Grok Bot users are running out of usage credits faster than expected, and Elon Musk says the rising cost of running AI agents is to blame.
SpaceX is tightening Grok Bot usage limits after Musk acknowledged that operating costs were “unsustainably high relative to revenue.” The move follows complaints from users on X who say they’re exhausting their allowances much faster than before, with some reporting that they reached their limits in a single night.
Grok Bot is one of many AI agents launched over the past few months that can be assigned long-running, repeatable tasks. What sets it apart from other bots is its integration with X, allowing it to search posts, timelines, mentions, and other information.
It is part of a push by SpaceX to compete more heavily in the enterprise AI market, where rivals Anthropic and OpenAI currently lead. Grok Bot has seen some early success, with SpaceX COO Gwynne Shotwell confirming 800,000 weekly paid users in October, although the throttling may see that usage drop. Grok Bot can also be used by individuals for consumer-related tasks, similar to Meta Muse, but it has not seen comparable success due to its lack of a free tier.
SpaceX trying to keep out of the red
While private AI labs OpenAI and Anthropic have been more laissez-faire with spending, SpaceX has made greater efforts to reduce costs and increase revenue. It has rented out a significant portion of its data center capacity to rivals Anthropic, Google, and Reflection AI, reportedly generating over $2 billion in monthly revenue.
It has also explored potentially turning Grok Bot into a routing tool for AI models, similar to OpenRouter, which would reduce output costs. Its latest model launch, Grok 4.7, also introduced higher charges for heavy token usage to prevent power users from hurting its profitability.
AI agents under more scrutiny
AI agents are currently under considerable scrutiny following several major security incidents involving rogue AI. OpenAI disclosed several incidents in which its AI escaped confinement and hacked into websites and applications, including Hugging Face, the Australian health system, and the US Education Department. Some of these disclosures came months after the incidents occurred.
The Federal Trade Commission has opened a probe into AI agents operated by Anthropic, OpenAI, and other firms, which appears to have been prepared before the Hugging Face disclosure. A European Commission office has also requested information on the security incidents, with an investigation by the bloc expected in the near future.
Given the level of autonomy these agents can have when granted permission, world leaders are scrambling to ensure that consumers and businesses are protected. At the same time, government online services have repeatedly been targeted, raising concerns that successful attacks could disrupt access to public sector resources for millions of people.
More must-read AI coverage
- SS&C Intralinks DealCentre AI vs. Datasite: Which platform is built for the future of dealmaking?
- SS&C Intralinks FundCentre AI vs. Juniper Square: Which platform better supports modern private markets fund managers?
- Why Data, Not Models, Determines AI Success
- The Rise of the AI-Native Factory: How Physical AI Is Transforming Manufacturing
What Grok Bot’s restrictions mean for AI users
For businesses and individuals relying on AI agents, Grok Bot’s reported usage restrictions highlight an important concern: Paying for an AI subscription doesn’t necessarily guarantee consistent access to its capabilities.
Users who depend on agents for research, monitoring, or other recurring tasks could find their workflows interrupted if they unexpectedly exhaust their usage allowances. Businesses may face additional challenges when budgeting for AI tools, particularly if providers adjust their limits or pricing as operating costs rise.
Before committing to an AI agent, users should understand how usage is calculated, what happens when limits are reached, and whether additional credits are available.
As AI agents become more integrated into everyday work, reliability and predictable pricing could prove just as important as the capabilities themselves. For customers, the question is increasingly whether these tools can deliver enough consistent value to justify their cost.
For more on why autonomous AI workloads are becoming increasingly expensive, read TechRepublic’s AI Agent Cloud Costs Are Making Enterprise Budgets Harder to Predict, which explores the challenges businesses face in controlling AI spending.