Vast Gap in AI Model Token Pricing Between China and the U.S.
Miles Bennett
For the same 1 million tokens of AI processing, OpenAI charges roughly $26 while Chinese models charge about ¥0.5 — Silicon Valley investor Chamath Palihapitiya warns this price war will trigger financial blow-ups among high-cost players.
How wide is the pricing gap?
Speaking on the *All-In Podcast* on July 15, Chamath Palihapitiya laid out the comparison: OpenAI charges about $26 per million tokens — a standard unit of AI text processing, roughly equivalent to 750,000 English words.
Chinese AI models offer the same volume for approximately ¥0.5 (about $0.07), putting the price gap at over 350×.
This means → for the same task, the Chinese option costs less than one-three-hundredth of the American one. These are no longer in the same pricing universe.
Who breaks first?
Palihapitiya's view is blunt: the AI "intelligence" price war is squeezing every player that entered at premium pricing.
In plain terms = if your business model assumes AI services command high prices, and a competitor delivers equivalent capability at near-zero cost, your revenue model collapses.
He believes financial blow-ups among these companies are a matter of when, not if.
What does this price war signal?
This reflects a widening split in US-China AI pricing logic: American leaders rely on high margins to fund R&D, while Chinese firms use rock-bottom pricing to grab market share.
This means → for AI companies that depend on API-call revenue, pricing power is eroding fast.
For everyday users and enterprise customers, plummeting AI costs are good news. For investors, it is time to re-examine valuations built on the assumption that AI services stay expensive.
Content is for reference only, not financial advice.