Last updated: 2026-08-09
Last updated: August 9, 2026. Short answer for most businesses: upgrade to Odoo 19 now. Odoo 20 has no release date, no stable build, and no first-patch track record yet. Waiting for it means running an aging, increasingly unsupported version for an unknown number of additional months for a version you would not want to be on-day-one anyway. But "most businesses" is not every business, so here is the actual decision framework, by current version, customisation depth, and urgency.
By your current version
| Current version | Recommendation | Why |
|---|---|---|
| Odoo 13-15 | Upgrade to 19 now, do not wait | These versions are old enough that security patching and community support are already thin. The upgrade distance to 19 is large regardless of whether you go now or after 20 ships, so there's no compounding benefit to waiting. |
| Odoo 16-17 | Upgrade to 19 now unless a specific Odoo 20 feature is a hard business requirement | You're not far from 19 already. Waiting for 20 just adds calendar time on a version with a shrinking support window, for uncertain gain. |
| Odoo 18 | Upgrade to 19 now | The jump is small, low-risk, and 19 is the current stable target either way. Skipping straight to 20 later is possible from 19, so this isn't wasted work. |
| Odoo 19 already | Wait and prepare | You're on the current stable version. No urgency. Use the time to prepare, see our upgrade planning guide. |
By customisation depth
The more custom modules and integrations you run, the more upgrade risk you carry regardless of which version you're moving to. If you have a heavily customised instance, our advice is the same either way: budget for a real staging-environment upgrade rehearsal, not a live cutover on faith. The specific version target (19 now, or 20 once stable) matters less than whether you've actually tested your customisations against it first. A heavily customised Odoo 15 instance jumping to 19 is a bigger project than a vanilla Odoo 18 instance jumping to 19, the delta in your own code matters more than the delta in Odoo's version number.
In practice we group customisation depth into three tiers when scoping an upgrade:
| Tier | What it looks like | Realistic upgrade effort |
|---|---|---|
| Light | A handful of view inheritances, some report template tweaks, no custom models | Days, mostly regression testing rather than code changes |
| Moderate | Custom models, a few server actions and automated actions, one or two third-party integrations via API | 1-3 weeks, expect some ORM and view compatibility fixes |
| Heavy | Dozens of custom modules, deep overrides of core business logic, multiple external systems talking to Odoo via XML-RPC or a custom REST layer | Multiple weeks to a couple of months, needs a proper staging rehearsal and a rollback plan, not a weekend project |
The tier you're in matters more for your timeline than which target version you're upgrading to. A heavy-tier instance moving from 17 to 19 and a heavy-tier instance moving from 19 to 20 face the same category of work: re-testing every override against a new ORM and view engine version, not just a version number bump.
By business urgency
- Security or compliance driven (running an EOL OS, unpatched dependencies, or compliance requirements tied to supported software versions): upgrade now, don't wait on Odoo 20 for any reason.
- Feature driven by something specifically in the Odoo 20 roadmap (for example, the native MCP server if you're building AI-agent integrations): worth waiting, but budget for the fact that "unveiled at Odoo Experience" and "stable enough for production" are different dates. Plan on waiting for at least the first patch release after general availability, not day one.
- No specific driver, general "should we upgrade at some point": upgrade to 19 now. There's no cost to being on the current stable version while you wait to see how Odoo 20's first few months go.
One urgency case worth calling out on its own: if you're currently on Odoo.sh and evaluating whether to stay there or move to your own server as part of this upgrade, that's a separate decision from the version question, see our Odoo.sh vs self-hosted comparison. Bundling a platform migration with a major version upgrade in the same change window is possible but doubles your risk surface, we generally recommend doing them as two separate, verified steps rather than one combined cutover.
What "wait for Odoo 20" actually costs you in practice
It's worth being concrete about what waiting means, rather than treating it as a free option. Every month spent on an older version while waiting for a not-yet-released one is a month of:
- Continuing to run on whatever OS and PostgreSQL version your current Odoo install depends on, which for older Odoo versions increasingly means an OS approaching or past its own end-of-life window.
- No new security patches for the version you're on beyond what Odoo's own support policy covers for that release.
- Losing ground on the upgrade itself, since Odoo's official upgrade tooling generally supports jumping from your current version to the latest stable, not an arbitrary number of versions back, so delaying doesn't make a future upgrade easier, it just delays when you pay the same cost.
None of this means Odoo 20 is a bad target eventually, it isn't. It means "wait" should be a deliberate, time-boxed decision tied to a specific driver (see the urgency list above), not a default while a shinier version is rumoured to be coming.
A decision worksheet, if you want to make the call in writing
For a decision like this, especially one that involves other stakeholders in the business who aren't reading this post, it's worth writing the actual reasoning down rather than making the call informally. A short internal note answering these four questions tends to settle the debate faster than an open-ended discussion: What version are we on today, and where does that place us in the table above? How deep is our customisation, honestly, using the three tiers above rather than a gut feeling? Is there a specific, named business driver for waiting, not a general sense that newer is better? And what does staying on our current version for another six to twelve months actually cost us in security patching and support terms? Answering these in writing, and sharing that answer with whoever else has a stake in the decision, usually converts a vague "should we wait" conversation into a clear, defensible call either way.
The mistake we see most often
Businesses delay an overdue upgrade because "the next version is coming soon and we don't want to do this twice." In practice this usually means staying on an aging version for 6-12 months longer than planned, because "coming soon" for a major ERP release routinely slips, and then rushing an upgrade under worse conditions once the pressure becomes unavoidable (an OS reaching end of life, a critical bug with no more patches, a compliance deadline). Upgrading to 19 now and upgrading again to 20 later, once 20 has a real track record, is very often less total risk and less total cost than one long delayed jump.
The second most common mistake is the opposite one: upgrading to whatever the newest version is the moment it's unveiled, without waiting for the first stable patch release. New major versions of any large software project, Odoo included, tend to have their roughest edges in the first few weeks after general availability, not at the conference announcement and not six months later. If you don't have a specific, time-sensitive reason to be on day one of a new major version, waiting for the first patch release is the lower-risk default, whether that's 19 or, eventually, 20.
FAQ
Will I have to upgrade twice if I go to 19 now and 20 later?
Yes, technically two separate upgrade projects. In our experience this is still lower risk than one large delayed jump from an old version straight to a brand new one with no patch history.
What if my business genuinely needs the native MCP server?
That's a real reason to wait, but plan for waiting past the initial Odoo Experience unveiling into the first stable patch release, not day one general availability. See our deep dive on what the MCP server actually ships with to confirm it covers your specific use case before committing to waiting for it.
Is Odoo 19 stable enough to trust right now?
Yes. It is the current, fully supported stable release, unveiled at Odoo Experience 2025 and matured through 2026.
Can I skip Odoo 19 entirely and go straight from an older version to Odoo 20?
Once Odoo 20 ships, Odoo's official upgrade path generally supports jumping from an older version directly to the current stable release, so in principle yes. We'd still recommend waiting for Odoo 20's first stable patch before treating it as an upgrade target, for the same reasons outlined above regardless of which version you're coming from.
Sources: Odoo official documentation.