Good Vs. Bad Acquisitions

M&A begins to ramp up, but not all of them will work.


M&A activity is beginning to heat up across the semiconductor industry, fueled by high market caps, low interest rates, and a slew of startups with innovative technology and limited market reach.

Some of these deals are gigantic, such as the pending acquisition of Arm by Nvidia, and the proposed purchase of Maxim Integrated by Analog Devices. Others are more modest, such as Arteris IP’s planned purchase of Magillem, which was announced this week. If approved by regulatory agencies, these and other deals will reshape the semiconductor industry for years to come.

Underlying all of this activity are four fundamental shifts in the industry:

  • Rapidly increasing amounts of data, including more images and video, are overwhelming the infrastructure for processing that data remotely. The solution is more processing closer to the source, and that has set off a scramble to stake a claim somewhere in the edge, which is vaguely defined as the space between the end point and the cloud.
  • The difficulty and expense in scaling at the most advanced process nodes is forcing chipmakers to unbundle SoCs into separate chips. While critical digital logic will still continue to scale, less-critical digital logic and analog circuitry can be developed at older nodes and packaged together using some advanced packaging schemes.
  • The growing reliance on advanced semiconductors in safety-critical systems, such as automotive, robotics and avionics, and the expectation those chips need to last for a decade or more, means that chips need to be monitored and updated throughout their extended lifetimes to ensure reliability. That requires a whole new infrastructure and the application of new and existing technologies.
  • The rollout of AI/ML everywhere has created a frenzy of activity around highly specialized accelerators, which can improve performance by orders of magnitude if they are optimized for specific tasks and algorithms.

While most companies can build these technologies internally, it’s much quicker to buy a company that already has them. That solves the problem of finding the right talent, which is in short supply globally, and of building a cohesive team, which can take months or years. Nevertheless, this is harder than it sounds, and not all acquisitions go smoothly.

Measuring success
One of the first priorities in any acquisition is establishing realistic goals and metrics for achieving those goals.

“Making acquisitions is one thing, but it’s the result afterwards is that that really matters,” said K. Charles Janac, chairman and CEO of Arteris IP. “If your company culture is very different from any other company culture, that makes it much more difficult. We’ve seen that with some acquisitions that have not been very profitable, and the company may end up divesting.”

Arteris IP’s pending acquisition of Magillem Design Services is aimed at expanding its footprint beyond the network on chip into what Janac calls SoC assembly. Magillem can package IP in an SoC into an IP-XACT format, which paves the way for uniform communication in an SoC or an advanced package.

“Magillem does RTL partitioning, so you can get the RTL ready for physical design,” he said. “The goal here is to follow the customer. So you grab innovation where you can find it and follow the customer. That increases the survivability of everyone. It also moves us closer to our goal of becoming one of the top 10 IP companies, which we expect to achieve over the next year.”

Expanding a company’s market opportunity is an effective way of achieving growth. “There are two trends here,” said Simon Rance, head of marketing at ClioSoft. “The more common one is where electronic design companies acquire electronic software companies to expand their ecosystem, control, and security across their electronic platforms. For example, it’s much easier to guarantee security if you control the design end-to-end. The second trend is that big electronic software companies acquire electronic design companies to unleash the power of their software. Most of today’s applications are software-driven and manage the decision-making. However, they are limited by their underlying electronic hardware design. Owning the electronic design will allow them to ensure they have enough compute power (microprocessing, graphic/video processing, quantum processing, ML processing, etc.) to unleash their software.”

EDA and IP companies have made hundreds of successful acquisitions over the past couple of decades. But not all acquisitions work as expected for a variety of reasons, ranging from cultural incompatibilities to market changes. Case in point: Intel bought McAfee for $7.7 billion in 2011 to bolster its security, only to spin it out six years later in a deal valued at $4.2 billion. Intel also bought Wind River for $884 million in 2009 for its real-time operating systems, and spun it out in 2018 for an undisclosed amount.

