xAI resets Grok Build limits after caching problems

xAI has announced a reset of usage limits for Grok Build Beta across all accounts. The cause was a set of problems in the caching system that made users hit their limits too quickly.

The limits problem

After Grok Build Beta launched, many users ran into an unexpected problem: they reached their usage limits very quickly. Users reported that after only a few attempts to use the tool, they saw a message saying their limits were exhausted.

This was especially frustrating for testers who wanted to find out what the new tool can do but could not explore it properly because of the restrictions. It got in the way of full testing of the functionality, of collecting feedback and of understanding Grok Build's real capabilities.

xAI actively collected complaints and reviews from users on social networks, Reddit and other channels, recognizing that the problem was widespread among testers and needed an urgent fix. For a developer it is very important to understand what exactly is getting in users' way, and xAI did not ignore the criticism.

The technical cause

xAI's engineering team ran a detailed analysis of the logging system and identified the root of the problem: faults in the caching system. The cache miscounted the resources used, which caused the limits to trigger prematurely. The system wrongly concluded that a user had already exhausted their limits, when in fact they had used only a small part of their allotted request budget.

Once the problem was found, the team focused on optimizing the cache so that the system would correctly track actual resource use. The team did significant work reworking the cache logic, added new verification mechanisms and tested the fixes across various usage scenarios. xAI engineers also checked whether the same problem recurs elsewhere in the system.

The fix and the limit reset

After fixing the problem, xAI decided to reset usage limits for all accounts, giving users a second chance to test the tool properly. Alongside this, all caching fixes were deployed and tested on all platforms:

• Full reset of the usage counter for all users
• A fixed caching system that counts resources more accurately
• Restored ability to test Grok Build Beta properly
• Resumed collection of feedback from the community

The company thanked users for sharing the problem and helping to identify a critical bug, and actively asks them to keep sending feedback and suggestions for improving the tool. This helps xAI make the final version of Grok Build genuinely useful and reliable.

What this means

This shows a sound, user-focused approach by xAI to developing and testing new features. A beta phase implies exactly this cycle: ship a feature quickly, collect feedback actively, find problems, fix them and run a new round of testing. xAI did not hide the problem or ignore complaints; it responded promptly to criticism and worked on a solution. That builds trust in the team and gives reason to hope for a high-quality final product.