Kimi K3 Overwhelmed by Demand, Suspends New Subscriptions

Moonshot AI's Kimi K3 has become the latest indicator that the global AI race is no longer a two-player game between the US and China — it's genuinely multi-polar, and demand is outstripping supply.

What Happened

Kimi K3 debuted last week and immediately stopped accepting new subscriptions because demand overwhelmed its compute capacity. The model is closing in on Anthropic's Fable 5 and OpenAI's GPT-5.6, and has surpassed both on at least one independent benchmark.

The Reaction

David Sacks, the Craft Ventures founder who previously served as the administration's AI czar, described Kimi K3's performance as "concerning," writing: "This is how you lose the AI race."

July 2026: The Month AI Became Multi-Polar

Kimi K3's launch is part of an extraordinary month for frontier AI:

Date	Model	Company
Jul 1	Claude Sonnet 5, Fable 5	Anthropic
Jul 2	NanoBanana 2 Lite, OmniFlash	Google DeepMind
Jul 9	GPT-5.6 (Sol, Terra, Luna)	OpenAI
Jul 16	Grok 4.5	xAI
Jul 16	Kimi K3	Moonshot AI
Jul 17	Gemini 3.5 Pro	Google DeepMind

That's six frontier model releases from five different organizations in roughly three weeks.

The Pricing Divergence

The model market is bifurcating dramatically:

• Gemini 3.5 Pro: $15 input / $60 output per million tokens
• Kimi K3: Competitive pricing (exact figures TBD)
• DeepSeek V4 (July 24): $0.87 per million output tokens — 69x cheaper than Gemini

What This Means

For enterprise buyers, July 2026 represents an unprecedented opportunity to compare multiple frontier models before committing to deployment contracts. For the AI industry, it signals that the era of one or two dominant model providers is over.

Key takeaway: The AI model market is now truly competitive. Chinese AI labs are shipping frontier-quality models. The winners will be determined by pricing, reliability, and ecosystem — not just benchmark scores.