Aquantia | Interview with its Co-Founder & VP of Technology – Ramin Farjad-Rad
In San Jose (CA), we meet Co-Founder & VP of Technology of Aquantia, Ramin Farjad-Rad. Ramin talks about his story how he came up with the idea and founded Aquantia, how the current business model works, as well as he provides some advice for young entrepreneurs.
Martin: Hi, today we are in the beautiful San Hose at the Aquantia office. Hi, Ramin, who are you and what do you do?
Ramin: I am Ramin Farjad-Radand I am the VP of technology at Aquantia corporation and I together with the team are in charge of developing interesting technologies that apply to the latest networking solution that the customers such as Cisco and so forth use to differentiate themselves in the market place.
BUSINESS MODEL OF AQUANTIA
Martin: What did you do before you started Aquantia?
Ramin: Well starting from beginning I got interested in communication transceivers back over at Stanford. Maybe part of my first summer internship that I was working at Soundlabs and I started working on a 1 Gig transceiver. The target was designing the whole thing in a CMOS technology which back then most people tried to implement these transceivers gigabit, multi gigabit and higher processors like Gully Morson and so forth which is really fast processing, but at the same time people could not really build them in high density to integrate with the microprocessors, big logics and so forth. Part of that work was extended and Sun was interested to fund the research I continued to actually turn it into a 10 gigabit per second kind of transceiver all in CMOS. Pretty much that research was successful, I managed to build that link in all in CMOS. Back then it was a core to max at CMOS which is more than a 10x larger than future sizes that are used today which is 28 or even 49 meter. But pretty much the technology showed that today it can implement up to 10 gigabits in CMOS technology that back then people struggled to build even up to 1 gig.
That was a big achievement in a sense not just, “Hey, I can do 10 Gig in CMOS” but because it opened the door for integration into much bigger chips that were already in CMOS technology or the processors or also the switches that back then let’s say in the early 2000; that was the beginning of internet opening up to the all public and bandwidths and data become like a necessity, people wanted faster and faster not just processor but also switches and routers and so forth. These switches needed to have very high thoughput and with a combination of IOs that was placed in CMOS to get over the switch the switch wasn’t really practical. They have to use multiple switches, build clone network it would call it, network topology or switch topology to be able to expand switch to turn it form n ports to multiple ports, n times or 10 times and so forth. It was expensive, high power and so forth and by being able to integrate this high speed link inside the same chip and it could manage to increase the throughput of the switch, being able to increase it by order of magnitude or even orders of magnitude. That’s why it plays a critical rule to the next step.
Back then in 2000, in fact ‘99 I teamed up with a couple people and tried to turn it into, actually turn this idea or turn this technology into actually a solution that will benefit the industry in terms of marketing. Through this whole process we’ve got challenges to be able to define the right product for the technology and we started a company called Veleo Communication whose main solution product was multi gigabit or even terabit switches. Back then we focused on optical market, sonic switching and so forth because the problem was really— in those days, the long hawl, why you doing to need networks and connecting cities at high speeds together and that was kind of the bottle neck that back then was mostly wired up to do phone networks and telephones at low data rates. Planning to do internet traffic, even forget about videos just normal internet traffic couldn’t really go through these low speed networking switches. So everybody was switching to optical communication, trying to build more and more connections and networks and then there came the need to do high throughput wirings and switching between all these links. And that is why optical communication was so hot back then before colabs later on in 2002-2003.
Martin: So basically what Aquantia is selling is some kind of devices with some switches?
Ramin: No. This was a Veleo Communication. It was my first company that I started back in ‘99. That was the focus of Veleo Communication, building terabit switches for optical market. And then after Veleo and because of the crash of the optical market and all of our customers which were big customers were Lucent and Nortel. At some point we have a combined you can say purchase order literal intents close to 500 million dollars’ worth of those switches but maybe within few months they went away. Nortel put hold on everything, Lucent basically shut down optical divisions and so forth. And all thing went away and expecting to be a multibillion dollar company, it didn’t turn out to be that way.
So we sold the company and the year after Kim and Alveta , two of my partners here in Aquantia, we started the company with the vision to being able to power the next generation of data centers and networking mainly from the connectivity prospective for the upcoming explosion of data that we could see. The data was increasing, not just communication of data but also the storage of the data and so forth, to the extent that the existing solutions, the connectivity solutions in the data centers and the storage centers and the computing centers and so forth was not really enough to answer this.
