- Leading industrial robotics giant Fanuc (TSE: 6954) recently acquired Life Robotics with hopes to expand its addressable market.
- The acquisition is uncharacteristic of an internally-focused Fanuc, but so are several of the company’s other strategic initiatives.
- With Fanuc’s openness to partnerships with leaders in other domains, the next 20 years of Fanuc’s developments is unlikely to resemble the last 20.
Fanuc (TSE: 6954) announced its acquisition of Life Robotics, a robot maker specializing in collaborative robots (or “cobots”), last week. Though Fanuc didn’t give out any details on acquisition cost, with Fanuc’s 380 billion yen ($3.55 billion USD) net cash position and Life Robotics’ 1.5 billion yen ($14 million USD) in capital stock (including capital reserves), the costs are presumably a drop in the bucket.
Those familiar with Fanuc would know that the acquisition is quite uncharacteristic of the company. In fact, this is the first acquisition for Fanuc in over 15 years.
However, this is merely the latest of a number of strategic moves that Fanuc has made over the past decade to expand its addressable market.
To understand where Fanuc’s going, I think it’s helpful to review the historical context that leads up to where we are today.
The company’s core strengths today were developed within the secluded “Fanuc Forest” at the bottom of Mount Fuji, where the company’s headquarters lie. Historically, Fanuc has remained mostly silent toward investors – even shutting down its corporate website back in 2011. This closed stance changed shortly after activist investor Daniel Loeb of Third Point Capital announced his stake in the company back in 2015, though Fanuc CEO Yoshiharu Inaba insists that Loeb’s encouragement had little to do with the change. Since then, Fanuc set up a Shareholder Relations division, which has done a pretty good job at keeping an open dialogue with investors.
Source: Fanuc (bird’s eye view of Fanuc HQ)
Misperceptions as a technological leader
It’s easy to categorize this mysterious robotics company as a technology leader as many have, but that isn’t quite accurate with Fanuc. I wrote about this over on Seeking Alpha last July, and I will summarize a bit of the takeaways here (since the article is behind a paywall now).
Nikkei BP published an interesting series on the inner workings of Fanuc back in 2015. Based on this series, one of my key takeaways was that part of Fanuc’s industry leading operating margins is a result of the company’s focus on standardization. In developing new products, Fanuc uses proven and reliable existing technologies over new technology. The rationale behind this is part standardization. Using the same parts for different robots means there is no need to open up a new production line for new parts, thereby conserving resources.
The two journalists who wrote the series talked about a couple of the conversations they had with Fanuc customers. What they discovered was that Fanuc doesn’t necessarily lead the industry as far as technology goes, but when other factors like reliability, quality, and price come into the mix, Fanuc’s value proposition is generally more attractive than its competitors’.
Fanuc isn’t secretive, it’s just hyper-focused
Some of Fanuc’s recent developments, which seemed out of line with the secretive image of Fanuc, are:
- Development of open platform FIELD system.
- Partnership with Preferred Networks (and others).
- Acquisition of Life Robotics.
A little bit about each one of the developments.
The FIELD system is Fanuc’s take on making factories smarter. It’s a system that connects every device inside a factory and manages the factory as a whole. This is in line with Fanuc’s pursuit of Zero Down Time (ZDT).
So for example, let’s say we have a robotic arm on our widget production line. Sure, this robotic arm has a regular maintenance schedule – we check it every 3 months. On occasion, however, the “elbow” part of this arm doesn’t move smoothly and causes a malfunction. We take the arm off the production line and now our production capacity for this specific process is reduced. Now some of the robots in the downstream processes have nothing to do because our broken robotic arm isn’t feeding any inventory. As you might imagine, this can get pretty expensive. When you spend big money on countless robots, you don’t want them sitting idle.
What the FIELD system would do, in this case, is report every nuance of a device. Sensors will be installed on each device so managers can quickly run a health check on them. When the elbow of the robotic arm encounters higher level of friction than normal, the sensors detect this and feeds the information to the FIELD system, which then alerts managers about the abnormality before the robotic arm actually malfunctions. Managers can then schedule maintenance when production is least impacted, or plan to temporarily replace the robotic arm with another. The downstream robots are kept busy as usual, factory improves operating efficiency, the company can afford to comfortably reduce its widget prices, and everybody is happy.
