The X Subscription Bundles Grok, Cursor and X Premium Into One Shared AI Usage Pool

SpaceX Super Intelligence announced late on Oct. 12, 2026 that it had “unified plans across Grok, Cursor, and X Premium” under a new offer called The X Subscription. The commercial shape is now visible on the linked pricing pages: one consumer ladder from Free to Ultra, individual paid tiers at $8, $30, $100 and $200 per month, and a shared AI usage pool that covers Grok Chat, Imagine, Voice, Bot and API access plus Cursor agent features, with allowances that refresh weekly. Video generation appears as a listed feature from Premium upward, not as a separate line inside the pool FAQ. Source: X announcement; x.ai pricing; Cursor pricing.

What actually changed

The important change is not a version bump. It is a billing and quota architecture change. Instead of evaluating Grok as a chatbot subscription, Cursor as a coding-agent subscription and X Premium as a social subscription, the fetched pages present them as one consumer ladder. The x.ai pricing page says “One subscription. Access to everything,” while the Cursor pricing page mirrors the same tiers, FAQ structure and pooled-usage language, including Cursor under the same plan names.

For readers who build with agents, the practical consequence is that the choice is no longer only “which model is best?” It becomes “which shared pool shape matches my weekly work?” The pages say every tier includes a pool of AI usage shared across Grok and Cursor, and that within Grok the pool covers Chat, Imagine, Voice, Bot and API access. They also say allowances refresh weekly, which matters for coding agents that can burn through a month of quota in one bad unattended run.

The tier ladder, as published

Tier	Price	Usage positioning	Practical fit
Free	$0/month	Limited AI usage, limited chat and agent requests	Trying the surface area
Plus	$8/month	Shared pool; Cursor included; Grok Chat, Voice, Imagine and Build; X Premium checkmark and creator tools	Social-first users with limited AI usage
Premium	$30/month	4x Plus; latest Grok models; Grok Bot; cloud agents and automations; code hosting and AI code reviews; AI images and videos; third-party models in Cursor billed separately	New AI users who also code
Super	$100/month	15x Plus; priority model access at peak times	Daily agent users
Ultra	$200/month	50x Plus; earliest access to products and models	Power users and fleet-like personal workloads

A key caveat sits inside the table: the Plus tier is listed as Grok models only in the Cursor FAQ. Premium and above add third-party models inside Cursor, but those third-party models are explicitly not included in the monthly price; they remain available at API list price, with spending caps recommended. That means the bundle can simplify the base bill while leaving a separate meter open for any third-party models used through Cursor.

What readers can do with it

1. Price the workload, not the logo. If your week is mostly Grok chat plus occasional Cursor edits, Plus or Premium may be enough. If you run daily coding agents, cloud automations, code reviews and Grok Bot, the pages themselves steer daily agent users toward Super and power users toward Ultra. This maps closely to the cost discipline in Cost per Completed Task Beats $/MTok: measure finished jobs, not sticker price.

2. Treat weekly refresh as a safety valve, not a budget. The published FAQ says the allowance refreshes weekly, so one heavy day should not wipe out the whole month. That is useful for agentic coding, but it is not a license to ignore meters. Keep the three-meter mindset from Three Meters for One Agent Job: subscription pool, API list-price usage for third-party models, and any product-specific limits still need an inventory.

3. Do not assume bundle-wide enterprise controls. The verified announcement and pricing text spell out tiers and pooling, but readers should still verify what happens to current SuperGrok, Cursor-only or X Premium billing cycles, team seats, privacy settings and API keys before switching. On the team side, the Cursor comparison table puts centralized billing and SAML/OIDC SSO on both Teams and Enterprise, while SCIM seat management, audit logs and granular spend controls are Enterprise-only columns. None of those should be assumed from the consumer bundle.

4. Watch the boundary between pooled and metered. The bundle pools base usage across Grok and Cursor, but third-party model usage in Cursor is separately billed. That creates a clear workflow rule: default to pooled Grok models for routine agent work, and escalate to third-party models only when the task justifies the separate meter.

Why it matters

This is part of the broader subscription squeeze already visible across AI tools: flat-rate plans keep absorbing more surfaces until the limit becomes the product. The difference here is scale and adjacency. A social platform, a frontier chatbot and a leading coding-agent harness are being sold as one ladder. For individual developers, that reduces subscription fragmentation. For operators, it concentrates risk: one account, one renewal, one weekly pool and one vendor can now gate chat, coding agents, API access and social identity. The relevant operator question is familiar from Cursor After SpaceX: when access changes, the command center needs a cutoff and migration plan, not surprise.

The right response is not panic or hype. It is a one-page decision: list current Grok, Cursor and X subscriptions; estimate weekly completed tasks rather than raw prompts; separate Grok-model tasks from third-party-model tasks; then choose the smallest tier whose weekly pool covers real usage. The announcement says the plans are unified. The pricing pages say the pool is shared. Your budget should still be itemized, as argued in The Subscription Squeeze.

Load X post

Loading this post connects to X.

The original source is available above.