🔍 Read the full analysis: What Do Grok’s Costs Mean For The Future Of AI Flat Rates? on ThorstenMeyerAI.com
Get smart everyday buys 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 the cost of serving xAI’s Grok Bot was “unsustainably high relative to revenue,” but provided no figures or details of a pricing or usage change. OpenAI made separate changes to account limits and shared usage allowances, highlighting a broader tension between predictable subscription prices and the variable costs of AI workloads.
Elon Musk said on November 1 that the cost of serving Grok Bot was “unsustainably high relative to revenue,” drawing attention to the limits of flat-priced AI subscriptions. He gave no cost figures and did not announce a specific price or usage change; separate changes to OpenAI’s account rules and shared allowances show how providers are managing usage while keeping monthly plans.
Musk responded on X to a user asking for a Grok Bot reset after reports that users had run through their usage quickly. His reply was one sentence: “The cost of serving Grok Bot was unsustainably high relative to revenue.” It concerned Grok Bot, xAI’s agent, rather than Grok chat generally. The post did not disclose costs, revenue, margins, or whether a usage pool had changed.
According to xAI’s pricing information, Grok Bot is included in its Premium, Super and Ultra plans, priced at $30, $100 and $200 a month, respectively. Its usage draws from a shared weekly pool that also covers Grok chat and certain other services. xAI describes consumption in terms of agent steps and tokens, with the amount used depending on the task and its complexity, rather than simply counting messages.
OpenAI made several changes in the same week, but they do not all amount to reductions. Its help information says it limits how many paid accounts can be used on one device and how many subscriptions can be purchased using one card or device; it does not publish the numerical limits. OpenAI also says that, from December, Plus, Pro and Business allowances for ChatGPT and Codex will be shared and measured in tokens. Unlimited Instant chats remain outside that allowance, according to the source material. Separately, OpenAI announced increased Work and Codex allowances for some Pro plans and cancelled a planned reduction for Pro 200.
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
Flat Prices Meet Variable AI Costs
A monthly plan gives customers a predictable bill, but it does not make every AI request cost the provider the same amount. A short text response and an agent task that runs for minutes or hours can use very different quantities of computing resources. That makes “unlimited” access difficult to sustain when a relatively small number of heavy users can consume far more than a typical subscriber.
The practical direction suggested by these changes is not necessarily the end of subscriptions. It is a move toward flat fees with boundaries: shared allowances, weekly pools, account controls, or separate metered usage for intensive work. For customers, that can make the value of a plan depend on the kinds of tasks they run, not just its advertised price. For businesses, a seat fee may no longer be enough to forecast spending; usage rules and overage terms matter too.
OpenAI’s published rate card illustrates why workloads can diverge: the source article says its most capable model consumes allowance per output token at roughly 100 times the rate of its cheapest model. That comparison is specific to the cited rate card, not a measure of total cost for every task or provider. Agent workloads add another variable because they may involve repeated steps and tool use. The result is a subscription market where price remains simple on the sign-up page while the usable amount of work can be more complex.
Products Worth Considering
Apple Mac mini (2026) with M5 Pro chip and AppleCare+ (3‑year)
FLEXISPOT Mochi-Pro Zero-Wall Oversized Power Recliner, Beige
Canon PIXMA PRO-200S Wireless Professional Inkjet Photo Printer
As an affiliate, we earn on qualifying purchases.
How AI Plans Add Usage Limits
The current debate follows earlier signs that AI companies were balancing subscription prices against unusually heavy use. OpenAI introduced a $200-a-month ChatGPT plan in December 2024 with “unlimited access” to its best models. The following month, chief executive Sam Altman wrote on X that the company was losing money on Pro subscriptions because people used them more than expected. That statement was about OpenAI’s experience at the time; it does not establish the finances of today’s plans or of xAI.
In the developments described here, OpenAI’s late-October changes included both more allowance on selected higher-priced plans and tighter controls on how accounts and subscriptions are used. From December, its shared ChatGPT and Codex allowance will bring those products under one usage measure for the listed plans. That is a change to how allowance is allocated, not evidence that all ChatGPT use is being metered in the same way.
Anthropic’s Claude subscriptions are also described in the source material as having usage limits. That provides relevant market context: the tension between a simple recurring price and caps on intensive use is not confined to Grok. Still, providers’ limits, model costs and plan terms differ, so one company’s change should not be treated as proof that every AI subscription will follow the same schedule.
“The cost of serving Grok Bot was unsustainably high relative to revenue.”
— Elon Musk, replying to a user on X on November 1, 2026
Grok’s Limits and Costs Remain Unspecified
Musk’s post provides no cost, revenue or margin figures. It also does not say what changed, whether xAI altered Grok Bot’s weekly pool, or whether users should expect a reset, a new cap or a price adjustment. His use of the past tense does not, by itself, confirm that a particular cost reduction or policy change has already taken place.
The source material says other outlets reported that an official Grok account attributed adjusted allocations to costs and referred to a separate dashboard issue that temporarily overstated usage. The author said they could not locate the post itself, so those details remain reported rather than independently confirmed here. OpenAI’s account limits also lack published numerical thresholds, leaving users without a precise public measure of how many accounts or subscriptions trigger them.
It is also unclear how much the announced OpenAI allowance changes will affect typical customers, since the source material does not provide a comparable usage distribution or examples of how many tasks each allowance supports. Nor do the cited developments establish that flat-rate plans as a whole are ending. They show specific providers changing or clarifying usage arrangements, not a universal industry policy.
Watch for Published Plan Details
The next stated milestone is December, when OpenAI says the listed Plus, Pro and Business plans will use a shared allowance across ChatGPT and Codex. Customers will need to check their account’s usage information and plan terms to see how that pool applies to their work. The source material does not give a date for any further Grok Bot change.
For xAI, the key information to watch for is a direct explanation of whether Grok Bot’s pool, pricing or reset schedule has changed. More public figures would be needed to assess Musk’s cost claim quantitatively. Until then, users should treat any specific estimate of Grok Bot’s losses or the size of a usage cut as unverified.
Organizations buying AI subscriptions can ask vendors to document included usage, model-specific rates, shared-pool rules, reset periods and charges for additional capacity. Those details can help distinguish a predictable seat price from a workload whose cost varies with model choice and agent activity. The announced changes do not establish what any particular customer will pay, so plan terms and actual usage remain the relevant measures.
Key Questions
Did Elon Musk announce a Grok Bot price increase?
No. Musk said serving Grok Bot was “unsustainably high relative to revenue,” but did not announce a price increase, a new limit or a specific policy change.
Does Musk’s post reveal how much Grok Bot costs xAI to run?
No figures were provided. The post gives no cost, revenue or margin data, so estimates of Grok Bot’s financial performance cannot be confirmed from it.
What is OpenAI changing about ChatGPT and Codex allowances?
According to OpenAI’s help information, from December Plus, Pro and Business plans will draw from one shared allowance across ChatGPT and Codex, measured in tokens. The source material says unlimited Instant chats remain outside that allowance.
Are AI flat-rate subscriptions ending?
The developments described do not show that subscriptions are ending. They point to providers retaining monthly plans while applying usage pools, allowances or account limits to some features and workloads.
Why can an AI agent use more subscription allowance than a chat?
An agent may take multiple steps and use tokens or tools while working on a task. xAI’s documentation says Grok Bot usage depends on the task and effort involved, so a message count alone may not reflect how much capacity a task uses.
Source: ThorstenMeyerAI.com
