Lambert pegs open model economy as Fireworks hits 40T/day

Fireworks is now serving more than 40T tokens per day to customers' customized models, growth the company says has run over 2x from this summer. That is the biggest hard number in Nathan Lambert's latest paid note at Interconnects, and in his own telling it is less a ceiling than the one vendor disclosure he has to work with.

Lambert's tracker, the Interconnects Dashboard, leans on public disclosures from three open model inference startups, Baseten, Fireworks and Together AI, plus the open model share of OpenRouter's reported token volume. Hyperscaler footprints on Amazon Bedrock and Azure, he notes, are major but undisclosed. The accompanying chart, titled 'Open model inference exponential growth,' is a log plot of those same startup disclosures over time.

'All the measurements we have on open model adoption are incredibly imprecise,' Lambert writes, calling the dashboard 'an aggregator of all of the best data on open model adoption and performance from across the web.' The rest of the note sits behind the paywall.