Cisco Didn't Pioneer a Single Adjacency. It Bought Its Way Out of Every Threat.

In September 1993, Cisco handed over two million of its own shares — worth about $97 million at the closing price that day¹ — for a small company called Crescendo Communications. The official story is that this was the moment a visionary router company saw the future of the network and boldly seized it. The truth is duller and far more instructive: Cisco's customers had started asking for switching gear, and Cisco didn't make any. It bought a company that did. That is not a leap. That is a flinch.

Repeat that flinch for thirty years across switching, optical, wireless, and finally software and security, and you have the whole Cisco expansion story — not a march into adjacent territory but a series of defensive purchases, each one triggered by a customer demand signal or a threat to the thing Cisco actually owned: the router.

Here is the thesis a smart friend could repeat at dinner. Cisco did not pioneer a single one of its great adjacencies. It was a fast follower in its own backyard, and its acquisition machine was less an engine of vision than a moat-widening reflex — buy the future every time it threatened to outrun the core.

Every new business arrived as a threat first

Look at the pattern and the reactive logic is hard to unsee. Crescendo wasn't a clean jump from routing to switching — Cisco was already a multi-protocol router company with WAN products. What it lacked was LAN switching, the gear enterprises increasingly wanted alongside their routers. So Cisco bought it, and the acquired product line went on to become its dominant enterprise switching franchise. The acquisition didn't open a frontier; it closed a gap a customer had pointed at.

The same reflex fired faster at the height of the boom. In August 1999, Cisco agreed to acquire optical-networking startup Cerent together with Monterey Networks for a combined $7.4 billion in stock³ — its move into optical transport, a layer of the network it had not built. Months later, in November 1999, it spent roughly $799 million on Aironet to acquire an enterprise wireless LAN business it likewise did not have.⁴ Optical and wireless were not visions Cisco hatched. They were transitions threatening the relevance of pure routing, and Cisco wrote checks to stay in the room.

The most expensive flinch of all came in 2024. Cisco closed its acquisition of Splunk at $157 per share in cash, an announced equity value of roughly $28 billion⁶ — though the final GAAP purchase consideration in its FY2024 10-K lands at $27.09 billion after equity-award adjustments.⁵ Splunk was the same move as Crescendo, just thirty years and a few hundred-fold larger: the value in networking was migrating to software and security analytics, and Cisco bought the franchise rather than build it.

Why the reflex worked — and why it eventually didn't

For a while the reflex was a superpower. A reactive acquirer that moves fast enough looks identical to a visionary, because the customer gets the product either way. By March 2000, Cisco had ridden this strategy to the top of the world: revenue had surged from roughly $2 billion in 1995 to about $19 billion in 2000, and the company briefly passed Microsoft and GE to become the most valuable on earth, with a market capitalization above $500 billion.⁸ The moat widened with every deal because each adjacency made Cisco harder to route around — the more layers of your network it supplied, the less reason you had to invite a competitor in for any of them.

But a reflex has a cost a vision doesn't, and the bubble exposed it. Buying every adjacency on rising stock and ordering parts to match a demand curve that only ever pointed up left Cisco fatally exposed when the curve broke. In fiscal 2001 it took a $2.77 billion inventory write-down — the price of building for a boom that vanished — and watched product gross margin collapse from 64.9% to 47.9% in a single year.⁷ A company genuinely steering toward the future hedges. A company reacting to the last signal it received simply over-buys the present, and the present can end.

But didn't it work? Cisco is still here

The honest objection is that calling Cisco a flincher is too cute. The company is enormous, durable, and embedded in the world's networks — and being a fast follower with a great balance sheet is a perfectly good strategy, arguably a better one than betting the company on visions that miss. Crescendo's switching line did become a franchise. The moat-widening did keep competitors out for decades. Reactive isn't the same as wrong.

Fair. But notice what the reactive engine did not produce: a return to glory. Cisco crossed above $500 billion in market value once, in March 2000, and has not seen that height since⁸ — even as revenue kept climbing. That gap is the signature of a defensive acquirer. Buying adjacencies protects the core; it rarely creates a new one large enough to relight the original fire. The Splunk-sized checks of recent years are the same admission Crescendo was: the next layer of value is being built somewhere else, and Cisco's surest way in is to write a check, not to invent it.

Cisco's expansion looks like vision from a distance and like a reflex up close. Every adjacency it ever entered was someone else's idea first — a startup's product, a customer's request, a transition that threatened the router. The genius was never seeing the future. It was being fast enough, and rich enough, to buy whoever already had. That builds an empire. It just doesn't build a second 2000.