SpaceXSI has folded four products into one bill. On October 12, 2026 the company introduced The X Subscription, a single plan covering Grok, Grok Bot, Cursor and X Premium. Existing subscribers can keep their current plan or switch, and paid plans share one usage pool. This is the first time a coding-agent product, a frontier chatbot and a social network share one subscription.
For anyone who pays for Cursor or SuperGrok, this is a pricing change worth reading carefully. It does not raise any existing plan, but it changes how limits work and what third-party models cost if you move.
A note on sourcing: the tier prices and usage multiples below match the official announcement and Tesla North's write-up, and we flag what is not yet confirmed.
TL;DR: the plan at a glance
| Tier | Price per month | What it reportedly includes |
|---|---|---|
| Free | $0 | Daily Grok access, a starter allowance for Cursor coding agents, standard X |
| Plus | $8 | Entry paid tier; Grok and X Premium features, the base usage unit |
| Premium | $30 | Replaces SuperGrok, Cursor Pro and X Premium Plus (about $90 a month separately); 4x Plus usage |
| Super | $100 | Same features, 15x Plus usage |
| Ultra | $200 | Same features, 50x Plus usage |
Other facts from the coverage: every paid plan includes X Premium features, Grok Bot comes with Premium and up, and teams can buy Premium, Super and Ultra per seat at the same prices, with SSO, admin controls and consolidated billing.
How did we get here?
The bundle was telegraphed for weeks. In mid-September SpaceXAI said it was aligning the Grok and Cursor subscriptions. On September 30 Bloomberg reported an internal document describing a single plan with a free tier, an $8 "Lite" tier and a $100 "Ultra" tier including the Grok Bot agent, covered by The Next Web. A leaked screen recording on October 4 showed four paid tiers from $8 to $200, which turned out closer to the final structure than Bloomberg's draft.
The corporate backdrop matters. SpaceX bought Cursor, and we covered the deal closing and Grok branding. The company later renamed its AI unit, a story we followed in the SpaceXSI rename and consciousness debate. A shared subscription is the commercial half of that integration, alongside agent products like Grok Bot.
Cost meter gauge rising toward its limit, illustrating shared usage limits in The X Subscription tiers
How does the shared usage pool work?
According to coverage of the launch, paid tiers share one pool of Grok model usage across the Grok apps, the Grok tab on X and Cursor, so anything left unused in one product carries over to the others. Usage is expressed as multiples of Plus rather than tokens or requests: Premium is 4x, Super is 15x and Ultra is 50x.
That framing makes the tiers hard to compare against competitors. We cannot tell you how many Cursor agent runs a Plus allowance buys, and neither could the launch coverage. Compare this with how Anthropic and OpenAI publish limits in our look at subscription value versus API cost and ChatGPT Pro's $200 reopening.
What changes for Cursor users?
Three things stand out for people who live in the Cursor editor.
- Weekly limits. Cursor limits move from monthly to weekly on the new plans. That smooths usage but can feel stricter for people who work in bursts.
- Third-party models cost API prices. Third-party models in the Cursor editor are available on Premium and higher and are billed separately at API list prices through top-ups. This is the biggest behavioral shift. Cursor already moved toward per-model accounting, as in its Auto and per-model pricing change, and this goes further.
- More coding usage for SpaceXSI models. Every plan reportedly gets at least twice the coding usage for SpaceXSI's own models compared with the equivalent Cursor plan.
One widely shared warning in launch discussion is the obvious one: if most of your usage is on third-party models, the new plans might cost you more than staying put. That is a reading of the terms, not an official statement, but the structure supports it.
Pricing tiers ladder of three boxes with the middle highlighted, comparing X Subscription plan levels
Should you switch? A quick decision guide
| If you are... | Likely best move |
|---|---|
| A Cursor Pro user who mostly runs Claude or GPT models | Stay on your current plan; the switch moves those models to API pricing |
| A SuperGrok subscriber who also uses X Premium Plus | Switch to Premium at $30; the bundle replaces roughly $90 of separate plans |
| A heavy Cursor user on Grok models | Test Super or Ultra; the shared pool and doubled coding usage favor you |
| A casual X user | Free or Plus; compare to what you already pay for X Premium |
| A team admin | Check per-seat pricing and consolidated billing before migrating |
Existing subscribers are not forced to move. Current SuperGrok, Cursor and X Premium plans continue unchanged, and a switch option is rolling out over the next few days in billing settings, reportedly with a prorated refund for the old plan and limited-time upgrade offers.
