Migrating to Atlassian Cloud?
Talk to a negotiation advisor. Thirty minutes, your estate, your edition mix and the migration levers ready before Atlassian sets the date.
Data Center list prices rose 15 to 40 percent on 17 February 2026, new Data Center sales close on 30 March 2026, and the loyalty discount window shuts on 30 June 2027. The migration is no longer optional. The bill is still negotiable.
Prepared by Redress Compliance · June 2026 · Representative Atlassian estate scenario (benchmark scenario, not a quote)
Atlassian has set the clock. Data Center prices rose 15 percent on standard list and up to 40 percent on legacy Advantage pricing on 17 February 2026, new sales close on 30 March 2026, and full end of life lands in March 2029. Renewal only buys time, at a rising price, on a retiring product.
Cloud is the destination, and the published price card is steep. Premium runs to 14.54 US dollars per user per month for Jira, before Confluence, Jira Service Management, Marketplace apps, and the Rovo tier stack on top. A full population move to Premium is the most expensive door in, and the one most resellers point you through.
The your side result is a bill that can fall 25 to 40 percent below the opening Cloud proposal. The levers are edition rationalisation, a Marketplace app inventory that clears blockers early, dual licensing so you never pay twice, and the loyalty discount taken on a rationalised baseline.
This paper documents the 2026 price card, the Data Center versus Cloud criteria, the edition mix model, the Marketplace and Rovo decisions, and the six migration contract levers. Every number carries a benchmark range from our advisory file. Your decision deadline is the loyalty window, 30 June 2027, not the price rise.
Start with the commercial model, because the proposal you receive is priced per product, per user, per month, and the line items stack faster than clients expect. Cloud is sold in three editions, and the price steps up sharply between them.
The per user rate also steps down at user count tier breakpoints. Pricing is banded, so an estate sized just above a band edge pays the higher marginal rate on the whole population. Knowing where the next band sits is a sizing lever, not a detail.
| Edition | Jira list, per user per month | What it adds | Best fit |
|---|---|---|---|
| Standard | About 7.75 to 8.15, annual billing | Core Jira, Confluence, or JSM. Community support, 250 GB storage. | The bulk of a deployed population that consumes core workflow. |
| Premium | Up to 14.54 | Sandbox, release tracks, audit log, IP allowlisting, 99.9 percent SLA, 24/7 support, advanced admin. | Power teams and admins that need the controls, not the whole estate. |
| Enterprise | Custom, annual only | Unlimited instances, data residency, Atlassian Analytics, Guard, central security. Threshold near 801 plus users. | Large regulated estates that need governance across many sites. |
Atlassian publishes these rates on its own Jira pricing page, and the editions repeat across Jira Software, Confluence, and Jira Service Management with different unit economics. JSM is priced per agent, not per user, which changes the math for service teams.
Build the Cloud bill bottom up before you read the vendor proposal. Price each product family at the edition you actually need, multiply by the population that needs it, and add Marketplace apps and any Rovo commitment as separate lines. The worked estate below shows the method.
We model a representative estate, Meridian Freight, with 1,000 Jira Software users, 1,000 Confluence users, and 120 Jira Service Management agents. This is a benchmark scenario, not a quote.
| Product line | Population | Rate, per user or agent per month | Annual list |
|---|---|---|---|
| Jira Software Premium | 1,000 users | $14.50 | $174,000 |
| Confluence Premium | 1,000 users | $11.00 | $132,000 |
| JSM Premium | 120 agents | $47.00 | $67,680 |
| Full Premium list | 2,000 seats, 120 agents | $373,680 |
Benchmark ranges: Redress Compliance advisory engagement file, 2024 to 2025. Confirmed against your estate during delivery.
The honest answer for most estates is move, but on your timeline, not Atlassian's. Data Center still works and still renews, yet the economics and the roadmap both point one way.
Three dates frame the decision. The Atlassian migration program confirms the direction, and partner notices confirm the pricing.
Standard list up 15 percent. Legacy Advantage pricing up 18 to 40 percent by user tier. Applies to renewals, upgrades, and new quotes.
New Data Center licenses can no longer be purchased. Existing customers keep the right to renew their current products.
Data Center support and renewals end. After this point there is no supported on premises path for Jira and Confluence.
Some populations have a real reason to delay. Use a written criteria framework so the decision is defensible, not reflexive.
Where one of these applies, negotiate a defined Data Center runway in writing, with a price hold across the runway and a committed migration date, rather than a silent annual renewal at the new list.
This is the single largest lever, and it is the one the opening proposal ignores. The proposal sizes the whole population at Premium because Premium is where the margin is. Most users never touch the Premium controls.
Map the deployed population to the edition it actually consumes. Reserve Premium for power teams and admins who need sandbox, release tracks, and advanced controls. Put the core population on Standard. Keep JSM agents on Premium where the SLA matters.
| Product line | Rationalised population | Rate, per user or agent per month | Annual |
|---|---|---|---|
| Jira Premium | 250 power users | $14.50 | $43,500 |
| Jira Standard | 750 users | $7.75 | $69,750 |
| Confluence Premium | 250 power users | $11.00 | $33,000 |
| Confluence Standard | 750 users | $6.05 | $54,450 |
| JSM Premium | 120 agents | $47.00 | $67,680 |
| Rationalised mix | 2,000 seats, 120 agents | $268,380 |
The rationalised mix lands at 268,380 dollars against the 373,680 dollar full Premium list, a saving of 105,300 dollars, or about 28 percent, before any negotiated discount. That saving recurs every year.
