Perpetual Was a Marketing Word

HOST: Welcome back to Stake and Rope, from Goat Security. Today: Tesco, the UK grocery chain, is moving forty thousand server workloads off VMware after suing Broadcom for what its legal filings call “abusive conduct.” Tesco bought perpetual licenses in 2021 with support through 2026 and an option to extend four more years. Broadcom acquired VMware in November 2023, refused to honor the deal, stopped supporting Tesco’s products in January, and is now being sued for at least a hundred million pounds in damages, three times over. The Register and Ars broke the migration filing this week. With me to sort the legal language from the predictable outcome: the Legacy Sysadmin, who has watched this acquisition play out under three different vendor logos. The Founder, who I assume has a defense of Broadcom’s strategy already drafted. The DBA, who has the unenviable job of actually understanding what migrating forty thousand workloads on a compressed timeline costs. Goat Farmer’s here too. Sysadmin, let’s start with you. What does this remind you of.

LEGACY SYSADMIN: [sighs] It doesn’t remind me of anything. I lived it. This is Sun-Oracle in 2010 with a different font on the invoice. Oracle bought Sun for seven point four billion, and within eighteen months they’d torched the Solaris support contracts, jacked the per-socket pricing on the database, and told every shop running SPARC that the migration window was however long it took them to write the check. I watched a regional bank spend four years getting off Solaris because they had thirty thousand workloads and Oracle wouldn’t honor the existing terms. Same playbook. HP did it to Autonomy customers. CA did it to every product they ever acquired. The acquisition isn’t the bug. The acquisition is the feature.

HOST: So when Tesco’s filing calls this “abusive conduct” —

LEGACY SYSADMIN: It’s not abusive. It’s contractual. Broadcom paid sixty-one billion dollars for VMware in 2023. You don’t spend sixty-one billion on a software company because you love the product. You spend it because you’ve modeled out three years of margin expansion through subscription conversion and license enforcement, and the model says it pencils. The “abuse” is them executing the model. The contract Tesco signed in 2021 just happens to be in the way of the model.

FOUNDER: Okay but hold on, I actually want to defend Broadcom here, because nobody else will and somebody has to.

HOST: Of course.

FOUNDER: Hock Tan is, like, one of the most underrated operators in tech. Genuinely. The Broadcom playbook is the playbook for legacy enterprise software. You acquire a company with sticky workloads, you sunset the long tail of unprofitable customers, you concentrate on the top ten percent who generate ninety percent of the revenue, and you convert them from perpetual licenses to subscription. That’s not abuse. That’s hygiene. That’s running the business. The fact that Tesco is mad about it doesn’t make it wrong.

DBA: Who signed the perpetual license contract in 2021.

FOUNDER: I mean — VMware did. Pre-acquisition. But —

DBA: Right. So somebody signed it. Somebody at Tesco read the terms, somebody at VMware countersigned, and four years of support plus an option for four more was the deal both sides agreed to. Broadcom bought the company. They bought the contracts with the company. That’s how acquisitions work.

FOUNDER: But the economics don’t —

DBA: [scoffs] Please… The economics aren’t a legal argument. Tesco didn’t sue Broadcom because the economics changed. They sued because Broadcom stopped delivering support they’d already paid for. That’s a different thing.

HOST: Founder, push on this. You said “rational” and you said “keeping your word.” Are those actually in conflict here?

FOUNDER: [chuckles] Okay, fair. Steel-manning it: Broadcom’s view is that the contracts were underpriced relative to the strategic value of the platform. They’re not refusing to deliver support, they’re trying to renegotiate. And Tesco, instead of coming to the table, is litigating.

LEGACY SYSADMIN: They turned down four offers. It’s in the filing. Tesco rejected at least four proposals from Broadcom to keep using VMware. Whatever Broadcom was offering, Tesco looked at it and decided forty thousand workload migrations on a two-year timeline was the cheaper option. Think about what that says about the offers.

HOST: DBA, you said migration. Let’s stay there for a minute. Forty thousand workloads. What does that actually mean.

