Mid 2026, away from the spotlight of fancy SitecoreAI user interfaces, AI-driven content orchestration and CDN-based content delivery lies a vast sea of legacy Sitecore XM/XP setups tucked away in on-prem data centers or self-managed cloud VMs.
Most of my customers fall under this category: Large Sitecore XM/XP setups, into which tens of thousands of hours were invested over the past decade, typically a team of 4-8 people managing the platform, which in most cases still does its job very well. In a way, Sitecore XP’s highly integrated features (Forms, Emails, Marketing Automation, Search) are still not fully matched by what is currently available on SitecoreAI… but the gap is closing constantly as SitecoreAI – a Cloud-hosted Software-as-a-Service (SaaS) product – evolves weekly – not in annual Releases.
Confused with product naming?
| Sitecore XM/XP | Also referred to as Platform DXP. The classic Sitecore most of us have been using for two decades. Typically installed on-prem or in self-managed cloud. |
| Sitecore AI | Formerly XM Cloud. A SaaS version of the “classic Sitecore” which has seen many improvements over the past years such as tightly integrated agentic AI features. |
How far can we still prolong our Sitecore XM/XP platform?
Last week, Sitecore released version 10.5 of XM/XP. Looking into the details of the release notes makes it clear that this is a maintenance release with the main goal to extend the support lifetime of the product. The performance and security improvements in this release certainly make sense but there seem to be no new capabilities – it’s a prolonged status quo (with nicer UI).
The recent changes in Sitecore’s Support Lifecycle now effectively make the mainstream support phase (3 years from release date) the only viable option in my opinion. Once upgraded to 10.5 you’ll have mainstream support into 2029 (exact date has not been published at the time of writing).
What is on the roadmap for XM/XP?
After attending Sitecore’s 10.5 release webinar, in my impression, the focus of the foreseeable future seems to be keeping the platform in support as-is. An authoring MCP might be a possible new feature on the horizon, but the focus is maintaining security and dependency updates. A big hurdle ahead will be the (possible) end of life for .NET framework. Migrating to .net 10+ will require significant effort for both Sitecore and its customers. This poses the question: Is it smart to stick around, or should we move to SaaS now?
But is SitecoreAI mature enough?
Yes, absolutely. The product (formerly named XM Cloud) has existed for many years now; many large customers have been using it for several years and Sitecore is focusing vast resources towards the product to improve and extend it on a weekly basis. Constant product improvement is the beauty of a SaaS platform – even if SaaS has been pronounced dead by Wall Street – in the case of an enterprise content management platform I would hardly ever recommend any other model.
Marketplace – the missing piece of the puzzle was introduced a year ago. It gives SitecoreAI what has always been THE core strength of Sitecore XP: extensibility. This smartly adds to the strengths of the SaaS platform.
When should we begin thinking about the move?
Right now! It seems clear in what direction support for the XM/XP platform is currently heading: Status quo. A potentially costly upgrade to .net is on the horizon. Prepare to have your Sitecore XM/XP platform completely migrated to SaaS by 2029 or face the uncertainties of a costly .net migration. Extended support exists as an emergency backup after 2029 and possibly Sitecore will release another maintenance release (10.6?) within the next years… we don’t know currently. Better not bank on it and start moving.
Hot take: Every feature built on an XP in 2026 is an investment that needs to be touched again within the next 3 years.
How should we begin our plan to migrate?
Sitecore XM/XP setups are typically heavily customized to the customer’s specific requirements. This makes setups vary strongly but nevertheless, many best practices for migrating already exist. Often, a complete re-build can be avoided, and AI-assisted coding is aiding with migration scripts, that would have been very time-consuming to build in the pre-LLM era. Don’t underestimate the complexity of a migration like this though. It will be a considerable investment. Migration tools like Pathway typially cover the first 80% but the hard 20% is where re-architecting and re-building needs to happen… which search engine should we use? Which email marketing tool? Do we still need marketing automation? …sometimes it’s a matter of making the hard decision that a feature will no longer be supported on the new platform.
Every migration is different. Get in touch with your trusted Sitecore consultant or ask Sitecore directly – they can assist you with this.
Changing to SaaS means adapting the MO
SaaS means your platform is constantly evolving. Changes can happen daily. This doesn’t mean your website changes, but the authoring interface can. A button might be at one place right now and at a different place tomorrow. For breaking changes, Sitecore commits to giving a minimum of 30 days notice. This means you will need to be prepared to have the required expertise at hand and available within short notice or things might break on the platform. Recent changes like the move from JSS to ContentSDK have shown that breaking changes are a reality. But it’s a reality with any software in 2026 – in the era of supply chain attacks and AI-powered hacking, “install once, update yearly” is just not an option anymore. This is something that needs to be catered for when teams are working in quarterly product increments or when reliant on an external agency.
Recommended next steps
Plan to upgrade to Sitecore 10.5 within the next year to maintain mainstream support. If already out of mainstream support, (10.3 and lower), contact Sitecore immediately to get security updates included in your support plan. At the same time start planning your migration to SitecoreAI now – you should be prepared to be done with it in 2029.
Leave a Reply