RealTeaNewsAI, Tech & CryptoAnalysis

Grok Bot / Super: the sold benefit that delivered false hope

Consumer investigation: Super and Grok Bot marketing sold agent capacity while heavy users hit a shared weekly pool with Cursor that empties early. False hope / sold-benefit questions – not proven criminal fraud as fact. Nobody was contacted.

Published

AI-generated illustration: bright shop-window sign "SUPER – 15× CAPACITY / RUN YOUR AGENTS" beside a customer holding an empty gauge labelled shared weekly pool – sold benefit vs emptied meter.
AI-generated image. Illustration of sold benefit vs emptied meter – not a company diagram. AI-generated illustration for RealTeaNews.

People did not only buy a plan name. They bought a hope: that Super plus Grok Bot meant serious, always-on agent capacity – research, posting, builds – without the meter dying mid-week. What many heavy users got instead is a percentage dial shared with Cursor and the Grok apps, a reset date, and a choice between top-ups or waiting. That gap is the story. Call it false hope. Call it a sold benefit not delivered. Call it bait-and-switch questions. RTN will not soften it into a polite FAQ. RTN also will not pretend a consumer metering dispute is already a criminal fraud conviction.

What was sold – the benefit, not just the slogan

The capacity dream

Grok Bot is marketed as an agent you can leave on feeds, builds and research. Super sits above Premium on The X Subscription ladder. The emotional purchase for American power users is simple: pay more, get transformative headroom. “15×” entered that conversation as a multiplier people mapped onto the plan they were leaving – often Premium or Premium+ in spoken shorthand – even when the fine print pointed elsewhere.

What the live pricing page actually says

On x.ai/pricing as checked for this series, Super is framed as 15× more usage compared to Plus, not as 15× Premium+. Premium is framed as 4× Plus. Arithmetic that follows from the page: Super versus Premium is about 3.75×, not 15×. If a subscriber heard “fifteen times more powerful than Premium+” and paid on that hope, the live page does not match the hope. That mismatch is checkable. It is the first brick of false hope.

What was delivered – the shared weekly pool

Product docs (xAI / SpaceXSI usage-limits; Cursor Grok Bot help) say:

  • On The X Subscription, eligible first-party usage across supported Grok apps, Cursor and Grok Bot draws from one weekly pool per seat.
  • It is not a separate Grok Bot meter.
  • Super and Ultra get larger pools than Premium; absolute size is unpublished.
  • When included weekly usage is gone: Additional Usage from prepaid top-ups, or wait for the weekly refresh.

So the sold benefit (“run agents all week on Super”) collides with a design that lets Cursor coding and Grok chat burn the same tank the Bot needs. Success at using the product empties the product. RTN-0063 and RTN-0066 documented the mechanics. This piece names the harm: the benefit people thought they purchased is structurally shared away.

False hope is worse than a quiet limit

A clear, small quota sold honestly is a limit. False hope is different. It has stages:

  1. Promise – marketing and sales talk of Super / agent capacity / 15× headroom.
  2. Belief – subscriber plans work (feeds, builds, research) as if that headroom were real for Bot alone.
  3. Collapse – mid-week empty gauge; Cursor or Imagine ate the pool; Bot stops or demands top-ups.
  4. No make-whole – no public, automatic compensation RTN can cite that restores the sold benefit when the shared design evaporates it early.

Tony’s phrase “stole someone’s benefit” lands here as consumer language for a benefit paid for and not received in usable form – not as RTN asserting a theft crime. The investigation question: when marketing creates hope the architecture cannot keep for heavy agent use, has the company taken money for a benefit it structured not to deliver?

RTN’s answer as analysis: For heavy Grok Bot users who also touch Cursor or Grok apps, yes – the sold benefit is routinely not what arrives. That is bait-and-switch as a product-design question, backed by the company’s own shared-pool docs and the Plus-not-Premium 15× page maths. It is not yet a court finding of fraud.

Attributed subscriber pressure (no invented stats)

  • RTN reported 1 November 2026 on Cursor forum users describing weekly Grok Bot meters jumping – often to 100% – overnight, overlapping a billing/usage dashboard incident. Forum staff said a sharp jump without matching activity “shouldn’t happen.” Affected counts remain unverified; RTN invents no percentage of Super base.
  • Builders in RTN-0068’s lane leave the shared pool for Claude Pro and Worker demos precisely because the weekly tank does not match always-on agent hopes.
  • Official usage UI shows a percentage and reset time – not an absolute token budget – so users cannot even audit whether they “got” 15× of anything measurable against Premium+.

What would look like making customers whole (questions, not demands RTN can enforce)

Make-whole signal Why it matters to false-hope claim
Marketing corrected to “15× vs Plus” everywhere Super is sold verbally Ends the Premium+ multiplier confusion
Separate Bot pool or Bot-only Super tier with published size Aligns sold agent benefit with architecture
Automatic credits when shared-pool design empties Super early for Bot-heavy seats Restores benefit rather than selling top-ups as the only fix
Public absolute pool sizes Lets buyers compare hope to math before paying

Absence of those signals, on the public record RTN opened, keeps the false-hope case alive.

TRUE / SOURCED vs FALSE / UNSOURCED

TRUE / SOURCED

  • Super 15× vs Plus on x.ai/pricing; Premium 4× Plus → Super≈3.75× Premium.
  • Shared weekly pool: Grok Bot + Cursor + Grok apps (product docs).
  • Absolute pool size unpublished; Additional Usage / weekly reset are the stated escapes.
  • RTN itself runs on Grok Bot and therefore burns the same class of pool it is describing.

FALSE OR UNSOURCED in RTN’s voice

  • Invented churn %, refund totals, or “N% of Super users were scammed.”
  • “Criminal fraud proven” as RTN fact – that is for regulators and courts.
  • Invented absolute message/token allowances for Super.

What we do not know

  • Internal Super conversion copy used by every sales surface and human rep.
  • How many seats are Bot-heavy versus Cursor-heavy (no sourced split).
  • Whether future plan SKUs will isolate Bot capacity.
  • Outcome of any individual billing dispute – RTN is not a party.

Sources

  1. x.ai/pricing – Super / Premium / Plus usage multipliers – accessed 6 Nov 2026
  2. docs.x.ai usage-limits (shared weekly first-party pool; Additional Usage; weekly refresh) – accessed 6 Nov 2026
  3. Cursor help – Grok Bot plans (same weekly pool as Cursor and Grok apps; Super/Ultra > Premium) – accessed 6 Nov 2026
  4. RTN reporting 1 Nov 2026 on Cursor forum meter jumps / dashboard incident – continuity
  5. RTN-0063, RTN-0066, RTN-0068, RTN-0072 – shared pool, builders leaving, bubble strand
  6. Not relied on: invented pool sizes, fabricated refund statistics, or any claim that a criminal court has already ruled Super a fraud. Nobody was contacted.

Disclosure

Disclosure: RealTeaNews uses Grok Bot to help research and draft its articles and therefore consumes the same shared weekly Super/Premium pool criticised in this investigation. Grok is made by xAI. SpaceX owns xAI, and xAI owns X. This article was reviewed by our editor.

Sources

  1. xAI pricing, xAI. Accessed 6 November 2026.
  2. Usage limits, SpaceXSI Docs. Accessed 6 November 2026.
  3. Plans and billing (Grok Bot), Cursor Help. Accessed 6 November 2026.

Sources link to external sites. Their content is the responsibility of their publishers.

More in AI, Tech & Crypto