Amin Shokrollahi | Founder and CEO | Kandou
Amin Shokrollahi is CEO and founder of Kandou from Lausanne, Switzerland, one of our newest members and a company developing high-speed, energy efficient, chip-to-chip link solutions. In other words, Kandou keeps consumers connected.
A featured speaker at the recent GSA Entrepreneurial Conference, Amin detailed his multi-cultural background and his move from academia to industry and entrepreneurship. He furthered his understanding of business and the electronics/semiconductor industry at Digital Fountain, now part of Qualcomm, where he worked to commercialize Raptor Codes, a method of transmitting data over massively impaired network connections. He continues to make his mark at Kandou and the soon-to-be-announced Matterhorn.
As Amin concluded his talk, he affirmed Kandou’s goal to be a Swiss and European success story. After reading the Q&A profile that follows, it’s easy to imagine Kandou as an international success story. Join us as we explore Matterhorn’s differentiation, Amin’s focus on getting Matterhorn to market, industry trends and advice for startup founders.
What is your background?
I have an extensive academic background. While I’m steeped today in the electronics industry, I’m a mathematician by training. I was born in Iran and moved to Germany when I was a teen. I studied Mathematics at Karlsruhe Institute of Technology (KIT), a university known for producing a host of noted inventors and entrepreneurs. I went on to study at the University of Bonn where I received my PhD in Mathematics and Computer Science. The University of California at Berkeley was my next stop where I did a post doctorate. Today, I am a professor of Mathematics and Computer Science at EPFL in addition to serving as Kandou’s CEO.
In your presentation, you said Kandou was born out of curiosity.
Yes, though the curiosity was purely coincidental, as often happens with breakthroughs. I stumbled on a problem that I didn’t know about because it wasn’t in my field. I took a good look at the problem of transmitting data on electrical wires with a fresh perspective and solved it with methods from my field.
I knew I had a solution that none of the information in literature from the electronics industry came close to. We solved a problem that allowed for a solution –– an abstract of an abstraction, if you will. Like everything, if you don’t step back and take a bird’s eye view, you may miss the answer. While I didn’t know electronics, I knew mathematics would make the difference and it did.
My ambition became making this solution part of every electronics product because we enable smaller, more energy-efficient and cost-effective electronic devices.
What problem is Kandou solving?
Matterhorn is a chip that will offer faster connectivity between electronic devices with lower power consumption and greater bandwidth, two recurring consumer complaints. It’s not two devices, but many connected together. This is not a wireless chip for 5G networks. Instead, it’s part of a hardwired cable connecting devices.
More specifically, Matterhorn will make the new connectivity standard USB4 a reality. We’re the only company other than Intel offering a USB4 retimer solution. Matterhorn meets all the figures of merit, according to the USB-IF standards body.
I compare the challenge to solve the problem to scaling a mountain, which is why we call our USB4 chip “Matterhorn.”
This is not Kandou’s first product.
That’s right. Our first is called Glasswing, a SerDes high-speed communications IP. Its chip-to-chip link technology is based on our CNRZ-5 Chord signalling architecture, delivering low-power, high-bandwidth signalling. It is an ultra-short reach (USR) SerDes IP proven in silicon and capable of providing 1Tbps bandwidth at under one watt to advance fundamental architectural shifts in power savings essential for data centers, as well as wireless, aerospace, and satellite communications.
In one example, a company implemented Glasswing in its chip design, shortening development of its products by at least a year and reducing the cost of manufacturing the product as well. Its project group split the chip into smaller chips and stitched them together. This technique reduced development time and increased the yield of the semiconductor wafer as well as cost. It also enabled the group to create a family of six different products while only designing two unique chips just by changing the configuration of different quantities of the chips in a shared package.
What’s next for Kandou?
Kandou’s link technologies can reduce power in systems and electronic devices by 50% or more. That will offer more robust user experiences such as higher definition displays, longer battery life, lower overall system power consumption, reduced manufacturing costs and better device connectivity.
Consumers can expect to see new types of electronics or semiconductor products with “Kandou Inside.” They could range from super high-res cameras and DNA sequencing methodologies orders of magnitude faster than what’s currently available today. Augmented reality imaging could be greatly enhanced with our chip so users won’t feel a difference in what’s real and what’s not. Today, AR equipment is lower quality than what processor frames will be able to do in the future.
Editor’s note: Amin switches topics here to discuss industry trends and the long-term effect Covid may have on the semiconductor industry. He concluded with advice for entrepreneurs.
What market trends are you seeing?
Covid created unprecedented challenges with fab capacity as demand for laptops and other consumer devices grew to levels fabs couldn’t meet. It affected all companies along the electronic products and semiconductor supply chain, including us, and shows us that we cannot rely solely on fabs. Instead, we need to plan ahead and be prepared for future constraints. That means home computer and consumer devices will need to be architected with different systems requirements such as chiplets with minor modifications per application.
Another trend is AI and ML chips. News that another AI or ML company closed another round of funding dominates technology news daily. All those AI and ML chips need accelerators that can be used for different applications, a good opportunity for companies like Kandou.
What is driving these trends?
Work from home drives several trends and we need to expect that it’s here to stay. Executives were forced to think differently to see the cost benefits of working from home. While there was concern about productivity, we were forced to give it a try. Now, we see productivity increases and, because in fact, there’s no longer a clear demarcation between work and personal life.
Did Covid affect product development or delivery?
Not in any significant way and it didn’t affect our product development schedule.
We are used to working from home and are not constrained by geography. Kandou is nimble and structured itself for a distributed remote workforce from our startup days when we set up processes and the software to make it happen. From the start of the pandemic, our software partners were supportive and made sure we had access to servers and whatever else was required to keep us productive and meeting our schedule.
The only shift we needed to consider was our lab to ensure we complied with prescribed protocols. Lab technicians work closely together in a less-than-perfectly ventilated area. It was not a scramble. We closed the lab for a bit and the technicians took the machines home with them.
All in all, we’re much further along today than I thought we would be 10 years ago.
What advice would you give an entrepreneur just starting out?
I will teasingly tell anyone considering starting a company, “Don’t do it!” Any entrepreneur needs to be really convinced starting a company is the right thing to do. It cannot be in a half-hearted way, either. If the answer is yes, then they need to be able to put up with plenty of frustration, especially starting from zero. It can be a slough and it’s hard to anticipate what’s coming, such as Covid. However, it’s exciting. Technology, especially networking technology, is difficult. The market is burdened with overhead and different than I expected. In my case, ignorance was bliss.