The decision about which technology an institution buys determines whose technology ecosystem it inhabits — with implications that extend far beyond the procurement itself.
Procurement as Strategic Alignment
Technology procurement decisions determine not just which technical capabilities an institution has access to but which vendor ecosystem's standards, products, and upgrade trajectory the institution is aligned with. The institution that standardises on a specific cloud provider's infrastructure is not merely purchasing compute capacity — it is aligning with that provider's product roadmap, pricing trajectory, data governance terms, and integration ecosystem. The switching costs that this alignment creates mean that the initial procurement decision shapes the institution's technology trajectory for years beyond the contract period.
The Power Asymmetry
The power asymmetry in technology markets — where a small number of very large vendors control the infrastructure on which a large number of smaller institutional buyers depend — shapes procurement decisions in ways that favour vendor interests over buyer interests in the long run. The vendor whose product is deeply integrated into institutional operations has leverage the buyer did not anticipate at procurement: switching costs have grown with integration depth, and the vendor can exploit that leverage in pricing renewals, in limiting interoperability with competitors, and in the terms attached to continued service provision.
The Mechanism, Made Concrete
This is not a hypothetical dynamic. When Broadcom acquired VMware, it restructured the pricing model for an already deeply embedded piece of enterprise infrastructure — one that institutions had spent years integrating into their virtualization and networking operations. Just before the acquisition VMware had an adjusted operating margin of just under 29 percent; by the end of 2024, Broadcom's CEO said the figure had hit 70 percent. That margin expansion was extracted from the installed base, not from new customers with a choice. AT&T's dispute with Broadcom made the mechanism explicit in court filings: AT&T accused VMware of proposing a 1,050 percent price increase under its new Broadcom ownership, and the telecoms giant's difficulty in migrating stemmed from how deeply VMware was integrated into its network operations, with a migration estimated to cost $40 to $50 million. The integration depth that made VMware valuable to AT&T is the same integration depth that made AT&T unable to walk away without absorbing a cost roughly on the scale of the price increase itself.
Cloud infrastructure pricing shows the same logic operating structurally rather than as a one-off renegotiation. Providers charge nothing to bring data in and meaningfully more to take it out — an asymmetry that is not an accident of network economics but a design choice. This asymmetry is a deliberate pricing strategy that creates economic lock-in, and it compounds with scale: the more data an institution has entrusted to a provider, the more expensive it becomes to leave, which is precisely the point at which switching should be evaluated on the merits rather than foreclosed by exit cost.
Boundary Conditions
The asymmetry is not permanent or immutable, and it is not uniform across vendor types. Regulators have begun treating egress and switching fees as a competition problem rather than a neutral cost of doing business: the EU Data Act requires cloud providers to make customer data portable and removes contractual and technical barriers that keep customers locked in, and from January 12, 2027, providers serving EU customers will be prohibited from charging any switching fees at all, including data egress charges incurred during a provider switch. That a legislature has judged this mechanism serious enough to regulate is itself evidence that the power asymmetry described here is real rather than rhetorical — but it also means the asymmetry is contingent on market structure and regulatory posture, not a fixed law of vendor relationships. The dynamic is sharpest where integration is deep and data-heavy — core infrastructure, enterprise virtualization, database platforms — and weaker for loosely coupled, easily substitutable SaaS tools where data export is cheap and switching does not require replatforming. An institution's exposure to this asymmetry is a function of how deeply a given system is woven into operations, not a uniform property of "the vendor relationship" as such.
Technology procurement is not a purchasing decision — it is a strategic alignment decision. The institution that makes it without the strategic analysis it requires is making a years-long commitment to a vendor's trajectory without understanding the terms of the commitment it has accepted.
Discussion