The Nvidia-Arm Deal: Behind the Scenes of Big Semiconductor Industry Mergers
Nvidia-Arm deal has been creating a lot of buzz in the news. Curious about what does go on behind the scenes after such mergers? Ever wonder what it is like for the employees? And what are the best practice approaches to deliver a successful merger?
So, Nvidia looks to soon be the new owner of Arm. While the bean-counters are quickly assessing the earnings potential of this significant business acquisition, the Technorati scratch their beards and discuss the technology implications for all of Arm’s licensees. However, behind the scenes, there are lots of people and their businesses who could be impacted by the outcome of this takeover. So, what does go on behind the scenes after such mergers? What is it like for the employees? And what are the best practice approaches to deliver a successful (for everyone) merger?
Developing these IP blocks for chips is complicated, expensive, and requires considerable engineering effort. With many capabilities being based on standards, such as USB and Ethernet, it makes little sense for companies to build their own versions of such interfaces if the result is the same as everyone else’s. Instead, you’ll probably benefit by licensing that block from someone that has already designed and tested it.
Microprocessors are an extreme version of IP blocks and are especially expensive to develop. Beyond the core they also need a compiler, which needs maintaining beyond the life of each iteration of the core. You also need to look beyond the core to improve how other blocks, like memories, caches, and interfaces, can efficiently interact with the core. These must also support your licensees’ needs when they attach their unique IP blocks, a not inconsiderable effort, and try to manufacture it at their fabrication facility. Thus a processing core is no longer an IP block — it is more of an entire ecosystem.
If you’re good at them and can market them well, it can be very profitable. If not, it can be very expensive to maintain a technological advantage. Since Nvidia reportedly wants to pay $40 billion for Arm [1], you can infer that they are good at what they do.
Small businesses often find a niche opportunity that has not been fully addressed by the big semiconductor vendors, as was the case of the Quantum Research Group Ltd. They had developed capacitive touch technologies during the mid-2000s, offering silicon devices that allowed designers to replicate the touch interfaces being made popular by Apple’s iPod [2]. Atmel’s announcement to buy Quantum in 2008 followed its announcement to license the company’s technology in 2007. Thus, their announcement to buy them would have been less of a shock to the employees. In fact, “big buys small” was probably seen very positively. The larger, more established semiconductor vendor Atmel opened up manufacturing resources, sales and marketing staff, and distribution channels that would have taken a smaller company, like Quantum, years to establish on their own. Their technology still exists today in the form of maxTouch [3] touchscreen controllers, which have since become part of Microchip’s product portfolio as the result of another big semiconductor industry merger that we will come to shortly.
When two big semiconductor heavyweights announce a merger, the impact on the employees of the company being taken over can be difficult to bear. This was the case in 2011 as Texas Instruments (TI) announced the $6.5 billion takeover of analog rival National Semiconductor (NSC) [4]. Employees, such as Günter Hohma who was Senior Distribution Manager at NSC at the time, shared some of the immediate worries: “The first and most important thoughts revolve around your immediate future and how that is linked to the security of your family.” Others state that they found themselves in a state of shock, as the deal had been kept so closely under wraps. Thoughts also quickly turn to practical matters, such as commute times to new offices should the existing office end up being closed.
Of course, not everyone is guaranteed a job under some of these mergers. Eyal Barzilay was a Software Manager working at MIPS Technologies when Imagination Technologies took them over in 2013. This is probably the most comparable example to the current Nvidia/Arm deal covered here since MIPS was also a processor core IP provider. “The engineers seemed to be more sure that their jobs were secure as they were the ones that had the knowledge that Imagination was buying,” he explained. “It was those that had roles in duplicate departments, such as HR (human resources) and finance, that were most worried about losing their jobs.”
One key aspect here seems to be the need to quickly establish clarity, something that falls to the HR team to resolve. Those for whom, unfortunately, no job exists in the new business must be quickly informed. For the remainder, HR and, sometimes, external consultants work with the employees of both companies to lay down future options. Obviously, the resultant business cannot support two heads of Automotive Marketing, for example. However, the merger may have resulted in significant market growth, or the addition of application-specific technology, allowing a new senior head role to be defined to take care of that niche. Inevitably, however, some employees will have to accept a demotion and perhaps even a pay cut as a result of the merger.
