Semiconductors have been the foundation of Silicon Valley for decades. The industry has grown to more than $335 billion in revenue, but growth has slowed. And in the past year, there’s been a tremendous amount of consolidation as chip companies that have been around for decades get acquired by larger competitors.
During 2015, more than $100 billion in mergers and acquisitions was announced, including Avago Technologies’ $37 billion purchase of Broadcom, Intel’s $17 billion acquisition of Altera, Microchip’s $3.6 billion acquisition of Atmel, and Western Digital’s $16 billion acquisition of SanDisk. Since 2013, nearly a quarter of the 30 companies that were once in the Philadelphia Semiconductor Index of publicly traded chip companies have been acquired. And now we’re seeing the fallout from some of those deals, with Intel announcing in April that it would lay off 12,000 employees.
That raises big questions about what comes next, with industry revenue growth stalling even though big trends such as the Internet of Things could lead to unprecedented volumes in sales of electronic devices. We don’t know the answers, but we have heard some concise views on the state of the industry from one expert.
Chanan Greenberg is vice president and general manager of high tech at Model N, a Redwood City, Calif.-based provider of revenue management for about 12 of the top 20 semiconductor companies. Model N’s clients include Advanced Micro Devices, Atmel, Intel, and more. We spoke with him about the trend toward consolidation in the industry and what chip makers can do about it.
Here’s an edited transcript of our interview.
Chanan Greenberg, vice president and general manager for high-tech at Model N.
Above: Chanan Greenberg, vice president and general manager for high tech at Model N.
Image Credit: Model N
VentureBeat: Tell us what you do in the chip industry.
Chanan Greenberg: My role at Model N is vice president and general manager of the high-tech business. I’ve been with Model N now for just over 10 years. Prior to that, I spent six years in the government contracting space, working with companies like Lockheed and Raytheon. Prior to that, I spent seven years developing multimedia products for high-tech OEMs, including IBM, The Learning Company, HP, and others. That’s my background in a nutshell.
Model N is a company which started out in Silicon Valley about 16 years ago. We’re a software company, publicly traded on the NYSE, with just over 800 employees. Our HQ is in Redwood City, not far from San Francisco. We have locations around the country — Denver, Boston, Chicago — as well as in London and Geneva. We also have a venture out in Hyderabad, India.
The company is focused on providing what we call “revenue management applications” for the leading brands in pharmaceuticals and semiconductors. Today, 17 of the top 25 pharmaceutical companies worldwide use Model N — Johnson and Johnson, Pfizer, Boston Scientific, many others in pharma and med tech. Twelve of the top 20 semiconductor companies use Model N, including Intel, AMD, Qualcomm, Broadcom, Micron, and the list goes on and on. Across those two verticals we have just north of 100 customers, and we have a pretty strong presence in the top tier of those companies as well as the middle range and the small range.
In semiconductor, we manage everything from the design cycle to revenue. That includes the design registration and approval processes, the quoting and pricing and contracts, directly and through channels. We have about 50,000 users in semiconductor companies in sales, operations, marketing, pricing, and finance departments. We’re also used heavily by the channels. Although the channels are not our customers, our system is made available to them by our customers. A company that has 3,000 people in our system, as much as 50 percent of them might actually be users from other places.
We also manage a lot of the backend channel finance processes – the ship and debit process, credit claims, inventory management, price protection on inventory, and varying values of inventory over time. Last year, we acquired a company that added to our portfolio the ability to do end-to-end channel data management, from the first engagement to the channel, defining the way the data is collected, collecting the data, cleansing the data, and enriching it with a database of more than three million end customers we maintain.
VentureBeat: How does that background lead to our topic of financial consolidation in the chip industry? How much has that been happening lately? Have you gained some insight there through the business you’re in?
Greenberg: A huge portion of semiconductor revenue is run through the Model N system. Every quote, every contract, every credit claim, every design restriction or design win is tracked through our system. This gives us significant visibility into what’s going on, and it has a direct impact on us. The integration or the consolidation we’re seeing across the board has a material impact on the way our system is used.
When NXP acquired Freescale, both of them Model N customers, we were heavily involved in all the processes associated with the integration of those companies. Especially when it comes to how they manage their design cycles, how they manage their prices, how they’ll handle the channels. It gives us an interesting vantage point and an active role in bringing these companies together.
We’re seeing a huge amount of visibility. The people we work with on a routine basis are the CFOs and heads of sales. I work routinely and directly with the EVP of worldwide sales at ST Microelectronics, with the VP of worldwide sales from Microchip, and so on. As a company, we’re directly impacted by this because this is determining how our software is used. If there are efficiency gains arising from these acquisitions, we see the impact that has on the number of people using the application, but more important, we see the impact it has on the overall operation of these companies.
