OpenRouter data showed token consumption up 224x and spending up 81x since early 2026; OpenAI at 52% tokens and 34% spending; Anthropic at 48% tokens and 66% spending.
FollowingDay 2194 sourcesupdated 7 Nov 2026
AI Token Consumption on OpenRouter Up 224x Since Early 2026
The gist
Multiple AI infrastructure platforms, covering both frontier and open models, are reporting similar trajectories of explosive token volume growth. The gap between OpenAI's 52% token share and 34% spending share versus Anthropic's 48% token share and 66% spending share shows the two companies are capturing different proportions of revenue relative to usage on OpenRouter.
When stories like this move, I write about them in an email each morning.
One free email each morning, around 8 AM US Eastern.
Check your email.
I have sent you a confirmation link. One click and you are in.
Not in your inbox? Look in spam or promotions. If it's there, mark it "Not spam" so future editions land in your inbox.
The full picture
Token consumption on OpenRouter grew 224 times and spending grew 81 times since the start of 2026. The platform crossed a 10 quadrillion token annual run rate, with volumes doubling every 10.5 weeks. In the last 9 weeks alone, OpenRouter processed more tokens than in all prior history. Weekly token volume grew five times in six months, from 5 trillion to 25 trillion tokens. OpenAI accounted for 52% of tokens but only 34% of spending on the platform; Anthropic captured 66% of spending with 48% of tokens. Fireworks AI reached 40 trillion tokens per day as of mid-July 2026, more than doubling from 15 trillion per day in April and quadrupling from 10 trillion per day in October 2025. Ollama CEO Jeffrey Morgan said open models now handle 80 to 90 percent of enterprise AI tokens, driven by economics: open models undercut frontier closed systems on price. Ollama Cloud saw 150 times growth in token usage since the start of 2026, with two distinct inflection points in early 2026 around coding agents and in April 2026 around the OpenClaw and Hermes agent wave.
How it developed
NVIDIA announces Vera CPU delivers 1.6x more agents per watt than fastest x86 CPU; early adopters report up to 2.2x speed improvements
Menlo Ventures reported average enterprise AI spend at 4.7x planned 2025 budget
OpenRouter crossed 10 quadrillion token annual run rate; volumes doubling every 10.5 weeks.
Ollama CEO Jeffrey Morgan reported open models handle 80-90% of enterprise AI tokens; Ollama Cloud at 150x growth since start of 2026.
Fireworks AI CEO Lin Qiao says platform processes roughly 15 trillion tokens per day, up from 13 trillion a few months prior
Sources
Milk Road AI on X
Rohan Paul on X
beth-kindig.medium.com
Related
- Related storyOpen-weight models' enterprise token share
- Open-weight AI models' token-to-spend split
- Chinese open models' lead over US labs
- Menlo Ventures enterprise AI spending report
- Stripe's reported acquisition of OpenRouter
BackgroundFrontier models, open models, and how they are compared
Get the daily email
When stories like this move, I write about them in an email each morning.
One free email each morning, around 8 AM US Eastern.
Check your email.
I have sent you a confirmation link. One click and you are in.
Not in your inbox? Look in spam or promotions. If it's there, mark it "Not spam" so future editions land in your inbox.