At this stage, the need for strong leadership and, again, the clarity it offers is reflected in a McKinsey&Company report on deal-making in the semiconductor industry [6]. They estimated that strong involvement of the CEO, another senior leader, or a dedicated “transformation manager” post-merger resulted in an increased success factor of 2.5 over those whose leaders were less involved in the transformation. The process also needs to be swift if the desired value creation from the merger is to be realized. Those that managed the transformation within 2 years were 2.5 times more likely to attain the value (i.e., financial) benefits that they had told investors they were pinning to the deal.
In order to quickly provide clarity, and in many ways protect the existing business that has been purchased, teams are formed that represent both businesses. These develop a list of responses to frequency asked questions (FAQs). To quote John McKenna, who was a Regional Distribution Manager at NSC during the TI/NSC merger, the key here is to “communicate, communicate and, did I mention, communicate?” Under such conditions, rumors are rife and quickly develop a life of their own. Establishing clarity at this stage is, therefore, essential.
This leads to another key player in acquisitions: distributors and catalog suppliers. Semiconductor vendors typically reserve direct sales relationships for those customers that purchase millions of dollars of products every year. For those under this threshold, their product purchasing needs can be fulfilled through distributors. The smallest quantities of part purchases are typically handled by the catalog suppliers — those businesses who used to be known for their weighty tomes that excited engineers-to-be who waited for their annual release with bated breath.
Both these partners not only handle the logistics of supply, they also offload some, if not all, of the technical support. With the high levels of integration being applied to even the simplest silicon devices, which is increasing their complexity, this is not an insignificant task. Both parties undertake significant investment in these relationships from their people and systems, to logistics and training. Each distributor also has its customer base, regional reach (local/global), application focus, and, as a result, a combination of strengths and weaknesses. Semiconductor vendors are therefore selective in choosing distribution partners. Thus, mergers and acquisitions (M&A) are a worrying piece of news for distributors as, inevitably, the risk of losing millions of dollars of revenue is high. One thing is for sure: consolidation in this space will happen and the challenge is managing it as best as possible.
Many others in the semiconductor ecosystem need to be kept in the loop. These range from software and tools partners, to consultants whose businesses may have been built upon the very existence of a particular microcontroller or other silicon solution. These may also benefit or lose out, depending on the decisions made by the acquiring business.
Whether these mergers result in success or not depends, however, on another factor: that of synergies that can be established between the acquirer and the target. Norbert Siedhoff, a Sales Coach and Consultant, explains it as follows: “Regardless of how you review the deal, from the perspective of the business, its customers, distributor relationships, or that of investors, you should be looking for synergies that deliver a 1 + 1 > 2 outcome in every area.” As the Managing Director of Microchip Technology GmbH up until 2019, he got to oversee many M&As, including the purchase of both Atmel (2016) and Microsemi Corporation (2018).
The purchase of Atmel was seen by many in the industry as possibly signaling the end of the 8-bit AVR technology. It was assumed that Microchip, with its well-established range of 8-bit PIC microcontrollers, would sideline the AVR. After all, what sense would there be in maintaining two processor cores (which, as we have already established, are expensive to develop and maintain)? However, with a massive community of “AVR Freaks,” this would have put some of the potential success of the deal at risk, as well as possibly tarnishing the Microchip brand.
The reason why this didn’t happen, and why the announcement was not linked to job loss announcements, possibly lies with Microchip’s strong “Guiding Values” [9]. These include “Customers Are Our Focus” and “Employees Are Our Greatest Strength”. While, purely financially speaking, the maintenance of two proprietary 8-bit cores makes little sense, looking beyond the spreadsheet of profit and loss, it is clear that by maintaining both the PIC and AVR cores, 1 + 1 > 2. Not only did Microchip retain the AVR family, it has actually continued to invest in it [10], satisfying customers of both technologies.
