Electric vehicle leader Tesla (TSLA) is slamming the brakes on internal AI spending even as it embraces artificial intelligence at full throttle. According to an internal memo, Tesla will impose a $200 weekly cap per employee on AI tool usage starting July 6, with any excess requiring manager approval. Notably, beta products developed by CEO Elon Musk’s xAI are exempt from these restrictions, enjoying privileged status.

The move stands in stark contrast to Tesla’s recent aggressive push to promote AI adoption. Just months ago, management was actively consolidating AI tools across departments and even establishing internal leaderboards to encourage employees to compete on token consumption as a way to boost productivity. The abrupt policy reversal reflects how even a mega-cap company betting its future on AI struggles to contain costs that spiral out of control with surging usage.

According to sources cited by The Information, over the past six months, some software engineers frequently racked up thousands of dollars in weekly AI token costs, far exceeding the new $200 threshold. This has forced Tesla to pivot from encouraging usage to strictly limiting expenses to prevent AI spending from ballooning unchecked.

Centralized Platform Barely Launched Before Costs Spun Out

To unify AI resource management, Tesla established a central AI platform called “Bottle Rocket” last year, allowing employees to access multiple large language models including OpenAI, Anthropic, xAI’s Grok, and the AI coding tool Cursor, along with some unreleased model versions. Prior to this, employees mostly used various AI services through personal accounts, resulting in fragmented management.

This spring, Tesla further advanced company-wide AI governance measures, restricting access to unauthorized AI models on company computers and internal networks, establishing unified security protocols that prohibit employees from inputting confidential company data into unapproved generative AI services, and reinforcing data security management through training programs. These measures mirror similar approaches taken by Google, Apple, and other tech giants in recent years.

However, the centralized infrastructure had barely taken shape before usage costs skyrocketed. To incentivize AI adoption, some teams even created internal leaderboards ranking employees by token consumption, which ultimately led to runaway usage growth. Sources revealed that once the new policy takes effect on July 6, any spending exceeding $200 per week will require prior manager approval.

xAI Tools Get a Free Pass, But Grok Fails to Win Employee Favor

The most striking aspect of the new policy is the special treatment granted to Musk’s xAI products. Reports indicate that costs incurred from xAI beta products are completely exempt from the $200 weekly cap. This effectively means that while Tesla squeezes overall AI spending, it is leaving a funding lifeline open for its affiliated xAI ecosystem.

Musk has been pushing Tesla employees to use tools within his business ecosystem for months. After xAI began working closely with Cursor parent company Anysphere in April, Musk sent a company-wide email encouraging employees to try Cursor’s Composer coding model. Additionally, SpaceX is reportedly planning to acquire Anysphere in an all-stock deal valued at $60 billion, expected to close this quarter. Tesla engineers have consequently become early testers of unreleased versions of Grok and Composer, with xAI product leads even personally hosting discussions on Tesla’s internal Teams channels to gather employee feedback.

Yet despite Musk’s aggressive promotion, internal adoption appears limited. Four sources indicated that Grok is unpopular among Tesla employees, with many actually preferring Anthropic’s Claude model, demonstrating that employees remain pragmatic in tool selection, prioritizing performance and familiarity.

Tech Industry AI Budgets Sound Collective Alarm

Tesla’s AI spending U-turn is hardly an isolated case. As generative AI’s token-based billing model exposes the cost of every single prompt, more and more U.S. companies are imposing constraints on AI usage. Uber had already exhausted its full-year 2026 AI budget by April, subsequently capping employee usage at $1,500 per month. Giants like Meta, Amazon, and Walmart have also set spending limits or required employees to switch to lower-cost models.

What sets Tesla apart is how rapidly the policy shift occurred. The company had lagged behind some tech giants in regulating employee AI usage, yet in a short span, it dramatically pivoted from encouraging heavy use to strictly limiting expenses.

Market analysts suggest that capping employee AI spending does not signal a pullback in Tesla’s broader AI ambitions. When reporting first-quarter earnings in April, Tesla announced it was raising its full-year 2026 capital expenditure budget to over $25 billion, an increase of roughly 25% from prior plans, with funds primarily directed toward autonomous driving technology, Robotaxi, Optimus humanoid robots, AI chips, and large-scale computing infrastructure. This underscores that the company still views AI as a core future growth engine—it simply seeks more efficient resource allocation between day-to-day operating costs and the pace of innovation.