Back then people were switching from 1 gigabit to 10 gigabit but the solutions for 10 gigabit was all pretty much an optical domain, meaning that you had to use optical transceivers, lasers and so forth and it ended up to be expensive solutions. And knowing that these port counts that were already about like making hundreds of millions if not billions in gigabit, switching them over to 10 gig or optical at price points that are hundreds of dollars was a big show stopper for many companies to scale to that level. Price was a big deal, especially when you are talking about such high switch counts and our mission was to create a technology to be able to implement it over the copper lines and away from using any optical components which was pretty much also the start of the 10 gigabits over copper, over basically twisted pair copper not shielded copper just to make everything cheaper, just to make the cable cheaper. And we of course added significant complexity on the chip side to make the other side cheaper and easier, meaning anything external, even only through connectors, cables and so forth was made to be as simple as possible even the connection of the cable to the port using like RG45 type of connector that just clicks, you can find it behind your computer or laptop that has this network port versus the optical solutions that nobody can just use like that. You have to align the fiber which is like a micrometer thick with the edge of laser to make sure that the light passes through. It takes an expert to do that every time so it wasn’t something that also could be basically expanded to the public or general market unless using a copper and ease of use of it. And pretty much, it puts all the burden on the technology and those required to implement it.
We knew that is a big challenge and from day one we started building a team that we believe were super stars to be able to excel in this market and be able to deliver it. While at the same time when we started we had technology that we believed will put us ahead of everybody else delivering this solution not just all in silicon but also make it in like low power, because once you assume that each server or each rack of servers that you have in data center could have tens of these ports and each rack will have tens of servers and each roll of a data center can have twenty or thirty of them and then in the whole data center there could be tens of thousands or a hundred thousand of these servers. If the power is not fully controlled, it will be like a big heating machine just for connectivity and so forth. And so power is also a significant part of the equation that we started to solve and when we entered the market we introduced ourselves as the low power leaders who can not only solve the connectivity problem but also the power problem for the system.
Martin: So if you would rephrase the value proposition it would basically mean that if you are looking at a data center you will improve the connectivity within the data center meaning that you will lower their energy cost thereby lowering total costs of having operating the data center and thereby transforming the data from one point to the other.
Ramin: Yes, exactly. First lowering the cost of installation, the price they pay to install each port and also lowering the maintenance cost, moving forward in terms of cost of the energy and so forth.
Martin: Can you elaborate a little bit about how defensive is this? Or how easy it is to defend this kind of model from other competitors?
Ramin: One of the reasons we approached this market as a startup knowing that especially if you are in hardware, networking hardware, especially semiconductor business center a lot of big companies are already in this market. We had to definitely find a path that we could differentiate ourselves and one of our biggest differentiations was of course the technology we could introduce into this market. And Tangabeste has been as they use to call it the toughest problem in the communication ICs back then and it turned out to be over yeas. Because when we started there were another three public companies that started in the same field together with us plus also the number of the startup companies, at some point were 7 companies doing the same thing. Over the years, Aquantia was the only company whose product was shipped in Cisco boxes and only one public company managed to get it right after few years after us to be our only competition today so far.
So one thing is to find the technology and to also have the solution, knowing that the technology that you have is differentiating enough that not anybody not so easily can come in and compete with you in that domain. Of course, the patents that protected us in that domain. On top of that I should add as an entrepreneur just having a cool technology doesn’t mean, or like differentiating technology doesn’t mean anything. You have to see the problems that your customers have or are actually going to have. The smart or most successful ones will see the waves that are coming and try to prepare for it. Try to build solutions for the waves that are coming. Once they are near then you can go to your customers and say, “Hey, don’t worry, I have a solution for you.” Helping the customer to win in the market, differentiating yourself and basically pass the big wave and they help you in return, creating a win-win situation in such a way that customers, especially the big ones really feel that they have to use you otherwise there are not many other solutions out there to do so.
Martin: But basically this means that you have some hardware solutions like switches or something for solving the connectivity issues in data centers and when you look at the customers, customer segments, who are they? Is it only like people who are operating the data center or are those companies like Facebook who has its own data center?
Ramin: Our customer’s customers would be the likes of Facebook and so forth. One of the big customers we can mention is Cisco systems. They actually build their network OEMs that build all the big switches and the network equipment.
Martin: And in terms of revenue model is it that you are selling them not only like revenue per item or is it something like a service model or—
Ramin: No, we build chips.
Martin: And you only sell them. You don’t rent them or something like that?
Ramin: No, just sell them the chips.
Martin: And how do you acquire customers. Are you using only a direct sales force or are you using distribution partners?
Ramin: No, actually because we are not in a consumer market, our customers are well defined. For example, there is usually in the space that we work in there is one dominant customer that owns maybe half the market or more. And there is another one who owns like 20 -30 percent and the rest combined will be another 20 something percent together. So we usually try to approach the first top two three because the amount of energy and even marketing and salesforce that is required to approach everyone for a small company is not really that efficient. So our approach is to go and target the big ones, try to understand their pain points and try to understand what are the things that they are working on and what are the problems that they are going to face in the next generation especially and how we can solve that. And convince them that we can actually solve the problem for them and work together with them to deliver the final solution.
