More Wirless Infrastructure

In my post looking at all the good that Microsoft did for Nokia Networks, I touched on a subject that I want to explore further- the wireless infrastructure market. I know in that post I warned that this was one of the dullest subjects in tech, but I could not help myself. Covering that market day to day can be very dull, but viewed from the perspective of a multi-year cycle there are some important things happening.

In the previous note I said:

I find this level of consolidation, in an immensely conservative industry, to be staggering. I would argue we are likely to see another round of consolidation somewhere down the line. But that is a topic for a different post. 

This is that other post.

First, let me define the market. It is the market for base stations and other pieces of network equipment that connect a cell phone to the Internet. This business emerged as vendors of specialty network equipment for mobie operators. Today there are five companies that really dominate the space:

  • Ericsson
  • Huawei
  • Nokia Networks + Alcatel-Lucent
  • Samsung
  • ZTE

Historically, base stations were worth a lot of money, but with price compression the value has shifted to other parts of the chain. Like so many things, the value here is increasingly being found in the software and integration parts of the sale. As a result, the big equipment vendors are in many ways becoming systems integrators that focus on the telecom vertical.

Selling equipment alone is no longer profitable. Base stations for 2G networks used to sell for $500,000, while today you can buy a 4G base station for as low as $10,000. That sounds bad, and it certainly caused big problems for Lucent, Motorola, Nortel and Siemens. But Ericsson barely slowed down. The remaining players in the space have  found other ways to deliver value to their operator customers. This has largely meant switching to sales of software, services and integration. That sounds nebulous, but is really not that different from a consumer paying $600 for a phone that only cost Apple $150 to build.

The ability of the remaing five vendors to adapt to this new value chain varies across a spectrum. Samsung, the smallest, is probably the most reliant on hardware. ZTE and Huawei are more weighted to the hardware end of the scale, but no one should underestimate the advances they are making in moving more towards software and services. Huawei, in particular, has become very active in promoting various open source software communities. This is classical strategy, destsbilize a rival’s competitive advantage by creating replacements.

In fact, another trend shaping the industry is the fact that the network iteslf is changing. Once upon a time we had different kinds of networks – a network for telephone companies, a network for cable companies, a network for academics, a network for governments, and a bunch of networks for businesses. Each of these had their own standards and suppliers. Today, we are pretty much down to two networks – Telecom (and cable) access networks and the Internet. Telecom operators, especially mobile operators, still spend a lot of money building their networks and thus get to design ‘custom’ gear for their industry, but pretty much everyone else uses the same gear.

I do not want to stray into the land of abbreviations, but if I did, here is the point where I would mention SDN and NFV.  Let’s not go down that path. Instead,  I will leave it with we are very likely seeing a move towards a complete compression of network types. It is very possible, that a few years from now, every type of network will run on the same sorts of equipment. The exact shape that takes is still very much up for debate, but the underlying trends are pretty clear.

For the big five telecom suppliers, this likley means that one of their big defensive barriers will disappear, and new entrants will start storming the citadel of operator capex dollars. Maybe those new entrants will be software companies running on commodity hardware, or maybe those new entrants will be named Cisco and HP.  Too soon to tell, but this will be one of the more interesting spectacles to watch in technology over the next decade.

This does not mean that the telecom majors are doomed. But I do think it will prevent them from enjoying any stability. Even the strongest of these companies, Ericsson, has some big holes in its product categories. These holes are big enough that they will likely either have to pay up for some big acquisitions (JNPR? FFIV? BRCD?) or end up being acquired themselves (very, very unlikely).  Nonetheless, Ericsson has the resources and relationships in the industry, it is hard to see them exiting. There is also the possibility that they jump a step ahead of everyone else taking computing in a different direction. They have not been standing still.

By contrast, Samsung looks very small. They have made a lot of progress in recent years, coming out of nowhere. I also think the parent chaebol sees this as a big addressable market for them to partially replace the inevitable decline of their handset business. So they have the motivation and resources to stick with it. Unfortunately, their product line is sub-scale and generally lacking in software. I think they either make a big acquisition bet or eventually fade away.

Huawei and ZTE have significant technical resources to stay in the race. They also have significant… how to put this…Institutional Resources which can back them up. For instance, their ability to get a $10 billion line of credit from China’s State-run banks to help customers finance equipment purchases. I think people tend to lump these companies together too often. They are run very differently, with very different corporate cultures. My sense is that Huawei is much further ahead in building software stacks, while ZTE has a much broader enterprise distribution channel. So maybe they eventually go their separate ways.

That leaves Nokia Networks (+ Siemens + Alcatel + Lucent), the company which sparked this series. If you line up all of the combined companies’ products, my guess is that you would find there are still some significant gaps in their line up. Importantly, Alcatel-Lucent does have some very interesting networking software initiatives. But right there is the problem – how do we line up the combined companies products? The deal probably does not close for another year, and then the real work begins. What gets cut? What gets fully funded? Integrating companies is complex and can be painful. Adding to the problem is the fact that Alcatel itself probably never fully digested Lucent. The Lucent acqusistion carried all kinds of baggage with it. My guess is that they are still working out accounting and compliance questions from ten years ago.  The combined company is going to look unstable for a number of years. So even though I celebrated their survival in my last post, I have to wonder how much longer they will remain independent.

One response to “More Wirless Infrastructure

Leave a Reply