TL;DR
Get everyday essentials delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
On Nov. 1, Elon Musk said Grok Bot had been too costly to serve relative to its revenue, but gave no figures or details of a change to user limits. That same week, OpenAI set new restrictions on sharing paid accounts and said several plans will use one allowance across ChatGPT and Codex from December. The developments point to more metering within fixed-price AI subscriptions, though neither company has said that its subscription model is ending.
Elon Musk said on Nov. 1 that the cost of serving Grok Bot had been “unsustainably high relative to revenue,” while OpenAI announced new limits on paid-account sharing and a shared ChatGPT and Codex allowance. The announcements add pressure to the idea that a fixed monthly AI subscription includes unlimited use, but neither company has said that its flat monthly prices are ending.
Musk replied on X to a user asking for a Grok Bot reset after reports that people had used up their allowance quickly. “The cost of serving Grok Bot was unsustainably high relative to revenue,” he wrote. The post did not include cost, revenue or margin figures, or explain whether a usage limit had changed. Grok Bot is xAI’s task-running agent, rather than Grok chat generally.
According to xAI’s pricing information cited in the source, Grok Bot is included with its Premium, Super and Ultra subscriptions. Its use comes from a shared weekly pool that also covers Grok chat and xAI models in the Cursor coding tool. xAI says usage depends on agent steps, tokens and the work requested, so a task’s cost to the allowance can vary.
OpenAI’s help center says paid accounts are subject to limits on how many accounts can be used on one device and how many subscriptions can be bought with one card or device. The restrictions apply to Go, Plus and Pro plans; OpenAI has not published a numeric limit. From December, Plus, Pro and Business customers are to draw from one token-based allowance across ChatGPT and Codex. The source says unlimited Instant chats remain outside that allowance.
OpenAI also increased some limits during the same week. On Oct. 29, Thibault Sottiaux, who works on Codex and ChatGPT, said on X that usage limits would double for Pro 100, Pro 200, Pro 500 and Business Premium. The source says OpenAI’s release notes specified increases for Pro 100 and Pro 500 and that a planned reduction for Pro 200 was canceled. Those changes accompanied, rather than replaced, the new metering and sharing rules.
AI subscriptions · What changes for you
The End of the AI Flat Rate?
Musk says serving Grok Bot cost too much. OpenAI is tightening account sharing and combining ChatGPT and Codex usage in a shared allowance. The shift points to more metering inside fixed-price plans, though neither company has announced an end to monthly subscriptions.
01 / At a glance
Two moves, one direction
Both announcements point to closer management of costly AI use. Neither says flat monthly prices are going away.
Agent work adds up
Grok Bot draws from a shared weekly pool. Usage depends on agent steps, tokens and the work requested—not just message count.
Sharing gets bounded
Go, Plus and Pro accounts face limits on accounts per device and subscriptions per card or device. No numeric threshold is public.
One pool across tools
From December, Plus, Pro and Business use one token-based allowance across ChatGPT and Codex. Unlimited Instant chats are described as outside it.
Choose a plan
A fixed monthly price sets the starting point.
Do different work
Chat, coding and multi-step agents consume resources differently.
Draw from an allowance
Tokens, model choice and task effort can affect capacity.
Watch published rules
Plan details determine how far your subscription goes.
02 / Why the meter matters
A message is not a unit of work
A short chat and a long-running coding agent can put very different demands on computing capacity.
Relative model cost · output tokens
The source says the most capable model can use an allowance about 100 times faster per output token than the cheapest model. That does not mean every task costs 100 times more.
What the change means
For individuals: whether an allowance fits depends on how often you use agents, coding tools and higher-cost models.
For teams: seat price may need to be weighed alongside metered usage, shared pools and account rules.
For households: multiple paid accounts on one device or card may be affected; the threshold is unpublished.
03 / Changes in context
More allowance can still mean more metering
OpenAI raised some limits during the same week it announced sharing rules and a planned shared usage pool.
| Development | What was announced | What remains unclear |
|---|---|---|
| Grok Bot costs | Musk said serving costs were unsustainably high relative to revenue. | No cost, revenue, margin, price or limit figures. |
| OpenAI account sharing | Restrictions apply to Go, Plus and Pro paid accounts. | No published account or subscription thresholds. |
| ChatGPT + Codex | Plus, Pro and Business move to one token-based allowance in December. | Allowance size and full meter details are not specified. |
| Usage increases | Oct. 29 post said limits would double for Pro 100/200/500 and Business Premium. Release notes specified Pro 100 and Pro 500 increases; a planned Pro 200 reduction was canceled. | These increases accompany the new rules; they do not replace them. |
04 / The economics behind flat fees
Subscriptions have always had a cost side
Today’s changes echo an earlier debate about heavy use, but past comments do not establish current plan economics.
“The cost of serving Grok Bot was unsustainably high relative to revenue.”
Elon Musk · X · Nov. 1, 2026
“We are currently losing money on OpenAI Pro subscriptions!”
Sam Altman · X · January 2025
“We are doubling all usage limits for all Pro 100/200/500 tiers and Business Premium.”
Thibault Sottiaux · X · Oct. 29
Keep the claims in scope. OpenAI introduced a $200-a-month plan in December 2024, advertising unlimited access to its best models. Altman’s January 2025 comment described the company’s experience at that time; it does not establish the economics of current plans. Claims about a separate Grok account post and a dashboard fault could not be verified from the supplied material.
