🔍 Read the full analysis: What Happens When AI Gets Costly? Musk On Grok And OpenAI’s New Rules on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Elon Musk said on Nov. 1 that the cost of serving xAI’s Grok Bot was “unsustainably high relative to revenue,” but gave no figures or details of a policy change. In the same week, OpenAI updated rules for paid accounts and said some plans will share a usage allowance across ChatGPT and Codex starting in December.
Elon Musk said on Nov. 1 that the cost of serving Grok Bot was “unsustainably high relative to revenue,” while OpenAI updated rules for paid accounts and announced that some ChatGPT and Codex plans will share one usage allowance starting in December. The developments point to tighter management of AI use, but neither company has disclosed enough financial or policy detail to establish exactly how costs or user limits are changing.
Musk’s comment followed an X user’s request for a Grok Bot reset, saying that many people had run out of usage quickly that week. Musk replied: “The cost of serving Grok Bot was unsustainably high relative to revenue.” The post named Grok Bot, xAI’s task-oriented agent, not Grok chat as a whole. Musk provided no cost, revenue or margin figures and did not announce a reset, price change or specific new limit.
According to xAI’s pricing information cited in the source material, Grok Bot is included with the Premium, Super and Ultra tiers of the X Subscription, priced at $30, $100 and $200 a month, respectively. Its use draws from a shared weekly pool that also covers Grok chat and xAI models in the Cursor coding tool. xAI’s documentation says usage depends on agent steps, tokens and the work requested, rather than simply the number of messages.
OpenAI’s changes include restrictions on how many paid subscriptions can be associated with a device or payment card. Its help page applies the rules to Go, Plus and Pro plans, but does not publish a numerical cap. OpenAI also says Plus, Pro and Business customers will draw from one shared, token-measured allowance across ChatGPT and Codex from December. The source says unlimited “Instant” chats remain outside that allowance.
OpenAI also expanded some high-tier allowances during the same week. Its release notes say Pro 100 and Pro 500 received twice their previous Work and Codex allowance, while a planned reduction for Pro 200 was cancelled. The changes are not all reductions: OpenAI increased some allowances while introducing shared metering and account restrictions.
Infographic 01 · One week, four moves
More allowance, more metering: the week the AI flat rate cracked
In four days, OpenAI raised some limits, tightened others and announced one shared meter, and Musk called the cost of serving Grok Bot unsustainably high relative to revenue.
“The cost of serving Grok Bot was unsustainably high relative to revenue”Elon Musk on X, 1 November 2026, 20:34 UTC, replying to a user who asked for a Grok Bot reset
OpenAI gives more on top plans
Pro 100 and Pro 500 get twice their Work and Codex allowance. A planned cut on Pro 200 is cancelled.
Limits per card and per device
ChatGPT may block a new subscription if too many are tied to one card or device. Applies to Go, Plus, Pro 100, 200 and 500. No number published.
One meter from December
Plus, Pro and Business: one token-based allowance across ChatGPT and Codex. Unlimited Instant chats stay outside.
Musk: Grok Bot cost too much
One sentence, past tense. No figures, no new limit, no word on resets or prices.
Paid resets on ChatGPT pull next week’s allowance forward; they do not add usage. Credits are a prepaid balance used after the included allowance.
Sources: Musk on X, 1 Nov 2026 · Tibo on X, 29 Oct 2026 · OpenAI: card and device limits · OpenAI: ChatGPT and Codex usage limits · OpenAI: paid resets.
Powered by Thorsten Meyer AI
AI Plans Meet Variable Usage Costs
The developments matter because a monthly subscription price does not mean every task costs a provider the same amount to serve. A short chat response and an agent that researches, browses or works through a coding task can use very different amounts of computing. That makes usage-based limits one way for providers to manage costs while continuing to offer fixed-price plans.
OpenAI’s published rate card, as described in the source material, shows that its most capable model uses allowance far faster per output token than its cheapest model. Agent tasks can also run for longer than an ordinary exchange. The result is that a flat monthly fee may coexist with model-specific allowances, shared pools and limits on account use. Users may need to check their usage before upgrading or buying additional access; businesses may need clear written terms on metering and overages.
