Why Cisco should merge with Dell

Disco – A real data center powerhouse

Cisco's two choices

Assuming that integrated stacks are good for vendors, and knowing that Cisco is in the middle of reorganizing itself and has said it will only stay in markets where it can be the number 1 or 2 player, then Cisco has two choices. It can sell off the UCS business and get out of the server racket, rebuilding bridges that it burned with the California machines and offering to provide integrated networking for IBM, HP, Dell, Fujitsu, and Oracle machines. Or, it can merge with Dell and actually become the number two server maker.

Cisco might be able to buy Dell, since it has $6.6bn in cash and $36.7bn in marketable securities. Even after a nice run-up in its stock last week after reporting much-improved profits on slim revenue growth, Dell has a market capitalization of only $30.5bn. Dell has $7.71bn in debts and $14.5bn in cash and equivalents, so that works out to $6.77bn of cash net of debt.

To my way of thinking, that means the underlying stock in Dell is worth only $23.7bn, but I am sure there are plenty of people on Wall Street who would argue that Dell has an enterprise value of closer to $40bn, and therefore an acquisition price of something more like $50bn.

Good luck trying to get that, even for a company that should bring in maybe $66bn in sales in its fiscal 2012 ending next January. Cisco might pull in something on the order of $44bn over the same term, so the combination would generate a $110bn IT powerhouse that would rival HP and IBM.

To do such a combination, Cisco and Dell would be wise to conserve their cash. They could pay out a big cash dividend to get shareholders on both sides of the deal excited about a merger. Chambers can be chairman of the combined company, and Michael Dell can be the CEO, and someone – make Steve Schuckenbrock of Dell and Gary Moore of Cisco wrestle for it – can be COO. Whoever loses runs the combined Dell+Cisco services business.

Dell would take over Cisco's nascent server operations and Cisco would take over Dell's equally juvenile networking business. There are plenty of executive VPs laying around the two companies to run the various units. They could think about whether or not they want to be in the PC business at all.

Let's call it Disco

What might Disco look like? It would have about $2.3bn a quarter in server sales, and probably something on the order of $500m in storage revenues per quarter that would very likely grow to around $1bn a quarter in fairly short order. Switching, once rationalized, would probably stabilize at around $3.5bn a quarter, and routing, where Cisco is not facing such staunch competition, would rake in about $2bn a quarter. If the combined company decided to sell PCs – rather than sell off the PC biz to a very eager Acer – then it could probably count on $4.5bn to $5bn a quarter in laptops and notebooks and about $3bn to $3.5bn per quarter on desktop PCs.

Dell's side of the server biz accounts for around $2bn in revenues per quarter, and Cisco is in the same range. There's another $1bn or so in Dell's software and random peripherals, and Cisco collaboration products and services would make up the balance of the Disco's $21bn in product revenues. Let's say the combined company can eliminate some back office overlap and data center costs (assuming Perot Systems can run data centers leanly), then the company could squeeze out some incremental profits.

The combination of Dell and Cisco would be able to peddle stacks of routers, switches, and custom servers to hyperscale data centers and service providers in one fell swoop. Similarly, Dell could converge the technologies in the UCS and PowerEdge server lines, and present enterprises with a single, integrated stack.

And in keeping with Dell's open tendencies, there is no reason to boot VMware's hypervisor or cloud extensions off the converged machines or standalone boxes, but it would probably be smart to back the KVM hypervisor and the OpenStack controller as the preferred cloudy software layer for the DEll+Cisco machines. Once OpenStack gets a virtual switch – the Open vSwitch is the obvious one to choose – neither company would need VMware, but could, of course, continue to sell whatever customers wanted.

You'll notice that I didn't mention the "Project Acadia" Virtual Computing Environment alliance between Cisco, EMC, and VMware and how the Disco merger might affect this alliance. It would kill it, dead. Such a three-way sales and marketing unit is tough to fly and runs counter to how IT vendors like to operate. They like to have control of one area, and they don't share well. That doesn't mean VCE is doomed or bad, but rather that it is just unnatural and requires coordinating the efforts of three companies with monstrous egos.

If there is a problem with the Disco combo, it is that the merged companies would have the same problem as HP: too much lean commodity hardware and not fatty software. Maybe the smart thing for Disco to do, then, would be to buy Red Hat with that big pile of cash. Red Hat has a market capitalization of $8.8bn and $860.6m in cash. Such an acquisition would be thinkable and doable for Disco.

If you give a mouse a cookie, he'll want a glass of milk. ®

Sponsored: Designing and building an open ITOA architecture