Siedhoff indicates that this approach can be applied across the board. For example, whether a distributor is retained or has to be cut has to be a question of focus, not just finance. “You may have three distributors doing $100m, $90m, and $5m of business for you. If the smallest is highly specialized, that relationship is worth more than the sum of the two businesses that make it up.” This implies that, by cutting that smaller partner, you’ll probably lose the $5m of revenue completely since your larger (financially), but less specialized, distributors won’t be able to take on that business on your behalf. Günter Hohma experienced precisely this “distributor’s worry” in his current role at Richardson RFPD, Inc., as Microchip announced the takeover of Microsemi, one of their semiconductor partners. “Our customers define us as ‘trusted advisors’ in the fields of RF and High Power. We were able to not only retain our Microsemi market share after their merger with Microchip, we were able to expand it with technically appropriate [Microchip] products.”
The staff, of course, are the critical element of any successful M&A, but merging cultures is exceptionally challenging. Here Microchip provides all new employees, both new hires and those acquired through M&As, team-building workshops. The mere thought of these two-week-long sessions at the company’s headquarters in Arizona can result in rolling eyes and reticence. After all, who wants to sit in a windowless room, air-conditioned down to 18°C (64°F) for 14 days? However, Siedhoff highlights that, by the end, all participants emerge enthused and eager, ready to put their new-found skills to work with their new colleagues. Rather than focus on products, these workshops focus on instilling the values of the company, how business is done, and ensure that, especially after a significant M&A, it results in “one company, one sales process, and one face to the customer.”
Why is Arm so important?
If you were to ask a crowd of people “Who here uses Arm?,” you might see a few hands raised. But if you were to ask “Who here uses a smartphone?,” you are actually finding the answer to the first question. Arm, in the parlance of the semiconductor industry, is an intellectual property (IP) supplier. This means that, rather than making chips like Infineon, NXP, or Renesas, they only provide part of the design for a chip. In their case, it is primarily a family of microprocessor cores. These designs are licensed to their customers. They, in turn, create microprocessors, microcontrollers, and other products based upon those cores.Developing these IP blocks for chips is complicated, expensive, and requires considerable engineering effort. With many capabilities being based on standards, such as USB and Ethernet, it makes little sense for companies to build their own versions of such interfaces if the result is the same as everyone else’s. Instead, you’ll probably benefit by licensing that block from someone that has already designed and tested it.
Microprocessors are an extreme version of IP blocks and are especially expensive to develop. Beyond the core they also need a compiler, which needs maintaining beyond the life of each iteration of the core. You also need to look beyond the core to improve how other blocks, like memories, caches, and interfaces, can efficiently interact with the core. These must also support your licensees’ needs when they attach their unique IP blocks, a not inconsiderable effort, and try to manufacture it at their fabrication facility. Thus a processing core is no longer an IP block — it is more of an entire ecosystem.
If you’re good at them and can market them well, it can be very profitable. If not, it can be very expensive to maintain a technological advantage. Since Nvidia reportedly wants to pay $40 billion for Arm [1], you can infer that they are good at what they do.
The employees
Not only do mergers of this kind send shock waves through the industry, they also are a shock to the employees of the companies involved. Whether this shock is underscored by positive emotions (perhaps relief that you’ll, at last, have big-company resources behind you) or negative ones depends largely on the constellation of the deal. Big companies buy smaller ones all the time, primarily to benefit from new technology.Small businesses often find a niche opportunity that has not been fully addressed by the big semiconductor vendors, as was the case of the Quantum Research Group Ltd. They had developed capacitive touch technologies during the mid-2000s, offering silicon devices that allowed designers to replicate the touch interfaces being made popular by Apple’s iPod [2]. Atmel’s announcement to buy Quantum in 2008 followed its announcement to license the company’s technology in 2007. Thus, their announcement to buy them would have been less of a shock to the employees. In fact, “big buys small” was probably seen very positively. The larger, more established semiconductor vendor Atmel opened up manufacturing resources, sales and marketing staff, and distribution channels that would have taken a smaller company, like Quantum, years to establish on their own. Their technology still exists today in the form of maxTouch [3] touchscreen controllers, which have since become part of Microchip’s product portfolio as the result of another big semiconductor industry merger that we will come to shortly.
When two big semiconductor heavyweights announce a merger, the impact on the employees of the company being taken over can be difficult to bear. This was the case in 2011 as Texas Instruments (TI) announced the $6.5 billion takeover of analog rival National Semiconductor (NSC) [4]. Employees, such as Günter Hohma who was Senior Distribution Manager at NSC at the time, shared some of the immediate worries: “The first and most important thoughts revolve around your immediate future and how that is linked to the security of your family.” Others state that they found themselves in a state of shock, as the deal had been kept so closely under wraps. Thoughts also quickly turn to practical matters, such as commute times to new offices should the existing office end up being closed.
