Co-packaged optics asks switch buyers to give up choosing their own optics suppliers

Bill Gartner, who runs optical systems and optics at Cisco, said the industry's response to the 2023 demo "ranged from indifference or a yawn to fairly negative" ^[1][2]. He explains the objection by starting at the front of the box. "There's a switch, which makes up the bulk of the router and its costs, and then there are ports on the face plate into which you plug these pluggable optics," he told DatacenterDynamics (DCD) ^[5]. "But a customer has a choice of which pluggable optic they want to use, and there are many providers of pluggable optics. That gives them negotiating leverage and the ability to make sure that they've got supply chain diversity, which gives them a lot of benefit." ^[4]

Buyers treat that port as a second purchase, with its own competition and its own backup supplier ^[4]. "Multi-vendor gives customers leverage, and customers like that leverage," Gartner said. "It gives them peace and comfort that they're not going to be held hostage by a single source." ^[6]

The case for the change is power. "What co-packaged optics does is move the conversion onto the actual silicon itself. It moves those lasers onto the silicon, eliminating the need for any external units," said Roy Chua, founder and industry analyst at AvidThink ^[12]. Co-packaging, he said, "allows for more consolidation while lowering costs and energy use" ^[13]. Chua does not put a number on the saving in those remarks.

What changes on the purchase order is who supplies the optics. Gartner's company is on both sides of that question: Cisco sells pluggables, and he said it is not in the habit of forcing customers to buy Cisco ^[7].

DCD wrote that the old indifference has vanished, "replaced by an industry ready to shrug off all concerns and welcome the technology with open arms" ^[17]. The same article says lock-in is still unresolved ^[8]. Chua listed the other early doubts. Manufacturing is complicated because the wave guides and modulators have to be made together, and "people felt it would not be sufficiently reliable and cost-effective enough to produce and take to market" ^[14][15]. I'd expect the team that negotiates optics contracts to be the last group to shrug.

I'd sort a co-packaged optics decision on two axes. The first is whether switch power and bandwidth is the limit your network actually hits. That is the bottleneck the concept was built around at the start of the decade ^[9]. The second is how much you get today from bidding optics across suppliers and holding a second source ^[4].

Where power binds and you already buy optics from one vendor, co-packaging costs little leverage and is an ordinary evaluation. Where power binds and your optics bids are competitive, the decision is hard. It turns on the vendor's written answer to two things: how a laser is replaced, and who else can supply the package ^[16][3]. A network that is not power-bound and saves money on optics bids should stay pluggable. One that fits neither description has no case for moving first.