Noises coming up from the basement of RF

The Journal reported yesterday that Broadcom has put its wireless business unit up for sale. If it comes to pass, this has some big implications for the wireless components industry.

Broadcom should be viewed as an iconic company in the semiconductor industry. For over a decade they have pioneered a new model for a semis industry that has matured, and is no longer the growth engine many view it as. The Broadcom model has been to acquire companies with leading market shares in specific niches, then ruthlessly cutting costs and divesting non-core assets. By focusing on profitability, they have been able to ramp up cash flow which gives them firepower for the next deal.

For those in Finance, this looks a lot like the playbook of a private equity firm. Broadcom was once a spin-out from HP, then known as Avago, and for many years has in fact been backed by PE firm Silver Lake. A few years ago, Broadcom essentially maxed out on semiconductor deals. After being blocked by the US government in its bid for Qualcomm, Broadcom had gotten so big that there were no semis companies left to buy which were sufficiently large to move the needle in Broadcom’s model. And so they moved on to software. (We have an unpublished post on this subject which we may dust off and post.)

According to the Journal, Broadcom has hired bankers to sell its wireless business unit. This is comprised of two businesses – connectivity (Wi-Fi and Bluetooth) and RF (filters and modules). The connectivity business comes from the original Broadcom and is one of the two leaders in the field of selling Wi-Fi for pretty much everything. The RF business is the root of Broadcom’s fortunes, from which everything else was made possible. Broadcom is one of two vendors for BAW (aka FBAR) filters which are needed for every mobile phone. It is hard to overstate the strategic value of this business. The fact that Broadcom is willing to sell says an immense amount about the industry.

We have to pause to consider the actual news. The Journal does not attribute the news to anyone, no source is mentioned in the original piece. If we had to guess, it sounds like the story may have leaked from the bankers, perhaps a bid to stir up more bidders. The rumored price tag is $10 billion, which implies the whole business unit is for sale (revenue in FY2018 was $6.4 billion). So take the whole story with a grain of salt at this stage. Maybe there is nothing to it, maybe just part of the unit is for sale. Or maybe it is true.

Assuming the story is accurate, the obvious question is why is Broadcom selling now? They clearly see their future in replicating their model in the software industry, where there are many more targets left to be consolidated. By selling wireless, they seem to be building up their warchest for future purchases in software.

But selling the wireless business implies that the company sees less value there. They have another hardware business, selling chips for networking gear – switches, routers and the like. This is another business where Broadcom enjoys healthy margins in tightly consolidated sectors. We may examine the prospects for that business in a future post, but it is probably safe to assume they think that side generates more value, at least for now.

Broadcom seems to be saying that the RF industry has reached its point of peak value. Most people in the industry believe we are living in a “Golden Age of RF”. With mobile 4G networks, the wireless operators expanded into dozens of new spectrum bands (last count is over 50), and that complexity required the use of advanced filters, which is Broadcom’s strength.

While the number of bands being used for mobile continues to grow, and along with it the value of RF components, there are some signs that things are changing. As with everything mobile the answer to that is 5G, the latest mobile standard.

The first phase of the 5G standard looks a lot like 4G, and as a result the demand for RF components remains very robust. But the bigger impact of 5G comes in a few years with the advent of mmWave technologies. (Here is some quick background on this.) mmWave is a very different type of radio than pretty much everything we have used so far, and so the requirements for RF are very different. In this context (and greatly oversimplifying) mmWave has a lot more spectrum room to maneuver and thus does not need the same RF filters. It turns out that a lot of work still needs to be done to determine what exactly the RF requirements for RF are, but it is dawning on the industry that the heady growth of the past ten or so years is fading. There is still a lot of money to be made selling RF filters, but the competitive landscape is changing.

mmWave mobile phones are still going to need some form of RF components, but if you want to sell those you will need to come up with a whole new set of products, an investment which will cost a lot of money.