Still, the available information does not establish that a broad shift away from subscriptions is underway. OpenAI’s increased allowances show providers may adjust access in either direction. Musk’s short post identifies a cost concern, but does not disclose what xAI plans to do about it.
Products Worth Considering
Apple Watch SE (2nd Gen) GPS 40mm with Starlight Aluminum Case and Sport Band (S/M)
DJI Osmo Mobile 8 Gimbal with Mic Mini Transmitter, Infinity Black
Amazon Kindle Scribe (32GB) with Premium Pen, Tungsten
As an affiliate, we earn on qualifying purchases.
Changes Across Three October Days
The cited developments came in a short period. On Oct. 29, OpenAI announced expanded Work and Codex allowances for Pro 100 and Pro 500 and said it would not proceed with a planned Pro 200 cut. Its help article on card and device limits was updated in late October. On Nov. 1, Musk responded to the Grok Bot user on X.
The shared ChatGPT and Codex allowance is scheduled to start in December for Plus, Pro and Business plans. The source material describes this as a single token-measured allowance, while unlimited Instant chats remain outside the meter. These changes come as providers offer products ranging from basic chat to agents and coding tools, whose resource use can vary considerably.
The source article also discloses that Anthropic’s Claude helped with its research and drafting, and says it reviewed Anthropic’s rules alongside those of OpenAI and xAI. No specific Anthropic policy change is detailed in the material provided, so it does not establish a parallel new announcement from that company.
““The cost of serving Grok Bot was unsustainably high relative to revenue.””
— Elon Musk, replying to an X user on Nov. 1
Costs and Limits Still Undisclosed
Musk did not provide Grok Bot cost or revenue figures, explain what made service unsustainable, or say whether xAI had changed the weekly pool. His use of the past tense does not, on its own, confirm that a specific limit was already reduced. The source material says other outlets reported a post from Grok’s official account about “adjusted allocations” and a dashboard fault, but the author could not locate the post; those details remain reported rather than confirmed here.
OpenAI’s help page does not state how many subscriptions one card or device may support, and the source does not spell out every operational detail of the December shared allowance. It is also unclear whether either company will change prices, offer resets or credits, or publish further usage rules. The announcements alone do not show whether the providers’ overall subscription economics are profitable.
December Metering and Further Details
OpenAI says the shared ChatGPT and Codex allowance for Plus, Pro and Business plans will begin in December. Users will be able to check the plan’s usage information, but the source material does not provide a complete breakdown of future caps or how customers will be notified as they approach them.
For Grok Bot, the next meaningful update would be a statement from xAI clarifying whether its weekly pool, pricing or reset policy has changed. Until then, Musk’s post confirms a concern about cost relative to revenue, not a particular new policy. Companies evaluating these services can ask providers to specify included usage, metering units and any additional charges in writing.
Key Questions
Did Musk announce a new Grok Bot limit?
No. Musk said serving Grok Bot was “unsustainably high relative to revenue,” but did not announce a new cap, reset, price or change to the weekly usage pool.
Which ChatGPT plans will share an allowance?
OpenAI says Plus, Pro and Business plans will share one token-measured allowance across ChatGPT and Codex starting in December. The source says unlimited Instant chats remain outside that meter.
Has OpenAI reduced all of its paid-plan allowances?
No. In the same week, OpenAI increased Work and Codex allowances for Pro 100 and Pro 500 and cancelled a planned Pro 200 reduction, while announcing shared metering and card and device restrictions.
How many subscriptions can use one card or device?
OpenAI’s help page says it limits paid subscriptions associated with a device or payment card, but the source material says the company did not publish a specific numerical cap.
Does the information prove AI subscriptions are ending?
No. The reported changes concern usage allowances, account restrictions and cost management. They do not establish that fixed-price AI subscriptions are ending or that either company has disclosed its subscription margins.
Source: ThorstenMeyerAI.com
