Grok Bot Is Free for Two Weeks: Enterprise Access, Autonomous Agents, Organization-Wide Invites, and Current Grok Plans

On September 3, 2026, SpaceXAI opened Grok Bot to enterprise customers and announced a limited promotion: Grok and Cursor Enterprise customers receive free Grok Bot usage for the next two weeks, while organizations can invite users who do not already hold an existing seat. The offer is therefore much broader than an individual trial, because a company can expose a larger internal group to persistent AI workers without first expanding its normal paid-seat footprint.

The word “free” needs a precise boundary. This is an enterprise Grok Bot promotion, not a temporary conversion of every paid Grok feature into a free consumer product. Grok itself already has a $0 consumer tier, while paid SuperGrok plans raise limits and unlock additional capabilities. Grok Bot is a separate agent product designed to execute work across tools, websites, and applications rather than simply answer within a chat window.

Each Bot runs on its own cloud computer, can sign into the accounts an organization authorizes, and can continue working when the user is no longer watching the session. SpaceXAI says enterprise deployment adds access, network, and audit controls so that persistent agents can be governed at organizational scale.

··········

GROK BOT IS FREE FOR ENTERPRISE CUSTOMERS FOR TWO WEEKS.

The promotion combines temporary free usage with unusually broad organization-level access.

SpaceXAI’s announcement specifically names Grok Enterprise and Cursor Enterprise customers. Those organizations can use Grok Bot without usage charges during the promotional window and can invite their whole organization, including people who do not currently have an eligible seat. That changes the economics of evaluation: a company can test agents across finance, engineering, operations, recruiting, or marketing without buying a separate subscription for every participant before it has established whether the workflow is useful.

The company describes the period as “the next two weeks” from the September 3 announcement. It does not publish a more granular cutoff time in the announcement, so the safe interpretation is a limited mid-September evaluation window rather than a permanent pricing change.

........

Element	Current position	Practical implication
Eligibility	Grok Enterprise and Cursor Enterprise customers	The promotion targets existing enterprise relationships rather than every Grok user.
Duration	Two weeks of free Grok Bot usage	The offer is a temporary adoption window, not a permanent free enterprise tier.
Organization invites	Users without an existing seat can be invited	Companies can evaluate agents with a broader employee group before expanding paid seats.
Enterprise controls	Access, network, and audit controls	The rollout is designed around governed deployment of persistent agents rather than unmanaged personal automation.

........

··········

GROK BOT IS BUILT AS A PERSISTENT AUTONOMOUS WORKER, NOT A CHAT SESSION.

A Bot has its own computer, its own task context, and the ability to continue working across external software.

The architectural distinction is important. A conventional assistant normally waits for the next user message and operates inside a bounded conversation. Grok Bot is designed around persistence: users create specialized Bots for jobs, message them like coworkers, and allow them to keep executing until the task is completed or a human decision is required. Each Bot can run independently from the others, so an organization can maintain separate workers for procurement, engineering review, recruiting, sales operations, or recurring administrative processes.

SpaceXAI says a Bot can use applications and websites through its own cloud computer, including systems that do not expose a clean API or MCP integration. The model can therefore interact with the same interfaces a person uses after the necessary accounts are connected. This expands the automation surface, but it also makes permission design more consequential because an agent’s effective capability is determined by the combination of model intelligence, credentials, network access, application permissions, and the persistence of the execution environment.

For enterprise deployments, SpaceXAI says each user’s work runs in a secure isolated environment and that a Bot has no account access by default; it reaches only the accounts that the user signs into. The new enterprise release adds controls for access, networking, and audits, which are particularly relevant when agents are allowed to work asynchronously and across multiple business systems.

The company has also published a concrete procurement example: an internal Grok Bot was given vendor spend, contracts, and usage data and reportedly identified more than $100,000 in direct savings. That case illustrates the intended operating model more clearly than a generic chatbot demo because the agent continuously compared spend, usage, renewal conditions, and market pricing to prepare actions for human review.

··········

CURRENT GROK PLANS SEPARATE FREE CONSUMER ACCESS FROM GROK BOT ACCESS.

The two-week enterprise offer sits on top of an existing product ladder in which Grok already has a permanent $0 entry tier.

SpaceXAI’s current pricing page lists a Free plan at $0 per month with real-time web and X search, Voice mode, connectors, and other baseline capabilities subject to usage limits. Grok Build is also available on every plan following the August 19 expansion. This means “Grok is free” and “Grok Bot is free for two weeks” refer to two different access mechanisms: the first is a standing consumer tier, while the second is a temporary enterprise agent promotion.

Paid access currently starts with SuperGrok at $30 per month. SpaceXAI lists Grok Bot access among SuperGrok capabilities, and the August 26 expansion also placed Grok Bot across SuperGrok, SuperGrok Plus, SuperGrok Heavy, Cursor Pro, Cursor Pro+, Cursor Ultra, and Cursor Teams Standard and Premium. Grok Bot usage is tracked separately from normal Grok or Cursor usage, so delegated agent work does not consume the ordinary usage allowance of those products.

........

Plan or access path	Current price / condition	Grok Bot position	Main distinction
Grok Free	$0/month	Not presented as standard Free-plan Bot access	Consumer AI access with usage limits; Grok Build is available on every plan.
SuperGrok	$30/month	Included	Higher limits, frontier-model access, media generation, connectors, and Grok Bot.
SuperGrok Plus	$100/month	Included through the paid Grok Bot expansion	Higher usage, 1080p video creation, priority access, and the SuperGrok feature set.
Grok / Cursor Enterprise promotion	Free Grok Bot usage for two weeks	Temporarily free for eligible enterprise customers	Organization-wide invitations can include users without an existing seat.

........

··········

THE TWO-WEEK WINDOW IS A LARGE-SCALE TEST OF PERSISTENT ENTERPRISE AGENTS.

The most consequential part of the promotion is not the temporary price reduction but the ability to expose autonomous workers to an entire organization.

Traditional software trials usually evaluate a feature at the level of one user or one team. Grok Bot’s enterprise offer can test a different adoption pattern: multiple employees can create specialized agents, connect them to different systems, and observe how persistent workers behave across heterogeneous operational tasks. That creates a much richer dataset for deciding whether the organization should later purchase broader access, but it also increases the importance of permissions, auditability, credential hygiene, and network restrictions during the trial itself.

The distinction between an assistant and an autonomous worker becomes operational rather than semantic. A model that can only produce text is constrained by the conversation. A Bot with a persistent cloud computer, credentials, scheduled routines, and permission to interact with business software can create real external effects. For enterprise buyers, evaluation therefore has to include not only answer quality but also failure recovery, approval boundaries, action logs, escalation behavior, access revocation, and the ability to determine exactly which systems an agent can reach.

SpaceXAI’s two-week free window is consequently best understood as a deployment and governance test as much as a product trial. Companies that use it only to ask Grok Bot questions will miss the main distinction. The relevant test is whether persistent agents can safely absorb repeatable operational work while remaining observable and controllable when a task crosses a boundary that should still belong to a human.

·····

FOLLOW US FOR MORE.

·····

·····

DATA STUDIOS

·····

[datastudios.org]