This FIELD system is a joint project with Fanuc, Cisco Systems (NASDAQ: CSCO), Rockwell Automation (NYSE: ROK), and Preferred Networks (private) at the core.
Preferred Networks is the Artificial Intelligence/Deep learning partner for Fanuc’s FIELD initiative. For those who have never heard of Preferred Networks, it is a small Japanese company that has become a big deal, particularly over the last few years. The company consists of about 100 people (mostly engineers) and refuses to take contract work, even from giants like Toyota Motors (TSE: 7203) and Fanuc. It only accepts partnerships as equals.
Now, I tend to get a little sheep-ish when it comes to news articles related to artificial intelligence, deep learning, and unidentified flying objects. To be clear, I’m not saying artificial intelligence and deep learning are nonsense, I just have a more ground level view on the technology compared to most, and I think Preferred Networks might share a similar view, as they seem more objectively minded than many.
A 2018 Nikkei series about Preferred Networks offered some insight on how the Fanuc partnership happened.
In early 2015, Preferred Networks CEO Toru Nishikawa visited Fanuc’s fully automated production line where robots were building robots. Fanuc CEO Yoshiharu Inaba points at a group of 5 robots in the production process and tells Toru:
“I’m wondering if we can figure out a way for 4 of those robots to cover for 1 robot in case there is a breakdown.”
That’s when Toru saw the opportunity for AI in an industrial setting. On the way back to Tokyo, Toru commented to his colleagues about quitting everything they were doing to focus on Fanuc.
At the time, Fanuc was already facing a big problem with its smart factory initiative: Processing big data.
The various sensors attached to each machine produces an enormous amount of data. That’s when Preferred Networks came in and suggested edge computing. Contrary to the popular Internet-of-Things (IoT) movement, where everybody is connecting everything to the cloud, Preferred Networks wanted to keep much of data processing local. Fanuc’s R&D Vice Manager Takayuki Tamai recalls that Preferred Networks was the only company he knew of that pushed for edge computing.
This was good enough for Fanuc, and Preferred Networks soon became the AI partner in the FIELD project.
Life Robotics is a university research project that turned into a venture. The company is famous for its “elbow-less” collaborative robot “CORO”, which takes up less space compared to robots with elbows.
CORO, with its elbowless, patented Transpander Technology, serves many purposes like part picking, loading/unloading, packing, etc. The small space requirements, flexibility, and safety of CORO has already opened up new opportunities. For example, Yoshinoya Holdings (TSE: 9861), a national restaurant chain specializing in beef rice bowls, started using CORO in its dishwashing process last year.
I think Fanuc’s acquisition of this small company says a lot about the future of industrial robotics and smart factories.
What does it mean to have smaller, collaborative robots with comparable functionality?
One thing most people who have never visited a factory with robots probably wouldn’t know is that many industrial robots are bolted down and inside a cage. This is mainly for safety reasons. Collaborative robots, on the other hand, often aren’t required to be bolted down or in a cage.
Partly because industrial robots consume space, energy, and historically have not been very flexible (bolted down, inside a cage, high technical requirements), it’s mostly been the larger companies that enjoyed the full extent of the benefits that robots have to offer. Now, with collaborative, smaller robot applications like CORO in the case of Yoshinoya’s dishwashing process, robots have a place in smaller settings.
According to Infinium Global Research, the global collaborative robot market is expected to grow at a 56.9% CAGR pace into a $4.28 billion dollar market by 2023. Meanwhile, the overall global industrial robotics market (inclusive of collaborative robots) is expected to grow at a 9.6% CAGR pace into a $71.7 billion dollar market by 2023.
This is hard to ignore for Fanuc. To be sure, Fanuc already has a collaborative robot offering (since 2015). Though most of Fanuc’s robots are painted yellow, the collaborative robots are painted green. The technical hurdle has come down considerably too, with functions like robot hand guidance which lets the operator manually show the robot what to do instead of writing code.
The bottom line
Putting all of this together, Fanuc drastically shifted its internally focused management style into far more open one in a few short years. This means the next 20 years of Fanuc’s development will have little resemblance to the last 20. Smart factories are a thing, so is artificial intelligence, and collaborative robots too. Despite Fanuc’s scale and currently strong competitive position in industrial robotics, the company realizes the need to open up its doors to similarly competitive companies in these new and unfamiliar domains.