Enflame's 200% Debut Is a Bet on Nvidia's Absence, Not Enflame's Business
Enflame jumped more than 200% in its Shanghai trading debut Friday, completing the run of China's four leading domestic AI chip startups. The pop says less about the company, which still isn't profitable, than about the captive market export controls have handed it.
Enflame's 200% debut pop in Shanghai on Friday isn't a verdict on Enflame. It's a verdict on Nvidia's absence.
The Tencent-backed chip designer closed its first day of trading up more than 200%, a number that would make most Silicon Valley bankers blush. It also completes a set. Four of China's leading domestic AI chip startups are now public companies, and each one has been greeted like a national champion. That's not an accident.
What the Number Actually Says
A 200% first-day jump tells you what buyers will pay when supply is thin and the story is thick. Enflame has yet to turn a profit. It's selling accelerators meant to slot in where Nvidia's H20 and whatever comes after it used to sit, and it's doing that in a market where export controls have carved out a captive customer base.
Granted, the demand is real. Chinese hyperscalers need compute, and Washington keeps narrowing the list of chips they're allowed to buy. Every round of restrictions hands another slice of the domestic market to whoever can build something close enough. That's the thesis behind all four of these listings.
But a thesis isn't a business. And 200% in a day is a valuation event, not an earnings event.
The Case for Skeptics
Let me steelman the bear case, because it's stronger than the tape suggests. Enflame is competing with Nvidia on the hardest part of the stack, which is software. CUDA has somewhere north of four million developers locked into it after more than fifteen years of compounding. Chinese chipmakers have their own toolchains, and every one of them admits the migration story is a work in progress.
There's also the margin question. Fabrication still runs through TSMC or SMIC, and SMIC's advanced nodes aren't cheap or abundant. A domestic chip that costs more to make and performs worse than the incumbent is a tough sell outside a policy-protected market.
So is the premium justified? The question worth asking: are you buying a chip company, or are you buying a subsidy with a ticker symbol?
My Verdict
I'll take the bulls' side over a two-year horizon, and I'll say it plainly. Not because Enflame is a great business today, but because the structural setup is hard to argue with. Export controls aren't loosening. Inference demand inside China isn't shrinking. And Beijing has made domestic silicon a funding priority it intends to carry through the awkward middle years.
That combination can float an unprofitable company a long way. Admittedly, the valuations are getting silly, and I'm not sold on all four newly public names justifying their multiples. History suggests otherwise once the lockups expire. Watch what happened to the earlier waves of Chinese chip listings after the first quarter of trading.
Still, the pattern matters more than any single ticker. China is rebuilding its compute stack from the bottom up, and it's using public markets to pay for the effort. Enflame is the fourth entrant, not the last.
Time will tell, though. The thing to watch next is Enflame's first earnings report as a public company. If revenue is climbing and losses are narrowing, Friday was a warm-up. If the top line is flat and the burn is ugly, that 200% was a transfer of wealth from retail buyers to early backers.
Color me skeptical on the current price. Not on the direction of travel.
