Macom Technology Solutions Holdings has announced its intention to acquire applied Micro for c.$770m – an important move in the large (but often overlooked) semiconductor market. Both companies are relatively new of course, but their relative success in terms of net income has been very different (see my Figure for a comparison based on a rolling 4-quarter analysis). Let’s look at the announced deal a bit more closely.
There has been a disparity in revenues as well as net income (see my Figure), as well as strategy. In particular:
- Macom’s has been mainly about M&A activity. It bought Mindspeed Technologies in 2011 before selling on its wireless products to Intel in 2014. In the same year it bought Nitronex – a specialist designer of Gallium nitride semiconductors, microwave company IKE Micro, BinOptics and Photonic Controls (a Silicon Photonics services company). In 2015 it bought the Japanese optical sub-assembly supplier, FiBest Limited as well as the diode business Aeroflex’s Metelics from Cobham.
- Applied Micro’s designs ‘connectivity’ products for enterprise and service provider networks, but it has also developed alternative server chips to challenge the domination of x86 products from Intel and (to a lesser extent) AMD. It acquired IBM’s PowerPC 400 design team, assets and Intellectual Property in 2004, although it hasn’t played a major role in the newer Open Power Foundation. It was the first company to implement the ARMv8-A architecture with its X-Gene Platform from 2011. These data center and embedded computing products Macom intends to offload once the deal is complete.
The deal will affect a number of interested parties. In particular:
- Winners
- Intel’s x86 will have less of a challenge from ARM alternatives.
- Macom itself, which will be able to expand its connectivity business.
- Losers
- potentially HPE, Dell and other OEMs who lose a chip designer for their low-power server, storage and other embedded enterprise computing developments.
- All server users if this proves to be another step in the on-going march of x86 products.
Of course things may turn out very differently – better if Applied Micro’s computing business is picked up by someone with the time and resources to continue to pioneer ARM server markets, or worse if this business – like Mindspeed – ends up being bought by Intel. It’s interesting that the ‘software defined’ storage and networking markets are booming, creating more sales of servers as a result – but that almost all of these disruptive products are based on Intel Xeon x86 chips.