The message states that Meta and Microsoft are urging employees to reduce the use of Claude and to make more use of their own AI in order to save costs.
The Block
47m ago
Ai Focus
According to The Information, Meta and Microsoft have instructed employees to reduce the use of AI models under Anthropic, and instead rely more on their own-developed AI tools to cut costs. Microsoft's internal annual expenditure on Anthropic AI was once estimated to be at least $1 billion, but recently this related budget has been cut by more than one-third; the number of internal Claude Code users has also decreased from about 60,000 to around 30,000.
Helpful
No.Help

IT News on October 6th: According to The Information, Meta and Microsoft have asked employees to reduce the use of AI models under Anthropic in order to cut costs, and have encouraged employees to make more use of their own developed AI tools.

It is reported that earlier this year, Microsoft executives estimated that the annual expenditure on using Anthropic AI within the company would reach at least $1 billion (Note from IT: The current exchange rate is approximately 6.714 billion RMB). Subsequently, Microsoft asked employees to reduce their reliance on Claude and make more use of its own-developed AI tools; recently, the related budget has been cut by more than one-third.

Microsoft's recent adjustment is mainly aimed at internal daily use and does not indicate a corresponding decline in the demand for its Anthropic model services provided to external clients. Currently, the company continues to offer customers Claude and other AI models through its enterprise-level cloud platform.

Internal data shows that the number of people using Claude Code has significantly decreased to about 30,000, whereas earlier this year it was around 60,000. In addition to the reduction in the total workforce due to a layoff of about 10% this spring, Meta's recent heavy promotion of its own AI tool is also a key reason for the decline in usage of Claude.

At present, Meta has launched several internal AI tools for employees to use. Among them, the number of internal users of MetaCode has exceeded 30,000; another tool, Muse Code, which is currently in testing with external customers, was launched in August, and the number of internal users has also surpassed 6,000. Muse Code is equipped with a large model developed in-house by Meta and is widely regarded as a direct competitor to Claude Code.

However, despite the significant reduction in the number of users within Claude Code of Meta, its expenditure scale remains quite considerable. It is reported that Meta spent over 105 million US dollars (approximately 705 million RMB at current exchange rates) on Claude Code in the past 28 days.

According to the prospectus for its initial public offering (IPO) previously submitted by Anthropic, two major customers contributed approximately 25% of its revenue; although their names are not directly disclosed, the market generally speculates that Meta and Microsoft are its core clients. In addition, Anthropic stated in the prospectus that most of its major customers have not signed long-term cooperation agreements with the company, and these customers may reduce or even cease their purchasing expenditures at any time.

Tip
$0
Like
0
Save
0
Views 22
HQYC reminds readers to view blockchain rationally, stay aware of risks, and beware of virtual token issuance and speculation. All content on this site represents market information or related viewpoints only and does not constitute any form of investment advice. If you find sensitive content, please click“Report”,and we will handle it promptly。
Submit
Comment 0
Hot
Latest
No comments yet. Be the first!
Related
The Pros and Cons of Layoffs to Be Determined Later
Some companies will issue notices of layoffs in advance, but it can take weeks or even months to determine which positions will be affected. Nike, Meta, and GitLab have all done this before. The article states that this approach may give employees time to find new jobs, but it could also increase anxiety throughout the company.
Businessinsider
·2026-10-06 17:53:37
6
ChainIT released the complete business architecture of Agentic Web3 in Singapore during TOKEN2049 week
ChainIT Inc released its “Agentic Web3 Complete Commerce” architecture during the week of TOKEN2049 in Singapore, and launched an executive white paper titled “ChainIT Transaction Truth: Stablecoins, Web3 Wallets, and Multi-Track Commerce.” The company stated that this architecture is designed to enable AI agents to conduct transactions through Web3 wallets, stablecoins, and traditional payment channels within a framework that ensures verified identities, authorized boundaries, runtime compliance, and controlled execution, while providing verifiable evidence for settlements and business outcomes.
PR Newswire
·2026-10-06 17:44:01
8
Ontinue Launches ION Dark Web Monitoring Service to Enhance Visibility of External Threats
Ontinue announces the launch of ION, for, and Dark Web Monitoring. This is a new managed add-on service that extends ION MXDR beyond corporate boundaries, continuously identifying exposed credentials, detecting brand impersonations, and detecting emerging external threats. The company states that this service combines investigation, automation, and response capabilities to transform external threat intelligence into actionable security results.
PR Newswire
·2026-10-06 17:33:16
8
Damming Robotics strengthens network security, promotes the deployment of trusted entities AI
Damion Robotics stated that its TM AI Cobot S series has obtained the IEC Cybersecurity Certification Level 2 (SL2) on April 2, 2023, after passing an independent assessment conducted by DEKRA. The company claims to be the world's first manufacturer of built-in vision AI collaborative robots to receive this certification, and this move will strengthen the security foundation for the deployment of collaborative robots in interconnected manufacturing environments.
PR Newswire
·2026-10-06 17:22:53
16
Bitcoin's "4chan prophecy" once predicted a peak in 2025; now it targets a bottom
An anonymous 4chan trader correctly predicted that Bitcoin would approach a market peak of $126,000 in October 2025, and now their model predicts that the next significant bottom will occur around October 5, 2026. The article states that this model also forecasts Bitcoin's next target range to be between $190,000 and $250,000, while Citigroup has raised its 12-month target price for Bitcoin from $82,000 to $113,000.
Coinpedia
·2026-10-06 17:11:56
23
View More