Through the consolidation of these massive companies, we’re seeing larger and larger companies with a longer list of products that they’re supporting. They have a more diverse customer base to support. That creates a lot of challenges and opportunities that we’re involved in addressing. To your first question about relevance, we’re deep in every one of these big deals. We’re one of the two or three pillars they run their business on. We’re very involved.
Chips require precision manufacturing and assembly.
Above: Chips require precision manufacturing and assembly.
Image Credit: Model N
VB: Why did all these deals start happening all at once? Was there some kind of trigger? I’ve covered the industry for 20 years. I’ve never seen such merger mania until the last year or two.
Greenberg: There are a few aspects of macroeconomics driving this, but I think there’s also an aspect of the maturity of the industry. It’s a normal process you’d expect in almost any industry. But there were macroeconomic conditions that made it necessary for the industry to make these changes.
Do we really need eight different vendors making memory, just as an example? Five or eight years ago we probably had 16 or 17 companies of variety of sizes in the memory business. There was a certain amount of expertise built up, but there were also a lot of commoditized capabilities. It was only natural to see a consolidation with large clusters like Samsung and Micron. That’s a natural progression.
You can say the same thing about the microcontroller space. There’s very few places, in fact, where the chip itself has become a huge differentiator, despite all the money that goes into the R&D of the chip. Most of the more broadly thinking, strategically thinking companies no longer think of the chip and the IP of that chip as their biggest differentiator. There’s been a commoditization of the technologies, a bifurcation of too many providers. It just didn’t make sense. There had to be consolidation of that aspect. It would happen in any industry.
There are macroeconomic aspects making it compelling. About two years ago, I attended the Global Semiconductor Alliance conference. It’s an interesting conference, to meet the leaders and listen to them. There was a forum of CEOs from Qualcomm, the CEO of Worldcom, a couple of others. They were all talking about how they see the future.
At some point, the EVP of strategy from TSMC got up and said, “Guys, I don’t know what you’re talking about, but the size of the industry has remained largely unchanged. We are a $300 billion industry. We’re stuck in the same range between $280 billion and $330 billion for the last five years. You talk about cloud computing and how great that is and all the money it’s going to create, but really it’s just putting the expense in a different bucket. The server farms inside the corporate environment are shrinking. We can all get excited about the iPad, but sales of laptops are declining. We have areas of growth, but they’re not outpacing areas of decline.”
The market as a whole has become stagnant, and there’s a natural economic reason for it. In the developed economies, like the U.S., the rate at which we consume new technologies is declining. I’m holding an iPhone 5C in my hand, and I’m not all that compelled to go to the iPhone 6 or whatever’s next from Apple. I don’t feel that I have to go from my iPad Air to the iPad Air Pro. That can wait another year or two. The rate of adoption is declining.
NXP's chips could be used in self-driving cars.
Above: NXP’s chips could be used in self-driving cars.
Image Credit: NXP
The buckets have changed, but the overall size of the pie hasn’t changed. We’ve moved from laptops to tablets, from server farms on premise to cloud farms and virtualization. That actually means less chips being sold.
VB: What does this mean?
If I look at Africa, India, China, the developing economies, while we expect great things from them, and they do have phenomenal growth already, what they’re buying are low-tech solutions. In some of these markets, moving from no phone to a flip phone is progress. They’re not all buying smartphones. Again, the amount of technology driving growth in those markets now is still not pulling the high end of the market and the bigger prizes.
You have this double whammy of general economic uncertainty, the usual complaints people have, but then you have some real issues of a slowdown in the developed economies. The developing economies are ramping up, but they’re not ramping up as quickly or on as high an end of the solutions as we’d like them to in the industry. That has contributed to a stagnation of growth. It depends on who you believe, but the industry is projecting three or four percent growth for next year. This is not very compelling to investors.
What you have is a situation where CEOs and CFOs look at their investors, who expect growth, and that growth isn’t coming from just organic sales. And so these companies are very proactively and very deliberately doing the next best thing, which is buying revenue. That’s driving this behavior. There are a few exceptions, but the first thing is, they have to show growth, and so they’re buying that revenue. They’re taking out competition and trying to make sure they’re buying and not being bought. That’s been the marching order for many, many companies. Find the revenue and buy it. And you’re right, because if you look at the rate of spending on M&As in 2014 versus what it was in 2015, I think it’s something like fivefold higher.
