Descriptions:
AI inference startup Baseten has raised $1.5 billion across two tranches — one at an $11 billion valuation and a follow-on at $13 billion — making it one of the largest funding rounds in the AI infrastructure space. Bloomberg Technology sat down with Baseten CEO Tuhin Srivastava and Ultimoto Capital partner Apoorv Agrawal, who co-led the round, to discuss what the capital will fund and why inference is drawing such aggressive investment.
Srivastava explains that the proceeds will go toward procuring compute and hiring infrastructure engineers to meet surging demand from enterprise customers running fine-tuned open-source models. To avoid supply-chain concentration risk, Baseten now sources compute from 18 different cloud providers across 19 clusters. The company serves customers like Cursor and Open Evidence, helping them deploy specialized open-source models — including recent frontier-quality releases like GLM 5.2 — without building internal infrastructure teams.
Agrawal makes a sweeping case that inference will become one of the largest markets globally, driven by multi-agent workloads where a single user request can trigger hundreds or thousands of downstream model calls. He frames the frontier model layer as a high-stakes “game of emperors” while positioning the infrastructure layer as the more durable value-accrual zone — echoing Satya Nadella’s argument that enterprises must compound their proprietary data and workflows rather than ceding intelligence to third-party model providers.
📺 Source: Bloomberg Technology · Published June 22, 2026
🏷️ Format: Interview
