The VMware Tax: What Vendor Lock-In Actually Costs You

The VMware Tax

6:31

I've run VMware infrastructure for years. Not recently — but enough of my career has run on top of it that I should have been paying closer attention to this story, and I wasn't. I saw the acquisition headlines in 2023 and moved on. I didn't clock how bad it actually got until this week, when I read that T-Mobile is suing Broadcom over VMware support for tens of thousands of virtual machines running across 303,140 CPU cores. That number stopped me. Then I found out T-Mobile isn't even the interesting part of the story — it's the third company doing this in two years.

The Receipts

Broadcom closed its $69 billion acquisition of VMware in November 2023. Within months, it killed perpetual licensing entirely, forcing every customer onto subscriptions, and bundled products into a small number of expensive suites whether customers wanted the extra components or not. In April 2025, it raised the minimum licensable deployment from 16 cores to 72 — instantly tripling the floor price for smaller shops that had no intention of running anything close to 72 cores.

The price increases that followed aren't rumor. European customers have reported increases up to 1,500%. AT&T's costs were set to spike 1,050% under Broadcom's new terms — a number confirmed in AT&T's own court filings before the two sides settled privately in 2024. General reports across the industry cite increases in the 150% to 300% range as typical, not exceptional.

Then the lawsuits started stacking up. AT&T sued and settled. Tesco is currently in litigation, in the process of moving 40,000 server workloads off VMware and accusing Broadcom of "abusive conduct." And now T-Mobile — which tried to pay $5.28 million just to extend support for a single additional year, got refused, then offered $20 million for two years of continued updates just to buy time to migrate over 1,000 applications safely. Broadcom said no to that too. A judge granted T-Mobile a temporary injunction to keep support running through August 2026 while the case plays out.

Broadcom's own numbers tell you why none of this is slowing them down: VMware revenue grew 13% year over year in Q1 of fiscal 2026, with total contract value exceeding $9.2 billion. Overall infrastructure software revenue is up sharply since the acquisition closed. The backlash is real. The lawsuits are real. The revenue growth is also real, at the same time. That's the whole story in one sentence — Broadcom is making more money by extracting more from customers who technically can't afford to leave.

Why "Just Migrate" Isn't an Answer

The obvious response is that customers should just move to Proxmox, Nutanix, Hyper-V, or one of the other alternatives gaining real traction since 2024. Plenty are trying. But T-Mobile's own complaint spells out why that's not a weekend project at enterprise scale: over 1,000 applications, tens of thousands of virtual machines, a network that can't tolerate downtime, and a multi-year technical migration that has to happen while the business keeps running. Broadcom knows this. Every price increase and every refused support renewal is priced against exactly how expensive and slow it is for a company that size to actually leave.

That's not a software licensing dispute. That's a company discovering it built its entire operational stack on top of a single vendor's goodwill, and finding out the goodwill expired the day that vendor got acquired by someone running a different playbook.

The Part That Should Worry Everyone Else

I've spent a lot of time lately talking through what happens when critical infrastructure — compute, data, network — sits entirely inside someone else's walled garden. Usually that conversation is about hyperscale cloud and AI workloads. VMware is the same failure mode wearing an older, more boring uniform, and that's exactly why it's worth paying attention to: the risk isn't unique to flashy AI infrastructure, it's baked into any critical system built on a single vendor's continued cooperation.

Broadcom didn't break any rules here. It found the rules nobody had bothered to protect themselves against — because for two decades, VMware was stable, reasonably priced, and not owned by a company with a documented playbook of buying mature software assets and extracting maximum value from captive install bases. That playbook worked. It's still working. T-Mobile, AT&T, and Tesco are the customers big enough and motivated enough to actually fight it in court. Every smaller company running the same VMware stack, without the legal budget to sue a $1.5 trillion company, just paid the increase and moved on.

That's the actual lesson, and it's not "avoid VMware." It's that infrastructure decisions made when a vendor is stable, well-liked, and reasonably priced need to be re-evaluated the moment ownership changes — because the terms you signed up for were never actually guaranteed, they were just convenient for as long as the previous owner needed them to be. Sovereignty over your own compute and data isn't a hypothetical resilience exercise anymore. It's the difference between being a customer and being a hostage, and right now a lot of companies are finding out which one they actually were.

---

Rich Washburn is a technologist and strategist working at the intersection of AI, infrastructure, and capital. He is Managing Partner and Chief AI Officer at Eliakim Capital and CIO of Data Power Supply.