As I say, there are some exceptions. Some really big companies out there have grown disproportionately and very quickly with revenue driven by one or two customers. The obvious culprit is a fruit company in Cupertino, which for some companies has become the main source of revenue. Those companies realize, first of all, they’re starting to plateau, and second, their revenue profile, from an investor community perspective, is very high risk.
What happens if you get designed out? Do you lose 80 percent of your revenue? Some of those companies are compelled to go and buy other companies, not only to show some additional growth, but mostly to diversify their customer base and revenue sources to reduce the risk of their revenue profile. That’s driving some of the acquisitions. But out of every 10 acquisitions, I’d say only two or three of them are motivated by that. The others are motivated simply by the need to grow.
VB: I never thought that some of these big acquisitions would result in more efficient companies.
Greenberg: I agree with you. That’s a huge issue. One other thing that dawns on me that I didn’t talk about as far as macroeconomics — There are all these new applications we expect to see take off, whether it’s wearables or self-driving cars or more consumption inside medical devices. All these things are happening, and they look promising, but for them to become a meaningful part of the revenue stream for a lot of these players in the space, it’s probably three to five years out. That’s the pace at which these solutions are developing, and the market is adopting them.
Companies simply can’t wait another three to five years with low single digit growth rates. The investor community would hammer them. They have to do something. That’s another motivation to go and address.
But the aspect you just raised about efficiency, you’re absolutely right. On paper it sounds great. If I’m ON Semiconductor and I acquire Aptina, maybe I can show on an Excel spreadsheet how I’m going to reduce GNA and get some operational efficiencies and improve my sales efficiency because I’ll have salespeople selling more product. It looks great on a spreadsheet.
In reality, you have an inherited sales force from Aptina that knows how to sell image sensors. They have no clue how to sell anything out of the 30,000 SKUs in the ON Semiconductor catalog. And you have a bunch of people at ON Semiconductor who have no idea how to sell image sensors. Plus, if I’m an ON Semiconductor salesman, I go engage with a customer who’s asking for a component, and I’m giving him a price. I’m not necessarily thinking about the entire block diagram of that board, looking at all the components that ON Semi may have organically or inorganically acquired and saying, “Hey, let me structure a better deal for you. I can provide you with the image sensor here, the microcontroller there, the memory over there.” What you’re actually creating are these silos of ineffective sales.
My impression is that, in the short term, these companies are actually going to see an increase in the cost of sales as a percent of revenue, rather than a decrease, because of these inefficiencies. That’s one of the areas, by the way, where Model N is focused on creating tools to help. The sales application we built specifically for semiconductor companies, which is built on the SalesForce 1 platform, we built a block diagram engine into the opportunity.
Even if I’m not very knowledgeable about everything my company has — let’s say I’m looking for a microchip because I’m talking to a manufacturer who’s making a smoke detector. He’s asking me for a microcontroller. Within my CRM system, inside my opportunity, based on that end application, we bring up a block diagram that says, “Here’s the component he asked for, plus six other components we have that you could propose to the customer. Try and grab those sockets on the board.” It helps the salesperson become a better solution seller. We believe that will drive some efficiency and allow the companies we’re working with to win market share at the expense of their competitors.
Efficiencies will take time to show, though. Companies are taking drastic immediate action, chopping down on any costs and senior management costs. I’m sure you’re aware of what was leaked about Atmel, the way they parted with all the C-level, SVP-level, VP-level, and most of the senior directors of the company in one day. Also they informed the employees that they’re not going to hold to everything they were told in terms of severance. They’re taking drastic steps.
But when you look at the technology, the next level of where they can get efficiency—“We make microcontrollers. So does Atmel. We compete in certain areas.” How do you rationalize the technology portfolio? That’s going to take them a while, probably a year, if not longer, to figure it all out and then act on it and see those gains happen. The rationalization of the product lines will take time.
Up front, it wasn’t about efficiencies. It was simply about top line growth. If you look at all these statements, all the 10Ks and all the leading CEOs of these companies, they’re talking about growth. They’re not looking at margins or efficiencies. If they don’t grow, they’re dead. The investor community, the valuations, will show that.
Intel's booth at CES 2016.
Above: Intel’s booth at CES 2016.
Image Credit: Dean Takahashi
VB: With Intel, they said they were going to cut 12,000 people. That’s 10 percent of the workforce. Does that signal something about what’s coming? Will we see cuts across the board in semiconductor companies on this level? Or is it mainly the ones that have acquired somebody big?
