CompoundSemi Online - The Original Compound Semiconductor Industry Newspaper

October 8, 2002...The announcement of Bookham's purchase of Nortel's optical assets confirmed what many of our core readership have thought was the real situation... that CoreTek's technology was deemed redundant to Bookham's and that their deal with Nortel was quietly contingent upon Nortel closing the CoreTek facility in Boston, Massachusetts USA. resulting in the the layoff of nearly all 160 CoreTek employees, versus allowing them to buy back their own company. Consensus among our very reliable sources as to why the closure versus letting the buyback go through, is that Bookham either thought that their existing tunable lasers and filters via the Marconi acquisition were superior products, or that Bookham felt that they did not have the in-house expertise to manage the CoreTek product line. Either way, the Boston loyalists are being dispersed, and with them, what most compound semi industry insiders regard as the best tunable VCSEL technology in the field.

The questions now to be asked are: Was the CoreTek IP submerged... or reserved? Either way... by whom, and for what reason(s)? Answers to these are likely to remain in the realm of speculation, but here are the facts as we currently know them. In September, when the closure of CoreTek first came to light, CoreTek actually had several million dollars in orders on the books for their filters, and had received yet another PO, amounting to $2 million, only a few days before Nortel flew someone down to Boston to abruptly pull the plug. The CoreTek VCSEL had also scored several design wins based on its full Telecordia qualification, fortified with multiple tens of thousands of hours of lifetime testing. Thus the formal excuse from Nortel that CoreTek's tunable VCSEL was "too advanced for the market" simply doesn't make sense, considering the orders it was commanding at the time. So, despite Nortel's admittedly flagging marketing efforts, CoreTek's two principal products had already been accepted, if not embraced, by the systems community.

When CoreTek was acquired, Nortel already had a low yield tunable DBR-type laser in the works, which was one of the more obvious reasons Nortel wanted to buy CoreTek. That low yield tunable program was to be killed off upon the acquisition CoreTek so that the CoreTek tunable VCSEL would, instead, get designed into the next generation Nortel systems. The internal deal carried with it the caveat that CoreTek meet certain technical and performance milestones. Marconi had announced a DBR-type laser (ref: Nov. '01 coverage) which likely moved over to Bookham when they acquired Marconi's Optical Components in July of this year (ref: July 4 coverage). Which technology was superior remains in the realm of the individual evaluator, but after the acquisition of CoreTek by Nortel, CoreTek did indeed complete all milestones (ahead of schedule), yet Nortel's DBR device lived on. Through the mysterious ways of Nortel's corporate politics, lo and behold, that DBR ended up the one designed into the follow-on generation of product, not CorTek's. Thus, the timeline got pushed out for Nortel to use its own newly acquired product in its own systems, although the new device actually met all the requirements.

We've all seen this type internal competition take place. Hold back your "advanced" solution until you've milked sales of what was already in the queue. Every team champions their design. But in this case, it appears all the due diligence was done and management simply wasn't putting it all on the table for the tech teams involved. The result was that the CoreTek VCSEL was now only incorporated into a 10Gb metro product by coupling it with a GaAs Mach-Zender that Nortel was making... but couldn't sell. This appeared to be the perfect stopgap device for CoreTek to sell into, and it remains a mystery why this product never flew. The oversight could be credited to either market or corporate politics, but it was clear from the technical side that the device met all of the internal requirements.

More came to light when, in early 2002 it came to the attention of key managers within CoreTek that Nortel had placed a $2M purchase order with Bookham for tunable filters. Hmmm... the plot thickens. When properly investigated, it turned out that Nortel's upper management had made a previous agreement with Bookham to fund development of a number of technologies, with an accompanying agreement to purchase tied tightly to it. So even though Nortel was losing money... and part of the rationale was that CoreTek was losing money... another $2 million had to leave the company for a product for which there was a existing in-house solution. Evidently, this was not viewed by anyone outside CoreTek to be a fundamental problem, but insiders, especially those that followed the Lucent/Agere saga, saw it as an indication that upper management was again asleep at the wheel. This was seen within CoreTek as another shining example of Nortel not understanding what it had in its prize newly purchased division: real products that were readily available for a current market.

So why didn't Nortel let CoreTek buy itself back? It tried to, and this brings us to an even more speculative side of the saga. If it is obvious that it will cost Nortel more to close down CoreTek than to keep it open, why wouldn't they just give it away to any interested entity? I was going to write just that in an editorial after hearing about the closure versus sell-off route, but something kept me from speculating without more information. With the sale to Bookham, it now makes sense, and becomes an important object lesson for others who may find themselves in a similar internal technology struggle. With the announcement this week of the Bookham deal, the answer is obvious. While Bookham already has its own versions of these products, allowing someone else to buy CoreTek would represent a serious challenge to Bookham for these devices. Some of the loyal core at the heart of CoreTek still feel that there might be some life to the possible management buyout, but with Nortel now becoming the majority shareholder in Bookham, how can CoreTek being owned by someone other than Nortel now be in Nortel's best interest?

Where is the CoreTek IP going to land? You had a leading edge product (publicly deemed "too advanced"), that met technical requirements, and proved it could sell. Did it go to Bookham for a near monopoly on the best DBR-technology around? Is Nortel saving it in case their 30% of Bookham doesn't pan out as they would expect it to? Will some CoreTek folks and a good IP attorney take the necessary forward steps to groom it to emerge, Phoenix-like, from the Massachusetts ashes? It would take some moxie, but wouldn't that be fun?

Sensors Unlimited Provides New Jersey Boost by Buying Itself Back

Bookham Agrees to Acquire Nortel's Opto Group, Minus CoreTek

Strong Showing of Blue Spectrum Laser-Based DVDs at CEATEC

Sirenza Puts Dibs on Vari-L

Structured Materials Awarded Tibbetts

SatCom Weighs In with $19.4 M in New Bookings

ASIP Emerges with Introduction of Initial 1310nm InP Integrated Photonics Device

The CoreTek Mystery Saga: Where's the IP?

Apologies to Our CSF.org/Primer Site Users