What are people asking about the plan?
Is the $30 Premium tier really a discount? On the arithmetic in the coverage, yes for a specific person: someone paying for SuperGrok at roughly $30, Cursor Pro and X Premium Plus separately reaches about $90 a month, against $30 for the bundle. The catch is the usage allowance. Premium is defined as 4x Plus, and nobody has published what Plus means in Cursor agent runs, so the saving is only real if the allowance covers your habits.
Do I need an X account to use Cursor now? The coverage says users switch by linking their Grok, Cursor and X accounts. Existing Cursor-only subscribers can stay as they are, so the link is a condition of migration, not of continuing to use the editor. Treat that as the reported behavior and confirm in your billing settings when the option reaches your account.
Is this good for teams? Teams get the same per-seat prices as individuals, plus SSO, admin controls and consolidated billing, and enterprise customers can manage Grok and Cursor from a single dashboard. The practical benefit is one invoice and one admin surface instead of three. The practical risk is the same shared-pool question: a team with a few heavy users and many light ones needs to know whether pools are per seat or pooled across seats before committing.
How does this compare with Claude or ChatGPT plans? Anthropic and OpenAI sell separate chat and coding experiences with their own limit systems, and both have been reshaping them through the year. The X Subscription differs in bundling a social platform and in explicitly metering rival models at API cost. If you want a model-agnostic editor, that is a reason for caution; if you want one vendor stack, it is a reason to look.
A worked example: who wins and who loses
Imagine three developers. The first spends most of the week in Cursor with Grok models for refactors and uses the Grok app for research. For them the shared pool is attractive: idle Grok capacity in the app is usable in the editor, and the doubled coding usage for SpaceXSI models stretches further.
The second developer uses Cursor mainly with Claude for review-heavy work. Under the new plans that traffic is billed at API list prices via top-ups, so a bill that was a flat monthly fee becomes a flat fee plus metered spend. This person has the clearest reason to stay on the legacy plan, which continues unchanged.
The third is a casual X user who occasionally asks Grok a question. The free tier and the $8 Plus tier are aimed at them, and the question is simply whether the Grok access justifies paying beyond what they already spend on X Premium.
What is still unclear
- The size of a Plus allowance. Multiples without a base unit make cross-vendor comparison impossible.
- Regional and app-store pricing. Earlier leaks showed higher in-app prices than web prices; the official position is not confirmed.
- Annual prices. Pre-launch posts listed annual prices, which we could not verify against the official announcement.
- Top-up rates. Third-party model billing is "API list prices," but the top-up mechanics and minimums are not detailed in the coverage we read.
- Naming. Some outlets say SpaceXAI and others SpaceXSI; the official account uses SpaceXSI.
Why this matters beyond one company
AI subscriptions are converging on a pattern: flat-looking tiers on top of metered usage pools, with third-party model access priced at cost. A bundle that includes a social network is unusual, but the engineering reality is the same one behind recent changes elsewhere, such as the Gemini free plan changes. If you build with several models, expect your coding editor to price them separately from the vendor's own model, because the vendor's own model is the one it can subsidize.
It also changes the Grok story. The Grok 4.7 launch put SpaceXSI's frontier model in front of developers through API pricing; The X Subscription is how it hopes to retain them through the editor and the consumer app together.
Bottom line
The X Subscription is a real consolidation, not a rebrand: five tiers, one usage pool, weekly Cursor limits and API pricing for rival models. Existing plans are untouched for now, so there is no rush. Switch if you are a Grok-heavy user who currently pays for three products; hold if you rely on Claude or GPT inside Cursor until the top-up rules are clearer.
Related reading
- SpaceX Cursor acquisition closed, Grok branding
- SpaceXSI rename and the consciousness debate
- Cursor Auto and per-model pricing
- Grok Bot persistent agents
- Grok 4.7 launch, benchmarks and pricing
- Claude vs ChatGPT subscription value
Prices and limits reflect launch-day coverage on October 12, 2026 and may change.