DBA: It means somebody at Tesco has a spreadsheet with forty thousand rows on it and every row is a server with a runbook, a backup schedule, a maintenance window, and at least one application owner who’s going to be unhappy. The filing says they’re moving to an unnamed virtualization platform that isn’t compatible with Veeam or Zerto, which is what they use for backup and replication. So it’s not forty thousand workloads. It’s forty thousand workloads plus a complete rebuild of the data protection layer. On a deadline. With third-party support on the source platform because Broadcom walked off the job in January.

HOST: What does that cost.

DBA: Depends on who you ask. If you ask the migration vendor it costs whatever they bid. If you ask the people doing the work it costs nights and weekends for two years. If you ask the business it costs whatever goes wrong while you’re moving — and something will go wrong, because you’re rebuilding the backup stack while the workloads are still running. Tesco is a grocery chain. Their tills, their supply chain, their warehouse systems, their loyalty program — that’s all in this estate. They’re not migrating dev VMs.

GOAT FARMER: [sighs] yeah… I don’t miss that.

FOUNDER: But here’s where I push back — this is a huge opportunity for whoever Tesco picked as the replacement. Like, genuinely. Whoever won that RFP just got the reference customer of the decade. Forty thousand workloads at a Tier 1 retailer. That’s a case study you can sell against VMware for the next five years. Whoever it is, they’re shipping the post-mortem podcast already.

DBA: They’re shipping nothing. They’re working.

FOUNDER: I mean from a marketing —

DBA: They’re working. The vendor that won the contract has a delivery team on site right now and they’re not writing case studies, they’re trying to figure out why the network team didn’t tell them about the iSCSI dependency on a legacy storage array that nobody documented. Marketing happens after. If it happens at all.

HOST: Sysadmin. You said the pattern repeats. Does it repeat because vendors are predatory, or because customers don’t learn.

LEGACY SYSADMIN: Both, but mostly the second one. Look — there’s a generation of procurement people who came up after the Sun-Oracle thing. They didn’t watch it. They read about it, maybe, in a Gartner report. So when they sign a perpetual license deal in 2021 with a thirty-year-old vendor, they’re not pricing in acquisition risk because they haven’t lived through an acquisition. The vendor offers four years of support with a four-year extension option and procurement thinks they’ve locked in eight years of stability. What they actually locked in was eight years of stability if the vendor still exists in its current form. Which it didn’t.

HOST: So the lesson is —

LEGACY SYSADMIN: The lesson is the same lesson it’s been for forty years. There’s no such thing as a perpetual license. There’s only a license that hasn’t been renegotiated yet. The day your vendor gets acquired, every contract you have with them is provisional. I watched Compaq absorb DEC, I watched HP absorb Compaq, I watched Oracle absorb Sun, I watched IBM absorb Red Hat — every one of those was somebody’s stability getting renegotiated.

GOAT FARMER:

Reason number 45. Operating systems come and go, but goats will probably never be orphaned, as they are expected to be produced by their manufacturer for quite some time to come.

HOST: Let me ask the harder question. Tesco reported ninety-eight billion dollars in revenue last year. They are not a small customer. If a company that size can’t get a vendor to honor a four-year-old contract, what does that tell us about the leverage anyone smaller has.

DBA: None. That’s what it tells us. The leverage is none. Tesco can afford to sue. They can afford the third-party support bridge. They can afford to run a parallel migration on a compressed timeline. Most of Broadcom’s mid-market customers can’t do any of those things. They’re paying the new invoice. That’s the whole strategy. You make an example of the few who fight and the rest pay.

FOUNDER: I mean, this is what private equity figured out fifteen years ago. Vista, Thoma Bravo — they didn’t invent this. Broadcom is just doing it at scale with a public-company balance sheet. The reason it keeps working is because it works.

LEGACY SYSADMIN: It works until it doesn’t. Oracle did the same thing to MySQL after the Sun acquisition and now Postgres has eaten half their market. CA did it for thirty years and got bought by Broadcom for parts. The model has a terminal phase. The terminal phase is when the customer base has migrated off and there’s nothing left to extract. We’re not at terminal phase on VMware yet. But Tesco is the leading indicator.

