Skip to content
← Companies

Baseten

HQ San FranciscoFounded 2019baseten.co ↗ (opens in a new tab)Engineering blog ↗ (opens in a new tab)Job board ↗ (opens in a new tab)

Baseten is an AI inference platform for serving and scaling open-source, custom, and fine-tuned models in production. It provides optimized model runtimes, autoscaling GPU infrastructure across clouds, and developer tooling so teams can deploy models with low latency and high reliability. Baseten serves AI-native startups and enterprises running demanding workloads such as LLMs, speech, and image generation.

Get Monday's brief

Free. We'll note that you care about Baseten.

What Baseten shipped, who joined or left, and what got funded or cut, summarized each week from public sources.

Latest brief

Baseten · Week of Oct 26–Nov 1, 2026

Baseten is reportedly raising about $2 billion at valuations of $26 billion and $30 billion

Bloomberg reported that Baseten is raising roughly $2 billion in two tranches, one pricing the inference company at $26 billion and the other at $30 billion, which would more than double the $13 billion valuation it reached in June.

Baseten has not confirmed the deal, so the terms remain reported rather than final (Anue).

Baseten named four senior engineering hires: a former Meta PyTorch inference acceleration lead as Head of Inference Platform, Anyscale's former Head of Engineering to lead forward-deployed engineering, a former Kindo CTO on the inference stack team, and a former NexHealth engineering head over Product Platform.

It says it keeps hiring across inference, model performance, infrastructure and forward-deployed roles (blog).

Base Labs released joey-1, a 2B-parameter open-weight model that returns a single answer token plus a calibrated probability, aimed at routing, moderation and classification steps inside agents; Baseten cites roughly 20 ms median latency and $0.01 per million tokens (blog).

Baseten and Postman partnered so Postman's Agent Stack handles agent tooling and governance while Baseten serves the models, giving Baseten a channel into Postman's enterprise developer base (blog).

Its engineers also rebuilt delta weight sync to apply updates against weights already on the GPU, cutting a roughly 770 GB checkpoint sync from about 36 seconds to around 15 and freeing host memory for KV cache (blog).

No Director, VP, or CTO postings in the week of Oct 26–Nov 1.

Roles 4

Roles at Baseten from its public job board, Director and above first. Most leadership roles are filled without a posting, so that list is often empty.

Director and above 0

No Director, VP, or CTO postings on Baseten's public job board right now.

Engineering management and team leads 4

Past briefs 1

Every weekly brief on Baseten, newest first.

  • Week of Oct 26–Nov 1Latest

    Baseten is reportedly raising about $2 billion at valuations of $26 billion and $30 billion

    7 stories

Highplane summarizes public sources in its own words. Something wrong? Write to hello@highplane.co.