“Acquisitions work because the people make them work,” said Simon Segars, CEO of Arm. “I don’t think anybody who’s sensible would ring up their favorite investment bank and say, ‘I’ve got X dollars to spend and I’m interested in this space, so bring me some targets.’ You can only acquire companies if you have strong relationships. The technology business is about people and relationships and aligning on a vision of the future. You don’t just do that from cold-calling someone and asking, ‘Hey, can I buy your company?'”

Defining success
Ultimately, success is relative to the goals of the company making the acquisition. Some acquisitions are done simply to eliminate competition or to acquire more engineering talent. Others are aimed at providing synergies for tapping into changes in existing markets.

“With the recent ADI/Maxim deal, following from a few years ago the ADI-Linear Technology deal, ADI now has many engineers that used to work at competitors all under one company,” said Carl Moore, yield management specialist at YieldHUB. “I would expect to see more system-level companies being integrated, too. For example, Maxim just acquired Trinamic, and Allegro Microsystems just acquired Voxtel. As chips and systems become more complex, the expertise on the apps and system design side becomes more valuable. This type of merger gives company a nice advantage as they can be the experts in the chip design as well as the system design.  If you understand the system better, you can design a better chip.”

Still other acquisitions are viewed as a shortcut into hot new markets, which may or may not pan out. It’s faster to buy than to build, even if it doesn’t always work out perfectly.

“It takes you three to four years to develop a product, and it takes you another three to four years to mature the market to develop the market for it,” said Arteris IP’s Janac. “You’re looking at six to seven years of development, and takes you another five or seven years to develop a royalty stream. So it’s much more efficient for a large company to buy technology than it is to develop it from scratch and a lot less risky.”

AI/ML provided something of a perfect storm here. The market exploded faster than any anticipated, and it continued to change as algorithms and hardware were constantly changed and updated.

“Big companies are trying to play catch up in the areas of AI and machine learning, and they will soon realize these areas are expanding too quickly,” said ClioSoft’s Rance. “They will need to ramp up the expertise in-house sooner than later to lead and differentiate in these markets. This, in turn, will drive an increase in acquisitions of smaller companies pioneering new technologies around AI, ML, and quantum computing. Some of the big companies are developing quantum computers, such as Google and IBM, but a lot more smaller companies with strong investment already are proving great progress. As quantum computers can drastically help in the development of new breakthroughs in science, medications and other areas such as machine learning to diagnose illnesses sooner, we will see a lot more activity around companies expanding in the area of quantum computing. Most big companies do not have in-house expertise in quantum computing, and we will soon see these startups and smaller companies acquired and technology integrated.”

Acquisitions also can be long-term, strategic investments for companies.

“Another area that could become more interesting in future is the high-end computing and cloud services merging with the system level semiconductor companies,” said yieldHUB’s Moore. “The system-level high-end computing as a foundation can be critical to the hardware and interface chips as we get more and more into massive databases and computational power. Sensors are becoming more complex, as well, and require a lot of calibrations and digital on the front end. You’ll see more mergers of companies that make sensors with high performance analog and digital ICs. The expertise in a sensor design is much different than a digital chip, so mergers like this leverage the overall product capabilities.”

Considerations before making acquisitions
As with most things in the high-tech world, up-front planning is critical. Due diligence in this industry often involves a working relationship that lasts for years before a deal is ultimately struck. That was true for Nvidia-Arm, ADI-Maxim Integrated, and Arteris IP-Magillem. But it’s also just a starting point.

“It’s always useful to look through three dimensions,” said Aart de Geus, chairman and co-CEO of Synopsys. “One is adjacency. That could be technical adjacency, customer adjacency, or channel adjacency. When you do an acquisition, if you can say, ‘Yes, they’re high adjacency,’ then the amount of risk is way, way lower. If we buy a shoe company, there is no technical or channel adjacency, and our customers would not come to us for that. It would broaden our TAM (total available market). There are acquisitions that broaden your opportunities, and acquisitions that are very positive about multiplying what you already do. We’ve done both. We never bought a shoe company, of course, but we did buy Coverity, which was in the business of reducing risk in software. From an adjacency point of view, the first thing I looked at was the technology, and it was very similar. We use similar algorithms, and have used them for over 10 years. From a channel and customer point of view, about half of their customers were companies we knew already because they embedded software. But the other half were people we never would have talked to. They’re in oil exploration, in finance, and in health. That was a substantial broadening of our potential TAM.”

