We had an opportunity to talk with Drew Henry, General Manager of the MCP Product Group at NVIDIA about the ULi acquisition. As is typical of most interviews, we send a list of questions to NVIDIA and see what comes back. This time we got more marketing talk and less specifics than we wanted, but there were still a handful of interesting points.
1. NVIDIA isn't specific on which technologies they aim to get from ULi. Perhaps that is because it would provide a clue into areas where NVIDIA is currently weak, and because there's always the possibility that a regulatory agency will determine that NVIDIA is not allowed to acquire ULi.
If I was forced to speculate, I would suggest mobile chipsets and SATA-II. Although NVIDIA offers nForce3 mobile chipsets, they've been unusually quiet for the last few years. From this interview, we now know that 2006 will be the year that NVIDIA will return to the mobile market in greater force. The ULi acquisition could be beneficial because ULi has a license for Pentium M bus as well as a fully developed chipset. NVIDIA could adapt and augment technologies developed at ULi.
The second area for potential technology transfer is with SATA-II. Although NVIDIA's SATA-II is good, Intel's implementations have been better. Although we haven't tested a modern ULi chipset ourselves, our colleagues at Anandtech recently published some benchmarks showing excellent SATA-II performance from the ULi chipset. Although it's unlikely that improved SATA-II performance was the driving force for the acquisition, I would be surprised if NVIDIA did not adapt ULi's SATA-II implementation into their next-generation product.
2. NVIDIA has motherboard teams in the US and India, but surprisingly they did not have a presence in Taiwan. The ULi acquisition should give NVIDIA additional strength in field engineering support in Taiwan as well as kickstart the Taiwanese division of NVIDIA's motherboard business. The full acquisition of ULi seems to be a more cost effective solution than trying to recruit an entire division worth of engineers in Taiwan from scratch.
3. NVIDIA "intends to continue ULi's business and [they] expect no change in service to their current customer base."
Many assumed that NVIDIA would choose to terminate ULi's deal with ATI. After all, ATI has been largely dependent on ULi southbridges for their motherboards and terminating this support could deliver a significant blow to ATI. That said, it's possible that the continued support ATi/ULi would, over time, give NVIDIA the monopoly on the know-how for producing high-quality south bridges. Likewise, because ATI presumably must design their northbridge to work with ULi southbridge and vice-versa, maintaining this support would give NVIDIA extra insight into the workings of their competitors. Third, continuing support for ATI may ensure continued support from ULi's other customers, which is one of the reasons for the acquisition in the first place.
The full interview is reproduced for you on the following pages: