Odyssey Closes $310M Series B at $1.45B Valuation, Secures Amazon Deal

AI lab Odyssey raised $310 million in a Series B round on June 17, 2026, pegging its valuation at $1.45 billion, according to Reuters. The round was accompanied by a deal with Amazon — terms undisclosed — which adds a significant strategic dimension beyond the capital itself.

The unicorn threshold is now almost unremarkable in AI infrastructure, where nine-figure raises have become routine cadence. What stands out here is the Amazon pairing. A hyperscaler deal embedded in a funding round typically signals more than a commercial agreement: it often means preferred cloud infrastructure commitments, co-development arrangements, or distribution leverage that a pure financial round cannot replicate. Whether this one carries those elements has not been confirmed, but the structure is worth watching.

Odyssey's focus areas have not been detailed in reporting to date, which makes precise competitive placement difficult. That said, a $1.45 billion valuation at Series B puts the company in territory occupied by well-funded foundation model developers and AI tooling platforms — not early-stage research shops. The scale of the raise suggests investors are pricing in either a near-term path to significant revenue or a position in a category they expect to consolidate quickly.

The Amazon angle may also reflect the ongoing contest among hyperscalers to lock in AI partnerships before the market's architecture firms up. AWS, Google Cloud, and Microsoft Azure have each made high-profile bets on specific AI labs — OpenAI's deep Microsoft tie-up being the most prominent — and a deal embedded at the funding stage is an efficient way to establish preferential relationships before valuation multiples climb further.

The $310 million figure is substantial but not exceptional by the standards of the current AI funding cycle. What the round does establish is that Odyssey has cleared the credibility bar required to attract both institutional capital at scale and a hyperscaler's direct interest simultaneously. That combination is harder to assemble than either component alone.