Although Acacia Communications was at the optimal point to exit the DSP merchant space, the split second after the announcement of the Cisco Systems takeover, the Web 2.0 operators changed their view of the former from being a key enabler to an undesirable inhibitor. Those system vendors, which were totally dependent on this solution, also had a similar type of reaction. Nevertheless, industry-wide, there was close to universal admiration expressed publicly about the transaction. Already in a weakened state, it is understandable that Infinera in its latest webinar, initially described its emphatic sureness that there was no negative impact from the sale (although arguably, it almost immediately, and indirectly, seemed to back somewhat away from that stance, as the presentation was a part of the slow process of deemphasizing its PICs). While ADVA Optical Networking, as a public company, was also compelled to articulate a positive stance about Cisco’s “inten[tion],” it did discuss its movement to greater vertical integration (fibeReality has addressed its activity in Ottawa). In addition, the vast majority of the analysts as well as the press chose to pretty much ignore both of these layers in the distribution structure, as they frequently give inadequate attention to the critical, demand side in their work. It tended to be enough for them that Cisco gave its assurance that nothing would change in terms of existing relationships (and even the company bending over backwards on the promise, apparently did not make them suspicious). There was certainly insufficient, in-depth investigation of the detrimental effects of the CoreOptics purchase (despite the same type of pledge made at that time), and while the historic, track record of Cisco in optics (most notably DWDM) relating to previous acquisitions, along with its poor execution with the technology in general, were acknowledged, it was expressed that somehow, everything would be magically different this time. Moreover, the PR rhetoric by the vendor, which implies that it has become an optical powerhouse, has undoubtedly approached the revolting point. The only irrefutable realizations about this deal (along with the Luxtera purchase) is that it should not be principally viewed as part of a long-lasting, offensive attack against other optical competitors, nor as a primary aspect of its overall corporate strategy.
In fibeReality’s opinion, the letter sent out to Acacia’s customer base was dead on arrival anyway, as all of the end-users, not just the hyperscalers, will refuse to have the supply chain held hostage by a system house’s rival, as the same reaction happened in the case of CoreOptics. For example, the 100G business from the now defunct, NSN, was destroyed because of the implications from that buyout several years ago. Specifically, to Cisco, large operators have enough problems in avoiding the supplier’s creative ways of trying to enclose them even deeper into its product line.
So, the Acacia move is exclusively about a calculation by Cisco to get its internal costs down, but its utility will be apt to max out with the 400G-ZR (in that way, it will be comparable to its investment in the CoreOptics gear at a lower data rate). As Infinera accurately stated on that webinar mentioned above, the new 400-gig interface will quickly become commoditized. Although one has to wonder just how much internal savings will be realized, as about 85 percent of Acacia’s business will go away, and there will be less of a unit cost advantage, resulting from lower volume production.
Actually, this anticipated, massive loss of revenue may help explain why the closing on the deal is as long as a year away (not just the expectation that China will drag its feet on approval), as Cisco may figure it could be in a position to renegotiate the buying price for Acacia. The premium amount for Acacia may have been partially created by other bidders, whether they were truly serious or not – as seemed to have occurred in the relatively expensive pickup of Luxtera. Just as we view the Acacia move as defensive in nature, we are now giving that categorization more weight for the Luxtera action as well – as we currently believe an important reason for obtaining the silicon photonics supplier was to simply keep it out of the hands of Broadcom.
While there are several motivations at Cisco for its two latest optical acquirements, and we do not rule out the possibility of them even being in conflict — we would not go as far as to say that the Acacia play represents Albert Einstein’s definition of insanity, as we are confident that the former knows what to expect in terms of a result this time. Yet, although it appears that winning brownie points with the hyperscalers is still a piece of a viable, loss-leader rationale in terms of catering to their misplaced whims on optics/electronics co-packaging, it probably has to be viewed at present (in light of Cisco’s latest detrimental deed), more akin to its Tail-f Systems purchase — taking out no less than the perception of a potentially long-term threat to itself. At the very least, Cisco feels comfortable enough with its forays into other business areas that it is willing to gamble away the possibility of future sales (at a minimum, involving stand-alone optical systems/modules) with the cloud providers on what we view as the unlikely possibility of it retaining a cost and/or technology leadership advantage indefinitely (with the exception of Ciena and Huawei Technologies).
It is hard to imagine that Google and Microsoft, if necessary, failing to make the required investments in the merchant DSP sector in order to prevent such an occurrence anyway, even to act as a future deterrence to future notions of cornering a critical portion of the marketplace. Regardless, we are inclined to foresee a resurgence anyway in this high-end DSP business. Post-CoreOptics sale, besides Acacia relatively quick entrance to the market, ClariPhy (now part of Inphi) eventually responded as well.
For our fast-growing, totally separate, quick update blog, which is exclusively on fibeReality’s LinkedIn page, please follow us here.
As always, fibeReality does not recommend any securities, and this writer does not invest in any companies being analyzed by us.
[written by Mark Lutkowitz]