This is especially painful for the AI concept stocks that are already expensive but whose fundamentals have not caught up.
Because in the past people bought them because they could not buy the real protagonists.
Now that the protagonists are taking the stage themselves, the supporting cast naturally has to be re-examined.
So for this wave of mega IPOs, the most important thing is not how much they jump on day one, but who becomes the future center of AI asset pricing.
If SpaceX succeeds, the market will reprice space infrastructure.
If OpenAI succeeds, the market will reprice the AI application gateway.
If Anthropic succeeds, the market will reprice enterprise AI commercialization.
In terms of investment priority, if we follow the logic in the two articles, Anthropic may be the one worth tracking most closely. Its capital efficiency is highest, its revenue growth is fastest, it is closest to profitability, and its compute constraint has not yet fully been released.
OpenAI comes second. Its brand is strongest, its user base is largest, and its ecosystem is biggest, but it burns cash the fastest and faces the most pressure from margins and inference costs. It is like a supercar with enormous horsepower: very fast, but fuel consumption is terrifying, and you need to be sure it can keep refueling all the way to the finish line.
SpaceX ranks third. Not because it is not great, but because it is too expensive, too complex, and too dependent on the Musk narrative. Starlink is a good business, and space infrastructure is a top-tier story, but the loss black hole after consolidating xAI will make the financial statements uglier. If it pulls back after the IPO, it may actually deserve a fresh look.
So the core conclusion is simple.
The three giant IPOs are not the end of the AI rally; they are a re-segmentation of it.
Companies with real revenue, real customers, real moats, and a path to cash flow will keep drawing capital.
Companies that have only the words "we also do AI" will likely be left behind by the market over time.
Retail investors usually lose money not when opportunities are absent, but when there are too many opportunities and every story sounds like the next Nvidia.
When you look at SpaceX, it feels like missing it means missing the future.
When you look at OpenAI, it feels like not buying it means missing the AI gateway.
When you look at Anthropic, it feels like it may be the next enterprise software giant.
But mature investing is not charging in just because the name is famous, and it is not dismissing everything just because the valuation is high.
It is about asking three questions.
First, how much of the future is already priced into today's valuation?
Second, can future profits catch up with today's story?
Third, if I am wrong, can I tolerate the volatility?
If you cannot answer those three questions, even a great company can become a hole in your account.
Baker dared to invest in SpaceX because he saw a future that had not yet been fully priced.
But today, ordinary investors may be facing a SpaceX whose future is already priced to the sky, an OpenAI pricing the AI gateway at a trillion dollars, and an Anthropic that looks closest to profitability but remains highly competitive.
This is not the time to rush in blind; it is the time to keep your eyes wide open.
Because the real opportunities are often not where the noise is loudest, but after the excitement fades and the market calms down, when price and value begin to line up again.
Alright, that is all for today. If you enjoyed this video, please like, save, and subscribe to the channel. I will keep bringing you the latest U.S. stock investment interpretations and updates. See you in the next video!
