And then there were ten

Yesterday’s big news was the combination of Analog Devices and Maxim, two of the largest analog chip makers left in the US. We have been commenting on the consolidation among US semiconductor companies for several years, and we see yesterday’s deal as one of the final plays to be made in the space. However, the deal itself is likely to have limited impact on the overall industry.

ADI and Maxim both make ‘analog’ semiconductors. That is to say chips that bridge the physical, real world and the digital world. Both companies’ websites contain a blizzard of products and industries. This tells us pretty much everything we need to know about the analog business. While we usually think of high-volume chips like CPUs for laptops or applications processors for mobile phones, analog products tend to be low volume and high margin. Each chip chip requires a lot of design work to semi-customize it for particular uses or customers. This is not a custom business, but it does operate under very different dynamics than the high volume chips.

This implies that there is not a lot of room for synergies. A point that Ben Thompson made in his good analysis of the deal (the link is behind a paywall, but you should really be subscribed to that), he pointed out that the deal is priced at a ~20% premium to Maxim’s share price implying limited cost savings.

ADI and Maxim are both solid companies, earning design wins in cars, base stations, X-ray machines and the concessional short-run smartphone socket. These companies represent the late stage of semis, they are grinders not high fliers. They are also both sub-scale. Both companies have been in a sort of limbo for many years, unable to grow much beyond the industry or eke out sustainable share gains. Normally companies in this position move to heavy M&A work, but there are very few targets left, except each other. The indisputable category leader in analog is Texas Instruments which is both much larger and much slower to do M&A. TI probably could not have bought either company for fear of attracting anti-trust obstacles.

There are a few other companies of any scale left here. On Semi is arguably an adjacent target. As is Silicon Labs, whose stock fell today possibly because the market was hoping ADI or Maxim bought them. Neither On nor SiLabs fits well into a natural M&A strategy. For On, cost synergies are probably limited as their product range is even more diverse than the other analog players. Silicon Labs has been investing heavily for years in its IoT platform. This is a smart strategy, but one which it will take years to show returns. The company is unlikely to sell until they know how well that bet is paying off.

The nature of the market for analog chips lends itself to fragmentation. A small team of designers can put together a chip fairly easily. Since these products use trailing edge fabs, the cost of taking a chip to manufacture are fairly low. This means that anyone with particular insight into a particular product area (e.g. former employees, research labs, “mom & pop” operations specializing in a particular niche) can build a competitive product. As a result, there are hundreds of tiny chip companies doing essentially custom work. Almost all of these companies have no viable path to scaling up into larger entities. They are the proverbial ‘lifestyle business’ , or at least as close to that as semis companies can manage. Doing M&A in this space is a headache. We have seen target lists with hundreds of names on them, none of which look particularly appealing. From a Wall Street, rational Capitalism viewpoint, this is a mess. From a US Strategic viewpoint this is something of a strategic reserve, the National Guard, if you want to torture the metaphor. Capable but far from ready for large scale conflict.

It is also worth pointing out that analog chips do not participate well in Moore’s Law. These chips have to interact with real world phenomenon which tends to limit the ability of the chip to shrink. No matter how small the circuitry some part of the chip has to handle radio waves or sound waves or power surges that dictate large chunks of chip area. So there is no miracle coming with the next process shrink to shake things up.

As we said at the outset, this means this deal is important but is unlikely to alter things much across the broader industry. There are only a handful of M&A deals left in analog space.

That being said, if we view this in the context of US-China relations, we remind people to consider analog in any strategic calculation. Most of the chips that go into planes and warships and satellites are analog chips. The US has no strategic deficit on this front. ADI and Maxim combining marginally shifts this equation, assuring the US has two full-scale analog providers, who have most of their fab assets in the US. Indeed, one question about the deal is will it be able to pass anti-trust muster in China. All these companies supply Huawei, but are far from dependent on them. As part of the Trade Deal announced last year, China indicated it would take a more permissible stance towards US M&A. As far as we know this deal has no clear, commercial anti-trust red flags for China or the US. So China’s review of this deal will be an indicator on where the Trade Deal stands.

One response to “And then there were ten

  1. Pingback: Do Nvidia’s competitors want it to fail with Arm? | DIGITS to DOLLARS·

Leave a Reply