Of course, not everyone is guaranteed a job under some of these mergers. Eyal Barzilay was a Software Manager working at MIPS Technologies when Imagination Technologies took them over in 2013. This is probably the most comparable example to the current Nvidia/Arm deal covered here since MIPS was also a processor core IP provider. “The engineers seemed to be more sure that their jobs were secure as they were the ones that had the knowledge that Imagination was buying,” he explained. “It was those that had roles in duplicate departments, such as HR (human resources) and finance, that were most worried about losing their jobs.”
One key aspect here seems to be the need to quickly establish clarity, something that falls to the HR team to resolve. Those for whom, unfortunately, no job exists in the new business must be quickly informed. For the remainder, HR and, sometimes, external consultants work with the employees of both companies to lay down future options. Obviously, the resultant business cannot support two heads of Automotive Marketing, for example. However, the merger may have resulted in significant market growth, or the addition of application-specific technology, allowing a new senior head role to be defined to take care of that niche. Inevitably, however, some employees will have to accept a demotion and perhaps even a pay cut as a result of the merger.
The outside world
It is not only the people within the organization that are impacted; there are plenty of people and businesses beyond the four walls that need to deal with the fall-out too. First and foremost are customers. Here the responses are as diverse as the people themselves and the businesses they represent. Industries such as industrial and automotive demand long-term stability of supply, have high quality expectations (especially in safety-critical applications), and may also have demanding just-in-time (JIT [5]) delivery requirements. In many cases, the merger can actually prove to be quite beneficial, especially if the acquirer is better established in a specific market segment (i.e., automotive) than the target of the acquisition was.At this stage, the need for strong leadership and, again, the clarity it offers is reflected in a McKinsey&Company report on deal-making in the semiconductor industry [6]. They estimated that strong involvement of the CEO, another senior leader, or a dedicated “transformation manager” post-merger resulted in an increased success factor of 2.5 over those whose leaders were less involved in the transformation. The process also needs to be swift if the desired value creation from the merger is to be realized. Those that managed the transformation within 2 years were 2.5 times more likely to attain the value (i.e., financial) benefits that they had told investors they were pinning to the deal.
In order to quickly provide clarity, and in many ways protect the existing business that has been purchased, teams are formed that represent both businesses. These develop a list of responses to frequency asked questions (FAQs). To quote John McKenna, who was a Regional Distribution Manager at NSC during the TI/NSC merger, the key here is to “communicate, communicate and, did I mention, communicate?” Under such conditions, rumors are rife and quickly develop a life of their own. Establishing clarity at this stage is, therefore, essential.
This leads to another key player in acquisitions: distributors and catalog suppliers. Semiconductor vendors typically reserve direct sales relationships for those customers that purchase millions of dollars of products every year. For those under this threshold, their product purchasing needs can be fulfilled through distributors. The smallest quantities of part purchases are typically handled by the catalog suppliers — those businesses who used to be known for their weighty tomes that excited engineers-to-be who waited for their annual release with bated breath.
Both these partners not only handle the logistics of supply, they also offload some, if not all, of the technical support. With the high levels of integration being applied to even the simplest silicon devices, which is increasing their complexity, this is not an insignificant task. Both parties undertake significant investment in these relationships from their people and systems, to logistics and training. Each distributor also has its customer base, regional reach (local/global), application focus, and, as a result, a combination of strengths and weaknesses. Semiconductor vendors are therefore selective in choosing distribution partners. Thus, mergers and acquisitions (M&A) are a worrying piece of news for distributors as, inevitably, the risk of losing millions of dollars of revenue is high. One thing is for sure: consolidation in this space will happen and the challenge is managing it as best as possible.
Many others in the semiconductor ecosystem need to be kept in the loop. These range from software and tools partners, to consultants whose businesses may have been built upon the very existence of a particular microcontroller or other silicon solution. These may also benefit or lose out, depending on the decisions made by the acquiring business.
What makes for a good semiconductor acquisition?
