Elon Musk said on X on 1 November that “The cost of serving Grok Bot was unsustainably high relative to revenue”. He wrote it at 20:34 UTC in a reply to a user who asked for a Grok Bot reset because “Lots of people say their usage ran out really fast this week”.
By our arithmetic, Musk replied about 23 minutes after the question. His post had about 144,795 views when we read it at 21:20 UTC, and the question about 116,106.
Business Pill · Usage-Based Pricing
An evergreen Business Engineer explainer on usage-based pricing. It explains the general idea only and does not discuss Grok Bot.
The key insight: As we read it, Musk’s reply puts Grok Bot’s limits down to cost against revenue, the trade-off a flat monthly plan creates when one customer’s agent use grows.
The question came from the account @DogecoinNorway at 20:11 UTC: “Any chance for a Grok Bot reset, @elonmusk?” Musk’s answer did not say whether a reset would happen, what the limits are now, or when they changed.
A post by Mark Kretschmann about 37 minutes later, by our arithmetic, quoted Musk’s reply and said it explained “the apparent cuts to usage limits”. It added that “Several users had complained that their usage meters suddenly jumped to 100% without warning”. Those complaints are the users’ and Kretschmann’s account; Musk’s post does not describe the meters.
Plans that include Grok Bot, in US dollars per month, per GIGAZINE, as cited in our 31 October piece (SuperGrok Heavy $300, Cursor Ultra $200, Cursor Teams Premium $120 per user). Musk said on 1 November that serving Grok Bot cost too much relative to revenue; he gave no figures.
What Was Posted on X
@DogecoinNorway The cost of serving Grok Bot was unsustainably high relative to revenue
Elon Musk’s 1 November post, at 20:34 UTC, replying to a user who asked for a Grok Bot reset. It reads: “The cost of serving Grok Bot was unsustainably high relative to revenue.” It gives no figures for cost, revenue or limits.
Any chance for a Grok Bot reset, @elonmusk Lots of people say their usage ran out really fast this week.
The question Musk answered, posted by @DogecoinNorway at 20:11 UTC on 1 November: “Any chance for a Grok Bot reset, @elonmusk? Lots of people say their usage ran out really fast this week.”
How Grok Bot Is Sold
In our Grok Bot 1.0 piece, access to Grok Bot 1.0 went to SuperGrok Heavy, Cursor Ultra and Cursor Teams Premium subscribers, at $300 a month, $200 a month and $120 per user a month, per GIGAZINE.
Our X Subscription pricing piece found Cursor’s pricing page describing the $8 entry tier as a “Shared AI usage pool across all products”, with no absolute usage amounts on the page. As we read it, a fixed monthly price on a shared pool means heavier agent use lowers the margin on each subscriber until the pool is tightened.
The Business Lever: How an Agent Is Set Up Moves Its Cost
Our earlier pieces show how much agent costs can differ between setups. In our shared agent memory piece, one shared history cost $914.36 per 398-task run against $161.77 with no memory, about 5.7 times as much by our arithmetic.
The other side is cheaper models. In our Revolut piece, Revolut put self-hosted GLM 4.5 at about 30 cents per million tokens, a tenth of its prior bill, per StartupHub’s account of a talk. As we read it, a subscription agent’s margin depends on which model serves each task as much as on the price of the plan.
The Structural Read
As we read it, a shared usage pool lets a provider cap what each subscriber can consume without changing the plan’s price.
By our arithmetic, Musk replied about 23 minutes after a user asked for a reset, and the reply had about 144,795 views when we read it.
As we read it, cheaper models are one lever and tighter limits another; Musk’s post names the cost problem but not which lever xAI pulled.
“The cost of serving Grok Bot was unsustainably high relative to revenue”
Three Implications
FOR GROK BOT SUBSCRIBERS Musk’s post does not say whether limits will be reset or what they now are, so the plans’ usage amounts remain unstated.
FOR COMPANIES SELLING AGENTS ON FLAT PLANS Our earlier reporting found one shared agent history costing $914.36 per 398-task run against $161.77 without memory; as we read it, how an agent uses memory can move serving cost.
FOR MODEL BUYERS Revolut put self-hosted GLM 4.5 at about 30 cents per million tokens, a tenth of its prior bill, per StartupHub’s account of a talk.
The framework says: “That ratio is the gross-margin equation of the agentic era.” The ratio it means is work produced per token burned.
As we read it, Musk’s reply is that equation seen from the seller’s side: revenue per subscriber against the tokens Grok Bot burns serving them. The framework predates his post and does not discuss Grok Bot.
Musk’s post gives no figures: it does not state Grok Bot’s serving cost, its revenue, the old or new usage limits, or the date any limit changed. We found no xAI page documenting a change in Grok Bot limits. The meter complaints are users’ complaints as Kretschmann described them, not statements from xAI. The 23-minute, 37-minute and 5.7x figures are our arithmetic, and the margin reading is ours.
This Week In The AI Supercycle
Anthropic’s Reality Check, The Capital Squeeze, & The New World Order
This week’s issue: for two years the AI build-out was priced as a growthstory. This week it started being priced as a credit story.
Elon Musk said on 1 November that the cost of serving Grok Bot was unsustainably high relative to revenue, replying to a user who said people’s usage was running out fast that week. GIGAZINE reported that Grok Bot 1.0 access went to plans costing $120 per user to $300 a month, and Musk’s post does not say what the limits now are.
A note on sourcing. Musk’s words, the question he answered and the quoting post are from X, read on 1 November at about 21:20 UTC; view counts are as of then. Plan prices and cost figures come from our earlier pieces and their sources. Our arithmetic is labelled.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
