Broadcom Didn't Buy VMware to Run It. It Bought VMware to Reprice It.

A UK university was paying about £40,000 a year to keep its VMware servers running. Then Broadcom bought VMware. To stay on the software it already depended on, the university was required to adopt the full VMware Cloud Foundation bundle — and its bill rose to nearly £500,000.⁶ Same servers. Same workloads. Roughly twelve times the invoice. The product didn't change. The owner did.

The official story is that this is the messy aftermath of a big merger — overzealous repricing, integration chaos, a new owner squeezing too hard. It is none of those things. The repricing was announced before the deal even closed, in the flat language of a federal filing, complete with a dollar target. Broadcom did not stumble into raising prices. It bought a company in order to.

Broadcom announced it would acquire VMware for approximately $61 billion in cash and stock on May 26, 2022.¹ In the same breath, it told the market the prize: about $8.5 billion of additional EBITDA inside three years.¹ That number is not a forecast of how well VMware's products would sell. It is a commitment to how much more profit could be extracted from the customers VMware already had. The thesis of the whole deal was written down on day one — and the price hikes are simply that thesis, arriving on schedule.

The repricing isn't the strategy. The bundle is.

Here is the part that gets read as villainy and is actually mechanism. In December 2023 — weeks after closing the deal — Broadcom did three things at once. It ended sales of perpetual VMware licenses. It killed off the ability to buy individual products à la carte. And it collapsed the catalog into four primary bundles.⁵ Each move sounds like a separate decision. Together they are a single machine. A perpetual license is a one-time payment that then generates only thin support revenue forever; a subscription is recurring revenue Broadcom can re-price at every renewal. Killing à-la-carte purchasing means a customer who wanted one tool now buys the whole bundle. And pricing the bundle on CPU cores means the bill scales with the size of a customer's data center, not their appetite for features. The price increase is the visible bruise. The bundle is the fist.

Why does this work, where a simple across-the-board price hike would have failed? Because of what VMware is. Its software sits underneath the running servers of banks, hospitals, telecoms, and universities — workloads that cannot be moved on a renewal deadline without months of risk and re-engineering. Broadcom's own filing shows roughly $20.5 billion of remaining performance obligations locked into firmly committed multi-year contracts across its segments.⁸ That is the leverage. The customer cannot say no quickly, and the alternatives cannot be installed quickly. So Broadcom can do something a new entrant never could: raise the price of staying, knowing that leaving is even more expensive in the short run.

The customers it loses were never the point

A normal company guards its customer base. Broadcom is openly willing to shed part of its. The bundle-and-core model is brutal on small and mid-market customers — the university paying £40,000 cannot absorb a jump to £500,000, and many in that tier will simply leave.⁶ That is not a flaw in the plan; it is the plan. The math only needs the large, deeply embedded accounts to stay and pay more. A telecom running tens of thousands of cores has no realistic exit on a renewal timeline, so its bill can rise dramatically while the relationship survives on inertia. Broadcom is trading a long tail of small accounts for a far fatter margin on the whale accounts that cannot move. It is pruning the orchard down to the trees that bear the most fruit.

AT&T is what that looks like when a whale fights back. In court filings, an AT&T executive's August 2024 email to Broadcom's CEO described the two sides at an 'impasse' over a VMware renewal and alleged a 1,050% price increase offer.⁴ Read the figure with care: the actual dollar amounts in that email were redacted in the public filing, so the 1,050% is an attributed assertion from a party in active litigation, not an audited number.⁴ But the direction is unmistakable, and the fact that an enterprise of AT&T's weight went to court rather than negotiate quietly tells you the increase was real and large. The whale was big enough to push back. Most customers are not.

And the numbers say it's working

The strategy is not a bet anymore; it is a result. In fiscal 2024, the first full year of integration, Broadcom's infrastructure software revenue reached $21.5 billion and total company revenue hit $51.6 billion, up 44% year over year, with adjusted EBITDA of $31.9 billion.³ A year later, the fiscal 2025 10-K reported total net revenue of $63.9 billion, up 24%, with the infrastructure software segment up 26% — driven, the filing says, primarily by VMware Cloud Foundation demand and the subscription transition.⁷ That last clause is the whole thesis confirmed in Broadcom's own words: the growth came not from selling more, but from the forced migration to the bundle.

Isn't this just gouging a captive market?

The fair objection is that this is rent extraction dressed up as strategy — Broadcom adds no new value, raises the price of software customers already had, and dares them to leave. There is real truth in that, and the AT&T fight shows the resentment is genuine. But the counter is sharper than 'it's working.' Broadcom is doing something its predecessors chose not to: VMware under prior ownership had priced for adoption and breadth, leaving enormous margin on the table inside accounts that would never realistically churn. Broadcom looked at the same customer base and saw it was underpriced relative to its switching costs. That is a defensible read of value, not just greed — the value being captured was always there; the previous owner simply declined to take it. The honest risk is that the strategy contains its own decay. Every customer Broadcom alienates spends the renewal cycle pricing out alternatives, and migration that was impossible in twelve months becomes feasible in thirty-six. The plan works as long as leaving stays more expensive than staying. Broadcom is betting it can collect for years before that gap closes. So far, the revenue says it's winning the bet — but it is a bet on a clock, not a moat that holds forever.

Most acquirers buy a company and ask how to grow it. Broadcom bought VMware and asked how much its existing customers could be made to pay to stay exactly where they already were. The £40,000 bill that became £500,000 was not an accident of integration; it was the deal's purpose, arriving on schedule. The genius — if that is the word — was never a product or a feature. It was recognizing that the most valuable thing VMware owned was the inability of its customers to leave, and pricing that inability for the first time.