Skip to content

All articles

Atlassian raises cloud prices and adds usage billing for AI

Atlassian increased list prices across its core cloud products and will introduce usage-based billing for AI and automation features in December 2026.

Atlassian increased list prices across its core cloud platforms on October 13, 2026, and confirmed it will introduce usage-based billing for artificial intelligence and automation features on December 3, 2026. For companies relying on tools like Jira and Confluence, software budgets will soon depend on a combination of higher base subscriptions and metered charges for new capabilities. This marks a shift from predictable per-user pricing to a model where active consumption dictates the final invoice. Business owners must now track both their headcount and their team's daily automation habits.

Higher base subscriptions

The October price adjustments affect almost every core Atlassian product. Adaptavist (opens in a new tab) reports that Jira Cloud and Confluence Cloud list prices increased by 7% to 10% for Standard and Premium plans. Enterprise plans for both products rose by 8%. Development and service tracking tools also cost more. Bitbucket Cloud Premium increased by 10%, while the five-user minimum requirement was removed from the plan structure. Service Collection, which includes Jira Service Management, saw price increases of 7.5% for Standard and Premium tiers, and 10% for Enterprise.

Other specialized tools face similar adjustments. Jira Product Discovery prices increased by 5%. Atlassian Guard rose by 3% to 5%, and Atlassian Government Cloud increased by 5% to 10%. These percentage increases translate into concrete budget changes. ONES.com (opens in a new tab) notes that a 200-user annual subscription for Jira and Confluence Standard increased from $28,500 to $30,500. A Premium subscription for the same user count increased from $55,000 to $59,000.

The shift to metered consumption

Starting on December 3, 2026, Atlassian will implement usage limits and billing for its artificial intelligence capabilities and automation steps. Subscriptions will include a pooled organization-level allowance, but usage beyond that threshold will carry variable costs.

SaaSJet (opens in a new tab) details the published extra-usage prices. Rovo artificial intelligence interactions will cost $0.01 per credit. Automation executions will cost $0.50 per 1,000 steps. Customer Service Management conversations resolved by an artificial intelligence agent without human escalation will cost $1 per resolution. Storing assets above the included allowance will cost $0.02 per object per month for the first 100,000 overage objects.

The allowances vary significantly by plan. A standalone Jira or Confluence Standard plan includes 25 Rovo credits per user per month. The Teamwork Collection, which bundles Jira, Confluence, Loom, and Rovo, provides 250 credits per user per month at the Standard level.

What it means for your business

The total cost of operating Atlassian software is no longer a simple per-user calculation. Finance and technology leads at small and mid-sized companies must now forecast four separate variables: the core subscription, marketplace applications, metered usage, and taxes.

The billing cycle dictates how seat counts affect the final invoice. Annual billing uses fixed user tiers. Adding a small group of users can push a company into the next tier, creating a sudden cost increase. For example, 480 Jira Standard users fall into the 500-user tier at $39,000 per year. Increasing the count to 565 users moves the subscription to the 600-user tier at $46,000. This $7,000 jump occurs even though the team only added 85 users.

Monthly billing is progressive and charges for the highest number of assigned seats reached during the billing period. Removing users mid-cycle does not lower that cycle's bill. Companies should remove inactive access before adding new users to avoid unnecessary prorated charges.

Marketplace applications add another layer of complexity. These extensions are typically priced for the total user count of the parent product, not just the people who actively use the application. A company with 100 Jira users will pay for a 100-user application license, even if only a fraction of the team uses the extension.

For companies building custom applications, Atlassian Forge pricing introduces a separate consumption model. Developers receive monthly free allowances for compute and storage resources. Usage above those thresholds is billed directly to the developer, separate from the customer's Rovo usage or marketplace subscription.

What to watch

Watch the December 3 rollout of usage-based billing. Administrators should inspect their platform usage settings to confirm their extra-usage configuration and spending caps. Official documentation currently offers conflicting information on whether extra usage is enabled by default.

Monitor the adoption of artificial intelligence features across your teams. A single administrator creating a complex automation or a service team heavily using asset management can quickly consume the centrally pooled allowance. Companies expecting high adoption should compare the cost of standalone subscriptions plus overage fees against the bundled Teamwork Collection.

Watch the long-term timeline for self-hosted deployments. Atlassian ended new-customer sales for Data Center products in March 2026. Specified existing customers can continue purchasing through March 30, 2028. The affected Data Center products will reach end of life and become read-only on March 28, 2029. Companies still running these deployments must plan their migration strategy well before the deadline.

Watch for changes in nonprofit pricing if your organization qualifies. Eligible nonprofits can receive a full discount on Teamwork Collection, Jira, Confluence, or Loom for up to 25 users. Larger deployments receive a 75% discount, while organizations in lower-income countries can access an 85% discount.

The shift in software pricing requires active management. Companies that map their actual usage and remove inactive seats before their next renewal will maintain control over their technology budgets.