05 / What is still unknown
Limits and costs remain unspecified
Public details are not yet enough to calculate how much work a typical plan will cover.
Did Musk announce a Grok Bot price increase?
No. He cited high serving costs relative to revenue but announced no price change.
Did xAI confirm a lower usage limit?
No. Musk did not say whether the weekly pool changed or by how much. A reported post about adjusted allocations could not be located.
What changes for OpenAI customers in December?
One shared allowance across ChatGPT and Codex for Plus, Pro and Business; Unlimited Instant chats are described as outside that meter.
Are fixed-price AI plans ending?
There is no announcement of that. The clearer trend is that a fixed fee can coexist with changing usage limits.
Next milestone: December. Look for OpenAI’s published allowance sizes, account-sharing thresholds and details on how usage is counted. xAI may clarify its weekly pool, but Musk gave no date or promise of further information. Until then, comparisons of plan capacity remain limited.
How AI Subscription Use Is Changing
The announcements matter because subscription price alone may not describe how much work a customer can do. A chat message and a long-running agent task can draw very different amounts of computing. The source notes that OpenAI’s rate card assigns different token costs to different models, with the most capable model using an allowance about 100 times faster per output token than the cheapest one. That comparison is tied to output tokens; it does not mean every task costs 100 times more.
For individuals, the practical question is whether their plan’s allowance fits their pattern of use, especially if they rely on agents or coding tools. For businesses, a seat price may need to be considered alongside metered use, shared pools and the rules for multiple accounts. OpenAI’s new terms may also affect households or teams that use a device or payment card for more than one paid account. The published material cited here does not give a numerical threshold.
These changes do not establish that fixed-price plans are disappearing. OpenAI raised some limits while bringing ChatGPT and Codex under a shared allowance, and Musk did not announce a new price or cap. The clearer development is that fixed monthly fees can coexist with usage limits, and those limits can change as companies manage the cost of serving demanding workloads.
Popular right now
Ninja CREAMi Scoop & Swirl 13-in-1 Ice Cream and Soft Serve Maker (NC701)
Samsung 75-Inch M80H Series Mini LED 4K Smart TV (2026)
Bose QuietComfort Wireless Earbuds with Active Noise Cancellation, Dewdrop Mint
As an affiliate, we earn on qualifying purchases.
The Cost Behind Flat Fees
OpenAI introduced a $200-a-month ChatGPT plan in December 2024, advertising unlimited access to its best models. The following month, chief executive Sam Altman wrote on X: “insane thing: we are currently losing money on openai pro subscriptions! people use it much more than we expected.” That earlier comment described OpenAI’s experience at the time; it does not establish the economics of current plans.
The current developments involve several kinds of AI use. Standard chat requests, coding assistance and agents that work through multi-step tasks can consume different quantities of tokens and computing resources. xAI’s own explanation, as quoted in the source, says Grok Bot use is measured by agent steps and tokens rather than message count, and that consumption varies with task type and effort.
OpenAI’s Oct. 29 allowance increases and its planned December shared meter show the two sides of subscription changes: a provider can grant more use on some plans while grouping services under one pool. Musk’s Nov. 1 post gives a cost explanation for Grok Bot, but does not disclose what xAI changed, if anything. The supplied source reports claims about a Grok account post on adjusted allocations and a dashboard fault, but says the original post could not be located; those claims remain unverified here.
“The cost of serving Grok Bot was unsustainably high relative to revenue.”
— Elon Musk, replying on X on Nov. 1
Limits and Costs Still Unspecified
Musk did not say how much Grok Bot costs to serve, how much revenue it generates, or whether the service is profitable. His wording referred to costs that “was” high, but reading that past tense as proof that xAI has already changed its system is an interpretation, not a confirmed announcement. He did not state whether users will receive a reset, face a lower weekly pool or see a price change.
OpenAI has not published the number of paid accounts allowed on a device or subscriptions allowed per payment card or device. The source also does not specify the size of the shared ChatGPT and Codex allowance or how the token meter will work in every case. Reports about a separate Grok account post and a dashboard display issue are not independently confirmed by the material provided.
Watch for Published Usage Rules
The scheduled milestone is OpenAI’s move to a shared ChatGPT and Codex allowance in December. Customers will need to consult the company’s help and usage pages for the allowance details and account-sharing thresholds. xAI may clarify whether Grok Bot’s weekly pool or access rules have changed, but Musk’s post did not give a date or promise further information.
Until either company publishes more detail, users and organizations have limited public information for comparing the actual capacity of plans. The next useful disclosures would include allowance sizes, how usage is counted, and whether heavy agent or coding workloads receive separate limits.
Key Questions
Did Musk announce that Grok Bot prices are increasing?
No. Musk cited high serving costs relative to revenue, but did not announce a price change.
Did xAI confirm that Grok Bot’s usage limit was cut?
Not in Musk’s post. He did not say whether the weekly pool changed or by how much. The source says a reported post about adjusted allocations could not be located, so that claim is unverified here.
What is OpenAI changing in December?
According to its help center as described in the source, Plus, Pro and Business plans will use one token-based allowance across ChatGPT and Codex. Unlimited Instant chats are described as outside that meter.
Are fixed-price AI subscriptions ending?
The announcements do not establish that. They show that fixed monthly plans can include shared allowances and usage limits, while some OpenAI plan limits were also increased.
Source: Thorsten Meyer AI
