Hi - I answer from the OpenSmartRoute documentation: routing, the API, plans and quotas, self-hosting. Ask away, or open a support ticket if you need a person.
Grounded in the docs - follow a source before acting on it.
Meta and Microsoft cut Claude spending as Anthropic becomes a competitor - OpenSmartRoute
Meta and Microsoft sharply cut internal use of Claude models. These two companies were major clients for Anthropic. They are now reducing their spending on the company's AI tools. Executives at both firms told teams to switch to other software. This shift marks a change in how they view their partners. The relationship has moved from support to competition.
At Microsoft, the decision was driven by budget cuts and internal strategy changes. Scott Guthrie is the CEO of Microsoft Cloud. Jay Parikh leads the AI business unit at the company. Both leaders instructed employees to stop using external models like Claude. They want staff to use GitHub Copilot instead. Microsoft also plans to use OpenAI's models for its own products. This move affects how many people work with large language models daily.
Microsoft had projected spending over $1 billion annually on Claude. That number represented a huge portion of their cloud division budget. Executives decided to slash these per-employee costs significantly. The monthly budget dropped from $100,000 to about $10,000 per person. This is a reduction of more than nine times the original amount. Such a cut shows how aggressively they are trimming AI expenses.
The decision impacts engineers who manage enterprise AI deployments. Managers must now justify new purchases with tighter budgets. Teams need to find cheaper alternatives for their daily workflows. The shift forces companies to rethink their vendor strategies quickly. It also changes who gets access to the latest model features.
At Meta, the situation looks similar but involves different tools and numbers. Claude Code is a specific product designed for coding tasks. The user base for this tool fell from roughly 60,000 to 30,000 users. This represents exactly half of the previous number of active users. Part of this drop happened because of layoffs within Meta itself. Another reason was the launch of new internal tools like Muse Code and MetaCode.
Meta spent over $105 million on Claude Code in one 28-day stretch. That is a massive amount for a single month of usage. The company had already introduced AI cost-cutting measures in June. These steps were taken before the user base started shrinking further. The focus has clearly shifted toward building proprietary solutions.
Mirror Particle raises capital to create an AI engine that simulates changing human motivations. The company rejects large language models in favor of a foundation model trained on longitudinal data.
Mistral released a preview of Mistral Large 4, a one-trillion-parameter model trained in Europe.
Cost savings are likely a primary driver behind these cuts. But data security concerns also play a major role in this decision. Meta wants to sell competing products that do not rely on Anthropic. They reportedly wanted to restrict Anthropic's access to their training data. This creates an irony when a company worries about its own data being used by a rival.
Anthropic has grown extremely fast since becoming a partner. This rapid expansion makes them a headache for incumbents like Meta and Microsoft. Tools like Claude Cowork and ChatGPT Work are turning into Office alternatives. These tools offer features that mimic the core functions of productivity suites. They allow users to create documents, spreadsheets, and presentations with AI help.
The competition is not just about code anymore. It extends to general business workflows and office management. Companies fear losing control over how their data flows through these systems. They worry about third-party access to sensitive internal information. This security risk pushes them toward closed ecosystems built by trusted partners.
Engineers running models face new challenges in this changing landscape. They must evaluate if the cost savings are worth the potential risks. Managers need to decide which vendors remain safe for enterprise use. The market is shifting from open partnerships to competitive dynamics. This trend could affect how many companies adopt AI tools soon.
Why it matters for engineers running models and managers buying AI services
This situation highlights the risk of relying on a single vendor for critical infrastructure. Companies must now consider competition when choosing their AI providers. Budget constraints force a re-evaluation of every dollar spent on cloud services. Security teams will scrutinize data access policies more closely than before.
What to do when your main model provider becomes a competitor
Check if your organization has multiple vendors for AI capabilities. Compare the features and costs of different platforms regularly. Review your data security agreements with current partners carefully. Consider building internal tools to reduce dependency on external APIs. Monitor industry trends to anticipate shifts in vendor relationships.
Meta and Microsoft sharply cut internal use of Claude models
Meta and Microsoft are two of Anthropic's biggest enterprise customers. They are now sharply cutting their internal use of Claude models. Executives at both firms told teams to switch to other software. This shift marks a change in how they view their partners. The relationship has moved from support to competition.
At Microsoft, the decision was driven by budget cuts and internal strategy changes. Scott Guthrie is the CEO of Microsoft Cloud. Jay Parikh leads the AI business unit at the company. Both leaders instructed employees to stop using external models like Claude. They want staff to use GitHub Copilot instead. Microsoft also plans to use OpenAI's models for its own products. This move affects how many people work with large language models daily.
Microsoft had projected spending over $1 billion annually on Claude. That number represented a huge portion of their cloud division budget. Executives decided to slash these per-employee costs significantly. The monthly budget dropped from $100,000 to about $10,000 per person. This is a reduction of more than nine times the original amount. Such a cut shows how aggressively they are trimming AI expenses.
