Every few decades, a technology comes along that makes investors lose their minds. The dot-com era had Pets.com. The 2010s had WeWork. And right now, AI is doing something that looks very familiar.

OpenAI $OPENAI ( 0.0% ) just closed a $122 billion funding round at an $852 billion valuation, the largest private funding round in Silicon Valley history.

At the same time, they are rationing GPU compute internally, shutting down their most popular product, and holding daily meetings just to decide which team gets computing power that day.

That does not look like a company in control. And for many analysts, it is the clearest sign yet that an AI bubble burst could be closer than most people think. But before you panic or go all-in, you need to understand what is actually happening underneath.

Let's be honest about one thing first. The technology is real. AI is genuinely changing how businesses operate and how code gets written. That part is not up for debate.

What is being questioned is the price tag attached to all of it.

OpenAI is valued at $852 billion. Their projected revenue for 2025 is around $20 billion. That means investors are paying roughly 42 times annual revenue for a company losing $14 billion per year, with no path to profit until 2029 or 2030.

During the dot-com bubble, companies traded at similar multiples based entirely on promises. OpenAI has real users (900 million weekly active users) and real revenue.

But 42 times revenue is the kind of number that gets repriced fast when growth slows even slightly.

Tech capital expenditure as a percentage of U.S. GDP has now reached 7.2%. During the dot-com peak in 2000, that number was 6.4%. We are now above that level.

What is worth noting is that the crypto market is also reacting to this pressure. Bitcoin $BTC ( ▼ 3.13% ) has seen sharp corrections every time the tech market shakes, because institutional money now flows through both channels at the same time.

When investors get worried about the AI bubble, they often cut crypto positions too just to reduce portfolio risk.

BCA Research has warned that the AI investment surge will likely hit hard limits by the end of 2026. If that slowdown happens at the same time as three massive IPOs hit the market at once, the pressure on valuations across the whole sector could be very serious.

The OpenAI funding round looked like a big win. Amazon pledged $50 billion, Nvidia committed $30 billion, SoftBank co-led the round. But if you look at who is writing these checks and why, the picture gets more complicated.

Amazon needs OpenAI's models running on AWS. Nvidia needs OpenAI consuming its GPUs at scale. SoftBank needs a flagship asset for its Vision Fund III story.

None of these are passive bets. These are infrastructure lock-in moves, defensive spending by companies that cannot afford to be left out.

Here is what rarely gets mentioned. OpenAI's $122 billion was already allocated before the wire transfer cleared. Every dollar has a place, including chips, data centers, and model training runs.

And even with that capital committed, they still cannot buy what they need fast enough. GPU lead times are running 36 to 52 weeks right now.

TSMC's advanced packaging capacity is a hard physical ceiling. High-bandwidth memory is in structural undersupply that money alone cannot fix.

So OpenAI raised the most money in startup history and is still holding daily meetings to ration compute. That is not abundance. That is how deep the problem actually goes.

This is one of the most revealing moves in the whole story.

Sora was OpenAI's AI video generation tool. It went viral the moment it launched. The demos were impressive. Then OpenAI quietly shut it down.

The reason was simple. Video generation requires 10 to 100 times more GPU compute per output than text generation. The head of Sora said it directly, the economics were completely unsustainable.

Every hour of compute keeping Sora alive was an hour taken away from ChatGPT and Codex, products with a clear revenue path heading into the expected 2026 IPO.

OpenAI shut down the product people loved to fund the products Wall Street wants to see. That is IPO discipline. But it also shows real financial pressure sitting underneath those record-breaking headline numbers.

Within the next 12 to 18 months, the public markets are expected to absorb three of the most expensive IPOs in history, at roughly the same time.

SpaceX is targeting a listing at a $1.75 trillion valuation. OpenAI is expected to follow later in 2026. Anthropic has hired law firm Wilson Sonsini to prepare for an IPO, targeting 2026 or 2027 at its current $380 billion valuation.

The combined capital demand from these three listings could reach hundreds of billions of dollars. The most likely source is existing tech holdings being trimmed, including Microsoft, Nvidia, and Alphabet. But Microsoft alone has already lost $440 billion in market cap over concerns about its cloud commitments to OpenAI.

One thing worth paying attention to is Ethereum $ETH ( ▼ 2.97% ) . Many large tech funds use ETH as a liquid reserve asset, and when IPO pressure forces them to free up capital, ETH is often one of the first things sold. This is a connection between the tech IPO cycle and crypto market movements that most people do not pay attention to.

If selling existing tech stocks to fund three mega-IPOs pushes those same stocks lower, you get a feedback loop that reprices the entire sector at once. That is the scenario analysts are watching most carefully heading into late 2026.

Most of the bubble conversation focuses on OpenAI. But Anthropic is also pre-IPO, also pre-profitable, and also asking investors to price a future that has not happened yet. The difference is in the numbers underneath.

OpenAI is projected to burn $143 to $150 billion before reaching profitability. Anthropic expects to reach positive cash flow by 2027 or 2028, burning around $20 billion to get there.

Since both companies crossed $1 billion in annualized revenue, Anthropic has grown at roughly 10 times per year. OpenAI has grown at 3.4 times per year over the same period.

Anthropic went from $1 billion to $19 billion ARR between early 2025 and March 2026. Over 300,000 businesses now use Claude, with 80% of revenue coming from enterprise and API channels.

Claude Code alone hit an estimated $2.5 billion annualized run-rate by early 2026. Anthropic's gross margin is projected at 50% in 2025 and 77% by 2028, compared to roughly 33% for OpenAI.

The efficiency gap between these two companies is real, and it matters a lot when public markets start asking hard questions about the path to profitability.

The honest answer to whether we are in an AI bubble is: probably yes. And that does not fully change what you should do next.

The railroad bubble of the 1800s wiped out most investors. But it also built the infrastructure that powered the American economy for a century. The dot-com crash destroyed trillions in paper wealth. But it also built the internet.

The infrastructure being built right now, including chips, data centers, and model training pipelines, is real and permanent.

The question is not whether AI changes things. It already is. The question is which companies will still be standing when the market finishes deciding what that transformation is actually worth.

OpenAI is betting $150 billion on owning the consumer layer. Anthropic is betting $20 billion on owning the enterprise layer. Both are pre-IPO. Both are pre-profitable. And both are about to find out what public markets think.

The build is happening either way. The only thing left to figure out is who owns it when the dust settles.

⚠️ Disclaimer: This newsletter is for informational purposes only, just for fun and knowledge. This is not investment advice. Your money, your responsibility!