Broadcom Doesn't Buy Companies to Build. It Buys Them to Mine.
Broadcom is read as a chipmaker that bought its way into software. It's something colder: a cash-extraction machine that acquires mission-critical monopolies, cuts to the bone, and reprices captive customers - delivering a record $31.9 billion in adjusted EBITDA in FY2024.
Comes with a free Profit-Engine Map template.
On November 22, 2023, VMware's stock stopped trading on the New York Stock Exchange.2 To an outsider it looked like one tech company joining another. To the people who ran the world's data centers on VMware's software, it was the moment their landlord changed - and the new landlord had a long, public record of exactly what it does with property it buys. Broadcom does not acquire companies the way a technology firm acquires capability. It acquires them the way a private-equity shop acquires a toll bridge: for the cash that flows over it, and for how much more it can charge once nobody can take another route.
The official story is that Broadcom is a chipmaker that pivoted into software. The truer story is that the modern Broadcom is itself a product of M&A - and that it has spent years perfecting one move, run on CA Technologies, run again, and run at its largest scale yet on VMware. Buy something mission-critical. Cut the cost that doesn't generate revenue. Reprice the customers who have nowhere to go. The financial proof of concept is not subtle.
The same move, run three times
Watch what happened to CA Technologies and the VMware future writes itself. Broadcom closed its $18.9 billion CA acquisition in November 2018 - its first real step into enterprise infrastructure software - and then laid off roughly 41% of CA's U.S. workforce.6 That is not a restructuring number; that is a thesis. CA sold mainframe and enterprise software that large institutions had wired into their operations over decades. Customers could not rip it out on a whim, which meant the cost of serving them - the sales effort, the speculative roadmap, the people chasing growth in markets Broadcom didn't care about - was pure subtraction. Cut it, and the revenue largely stays while the cost falls away. The margin appears as if from nowhere. It was never nowhere; it was sitting inside an org chart that a captive customer base made redundant.
VMware is the same machine pointed at a bigger target. At close, VMware shareholders took roughly $30.8 billion in cash and 544 million Broadcom shares worth $53.4 billion, financed in part by $30.4 billion in term loans.3 Debt that size demands cash, fast - and Broadcom told investors exactly where it would come from. In its own announcement it targeted adding approximately $8.5 billion of pro forma EBITDA from VMware within three years of closing.1 That sentence is the whole strategy in one number. It is not a revenue ambition. It is an extraction target.
| A technology acquirer | Broadcom | |
|---|---|---|
| What it buys | Capability and talent | A mission-critical, hard-to-leave customer base |
| What it does first | Integrate, invest, grow | Cut non-core cost; sell off the periphery |
| Source of returns | New revenue over years | Margin freed by removing cost the captive base made redundant |
| The headline metric | Revenue growth | Pro forma EBITDA added |
Because the customers are locked in, churn from a price increase stays low and the revenue line largely holds. Because growth-chasing functions can be cut without losing that revenue, cost falls hard. The gap between the two is the prize - and on VMware, Broadcom told the market that gap was worth roughly $8.5 billion of EBITDA inside three years.1 The peripheral assets that don't fit, like VMware's End-User Computing division sold to KKR for about $4 billion, are simply auctioned off.8
Why the margin headline isn't the whole truth
Broadcom likes to lead with adjusted EBITDA, and the number is genuinely enormous: Q4 FY2024 came in at $9.089 billion, or 65% of quarterly revenue, with guidance for the next quarter near 66%.5 That figure is real, but it is also flattering by design. Adjusted EBITDA strips out the amortization of intangibles - and a deal like VMware loads the books with intangibles to amortize for years. The GAAP picture is far heavier. Even so, the cash is not an accounting illusion: Broadcom generated $27.5 billion in cash from operations in FY2024 on $63.9 billion of revenue and a 68% gross margin.4 Whatever you call the margin, the cash arrives. That is the point of buying toll bridges.
“...responsibly financed acquisitions of category-leading businesses and technologies... investing in research and development to ensure our products retain technology leadership.”7
Notice what that strategy language admits and what it carefully avoids. It says acquisitions and R&D in the same breath - which matters, because the lazy version of this story is that Broadcom guts all research. It doesn't. It funds R&D on the core products that actually generate the cash, and starves the speculative roadmaps that don't.7 'Category-leading' is doing quiet work too. Broadcom does not buy contenders. It buys things that already won their category and are therefore expensive to leave - which is precisely what makes the cost-cutting safe and the repricing possible.
