The vendor contract time bomb

The estate's biggest risks are filed in procurement, not running in production. Reading the contracts folder like an engineer.

Pixel drawing of a contract with a small alarm clock counting beside it

Some of the worst outages we have worked never touched a server. A support contract lapsed, an audit clause fired, a vendor got acquired and repriced the renewal at multiples. The estate's time bombs are filed in procurement, ticking on renewal calendars nobody in engineering has seen. Read the folder like an engineer. It is inventory, same as the servers.

What to pull, contract by contract:

  1. Renewal dates and notice windows. The auto-renew with a 90-day cancellation notice is how estates end up paying another year for a system already replaced. Put every notice window on the same calendar the engineers can see.
  2. Change-of-control and repricing rights. The acquiring company inherits your vendor's obligations exactly as far as the paper says, and no further. The industry's standing case study is Broadcom's VMware, where subscription-only bundles and new core minimums pushed renewals up by hundreds of percent for many shops, entirely legally, entirely on schedule. Ask of every critical vendor: if this pricing model arrived at our renewal, what is our move? "Pay it" is an answer. Having no answer is the bomb.
  3. Audit clauses and their tripwires. Licensing regimes with per-core, per-employee or virtualization-hostile terms turn infrastructure changes into financial events. The engineer who vMotions a licensed workload onto a bigger host can detonate a true-up. Engineering needs to know which changes are contractually expensive before making them, which means the clauses need translating out of legalese and into the runbook.
  4. End-of-support riders. What the vendor owes you when the product ages out, and what extended support actually costs. SAP's ECC timeline is the reference case: mainstream maintenance ends in 2027, with extended maintenance at a premium, which is a repricing everyone got a decade's notice on and most of the market will still meet unprepared.

The defusal kit is boring on purpose. A single renewals calendar, engineering-visible. A second-source answer per critical vendor, refreshed annually, even if the answer is "none exists, and here is the exposure in dollars." Exit costs estimated while relations are good; the day you need the number is the worst day to compute it. And contract review in the change process for anything licensed: procurement reads the infra roadmap, engineering reads the renewal pipeline, quarterly, same table.

Vendors are not villains here. They are counterparties, optimizing their side of paper you both signed. The bomb is never the clause. It is the clause nobody on your side had read.