SharePoint 2016 and 2019 lost support on July 14

Both SharePoint Server 2016 and SharePoint Server 2019 left extended support on July 14, 2026. Two versions, one date, and it was two weeks ago. No more security updates, no bug fixes, no paid assisted support. The farm still runs. That is the problem with this deadline: nothing breaks on the day, so nothing gets escalated.

We walk estates for a living, and the SharePoint farm is almost never on the first inventory anyone hands us. It surfaces on the third pass, usually because a finance user mentions the site where the quarter-end templates live. Somebody stood it up in 2017. That person works somewhere else now.

Find the farms first.

1. Query the estate for anything answering on 80 and 443 with a SharePoint header, then reconcile that list against the CMDB. Every estate we have walked had at least one farm the CMDB did not list. Count the web front ends, the application servers, and the SQL instances behind them, because the SQL tier runs its own support clock and it is frequently 2016 as well.
2. Name an owner per farm this week, a person and not a team. A team is not an owner. If nobody will take it, that is the finding, and it goes in the report next to the name of whoever declined.
3. Pull the content database sizes and the last-modified dates. Farms sort into two piles: still edited, and an archive that happens to have a login page. The second pile is a records project, not a migration project, and it costs a fraction of the first.
4. Count the customizations before promising anything. Full-trust farm solutions, InfoPath forms, SharePoint Designer workflows and custom web parts turn a two-quarter migration into a rewrite. Inventory them per site collection, by count, with the name of the business process each one serves.

Two supported destinations exist, and choosing between them is a budget decision more than an architecture one.

SharePoint Server Subscription Edition keeps the farm on premises and keeps it patched. It is the shorter move: a supported upgrade path exists from both 2016 and 2019, the operating model stays familiar to the people who already run it, and the licensing turns into a subscription line instead of a capital one. It buys years. It does not answer where this workload belongs in 2030.

SharePoint Online ends the patching conversation permanently and opens a different one about identity, external sharing policy, and the customizations that will not follow. It is the longer move, and it is where Microsoft is steering.

Pick per farm, not per estate. The finance archive can move to SharePoint Online next quarter. The farm carrying full-trust solutions and an integration into the shop-floor system may correctly sit on Subscription Edition until that integration gets retired on purpose.

One estimate, with the assumption stated: for a farm with real customizations and more than a terabyte of content, plan three quarters and assume discovery alone consumes one of them. That figure comes from our inventory work, not from a vendor. A farm that is a document dump with no code in it is weeks, not quarters.

The unglamorous part carries the most risk. An unpatched SharePoint farm holds contracts, HR letters, board decks and the spreadsheets that price the product, it sits inside the network, and its service account usually holds more rights than anyone intended. Compensating controls are a legitimate answer for a bounded window: restrict it to internal traffic, tighten the service account, verify the backups and date the verification. Compensating controls are not a plan.

Put the dates on the wall. Every farm gets an owner, a destination and a quarter, written where the steering committee sees all three at once. The farms that lost support on July 14 do not need a decision this month. They needed one before July 14, and the next best time is this week.