Private equity funds do not like to invest in R&D, it diverts cash from their own pockets. This alone may be reason for Broadcom to consider selling the unit. However, there is another major factor at work here.

It turns out there is a company that has already invested heavily in building out a suite of mmWave RF products – Qualcomm. That company is already selling its first generation of mmWave RF modules, at a time when most of the other vendors have barely begun designing products. To stay competitive, Broadcom will have to invest likely billions into mmWave.

By contrast, Qualcomm is already way ahead. Qualcomm has struggled for a decade to get into RF. Part of the reason they have been promoting 5G so heavily rests in the fact that the transition to mmWave will give them a major lead. Further compounding this is the fact that Qualcomm’s RF business finally looks to be in good shape, and not just for mmWave.

Qualcomm’s entrance into RF has long been the bogeyman of RF suppliers. When they first announced their entrance into RF ten years ago, incumbents Skyworks’ and Qorvo’s stock price fell ~10% in a day. The fear was that Qualcomm’s dominance in the rest of the handset would give them a massive advantage in RF. For a variety of reasons, that threat never materialized and both company’s stocks have done very well since then. But just because you turned on the lights and checked the closets does not mean that bogeyman really went away. And just like a Halloween or Elm Street movie, it turns out he may have just been sharpening his knives down in the basement.

After hearing muffled noises from downstairs for years, it now seems like something is coming up from the cellar.

Which brings us back to Broadcom. Surveying the industry, with Qualcomm now at peace (or at least a truce) with Apple and its RF products rapidly gaining share, Broadcom realized that discretion is the better, lucrative, part of valor.

That leaves us with speculation as to who could buy Broadcom’s business. At $10 billion, there are really only a handful of buyers. One option would be to sell the unit to other private equity buyers. There is some logic to this, buying a low-growth, profitable business destined for some turbulence ahead. The trick to this will be reducing costs sufficiently to meet loan payments, but Broadcom has probably already taken most of the cost out already.

Among strategic buyers, there are probably a handful of Chinese companies who would like to buy this but cannot for reasons of geopolitics. Most of the other RF suppliers are probably too small. Skyworks has a market cap of ~$20 billion and Qorvo ~$14 billion. Neither has a lot of debt, but this would be a massive undertaking for either one.

That leaves Qualcomm, Apple and Murata. Murata is a major Japanese components manufacturer, many of whose products sit right next to Broadcom’s. At ~$40 billion market cap, they could afford a $10 billion acquisition. They are smart acquirers, but also fairly conservative. So it is unclear if they have the stomach for a deal this size.

On Twitter yesterday, we saw many people list Apple as a top candidate for acquiror. They are the largest customer for this business, but would risk diminishing its value as all the other customers may flee rather than buy from a competitor.

Qualcomm makes strategic sense, as they still lack a viable BAW filter option and would benefit greatly from acquiring Broadcom’s products and team. To say nothing of the emotional victory this would provide.

With all three of these there are some anti-trust questions that will arise, especially for Qualcomm who is already under the US FTC’s spotlight.

In theory, there may be some other chip vendors who could take the bait. Intel is probably done with all things mobile at this point. For real this time. So strike them from the list. This could also fit strategically with Texas Instruments or ADI, both of whom could afford it. However, neither company has shown much appetite in the past decade for: a) large acquisitions and b) anything touching consumer devices. TI actually exited its Wi-Fi business years ago (selling it to Apple). NXP also has a potential fit here. They could afford this, just, but it would also mark a major strategic step that the company may not want to take.

So unless the Journal story scares up some other buyers, the universe of bidders is pretty clear.

That all being said, one thing is clear. If this sale actually takes place, we believe strongly that it is a signal that the RF industry is about to change, a lot.

One response to “Noises coming up from the basement of RF

  1. Pingback: The State of RF Semis in 2020 | DIGITS to DOLLARS·

Leave a Reply