The decision impacts engineers who manage enterprise AI deployments. Managers must now justify new purchases with tighter budgets. Teams need to find cheaper alternatives for their daily workflows. The shift forces companies to rethink their vendor strategies quickly. It also changes who gets access to the latest model features.
At Meta, the situation looks similar but involves different tools and numbers. Claude Code is a specific product designed for coding tasks. The user base for this tool fell from roughly 60,000 to 30,000 users. This represents exactly half of the previous number of active users. Part of this drop happened because of layoffs within Meta itself. Another reason was the launch of new internal tools like Muse Code and MetaCode.
Meta spent over $105 million on Claude Code in one 28-day stretch. That is a massive amount for a single month of usage. The company had already introduced AI cost-cutting measures in June. These steps were taken before the user base started shrinking further. The focus has clearly shifted toward building proprietary solutions.
Cost savings are likely a primary driver behind these cuts. But data security concerns also play a major role in this decision. Meta wants to sell competing products that do not rely on Anthropic. They reportedly wanted to restrict Anthropic's access to their training data. This creates an irony when a company worries about its own data being used by a rival.
Anthropic has grown extremely fast since becoming a partner. This rapid expansion makes them a headache for incumbents like Meta and Microsoft. Tools like Claude Cowork and ChatGPT Work are turning into Office alternatives. These tools offer features that mimic the core functions of productivity suites. They allow users to create documents, spreadsheets, and presentations with AI help.
The competition is not just about code anymore. It extends to general business workflows and office management. Companies fear losing control over how their data flows through these systems. They worry about third-party access to sensitive internal information. This security risk pushes them toward closed ecosystems built by trusted partners.
Engineers running models face new challenges in this changing landscape. They must evaluate if the cost savings are worth the potential risks. Managers need to decide which vendors remain safe for enterprise use. The market is shifting from open partnerships to competitive dynamics. This trend could affect how many companies adopt AI tools soon.
Cost savings and data security concerns drive these major clients away
Cost savings are one likely driver, but not the only one. Meta has little reason to keep feeding Anthropic when it wants to sell competing products. They reportedly also wanted to restrict Anthropic's access to its training data. The irony of Meta worrying about someone else using its data is hard to miss.
Anthropic has also grown extremely fast and is becoming a headache for incumbents. Even Microsoft has reason to worry, since tools like Claude Cowork and ChatGPT Work are increasingly turning into Office alternatives. These tools offer features that mimic the core functions of productivity suites. They allow users to create documents, spreadsheets, and presentations with AI help.
The competition is not just about code anymore. It extends to general business workflows and office management. Companies fear losing control over how their data flows through these systems. They worry about third-party access to sensitive internal information. This security risk pushes them toward closed ecosystems built by trusted partners.
Engineers running models face new challenges in this changing landscape. They must evaluate if the cost savings are worth the potential risks. Managers need to decide which vendors remain safe for enterprise use. The market is shifting from open partnerships to competitive dynamics. This trend could affect how many companies adopt AI tools soon.
Anthropic's rapid growth threatens incumbents with Office-like alternatives
AI News Without the Hype – Curated by Humans
Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive "AI Radar" frontier report six times a year, full archive access, and access to our comment section.
Anthropic has grown extremely fast since becoming a partner. This rapid expansion makes them a headache for incumbents like Meta and Microsoft. Tools like Claude Cowork and ChatGPT Work are turning into Office alternatives. These tools offer features that mimic the core functions of productivity suites. They allow users to create documents, spreadsheets, and presentations with AI help.
The competition is not just about code anymore. It extends to general business workflows and office management. Companies fear losing control over how their data flows through these systems. They worry about third-party access to sensitive internal information. This security risk pushes them toward closed ecosystems built by trusted partners.
Engineers running models face new challenges in this changing landscape. They must evaluate if the cost savings are worth the potential risks. Managers need to decide which vendors remain safe for enterprise use. The market is shifting from open partnerships to competitive dynamics. This trend could affect how many companies adopt AI tools soon.
Why it matters for engineers running models and managers buying AI services
This situation highlights the risk of relying on a single vendor for critical infrastructure. Companies must now consider competition when choosing their AI providers. Budget constraints force a re-evaluation of every dollar spent on cloud services. Security teams will scrutinize data access policies more closely than before.
Engineers need to understand how these shifts affect their daily work. They manage models that power applications for thousands of users. A sudden cut in funding can stop new features from being built. Managers must decide which vendors remain safe for enterprise use. The market is shifting from open partnerships to competitive dynamics. This trend could affect how many companies adopt AI tools soon.
What to do when your main model provider becomes a competitor
Check if your organization has multiple vendors for AI capabilities. Compare the features and costs of different platforms regularly. Review your data security agreements with current partners carefully. Consider building internal tools to reduce dependency on external APIs. Monitor industry trends to anticipate shifts in vendor relationships.
You can start by auditing your current contract terms today. Look for clauses that allow you to switch vendors easily. Test alternative models to see if they meet your needs. Talk to your legal team about data ownership and usage rights. Keep a close eye on announcements from both Anthropic and your competitors.
