%title%
Applied AI
The AI business continues to create and shift alliances between almost every leading tech firm. Companies may be incentivized to work together to fill each other’s AI or capital needs even as they compete more fiercely for consumer or enterprise dollars. That’s what’s happening with Microsoft and Anthropic. The two firms have recently become fast friends, with Microsoft on track to spend more than $500 million annually on Anthropic models to power its AI software, Copilot, as we reported Wednesday. 
Jan 15, 2026

Applied AI

Aaron Holmes headshot
Supported by Sponsor Logo

Hi! If you’re finding value in our Applied AI newsletter, I encourage you to consider subscribing to The Information. It contains exclusive reporting on the most important stories in tech, like this story from Qianer on TSMC's chip demands. Save up to $250 on your first year of access.


Welcome back!

The AI business continues to create and shift alliances between almost every leading tech firm. Companies may be incentivized to work together to fill each other’s AI or capital needs even as they compete more fiercely for consumer or enterprise dollars.

That’s what’s happening with Microsoft and Anthropic. The two firms have recently become fast friends, with Microsoft on track to spend more than $500 million annually on Anthropic models to power its AI software, Copilot, as we reported Wednesday. 

Now, Microsoft is racing to sell Copilot to the same businesses Anthropic is targeting with its own products.

Anthropic this week released Cowork, which aims to automate white-collar tasks—basically anything someone would do on their computer for work. For instance, you can point Cowork to a file full of screenshots of receipts and direct it to create a spreadsheet listing the expenses shown in the photos, Anthropic says. In another example Anthropic shared, Cowork parses transcripts of a manager’s meetings from the past week, generating a summary of the most important tasks they need to complete and creating a slide deck recapping their team’s goals and objectives, which the manager can share with staff.

Cowork can execute tasks involving any applications a person uses, and it can work in the background, letting the person do other things while the AI does its thing.

After years in which every major tech and AI firm—Microsoft and Anthropic included—have touted such possibilities thanks to generative AI, will Cowork finally do the trick? We don’t know yet, but it can’t be lost on Microsoft that the name Cowork shares a couple of key letters with the name of Microsoft’s AI brand. And CEO Satya Nadella previously referred to Copilot’s ambition to be a “digital coworker.”

Part of the reason for the hype around Cowork is it’s a direct offshoot of Anthropic’s wildly popular coding tool, Claude Code, which has captured the AI attention economy much the way ChatGPT did three years ago. (To understand why, see this piece.) 

Microsoft’s next challenge will be to harness Anthropic’s technology to differentiate Copilot from Cowork.

The situation heralds back to Microsoft’s prior history with OpenAI. Around the time that Microsoft invested $10 billion in OpenAI three years ago, executives thought adding the startup’s models to Bing to create a chatbot-like experience would finally give Microsoft’s search engine an edge over Google; instead, ChatGPT, released just weeks before the new Bing, stole Bing’s thunder.

And lately, Microsoft’s Office business has faced growing competition from ChatGPT, even as Nadella races to use OpenAI models to boost Copilot features in Office apps, such as making calculations in Excel or generating Powerpoint decks based on written prompts.

At the end of the day, Microsoft has many ways to win in AI—it’s still generating billions of dollars in new revenue a year from its OpenAI partnership, primarily from the startup renting Microsoft cloud servers to run its AI.

Anthropic, similarly, agreed last year to spend billions of dollars renting Microsoft servers. 

The advent of Cowork means chief information and technology officers will have yet another option to choose from in the growing menu of workplace AI products. Many of those tools, offered by just about every enterprise software firm on the market, are promising to automate the same kind of work.

Unlike those traditional enterprise firms, Anthropic can boast that it created the underlying AI models powering the automation and therefore can handle more cutting-edge tasks. On the other hand, Microsoft has existing contracts with most of the largest companies in the world, and it usually harps on security as a core concern customers should have if they work directly with young startups. (See Tuesday’s column for more on lingering security concerns involving products such as Claude Code.) 

So Microsoft could still have a leg up in this race. By working closely with Microsoft, Anthropic is hedging bets. 

A message from Google Cloud

5 trends defining AI agents in 2026

We interviewed Google experts and surveyed 3,466 global execs to find out how businesses are putting AI agents to work.

Explore how to go from one-off prompts to agents that run entire workflows. Discover real examples of how agents improve customer service, code quality, and threat detection. And learn why teaching your team is the only way this actually works.

Uncover the 5 major trends you need to define your business strategy in 2026.

Thank you for reading the Applied AI newsletter! I’d love your feedback, ideas and tips: aaron@theinformation.com.

If you think someone else might enjoy this newsletter, please pass it forward or they can sign up here.

Opportunities

Group subscriptions

Empower your teams to stay ahead of market trends with the most trusted tech journalism.

Learn more


Brand partnerships

Reach The Information’s influential audience with your message.

Connect with our team

About Applied AI

A new The Information franchise that will take you inside how businesses are using AI to automate all kinds of work.

Read the archives

Follow us
X
LinkedIn
Facebook
Threads
Instagram
Sent to niepodam@niepodam.­pl | Manage your preferences or unsubscribe | Help The Information · 251 Rhode Island Street, Suite 107, San Francisco, CA 94103