After users on X complained that their Grok Bot usage ran out early this cycle, Elon Musk replied that "the cost of serving Grok Bot was unsustainably high relative to revenue", a reply with about 13K likes and 5.5M views. Replit CEO Amjad Masad then quoted the news, saying people enjoying subsidized AI subscriptions should "get ready to get rug pulled", sparking a debate over whether flat-rate AI plans can last. This is an X discussion with personal views; xAI has not published a formal pricing or limit announcement.
Key Takeaways
- ✓Trigger: user @DogecoinNorway said their Grok Bot usage hit 100% with 4 days left until reset, a post with about 1.24M views.
- ✓Musk replied that the cost of serving Grok Bot was unsustainably high relative to revenue, with about 13.2K likes, 5.51M views and 1,900+ quotes.
- ✓Amjad Masad (Replit CEO) quoted it, warning subsidized AI subscriptions are about to get rug pulled, with about 2,347 likes and 146K views; in replies he said Replit never subsidized.
- ✓xAI has not published a formal limit or pricing change; rely on official notices for specifics.
Key Decision Metrics at a Glance
Turn your technical choice into a development budget
Compare 40 dev plans & simulate token costs vs $20/mo subscriptions
In-Depth Technical Analysis
Background
Over the past year, AI products have competed with flat monthly plans around $20, but a single agent task can burn a lot of tokens, and heavy users' inference costs can easily exceed the subscription fee. On November 1, X user @DogecoinNorway complained that their Grok Bot usage was already at 100% with 4 days left until reset, and asked Elon Musk for a reset. Musk replied: "The cost of serving Grok Bot was unsustainably high relative to revenue." That reply drew about 13.2K likes and 5.51M views.
What the debate is about
The focus is cost structure rather than any feature. Replit CEO Amjad Masad quoted Sawyer Merritt's summary of the exchange and wrote, "If you're enjoying a subsidized AI subscription get ready to get rug pulled," drawing about 2,347 likes and 146K views. When asked about Replit, he said Replit "never subsidized" and that its customers are people serious about what they build and willing to pay for reliable hosting. He also argued the bigger issue is a compute shortage, which means compute goes to customers paying full price. These are personal views; xAI has not published a formal pricing or limit change.
Benchmarks
This is an X discussion with no model benchmarks involved. No official cost, margin or limit figures were published. The numbers above are only the posts' public engagement counts.
What developers can do
If your workflow depends on a flat-rate AI tool: check the usage dashboard regularly instead of finding out when the limit is gone; break long-running agent tasks into smaller steps and move simple steps to cheaper models; and keep a pay-as-you-go API as a fallback for critical flows so a sudden limit change does not stop your work. For any pricing change, rely on official announcements from xAI and other vendors.
Benchmark side-by-side against alternatives, or calculate monthly token cost vs subscription break-even.
Discussion & Comments
0Sign in to join the discussion
Connect with AI developers to exchange benchmark insights.