Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
Microsoft is integrating Anthropic’s models into an upgraded version of its suite of workplace tools as it moves to reduce its dependence on OpenAI for its AI products.
The tech giant on Monday said that it would embed Anthropic’s Cowork, a user-friendly version of the group’s popular coding platform Claude Code, into its Microsoft “Copilot” virtual assistant.
Anthropic’s Claude will also be available in Microsoft’s Copilot chatbot alongside OpenAI’s most advanced models.
“We’re excited about the things we can do with Anthropic,” Judson Althoff, chief executive of commercial at Microsoft, told the FT. “We are increasingly working [with them] across organisations from engineering through to go-to-market.”
Microsoft’s deepening partnership with Anthropic comes after it agreed to invest up to $5bn in the start-up as part of a $30bn computing deal disclosed in November.
Microsoft said that Cowork would enable customers to use Anthropic’s AI for tasks such as building presentations, gathering financial data, emailing team members and coordinating teams’ diaries for meetings.
The move to integrate Anthropic products across a wide spread of Microsoft services further loosens OpenAI’s grip over the software giant’s AI services, which have been underpinned by its models for several years.
Executives at the Redmond, Washington-based software group have been seeking to reduce the group’s reliance on the $730bn ChatGPT maker. Microsoft AI chief Mustafa Suleyman said last month the group was pursuing “true self-sufficiency” by also building its own set of powerful models.
Microsoft remains a major shareholder in OpenAI, holding a 27 per cent stake in the model maker, having invested nearly $14bn since 2019.
OpenAI last week reached a deal with Microsoft rival Amazon as part of its latest funding round that will see the start-up purchase vast amounts of data centre capacity from the Seattle-based group. The two companies will also partner to create a new service to help Amazon Web Services customers build AI agents and applications.
Microsoft has in the past month sought to assure investors about its relationship with OpenAI amid shifts in the market for workplace tools and a surge in its data centre spending. Shares in the group are down more than 13 per cent so far this year.
Anthropic released Cowork in January, promising customers the ability to use its Claude Code tool without prior coding experience. The group went further last month by launching “plug-ins” for the tool that could be used across sectors such as law, sales and customer service.
Those releases rattled public markets leading to widespread sell-offs in a range of industries — including software, insurance and financial services — as Wall Street reacted to the threat from AI to broad swaths of white-collar work.
“The power of these [models and tools] being infused into a core platform has real durability,” Althoff said. “The idea that AI is going to eat SaaS or all information work . . . we [just] don’t think that’s [the case].”
The Copilot Cowork service will be made available this month to a limited set of customers before it becomes generally available later this year.
Read the full article here