Last week’s announcement by Avago that it would buy Broadcom presented this year’s second major example of how the semiconductor industry is seeking consolidation.
At one level, the acquisition is a symptom of the need among second-tier semiconductor to find strength in numbers. The merger of NXP Semiconductor and Freescale Semiconductors is another example, although the Avago-Broadcom combination will result in a much larger entity.
Assuming Avago does not sell off a significant number of Broadcom operations to other suppliers post-acquisition, the new company will have more than $15bn in sales, compared to the $10bn of the new NXP, effectively bracketing the $13bn Texas Instruments.
Unlike NXP and Freescale, which today compete in microcontrollers and RF among other markets, Avago and Broadcom barely overlap at all in terms of product types. It limits the amount of design-team layoffs and divestments that will be needed. The same day as the Avago and Broadcom deal was announced, NXP said it would sell its RF power amplifier operation to JAC Capital, letting the combined company concentrate on Freescale’s offerings in that area.
However, the lack of overlap in such an acquisition risks creating a sprawling, unfocused company that fights on many different fronts. Arago-Broadcom may be different and in many ways the strategy is similar to that of Maxim Integrated in recent years, albeit on a much larger scale. Starting as an analog-focused company Maxim has been buying into digital technologies in order to create a mixed-signal SoC business.
The split between the core Avago product line – not including the recent LSI Logic and PLX acquisitions – and that of Broadcom is even larger than that of Maxim and its acquisitions. Much of Avago’s core product line relies on much more specialized processes than those used by CMOS mixed-signal parts. They include parts such as resonating filters for cellular communications, laser diodes for networking infrastructure, and RF amplifiers based on III-V materials.
These physical-layer products provide an integration strategy different to the one Broadcom has pursued historically as the company faces being locked out of further digital SoC integration by larger competitors, such as Intel and Qualcomm. Being able to attach storage and network control direct to an multi-gigabit optical Ethernet transceiver is one way to increase the density of servers and an opportunity for Avago-Broadcom to capture a large chunk of the server’s motherboard sockets even though the company might not be able to claim the host processor.
Broadcom has a strong position in carrier Ethernet and the controllers behind optical transport switches. As software-defined networking (SDN) and network function virtualization (NFV) bites, the crossover between compute servers and network equipment will continue to increase. The ability to co-optimize light and packet processing may provide a stronger position to help drive future high-speed standards.
Avago’s strong foothold in optical communications also supports a likely move towards photons instead of electrons for high-speed but low-power intrasystem transactions and, potentially, providing a way of maximizing bandwidth across blades.
The combined company may not be able to do the same for cellular handsets, where Broadcom’s market share has dropped away. Cellular infrastructure though may provide opportunities for integration, particularly for the tiny basestations that will be needed to support further urban-LTE and public WiFi rollouts, in anticipation of the mixed-network 5G environment.
It might be the combination that delivers on the more-than-Moore integration promise as the market comes to terms with the implications of continuing monolithic-SoC integration.