Baseten, which runs AI models in production for software companies, is set to raise about $2 billion at a valuation of up to $30 billion, Bloomberg reported on October 30, citing people familiar with the discussions. Sands Capital, already an investor, is expected to lead.

The price would be 2.3 times the $13 billion valuation Baseten announced with its Series F only 130 days earlier, and six times the $5 billion of its Series E in January. Baseten did not respond to Bloomberg's requests for comment, and Sands Capital declined to comment, Investing.com reported in its summary of the story. Baseten had not posted about the round on its X account by 11:30 p.m. UTC.

$2BReported round size
$30BReported top valuation
2.3xStep-up from June's $13B
130 daysSince the Series F

What Bloomberg reported

The money would come in two tranches, the first valuing Baseten at $26 billion and the second at $30 billion, according to Bloomberg's account as summarized by Investing.com. The account also names Altimeter Capital, IVP and Spark Capital as existing investors. Rebecca Torrence, a Bloomberg reporter who writes about startups and venture capital, set out the terms in her own post shortly after the story ran.

Post on X by @RebeccaTorrenc5

Scoop! Baseten is set for a $30 billion valuation in its latest fundraise. The AI inference startup is raising in 2 tranches -- one at $26 billion and the other at $30 billion, per our sources. Sands Capital is leading. Round total = about $2 billion. w/ @BTLipschultz 👇

The round is not final and has not been announced. Until Baseten confirms it, the amount, the lead and both prices are Bloomberg's reporting, not the company's.

What Baseten sells

Baseten sells inference: the computing that runs a trained AI model each time a product calls it. Its website says it serves "open-source, custom, and fine-tuned AI models" on infrastructure built for inference at scale, and names customers including Canva, Cursor, Notion, HubSpot, Ramp and Wayfair. Its researchers and engineers also help customers post-train models for a specific job before Baseten hosts them.

In its Series F announcement on June 22, the company said "our revenue has grown 20x, and inference volume has grown 40x" over the previous year. It did not give a revenue figure, so no revenue multiple can be calculated from what Baseten has disclosed. It said the $1.5 billion would go to "the compute, software, and talent required to support our customers."

How fast the price has climbed

Baseten's own announcements give the last two marks. The reported round would be its third priced round of 2026.

Date

Round

Valuation

Jan. 23

Series E, $300M

$5B

June 22

Series F, $1.5B

$13B

Oct. 30

Reported, about $2B

$26B to $30B

Sources: Baseten's Series E and Series F posts; Bloomberg via Investing.com. Step-ups calculated by SaasBuzz.

IVP and CapitalG led the Series E. The Series F was led by Altimeter Capital, Conviction Partners and Spark Capital and co-led by Sands Capital and Wellington Management, and Baseten called it "our fourth fundraise in 18 months." Measured from the $13 billion mark, the reported top price adds $17 billion of paper value in 130 days, about $131 million a day. Even the lower $26 billion tranche would double it.

What it means for buyers of AI infrastructure

Investors had already priced one company that serves models, rather than training them, above $15 billion. Baseten's rival Fireworks announced on July 15 a $1.505 billion Series D at a $17.5 billion valuation, led by Atreides Management, Index Ventures and TCV, and said it had passed $1 billion in annualized revenue. At $30 billion, Baseten would be valued at 1.7 times Fireworks' July mark without having disclosed a revenue figure of its own.

For software companies that buy inference, the rounds mean their suppliers have money to add capacity, which is what Baseten said its last round was for. For investors, the question the numbers leave open is revenue. Fireworks has published one; Baseten has published only growth multiples. The week's other AI money is in our weekly roundup, and the training side of the spending is in our analysis of how AI data centers are being financed.