Isn't this just disciplined capital allocation?
The fair objection is that none of this is sinister - it is what good operators do. Plenty of acquired software was bloated, chasing growth it would never reach, employing people on projects no customer wanted. Cutting that is not vandalism; it is honesty, and the freed cash funds the chip R&D and dividends that reward shareholders. There is real truth here. Broadcom is not lighting value on fire; it is harvesting value that prior management left on the table. The honest counter, though, is about who pays for the harvest. The retained revenue holds not because customers are delighted but because they are trapped - the same lock-in that makes the cost-cutting safe is what makes the price increases stick. That is a durable model, but it is durable in the way a tollbooth is durable: it depends on there being no other road. The long-term risk is the obvious one. Squeeze captive customers hard enough and you turn lock-in into motivation, funding the very migration off your platform that the model assumes will never come. Broadcom is betting that switching costs outlast resentment. So far, the cash flow says it is winning that bet.
The most profitable acquisitions aren't the most innovative ones - they're the ones the customers can't walk away from. A mission-critical product with high switching costs gives an acquirer two levers at once: it can cut the cost of serving a base that won't churn, and it can raise prices on a base that can't leave. The technology is almost incidental; the captivity is the asset. The catch, and it's a real one, is that every turn of the screw raises the customer's incentive to escape. Extraction works right up until the migration it provokes becomes cheaper than the next price increase - so the model lives or dies on switching costs outrunning resentment, not on the cleverness of the code.
Broadcom is not a technology company that happens to acquire. It is an acquisition machine that happens to own technology - and the technology's job is mainly to be indispensable. CA proved the playbook on mainframe software. VMware is the same move at five times the scale, financed with debt that the extraction itself is engineered to repay. The genius is not invention. It is choosing, again and again, to buy the one thing a customer can't quit, and then quietly raising the rent. The bill for that genius lands somewhere - and it lands on the people who built their data centers on a foundation they no longer control.
Profit-Engine Map
A one-page map that pulls a business apart into the hook that gets the customer in the door and the engine that quietly earns the margin. Use it to see where the real profit lives, how the two halves are wired together, and what breaks if the link is cut. Blank to dissect your own P&L; filled as the worked example of a business whose advertised product is not where it makes its money.
The worked example unlocks with a subscription. See plans →
Sources
Where this comes from — the filings, records, and reporting behind it.
- 1Broadcom and VMware announced the acquisition on May 26, 2022 at an equity value of approximately $61 billion in cash and stock, with Broadcom also assuming approximately $8 billion of VMware net debt; Broadcom targeted adding approximately $8.5 billion of pro forma EBITDA from the acquisition within three years post-closing.
- 2Broadcom completed the acquisition of VMware on November 22, 2023; VMware's common stock ceased trading on the NYSE upon closing.
- 3At close, VMware shareholders received approximately $30.8 billion in cash and 544 million Broadcom shares valued at $53.4 billion; financing included $30.4 billion in term loans.
- 4In FY2024 (ending November 3, 2024), Broadcom reported total net revenue of $63.887 billion (up 24% year-over-year), GAAP gross margin of 68% of revenue, and generated $27.537 billion in cash from operations.
- 5In FY2024, Broadcom's adjusted EBITDA reached a record $31.9 billion (up 37% year-over-year); Q4 FY2024 adjusted EBITDA was $9.089 billion, or 65% of quarterly revenue. Q1 FY2025 adjusted EBITDA guidance was approximately 66% of projected revenue.
- 6Broadcom completed the acquisition of CA Technologies on November 5, 2018 for approximately $18.9 billion, marking its entry into enterprise infrastructure software; post-acquisition, Broadcom laid off approximately 40.9% of CA's U.S. workforce.
- 7Broadcom's stated corporate strategy, as written in its own 10-K filings, is to pursue 'responsibly financed acquisitions of category-leading businesses and technologies' combined with R&D investment 'to ensure products retain technology leadership' — not to eliminate R&D wholesale.SEC / Broadcom Inc., Form 10-K FY2023 ↗ · 2023-12-14
- 8Broadcom sold VMware's End-User Computing (EUC) division to KKR for approximately $4 billion, announced in February 2024, as part of its post-acquisition portfolio rationalization.