The past decade has seen significant M&A activity, with over $310 billion of agreements being signed off, $107.7 billion of which were undertaken in 2015 [7]. Some of these come about because the acquirer wants to benefit from the technology (as well as the customer base) of the target organization. As already highlighted with the Arm and MIPS examples, the development of IP is a significant undertaking, and this is only part of the story. In addition, there are the challenges of keeping fabrication facilities up to date, developing new device packaging, and innovating in specific, emerging application spaces, such as capacitive touch or Internet of Things (IoT). Patent applications are one way to measure technical prowess, and it is perhaps not insignificant that U.S. patent publications peaked in 2007, around eight years prior to the biggest year of semiconductor M&As. According to a paper by B. Shao, K. Asatani, and I. Sakata [8], who compared M&A activity with these patent publications, M&As are a “resource reallocation method [to] integrate different unique resources among companies, no matter [if they are] horizontal or vertical.” M&A deals they could define were broadly split into three areas:- Semiconductor manufacturing process — the acquirer was looking to apply the new manufacturing processes attained to enlarge their business.
- Chips — the acquirer was looking for new device manufacturing process technologies (wire bonding, packaging, substrate).
- Monitors — the acquirer was looking for new technology relating to visual interfaces of smart devices (innovation in relation to pixels, graphics layers, imaging, sensors).
Whether these mergers result in success or not depends, however, on another factor: that of synergies that can be established between the acquirer and the target. Norbert Siedhoff, a Sales Coach and Consultant, explains it as follows: “Regardless of how you review the deal, from the perspective of the business, its customers, distributor relationships, or that of investors, you should be looking for synergies that deliver a 1 + 1 > 2 outcome in every area.” As the Managing Director of Microchip Technology GmbH up until 2019, he got to oversee many M&As, including the purchase of both Atmel (2016) and Microsemi Corporation (2018).
The purchase of Atmel was seen by many in the industry as possibly signaling the end of the 8-bit AVR technology. It was assumed that Microchip, with its well-established range of 8-bit PIC microcontrollers, would sideline the AVR. After all, what sense would there be in maintaining two processor cores (which, as we have already established, are expensive to develop and maintain)? However, with a massive community of “AVR Freaks,” this would have put some of the potential success of the deal at risk, as well as possibly tarnishing the Microchip brand.
The reason why this didn’t happen, and why the announcement was not linked to job loss announcements, possibly lies with Microchip’s strong “Guiding Values” [9]. These include “Customers Are Our Focus” and “Employees Are Our Greatest Strength”. While, purely financially speaking, the maintenance of two proprietary 8-bit cores makes little sense, looking beyond the spreadsheet of profit and loss, it is clear that by maintaining both the PIC and AVR cores, 1 + 1 > 2. Not only did Microchip retain the AVR family, it has actually continued to invest in it [10], satisfying customers of both technologies.
Siedhoff indicates that this approach can be applied across the board. For example, whether a distributor is retained or has to be cut has to be a question of focus, not just finance. “You may have three distributors doing $100m, $90m, and $5m of business for you. If the smallest is highly specialized, that relationship is worth more than the sum of the two businesses that make it up.” This implies that, by cutting that smaller partner, you’ll probably lose the $5m of revenue completely since your larger (financially), but less specialized, distributors won’t be able to take on that business on your behalf. Günter Hohma experienced precisely this “distributor’s worry” in his current role at Richardson RFPD, Inc., as Microchip announced the takeover of Microsemi, one of their semiconductor partners. “Our customers define us as ‘trusted advisors’ in the fields of RF and High Power. We were able to not only retain our Microsemi market share after their merger with Microchip, we were able to expand it with technically appropriate [Microchip] products.”
The staff, of course, are the critical element of any successful M&A, but merging cultures is exceptionally challenging. Here Microchip provides all new employees, both new hires and those acquired through M&As, team-building workshops. The mere thought of these two-week-long sessions at the company’s headquarters in Arizona can result in rolling eyes and reticence. After all, who wants to sit in a windowless room, air-conditioned down to 18°C (64°F) for 14 days? However, Siedhoff highlights that, by the end, all participants emerge enthused and eager, ready to put their new-found skills to work with their new colleagues. Rather than focus on products, these workshops focus on instilling the values of the company, how business is done, and ensure that, especially after a significant M&A, it results in “one company, one sales process, and one face to the customer.”
