Broadcom andVMware

The stated rationale

The announcement described a combination of strengths. Hock Tan, Broadcom's chief executive, said that “this transaction combines our leading semiconductor and infrastructure software businesses.” VMware's chief executive, Raghu Raghuram, said that “collectively, we will deliver even more choice, value and innovation to customers” (joint press release). On that account the deal adds product breadth and the customers benefit.

What VMware sells and what Broadcom does with such assets

VMware's core product allows one physical server to run as several independent machines. It is the basis on which much corporate computing has been organized for two decades. It sits beneath the applications a company runs and not beside them. Replacing it means moving every workload that depends on it, testing each one and accepting the risk of the move. Most organizations conclude that the project would cost more than the licenses.

The release's own headline describes Broadcom's method: the purchase “advances Broadcom's strategy to build the world's leading infrastructure technology company, with track record of acquiring established, mission-critical platforms.” The record includes CA Technologies and the enterprise business of Symantec. In each, Broadcom reduced research spending and smaller product lines, focused commercial effort on its largest customers and raised prices. The return in that model does not require the acquired business to grow. It requires existing customers to stay, and the cost of replacing the product is what keeps them.

What the record shows

The transaction closed on 22 November 2023 after review in several jurisdictions. Since then, complaints have centered on price. CISPE, the association of European cloud providers, challenged the European Commission's approval in court and later filed a competition complaint that put the cumulative increase for European providers above 1,000 percent once price rises, bundling and minimum purchase terms are counted. Other reporting has put increases for some customers between 800 and 1,500 percent (Computing, heise). Broadcom disputes the complaint and describes CISPE as funded by large cloud providers.

These sources show that prices rose and that some customers objected loudly. They do not show how many customers left or what share of revenue was lost, and that is the number the thesis turns on.

The strongest objection

The objection is that the price increases are not evidence of a problem. If Broadcom's customers could leave, many would have, and the fact that revenue is retained after the increases shows the product is worth the new price. Broadcom's twenty-year record of profitable acquisitions supports that view.

I agree that the approach is profitable. The question is whether it has a delayed cost. Enterprise buyers have long memories and wide discretion over whom they invite to bid next time. Such a cost would not appear in a quarterly figure. It would appear in processes Broadcom is not asked to join, years later, and would be hard to attribute. The evidence that would decide it is documented migration away from VMware, the outcome of the European complaint and court case, and how customers describe Broadcom as a vendor in later purchasing decisions.

What a reader should take from it

When a buyer pays a large premium for infrastructure software, ask whether the underwriting depends on growth or on retention. A deal underwritten on retention can proceed in conditions that stop other buyers, because its return depends on a switching cost and not on a forecast. The same feature is what makes the later customer relationship adversarial.