Moonshot AI at $50 Billion: Cheap Only If You Squint at the Comparison

Two of China's most watched AI labs are chasing money at the same time, and the price tags on the table say more about investor appetite than about the businesses underneath.

The Sequence

Start with DeepSeek. Reuters reported the lab is set to close a funding round above 80 billion yuan, or about $11.9 billion, some time this month. Then look at Moonshot AI, its closest domestic rival, which filed confidentially for a Hong Kong IPO after a private round valued it at $50 billion, according to Bloomberg.

Here's the part that matters. Both labs report roughly $1 billion in annualized revenue. So investors are paying somewhere between 50 and 74 times that figure, depending on how you anchor the math. That's steeper than what OpenAI or Anthropic command at valuations near a trillion dollars.

Read that again. The smaller Chinese labs are being priced at a richer multiple than the American giants they're trying to catch.

The Wider Impact

The question worth asking: is that a bargain or a warning? Proponents will tell you Moonshot at $50 billion is cheap because OpenAI is worth so much more. Skeptics will point out that both companies book about the same revenue, so the comparison cuts the other way. And you can't call something cheap just because a bigger number exists elsewhere.

Admittedly, multiples in early-stage AI aren't really about this year's revenue. They're about who wins the next three years. That's the thesis behind every one of these rounds, and it's the narrative both labs are selling to their backers. But the track record for that kind of pricing is mixed at best when a sector gets this crowded this fast.

What actually shifted this month is the exit path. Moonshot's confidential Hong Kong filing is a real signal. It tells you the founders think public markets will pay up, and that they'd rather test that now than wait around for another private round. DeepSeek is doing the opposite, staying private and stacking capital.

Two strategies, same moment. One lab raises, the other lists.

For anyone watching AI-linked tokens, this matters more than it looks. When equity investors pay up to 74x revenue for a model lab, every AI token with a whitepaper and a GitHub repo gets a bump in perceived legitimacy. And when those same investors eventually mark the position down, the token market usually feels it first. That's been the pattern in every hype cycle I've covered, and I don't see why this one breaks it.

What Comes Next

Watch the DeepSeek round close this month. Reuters says it's coming, and the final number sets the ceiling for what every other Chinese lab can ask for. Granted, exact terms have a way of drifting, so treat the headline figure as a starting point, not gospel.

Then watch the Hong Kong listing process. Confidential filings typically surface as a prospectus within a few months, and that document will be the first honest look at Moonshot's margins, its compute burn, and how concentrated its customers are.

If the prospectus shows revenue climbing past $1 billion while costs stay flat, the $50 billion tag starts to look defensible. If it shows heavy compute spend eating straight through the top line, the IPO prices soft and the whole comparison falls apart.

So here's the stance. A 50 to 74x multiple on $1 billion of revenue assumes a growth curve that neither lab has publicly proven yet. To be fair, public proof isn't how private AI rounds work. Time will tell, though. The funding close and the prospectus are the two data points that settle this, and both land within the next few months.