The Information AM

Save 25% on an annual subscription to read the most important news about technology and business first. For even more, Save $250 on The Information Pro for unlimited access to our proprietary org charts, databases and surveys.

1.	Microsoft Cloud Exec Joins Anthropic as Head of Infrastructure By Aaron Holmes Source: The Information

Longtime Microsoft executive Eric Boyd is joining Anthropic as the AI lab’s head of infrastructure, Boyd said on Tuesday.

Boyd has spent more than 18 years at Microsoft, and for the past several years has led engineering of hardware and software needed to host OpenAI and Anthropic models on its Azure cloud platform. In that role, he helped plan Microsoft’s initial construction of massive computing clusters that were needed to train and run OpenAI’s models, as well as the computing infrastructure that Microsoft used to run its own AI-powered Copilot products.

In his new role, Boyd will lead Anthropic’s infrastructure team at a time when the company is facing a computing crunch in the face of surging usage of its models. Anthropic has recently discussed leasing its own data centers and has been hiring a team of cloud veterans including former Google Cloud executives, The Information previously reported.

Anthropic’s annualized revenue has surged to over $30 billion, the company said on Monday, up from $9 billion at the end of last year. But the company also said it was dealing with limited computing capacity this week when it started charging users more money to connect its Claude Code agent to services like OpenClaw, which drive heavy usage of its models.

2.	Intel Says It’s Joining SpaceX and Tesla ‘Terafab’ Project By Theo Wayt Source: The Information

Intel says it’s now involved in a chip-making project led by Elon Musk’s companies SpaceX and Tesla, though neither company provided details on how they will work together or a timeline for the project.

Musk announced plans for Terafab at an event in Texas last month, saying that it will eventually manufacture chips for Tesla’s vehicles and robots, as well as SpaceX satellites. He did not provide a timeline or estimated cost for the project. Tesla currently uses AI4 chips manufactured by Samsung in its vehicles, while SpaceX subsidiary xAI is a large Nvidia customer.

In a Tuesday morning post on X, Intel shared an image of Musk and Intel CEO Lip-Bu Tan shaking hands and said it would “join” the project, known as Terafab. Intel said its ability to “design, fabricate, and package ultra-high-performance chips at scale” will “help accelerate” Terafab. Tesla and SpaceX did respond to requests for comment. Intel declined to comment beyond the X  post. Intel shares were up 2% early Tuesday.

3.	Anthropic Announces AI Cybersecurity Project Powered By Claude Mythos Model By Stephanie Palazzolo Source: The Information

Anthropic announced on Tuesday that it will be providing access to its unreleased Claude Mythos model to more than 40 organizations, including Apple, JPMorgan Chase and the Linux Foundation, so they can test their software for security vulnerabilities. Anthropic says that Claude Mythos is significantly better at identifying and exploiting vulnerabilities compared to prior models.

Anthropic said in the blog post that in recent weeks, it was able to use the AI to identify “thousands of zero-day vulnerabilities” in “every major operating system and every major web browser” including Linux and OpenBSD.

Anthropic said that it does not plan to make the version of Claude Mythos it’s sharing with these partners generally available, but it hopes to eventually “enable…users to safely deploy Mythos-class models at scale” once it develops the necessary safeguards.

The company also said that it would donate $100 million worth of model usage credits to Project Glasswing, its name for the partnership with companies testing their software. Anthropic is also donating $4 million in cash to open-source security organizations. After those donations run out, Claude Mythos Preview will be available to partners at $25 per million input tokens and $125 per million output tokens—five times more expensive than its most advanced model today.

4.	Amazon Reduces Package Volume with USPS By Catherine Perloff Source: The Information

Amazon has reduced the volume of packages it will ship through the US Postal Service to 80% of the current volume, under a new agreement the two sides struck.

“We’re pleased to have reached a new agreement with USPS that furthers our longstanding partnership,” an Amazon spokesperson said. The company didn’t announce the 80% figure, which was first reported by Reuters.

Amazon and the postal service have been at loggerheads for several months renegotiating the contract, which was due to expire in October. The USPS wanted to change the terms under which it delivers packages for retailers. Instead of signing long term contracts, it wanted to have the flexibility to let retailers bid for the USPS’ services on an auction basis. Amazon wasn’t in favor of the switch.

Amazon makes up $6 billion of the postal service’s revenue, Reuters reported. Postmaster general David Steiner told Congress in March that the service is poised to run out of cash in less than 12 months.

5.	AWS CEO Says Claude Code Won’t Replace SaaS By Laura Bratton Source: The Information

Amazon Web Services CEO Matt Garman on Tuesday played down the idea that Anthropic’s highly-praised Claude Code tool could undercut existing enterprise software companies. But Garman warned that if incumbent software firms try to “protect what they have and not lean in” to AI, “they’re in trouble.”

Speaking during the HumanX conference in San Francisco, Garman said the idea that companies could use Claude Code to write software to handle customer relations management—a widely used enterprise software product sold by firms like Salesforce—was “overblown.”

He acknowledged that AI could be “enormously disruptive” to software but it was also a “huge opportunity,” given they “know a lot about the areas they operate in [and] they have a large existing customer base” as well as a “huge trove of existing data.”

“They know more about their area, the edges of their software…and so they are in a better position to build the next generation of AI enabled products,“ he said. Still, he said that everybody running an existing enterprise software firm “should really be questioning what are the things I should do differently.”

Garman’s comments are noteworthy given that AWS’ parent Amazon is a shareholder in Anthropic and AWS is a major cloud provider of Anthropic models. AWS also offers businesses access to a wide range of enterprise software products sold by other firms.