We have implemented this model at least twice in the life of the company and that has been the part of the success of Aquantia, a big part of the success in Aquantia that basically helped us maneuver all the hard times and the financial crisis by having strong partners that actually needed our technology so they ensure that we stay and survive and flourish and expand and being able to help them move forward.
Martin: And when you look at the potential customers that you could have and you said they are very concentrated, is it that you would like to sell your solution to any potential customer who would like to buy this or is it more that you go to the top two customers and say, “Guys, if you want to differentiate from the others 30 percent or so then buy with us and we give you some kind of exclusivity”?
Ramin: Depends. Of course as I mentioned, we don’t have bandwidth in the beginning to approach everyone so we usually we work with top two, three big customers in the field and making sure that pipe thing and all the issues that would happen. And one of the things that being in an industry means you pass the quality assurance of the biggest company in the field then all the other companies will just wave any qualification or so forth and they just comfortably come to you and buy from you. So rather than having to deal with twenty companies and try to prove to them that it is good, everything is perfect and so forth working with one company and getting that stamp of approval and other guys will automatically come to you, because they say “If they selected you to be their choice then it is already guaranteed that it passes our standard level”.
Martin: Are you currently having an international customer base or is it mainly US focused?
Ramin: Currently we have also international.
Martin: And you said that you have very few customers. How do you prevent the situation happening like your former company? Because your former company you also had very clamped, very clustered customer portfolio which was risky in that specific time. How do you prevent something like this in the future?
Ramin: One of the things that we have to also be careful is also that when we approached the optical markets that was our only specific solution that we had, it was the only line of product that the company relied on. And that specific solution with a certain optical market going down, we didn’t have any other product to be able to rely on. Even after that we tried to leverage that technology to a different market – from optical we could move into IP switching for example, and the company back then tried to do that. But because this whole process happened so late by the time that we switched and tried to address the same technology, it was a great technology, into a different type of application was too late for us to be able to solve the product and survive on it.
One thing is that what we have done only to guarantee the don’t leverage the technology itonly one application. If that application goes down, then you have no more time to try to leverage it or maneuver into something else. For our application from the beginning we tried to secure at least two markets and that is why even if we had like hiccups is one from time to time we could rely on the other and making sure that it is just like not just one big customer.
Martin: Ramin, walk me through first twelve months of Aquantia. How really was it like if you would have to explain it to somebody transparently and say, “Okay, this is what I was doing by day, this was what we were focusing on”, just vividly?
Ramin: First twelve months I would say depends on when the start was… When we started, when I quit my job at the previous company, we had the promise from our investors that if we can prove to them the ideas we have can be manufactured and can enlarge volume and be profitable and so forth they will invest in us. They had given us very little money for the period of whatever, until we prove this. Those were really challenging times because we had to cut back and just work maybe 70 – 80 hours a week and try to make sure that we reach as fast as possible.
At the same time, it was a fairly big task so we had to get other people that we respect both technically and also have worked with before and synergy was right to get involved. But we couldn’t expect anybody to quit his job or anything so we had to work with them – first excite them about the potential of the technology and the potential of the market that this technology can address and at the same time encourage them to almost as hard as we did after work. Pulling everything together and be able to present that at the end, we can actually deliver this.
Martin: So what was the task – only to develop a prototype of the hardware switch or was it really to have the prototype and convince a customer and only then the investor would invest?
Ramin: It was more of proof of concept, it was not necessarily customers involved. So usually VCs they have their own technical team that is not just necessarily people who work at the VCs, the other high tech entrepreneurs in that specific field that they highly respect and they ask them to come and do the due diligence on us.
So we went through maybe three different day long due diligence by three different groups and finally we got the checkmark and we got the money. That was a good sign but more work started afterwards because building the team, building the whole infrastructure for the company that we can expand, make it scalable enough and so forth. It started from doing technical work to building the actual company for us.
One of the big lessons I’ve learned from my previous company is that technology is very important, we need to know we have something great, but first you have to know what is the right application for it which in this case we found what would be the application. And the second of course is the team. Building the team becomes I would say one of the most important if not the most important factor that is in entrepreneurs control. The other one is the market, but we can’t control the market. Markets go up and down and you have to just be smart about it and how you strategize your path, making sure that you will be safe in times and storms and so forth. But building the team is the most important thing. As a junior entrepreneur in my first company, we didn’t have much experience. We thought that if somebody is smart enough and can answer all our questions just hire him. And later we ran into a lot of issues, efficiency issues. A lot of those people that we hired were really smart but at the same time it was “My way or the highway” – they didn’t want to work with each other.