There also is a question of how tightly to integrate acquired companies. This requires a balancing act, which not all companies are prepared for. It depends upon the acquiring company, whether that company has a history of working well with startups, and how much overlap there is between the teams in each company.

“I’ve learned that lesson multiple times, said de Geus. “In one case I sort of bought into the argument of the acquired company that it had its own brand and it would be more efficient to leave it alone. And then they started to behave as if they had no responsibility whatsoever for the success of the overall entity. They had their own business cards and they could never speak for the overall company. It became completely clear that the existing company needed to go away in a very short amount of time because we needed a common set of values and behavior. When you acquire a company, things change, and change is never simple. But we also need to see it as a great opportunity to question the heck out of what we’re doing and to add new DNA that sees the world a little big differently.”

There is a range of opinions about this, and what works for one company doesn’t necessarily work for another. “The best acquisitions are one of two types,” said Wally Rhines, CEO Emeritus at Mentor, a Siemens Business. “One is an area that is complementary, but does not compete with anything you have. You leave that running with its culture and with a high degree of independence. The other thing that is very successful is when the acquiree becomes the acquirer. So you have a losing business and you acquire a winner and take your business and put it under the acquiree, then that tends to work well, too. And if you have overlap, take the overlapping piece and put it under the acquiree. You cannot stall yourself with turf wars over why one technology is better or who’s in charge. If you decide that it’s a bolt-on, where it’s just going to complement your business, and if that’s the understanding with the people you acquired, that’s different.”

Put simply, the best acquisitions are ones where both parties gain something and are knowledgeable about the other’s intentions. “One of the reasons EDA companies can succeed is that startups in EDA get to a certain size and then they’re limited by their distribution, bandwidth, marketing, sales and support,” said Rhines. “They know that to grow larger, they’ve got to hire salespeople around the world. And then you have a whole bunch of engineers who don’t make a lot in the acquisition but who are wedded to the product, and now they have a sales force of 1,000 people up pushing their product, and so they’re happy, too. That’s the kind of merger that is very successful.”

Not every acquisition succeeds, even if the goals are clearly defined. Likewise, not every acquisition fails, even if the pieces don’t align well at first. There are many reasons for acquisitions, and not all of them are well thought out.

Due diligence will expose some weaknesses, but not all. After that, it’s a matter of either living with the deal and fixing what’s broken, or trying to integrate a company’s people, processes and IP.

“We all know there are things for sale, and you never quite know the reason until about seven seconds after they cashed the check,” said de Geus. “Then you find out why. I’m overstating it a bit, but there’s some truth to that. Due diligence is meant to unearth surprises before they occur. But while due diligence is important, alignment of DNA beforehand is much more important. If you align on the basis of building trust before you get married, you can handle difficult things in due diligence. If you don’t have the trust, and now you find out they said they were going to do a mega deal — but it wasn’t nearly what they said it would be — it doesn’t feel so great to have entered the marriage. Our emphasis is on how can you get to a level of trust so that due diligence doesn’t uncover any surprises. And by the way, that goes in both directions. It’s the same for us. If we acquire somebody, we don’t want to surprise the heck out of them two days after they become part of the family.”

—Susan Rambo contributed to this report.

Deals That Change The Chip Industry
Nvidia-Arm is just the beginning; more acquisitions are on the horizon.
Israel: Startup Powerhouse
Education, entrepreneurship, and plenty of talent, but no giant companies.
Nvidia To Buy Arm For $40B
Analysis: What’s driving the deal, and what to watch for.
The Next Wave Of Consolidation
Economic conditions are all there. So what’s the holdup?
New Architectures, Much Faster Chips
Massive innovation to drive orders of magnitude improvements in performance.

Leave a Reply

(Note: This name will be displayed publicly)