By Krystal Hu, Greg Bensinger and Jody Godoy
July 16 (Reuters) - The U.S. Federal Trade Commission
has asked Amazon.com AMZN.O to provide more details on its
deal to hire top executives and researchers from artificial
intelligence startup Adept, a person familiar with the matter
told Reuters.
The request reflects the FTC's growing concern about how AI
deals have been put together and follows a broader review of
partnerships between Big Tech and prominent AI startups.
The informal inquiry into Amazon, which has not previously
been reported, centers on last month's announcement that Adept
Chief Executive David Luan and others were leaving to join
Amazon, which would also license some of the startup's
technology.
Such inquiries do not necessarily result in an official
investigation or enforcement action.
Amazon is trying to catch up with rivals Google GOOGL.O
and Microsoft MSFT.O , which partners with OpenAI, in
developing its own large language models that can respond almost
instantaneously to complicated prompts or queries.
Reuters previously reported on Amazon's efforts to establish
a new organization called the Artificial General Intelligence
(AGI) team, focused on developing large language models. Luan is
now running the “AGI Autonomy” team, consisting of many former
Adept employees, and reporting to Rohit Prasad, head of the AGI
team.
Founded in 2022, Adept made a splash by raising over $400
million from venture capital investors with the aim to train
large language models to perform general tasks for enterprise
clients. Once valued at over $1 billion, it released some
open-source models but failed to launch successful commercial
products. It's unclear whether Amazon has compensated Adept
investors, or the details of the licensing fees.
Amazon, Adept and the FTC declined to comment.
The regulator is already investigating a similar move by
Microsoft, which hired away much of another startup, Inflection
AI's leadership and employees and agreed to pay a roughly $650
million licensing fee. The FTC is looking into whether the deal
was a play to skirt merger disclosure requirements, a source
told Reuters last month.
This is not Amazon's first AI startup bet. Amazon has
invested $4 billion in AI startup Anthropic since September,
taking a minority stake in the San Francisco company.
The FTC earlier this year launched a study of investments
and partnerships in the AI space, demanding information on
Microsoft's relationship with OpenAI, and Anthropic's tie ups
with Google GOOGL.O and Amazon.
The extensive document request, made in January, seeks
details on how AI company partnerships with Big Tech influence
strategy, pricing decisions, access to products and services,
and personnel decisions.
U.S. antitrust enforcers also expressed concern about Big
Tech companies wielding their existing advantages in AI to shut
out smaller competitors. The FTC and Justice Department have
staked out responsibility for potential probes into Microsoft,
OpenAI and chipmaker Nvidia.
(Reporting by Krystal Hu and Greg Bensinger in San Francisco,
Jody Godoy in New York; Editing by Kim Coghill)
((Jody.Godoy@thomsonreuters.com;))