Martin: OK, so they didn’t cooperate.
Ramin: The cooperation was not really great and that inefficiency basically took a big toll on us in the first company that we had. Second time being through that we began making sure that they are not just good technically, sometimes even rejected some people who were really smart, really sharp, intelligent people, genius people just because we felt that the attitude does not really match with the rest of the team that we had.
And to be honest maybe the first 50 people that we hired were the people we either have worked with before or one of the people that we had hired and worked with well enough to know he/she is not just great from technology and expertise but also great from personality and making sure that the synergy is perfect.
Martin: And you minimize the risk because at least either you or your friend know that this guy is working.
Ramin: Exactly. That was very important this goes by far more than getting reference checks that is limited to a certain extent. But working with somebody is like going on a trip with somebody friend you know much better seeing him on and off. Being a roommate with someone and you even know that person better so working with somebody is like being a roommate for year. Usually in companies you spend more time than you spend at home and the level of interaction is significant.
That was also a key part of our selection for the team and some of the people that we hired honestly are key members, technical members that have been with us from the past ten years, ever since we hired them. And they have been with the company ever since.
ADVICE TO ENTREPRENEURS FROM RAMIN FARJAD-RAD
Martin: Ramin, let’s go back 10 – 15 years or so of your time and really imagine: what would be great if you have known this before you start doing any company? So if you go back and say “Oh, if somebody would have told me this kind of lesson this would have helped me a lot.”
Ramin: My first company Veleo thought me many, many lessons that we tried to implement as much as possible in this company. At the same time one of the things in addition to the fact that you have to be extra careful once you get the money to build the right team because the team is the most important asset that is in your control. The most important factor for your success that is in your control and the technology can actually change even the applications they use for the technology can change as you move forward.
Believe it or not as I mentioned in the first six months of the company before our funding we were supposed to prove that the technology we are so excited about is the best thing since white bread and a year after we dumped the whole thing because we found something better. We didn’t stay as “This is our technology, we are going to push for it to make it happen.”
Entrepreneurs really have to be flexible. They shouldn’t be attached to original ideas if they feel there are other alternatives. They have to be open minded to any idea and really question everything and even it is your baby technology always have to question it can it be better, can it do anything better, what is out there that can beat this and so forth. That is very important not to get lost or drinking you are cool; ain’t thinking you are the best.
That is I would say one of the most important things and the other thing I should mention is the product and the application that use the technology. You have to make sure it is differentiating enough for all the big companies, because they will be very reluctant to pick somebody with a small company like a 50-man team to deliver a critical technology to them. It becomes so risky. If you put yourself in the shoes of some executive VP making a decision in let’s say at Cisco that they want to use their technology. And if Aquantia fails, the executive VP will be in serious trouble. But if give it to some joint company like a multibillion dollar public company that god knows how many people working for them and has been around for like 30 years. If they failed that company fails and the executive VP fails. There is a lot of pressure in the decision making in the big company to go with you. So they have to have very convincing reason in a way that you have to create great enough differentiation for them so they say “If I don’t go with them there is nothing out there and then my competition may pick them and then they may get ahead.” So they take the risk for the right reward. And you have to continue making that differentiation sustain because you start going up on a path to say “I am different from everyone” and then as soon as the big guys notice that you are up there they start development right away because they don’t want to fall behind and you have to continue that and continuously offer the differentiation that these big customers feel like they have to stick with you and never go back and try to consider the alternative.
One of the things that a lot of companies fall into is the fact that after some point and after they feel they have a solution or a technology, then the other guys catch up, they try to add I would say maybe change the prices so they offer you lower prices and to add a little bit features here and there to make it attractive. And if you go on and see if the features that are added are not something significant, then the other guys can also add it. So this competition, competition on the small steps rather than differentiating actually causes everybody to be the same. They say I have added this feature that do this extra thing and if it is easy other guys will add it. It becomes more and more complex product that everybody has. So differentiation actually becomes nothing.
But the differentiation has to be in such way that it is like great leaps that gets others by surprise and before they can catch up to it you already got the next marker. And this takes agile and I would say knowledgeable team with enough experience both on the technology and the business to know what is the next great technology and what are the problems that are coming towards my big customers. Even we have to know better than them what the problems are and try to solve them for the next wave. That is how we survived and grew.
Martin: Ramin, thank you so much for you time.
Ramin: Thank you.
Martin: It was a pleasure. Next time you are thinking about starting a company you need to be very crystal clear about your value proposition so that you make sure that you differentiate it from all the other people and don’t be better only by 10 percent but maybe like 10x. Thanks.