Moonshot AI eyes $2B annualized revenue as Kimi K3 lifts sales
What's the deal? Moonshot AI is targeting $2 billion in annualized revenue by the end of 2026, more than doubling its current pace. The Beijing-based startup told investors its annual recurring revenue topped $1 billion in August, up from $300 million in June, according to people familiar with the matter.
Why now? The jump followed July's release of Kimi K3, a model with 2.8 trillion parameters that topped performance benchmarks while costing far less than top US rivals. Its debut drew global attention and validated an ambitious bet on building an unusually large model.
What's the endgame? Moonshot found early success with coders and solo entrepreneurs, and is now chasing bigger enterprise clients for growth. It recently assembled engineers to help firms such as AsiaInfo TechnologiesDealroom has a profile for this one. Try Dealroom → and Kingsoft Cloud deploy its models.
Cloud providers offer another revenue stream. Kimi K3's license requires companies hosting the model commercially to strike a separate deal once their cumulative revenue passes $20 million over 12 months. Moonshot is in talks with Microsoft, Amazon, and GoogleDealroom has a profile for this one. Try Dealroom → for a 30% revenue share, Reuters reported last month.
The competitive picture: Moonshot is becoming a formidable rival to listed peers Z.ai and MiniMaxDealroom has a profile for this one. Try Dealroom →. MiniMax's annual recurring revenue hit $800 million in August; Z.ai's surged to $1.6 billion that month after passing $1 billion in July. Moonshot is raising funds at a $50 billion valuation, roughly matching Z.ai, ahead of a Hong Kong IPO as soon as this year.
What could go wrong? Chinese AI labs still trail US frontier firms, where Anthropic and OpenAI reached $65 billion and $40 billion in annualized revenue by mid-2026. Moonshot also faces allegations of illicit distillation from top US labs; this week, Anthropic accused it of covertly routing user requests to Claude models and passing off the responses as its own.
The signal: A single breakout model can supercharge an AI lab's prospects overnight. But labs everywhere face persistent investor concern over whether massive hardware and infrastructure bets will ultimately pay off.
Read more: Bloomberg
Image credit: Moonshot AI
