SpaceXSI launches unified 'The X Subscription' bundling Grok, Cursor, and X Premium

The confirmable part of this launch is actually very thin. The only first-party trace in the source is a post from the SpaceX Super Intelligence official account: "We've unified plans across Grok, Cursor, and X Premium. Introducing The X Subscription."—this sentence confirms the merger itself, and that the three brands Grok, Cursor, and X Premium are included in the same subscription.

Beyond that, the five-tier pricing, existing-user migration rules, refund method, and usage conversion all come from the same relayed post on X (article_id 22178). That post does not attach the first-party announcement text or full terms, so these details are treated throughout this article as a relay of the official line, not as independently confirmed facts.

In other words, readers need to distinguish two layers: the launch of the merged subscription is confirmed; how exactly money is calculated, how migration works, and how usage is metered after the merger currently exist only as a relay.

According to the relayed account, the unified subscription covers four business lines—conversation, agents, coding, and social—corresponding to the previously separate Grok, Grok Bot, Cursor, and X Premium. Tiers start at Free, then Plus ($8), Premium ($30), Super ($100), and Ultra ($200).

The key to this structure is not the number of tiers, but that it changes subscriptions originally billed by product into billing by person. The value reference given in the relay is: the Premium tier ($30) merges the three benefits of the former SuperGrok, Cursor Pro, and X Premium+, which previously totaled $90/month, with the usage pool refreshing weekly.

It should be noted that both the $90/month comparison figure and the "refreshes weekly" metering cycle come from the same relayed post, belonging to a relay of the official line; the source does not provide a corresponding first-party price list.

The migration design given in the relay is dual-track: old plans are retained as-is with no deadline, existing plans, prices, and features remain unchanged, and old plans only cease accepting new users; at the same time, the official line explicitly states it "will eventually migrate remaining users," but there is no timeline, with a promise of advance notice, and switching is entirely voluntary.

The switch entry is located in account billing settings, rolled out gradually; once visible, users follow prompts to link their X, Grok, and Cursor accounts, and the system recommends a tier based on existing subscriptions; Cursor users have a dedicated entry at cursor.com/unify, and the new subscription takes effect immediately after switching.

On refund arrangements, the relay says that upon switching, the old plan is refunded prorated for remaining time, annual subscribers "keep every day already paid for," no changes are made before the cycle ends, and a limited-time offer is attached—but the specific amount of the offer is shown only on the individual billing settings page and cannot be known from public sources.

There is a notable tension here: the official line simultaneously says "entirely voluntary" and "will eventually migrate remaining users." The former describes the current choice, the latter describes the endgame direction, and the time and method between the two are not disclosed in the source.

Among all the terms, the one with the most direct impact on existing Cursor users is: third-party models (Claude, GPT, etc.) included in the original Cursor plan remain valid before switching; after switching, they are not included in the shared usage pool, require separate top-ups at API rates, and are available only on Premium and above.

This means the "merger" is not a flat transfer of all capabilities into a single pool. After switching, xAI in-house models and third-party models are handled on separate accounts: the former enters the shared usage pool, the latter changes to separate top-ups at API rates. The conversion reference given in the relay is that each tier's xAI in-house model coding usage is "at least 2x the corresponding old Cursor plan."

This 2x is a relay of the official line, and the comparison baseline is "the corresponding old Cursor plan"; the source does not give absolute quotas for each tier. The specific rates for separate top-ups of third-party models at API rates are likewise undisclosed.

The comments attached to the relayed post expose two gaps not yet covered by the official line. First, some users worry that quotas share a single pool, so heavy coding might leave Grok short—this directly points to the actual quota and exhaustion rules of the shared usage pool, which the source does not disclose.

Second, some users ask whether the $20 Cursor plan for new users is now immediately discontinued. The source contains only this question, with no official answer; combined with the statement that "old plans only cease accepting new users," the actual execution status of this point cannot be determined from the available material.

Both of these are user-side questions, not official information, but they mark the blurriest areas of this merger plan in public information: how the pool is divided, and when the old entry points close.

Placing the above content back against evidence strength, the verifiability of this launch is low. Pricing, migration policy, and usage conversion are all relays of the official line, belonging to source claims rather than independent confirmation; key parameters such as the limited-time offer amount, shared pool quota, third-party model rates, and migration timeline are all missing from the source.

Therefore a more prudent reading is: the merged subscription has already happened, the five-tier structure and migration framework are the direction released officially, but the specific terms still need to be based on the first-party announcement or the actual display on the account billing page.

For users currently using Cursor or Grok, what really needs watching is not the tier names, but whether a switch entry appears in their account billing settings, what the recommended tier is, and how third-party model usage is billed after switching—these are the actual variables that determine cost.