Greenberg: Intel did buy Altera. It’s not a material acquisition, but it’s still sizable. Every company has its own story. Those that did the acquisitions have announced that they’re going to be making what they call productivity or efficiency moves. Every one has been promising investors anywhere between $200 million and $1 billion in savings. Those are going to come first and foremost from laying off people. It’ll be a while before those savings also manifest from how effective and rationalized the product line is, how effective the sales organization is. That will take longer. Laying off people is the fastest way. That trend will continue in the ones doing acquisitions.
In Intel’s case, if you look at where they’re doing the layoffs, it’s not the areas where they believe the business is growing or where they’re betting on the future. There were no material layoffs in the IOT group I know if, which is the fastest-growing division there right now. It mostly hit the people who were focused on PC. Given the shrinkage of the PC market in the last few years — even if you didn’t have everything I said about macroeconomics and the need to grow, that was a move that had to happen anyway.
VB: What do you see on a regional basis? When I was covering the chip industry decades ago, there was always a huge amount of regional pride that some of these countries had. They had some of the big surviving chip companies in their territory. Now companies are getting taken over. I wonder what the modern thinking is like as far as how many regions will still have this key industry present. An example I think of is Micron, which was always considered a United States strategic asset.
Greenberg: That’s still the thinking. I don’t think any foreign entity would ever be able to buy Micron. That doesn’t change much. Micron is definitely doing acquisitions. They’re broadening and differentiating their portfolio. I would find it very difficult to ever see the U.S. government approving a deal to acquire someone like Micron.
This isn’t just our country. I see that in other countries. China is fascinating from multiple angles. Last year was the first time that the number one commodity imported by China was not oil, but rather semiconductors. That’s from a monetary value perspective, not units. When that happens, it gets government attention.
The government has publicly said that they’ve put aside a $100 billion fund — sources closer to the details say it’s more like $5-20 billion, but nevertheless a sizable fund — designed over the next five years to improve China’s independence in this area for internal consumption. China’s internal consumption of semiconductors has become a sizable market, a very viable market for them to create their own companies for internal consumption, not just manufacturing for consumption in the U.S.
A lot of countries view these capabilities as strategic. The government committee in China that’s working on rationalizing which companies they should buy is meticulously going across different capabilities, different technology stacks. They’re very deliberate in how they’re going about buying this. It’s not a revenue-buying initiative. It’s a strategic capability acquisition process that they’re going through. If they’re thinking that way, I can’t envision a situation where the U.S. government would think any differently.
This is why you have ST Microelectronics in Europe. If you look at their gross margins, you might say that they’re not the most efficiently run company. If you look at their prices in the marketplace, you’d see that they’re relatively low. If you look at the shareholder structure of that company, you have the French, Swiss, and Italian governments holding sizable stakes. They’re a strategic asset for Europe. That aspect will not go away.
An electronics market in Shenzhen
Above: An electronics market in Shenzhen
Image Credit: Andrew Bunnie Huang
VB: Are there any other topics that might be worth mentioning in this context?
Greenberg: My high-level take about the state of the industry — I expect the acquisitions to continue. There are two really key trends on the business side that you need to pay attention to. One is, “How good are my salespeople? Do they know how to sell my stuff? How do I enable that?” That’s one way to get reasonably good efficiency gains if you have the right capabilities and tools in place.
The second thing is the importance of channels. If you take the example of Qualcomm, their revenue stream prior to the acquisition and CSR was driven by a very limited, single-digit number of customers driving a huge portion of their revenue. The relationship was very personal, very hands-on and very direct. They’re now in a world they weren’t used to, where they have to serve hundreds or even thousands of customers. They don’t have the logistics, the supply chain, the infrastructure to do anything around that. The reliance of the company on channels is growing. This is true across the board.
In the industry, there’s always this love-hate relationship with the channels. The semis and component makers view the channels as takers. The channels are saying, “Well, without our logistics you wouldn’t be able to serve this business.” There’s a bit of a tug of war going on now. But the reliance on channels now is absolutely critical.
Again, if you look at a company with a broader base of customers, a Microchip, they have probably about 800 to 1000 customers that they work with directly and that are very important to their business. They serve more than 120,000 end customers. There’s no way on the planet that they’d have the logistics to handle that if they did not have the network of channels and distributors they work with. As this trend continues, as the number of end customers grows, as the diversity of applications grows, you’ll rely even more on the channel.
How much data do you get out of the channel? A big portion of your business is driven by the channel. If your data from there is subpar compared to your direct business, you’re managing a big portion of your business based on iffy information. Handling channels is going to become even more important in the next few years.