Particle.news
Download on the App Store
3 ARTICLES
·
last mo.

Sail Research Raises $80 Million to Power Long‑Running AI Agents

Its throughput‑first inference stack with stateful Sailboxes aims to cut costs for hours‑long agent workloads.

Neil Movva met his cofounder Samin Menon their first day at Stanford University where they took exactly the same course load.
Image
Image

Overview

  • Sail Research announced its public launch and an $80 million seed and Series A on Thursday, June 25, 2026, with Kleiner Perkins leading the Series A and Sequoia leading the Seed.
  • The company says it rebuilt the inference stack from the chip level up and launched Sailboxes, persistent sandboxed runtimes that keep agents running for hours or days while charging only for active work.
  • Sail reports rapid customer adoption and token scale, saying early users have run trillions of tokens and that some workloads see roughly 3x to 10x lower costs versus existing providers.
  • The platform is compatible with OpenAI‑style APIs and several open‑source models, making it suitable for background, long‑horizon tasks like deep code review but not for latency‑sensitive live chat or voice applications.
  • Investors view Sail as a bet on surging token demand, yet analysts warn frontier labs and big incumbents could replicate or commoditize the throughput layer, which would blunt Sail's competitive edge.