Edition rationalisation beats the discount.
Across the Cloud migrations we benchmarked in 2024 to 2025, mapping editions to actual consumption cut the recurring bill 22 to 34 percent before any negotiated discount.
Opening proposals over provision Premium.
Roughly two in three opening Cloud proposals we reviewed sized the entire population at Premium when 60 to 80 percent of users needed only Standard.
Benchmark ranges: Redress Compliance advisory engagement file, 2024 to 2025.
Apps, not data, are what stall migrations. The Jira and Confluence Cloud Migration Assistants move issues and pages reliably. The third party apps bolted onto Server or Data Center are where the cutover slips.
Run an app inventory audit before you commit to a Cloud date. Classify every installed app into one of four readiness states, and price the substitution work into the migration plan.
| Readiness state | What it means | Counter move |
|---|---|---|
| Cloud native exists | The vendor ships a Cloud version with a migration path. | Migrate in place, confirm the Cloud price, which often differs from Server. |
| Substitute available | No Cloud version, but a competing app or native Cloud feature covers the need. | Switch, and negotiate the new app into the Atlassian deal where possible. |
| Native feature absorbs it | A Premium or Enterprise feature replaces the app entirely. | Retire the app, count the saving against the edition step up. |
| No path | Bespoke or abandoned app with no Cloud equivalent. | Rebuild, defer that population, or negotiate a Marketplace readiness exception. |
Marketplace economics also shift on Cloud. Many apps move from a one time tier price to a per user monthly fee, so a Server app that felt free can become a material recurring line. Audit the app bill, not just the Atlassian bill.
Commit to almost nothing in writing, because the value is already bundled and the pricing is still moving. Atlassian Intelligence and Rovo search, chat, and agents now ship into every paid Cloud subscription at no separate line item.
Each licensed seat earns a monthly Rovo credit pool that scales with the edition. Credits pool across the organisation and do not roll over.
| Plan | Rovo credits per seat per month | Reset behavior |
|---|---|---|
| Standard | 25 | Pool org wide, reset monthly, no roll over. |
| Premium | 70 | Pool org wide, reset monthly, no roll over. |
| Enterprise | 150 | Pool org wide, reset monthly, no roll over. |
The trap is the overage. As of May 2026, Atlassian is not billing for usage above the included allowance, with at least 90 days notice promised before that changes. Treat any priced Rovo commitment, including the separate 20 dollar Rovo Dev surface, as a future cost line tied to a usage review, not a multi year minimum.
The migration is a contract event, and the contract is where the savings are protected. Six levers do most of the work. Put each one in writing before signing.
| Lever | What it does | Why it matters |
|---|---|---|
| Migration timeline | A defined cutover window that you control, protected from a forced date. | Stops a roadmap deadline becoming a rushed, full price renewal. |
| Dual licensing | Atlassian credits the Data Center renewal so you do not pay twice during transition, up to 12 months. | Removes the double cost that makes phased migrations look expensive. |
| User count grandfather | Holds the priced population for the term even as headcount drifts. | Prevents quiet true ups inflating the bill mid term. |
| Edition substitution | The right to move seats between Standard and Premium as needs change. | Locks in the rationalised mix without penalty if usage shifts. |
| Marketplace readiness exception | A defined runway for populations blocked by an app with no Cloud path. | Avoids paying for Cloud seats that cannot yet move. |
| Price hold and escalation cap | A fixed rate for the term and a capped annual uplift after it. | Defends against the next list increase landing on your renewal. |
Atlassian offers a free Cloud trial for the remaining duration of your Data Center term, up to 12 months. Combined with dual licensing, this lets you run Cloud and Data Center in parallel without double paying, and it moves the negotiation onto your calendar.
The Atlassian loyalty discount, 10 to 20 percent and higher in larger deals, is tied to committing to Cloud Enterprise by 30 June 2027. It is real money, but it is leverage that decays. Take it on a rationalised baseline, and use the deadline to compress the negotiation, not to justify over buying.
Atlassian rarely sits alone in the estate. It overlaps with the wider engineering and knowledge worker stack, and the Cloud commitment should be planned against that whole estate, not signed in isolation.
Sequence the migration in three phases so the commercial commitment follows the technical reality, not the other way around.
Inventory users, editions, and apps. Build the bottom up cost model. Classify every Marketplace app by readiness state.
Run the free Cloud Migration Trial. Negotiate the rationalised baseline, the six levers, and the loyalty discount before the cutover.
Migrate by population, retire Data Center, and set a usage review for Rovo and edition mix at each renewal.
Align the Atlassian term with adjacent renewals where it helps consolidate leverage, and avoid stacking every software renewal in one quarter, which hands the vendors a deadline you cannot move.
Treat the 2026 Data Center deadlines as a timing signal, not a reason to panic renew or over buy. Move to Cloud on a rationalised baseline, protected by the six contract levers, with the loyalty discount taken last.
We are glad to tie a meaningful part of the fee to delivered value.
Talk to a negotiation advisor. Thirty minutes, your estate, your edition mix and the migration levers ready before Atlassian sets the date.
One letter a month. Negotiation moves, audit signals, and price book shifts.