卖什么
GPUs for Agents
商业模式拆解
买家：谁掌握预算？
未核实ML engineers, researchers and AI startup teams spending their own GPU budget (CTO or head of ML for larger teams).
触发点：什么让他们现在购买？
GPU cost eating runway, or need for many parallel RL/eval sandboxes for agents.
工作流：什么输入变成什么输出？
Create a GPU instance from CLI or agent (tnr create), SSH in, snapshot disk when done; or fan out parallel sandbox rollouts from Python for RL environments, evals and post-training.
经济账：省了什么，怎么收费？
Usage-based GPU cloud billed per minute (instances) or per second (sandboxes), positioned as cheaper than AWS, Azure, GCP and CoreWeave (H100 $3.20 vs $6.16-$10.98 list).
渠道：怎么接触到客户？
Self-serve signup with CLI, VS Code extension, API and MCP server; sandbox access by request; reserved capacity via contact; student and referral programs.
融资历史与报道
| 融资日期 | 轮次 | 金额 | 投资方 | 来源 |
|---|---|---|---|---|
| — | YC 2024 夏季 | — | Y Combinator | ycombinator.com |
档案汇总自公开报道和公司官网。商业模式部分是 MaxBC 对这些材料的解读，不是已确认的事实；未知之处留空，不作猜测。 更新于 2026-10-12.