Starting Saturday, Tesla employees who want to keep using ChatGPT, Anthropic’s Claude or other paid AI tools will have to make a choice. Spend more than $200 a week — roughly ₹19,000 — and they need a manager’s sign-off. Or switch to Grok, the chatbot built by Elon Musk’s other company, xAI, which is exempt from the cap entirely.
The internal memo, first reported by The Information and dated July 5, marks a sudden about-face. For the past six months, Tesla had been pushing staff to use generative AI.
The company built an internal platform called Bottle Rocket that consolidated access to models from OpenAI, Anthropic, xAI and Cursor. Some teams even set up dashboards that tracked token consumption — a move that effectively encouraged engineers to burn through more tokens, not fewer. It worked.
Too well. Sources familiar with the matter say some software engineers were spending thousands of dollars each week on AI tokens.
Those are the usage-based units that determine how much it costs to query a model. A single heavy user could rack up more in a day than the new weekly limit allows in a month. The $200 cap applies to all employees.
The only exception: beta versions of xAI’s products. That carve-out means engineers who need high-volume AI access can keep using Grok and Cursor’s Composer model without asking permission. Competing platforms like OpenAI’s GPT-4 and Anthropic’s Claude are now effectively rationed.
For Tesla’s India operations, the math is tight. At an exchange rate of roughly ₹95.37 per U.S. dollar, $200 converts to about ₹19,046 a week.
That is not a lot of tokens. A single complex coding query on a premium model can cost several dollars. An engineer debugging a large codebase could burn through ₹19,000 in a day or two.
The policy shift reflects a wider reckoning across the tech industry. Companies spent 2023 and early 2024 racing to integrate generative AI into every workflow.
Now the bills are coming due. AI model providers charge by the token, and heavy usage adds up fast. Tesla’s internal data apparently showed that the spending had grown faster than the company expected.
What the memo does not say is whether the cap is temporary or permanent. It does not say what happens to teams that already built workflows dependent on expensive models. It does not say whether the exemption for xAI’s beta products will remain in place once those products leave beta.
What it does say: starting July 6, anyone who wants to spend more than $200 a week on AI tokens needs approval. That puts a manager — not an engineer — in control of how much the company spends on AI.
It also puts xAI in a privileged position inside a company run by xAI’s founder. Tesla did not respond to requests for comment. The memo speaks for itself.
{“@context”: “https://schema.org”, “@type”: “NewsArticle”, “headline”: “Tesla exempts Musk’s xAI chatbot Grok from new AI spending cap for staff”, “description”: “Starting Saturday, Tesla employees who want to keep using ChatGPT, Anthropic’s Claude or other paid AI tools will have to make a choice. Spend more…”, “datePublished”: “2026-07-05T17:21:00+00:00T09:30:00+00:00”, “author”: {“@type”: “Person”, “name”: “Sebastian Smith”}, “publisher”: {“@type”: “Organization”, “name”: “InfoPulse Today”, “url”: “https://newsfeed.2bme.xyz/”}}