HOST: Take a step back with me. Sysadmin, you said sixty-one billion. Walk me through the actual logic of that number.

LEGACY SYSADMIN: Broadcom paid sixty-one billion for VMware in November 2023. To make that pencil at the kind of returns Broadcom promises shareholders, you need to expand operating margins by something like fifteen, twenty points over three years. VMware under its old management ran at maybe twenty-five percent operating margin. Broadcom’s target for acquired software businesses is closer to seventy. You don’t get from twenty-five to seventy by being nicer to customers. You get there by cutting R&D, sunsetting unprofitable SKUs, eliminating the channel where you can, and forcing every customer who isn’t in the top revenue tier to either pay more or leave. The Tesco situation isn’t an accident of the integration. It’s the integration. The CFO modeled it. The model said: some customers will sue, factor in legal reserves, the net is still positive. And the net is still positive. Even with a hundred million pound claim from Tesco, even three times over, Broadcom’s VMware business unit is more profitable today than VMware was as an independent company. That’s the whole game.

HOST: Bring us back to the operational picture. DBA, the filing says Tesco is “sprinting” off the platform. What does sprinting mean at this scale.

DBA: It means they’re cutting corners they wouldn’t cut in a normal migration. Normal migration of a workload that size, you’d take three to five years, you’d do it application by application, you’d refactor the things that need refactoring on the way over. A sprint means you’re doing lift-and-shift wherever you can, you’re accepting reduced functionality on the destination platform — the filing literally says “reduced functionality” — and you’re booking the technical debt to deal with later. Two years from now, somebody at Tesco is going to be standing in front of a steering committee explaining why fifteen thousand workloads are running on a platform that doesn’t support a feature their old monitoring stack depended on. That’s the bill. It just hasn’t come due.

HOST: Closing thoughts. Goat Farmer, you first.

GOAT FARMER:

Reason number 80. Ever heard of a proprietary goat?

HOST: Founder.

FOUNDER: [chuckles] Look, I know I’m the heel on this one, but I’m going to say it anyway. Broadcom is doing exactly what shareholders pay them to do. Tesco signed a contract with a company that no longer exists. The new company has a different risk model, a different return target, and a different relationship with the customer base. That’s not abuse. That’s M&A. The lesson for every CIO listening to this is: stop signing perpetual licenses with mid-cap software companies that have any chance of being acquired, and start pricing acquisition risk into your contracts. The legal language exists. You just have to ask for it.

HOST: Sysadmin.

LEGACY SYSADMIN: The pattern repeats because every generation thinks it’s the one that won’t be repeating it. Tesco’s procurement team in 2021 didn’t think they were the next Sun-Oracle. The next company to sign a sixty-month deal with a strategic software vendor isn’t going to think they’re the next Tesco. That’s how it keeps working. The vendors don’t have to be sophisticated. They just have to be patient. Every customer eventually forgets, and a fresh acquisition cycle starts the clock over. I’ve watched it happen four times. I’ll probably watch it happen twice more before I’m done. The only people who learn are the ones who got burned, and they aren’t the ones making the next decision.

HOST: DBA. Land it.

DBA: [scoffs] Forty thousand workloads in two years with incompatible backup tooling and third-party support on the source platform. That’s not a migration plan. That’s an evacuation. Somebody at Tesco is working a holiday weekend in 2027 to make a deadline that exists because Broadcom said so and a judge hasn’t ruled yet. The lawsuit is going to trial in 2028. The migration finishes in 2027. The migration finishes first because the migration has to finish first — the lawsuit is just how you recover the damages after. That’s the actual structure. The legal filing is a press release with a docket number. The work is the work. And the work is going to get done by a team of people who didn’t sign the original contract, didn’t sign the lawsuit, and aren’t going to see any of the hundred million pounds. They’re going to migrate forty thousand workloads on a deadline because somebody told them to. That’s the cost. That’s where it lands.

HOST: Broadcom is doing what shareholders paid them to do, and Tesco is doing what customers always do, which is discover that perpetual was a marketing word. We’ll see you next time.