In this interview from theCUBE + NYSE Wired: AI Factories - Data Centers of the Future, Faraj Aalaei, founder and chief executive officer of Cognichip, joins theCUBE + NYSE Wired's Gemma Allen to discuss how artificial chip intelligence is set to collapse the cost and time of semiconductor design. Aalaei draws on decades of experience building and taking public two chip companies to explain why the industry is stuck — with complex devices taking five or more years to reach production and costing hundreds of millions of dollars even at medium complexity. He introduces Cognichip's core innovation, Artificial Chip Intelligence (ACI), a purpose-trained AI system designed to compress that design cycle by a factor of 10 and enable bespoke chips tailored to specific workloads, from self-driving vehicles to supercomputing.
The conversation also explores how Cognichip's platform blends proprietary and open-source AI, uniting specialists in AI science, chip design and software into a single integrated product team. Aalaei underscores the severity of the industry's contraction, noting that annual pure-play semiconductor startup launches have fallen from roughly 200 to single digits since 2000. He reveals that more than 30 of the world's largest chip companies are already in active trials of the system, with early results exceeding expectations. He also details the company's $93 million raised to date — a seed round from Lux Capital and Mayfield, followed by a $60 million Series A led by Seligman Ventures and SBI — to accelerate commercialization. From restoring the conditions that once let a four-person team build a chip company to giving scientists with no hardware expertise a path to custom silicon, Aalaei outlines why Cognichip is positioned to fundamentally reshape how the next generation of chips gets built.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
theCUBE + NYSE Wired: The AI Factory - Data Center of the Future. If you don’t think you received an email check your
spam folder.
Sign in to AI Factories - Data Centers of the Future.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Register for AI Factories - Data Centers of the Future
Please fill out the information below. You will receive an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for AI Factories - Data Centers of the Future.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Select your Interests!
add
Upload your photo
Uploading..
OR
Connect via Twitter
Connect via Linkedin
EDIT PASSWORD
Share
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
theCUBE + NYSE Wired: The AI Factory - Data Center of the Future. If you don’t think you received an email check your
spam folder.
Sign in to AI Factories - Data Centers of the Future.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open the link to automatically sign into the site.
Sign in to gain access to theCUBE + NYSE Wired: The AI Factory - Data Center of the Future
Please sign in with LinkedIn to continue to theCUBE + NYSE Wired: The AI Factory - Data Center of the Future. Signing in with LinkedIn ensures a professional environment.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
Faraj Aalaei, Cognichip
In this interview from theCUBE + NYSE Wired: AI Factories - Data Centers of the Future, Faraj Aalaei, founder and chief executive officer of Cognichip, joins theCUBE + NYSE Wired's Gemma Allen to discuss how artificial chip intelligence is set to collapse the cost and time of semiconductor design. Aalaei draws on decades of experience building and taking public two chip companies to explain why the industry is stuck — with complex devices taking five or more years to reach production and costing hundreds of millions of dollars even at medium complexity. He introduces Cognichip's core innovation, Artificial Chip Intelligence (ACI), a purpose-trained AI system designed to compress that design cycle by a factor of 10 and enable bespoke chips tailored to specific workloads, from self-driving vehicles to supercomputing.
The conversation also explores how Cognichip's platform blends proprietary and open-source AI, uniting specialists in AI science, chip design and software into a single integrated product team. Aalaei underscores the severity of the industry's contraction, noting that annual pure-play semiconductor startup launches have fallen from roughly 200 to single digits since 2000. He reveals that more than 30 of the world's largest chip companies are already in active trials of the system, with early results exceeding expectations. He also details the company's $93 million raised to date — a seed round from Lux Capital and Mayfield, followed by a $60 million Series A led by Seligman Ventures and SBI — to accelerate commercialization. From restoring the conditions that once let a four-person team build a chip company to giving scientists with no hardware expertise a path to custom silicon, Aalaei outlines why Cognichip is positioned to fundamentally reshape how the next generation of chips gets built.
>> Palo Alto studio connecting Silicon Valley and Wall Street. I'm John Furrier, host of theCUBE, here with Dave Vallante, my co-host.
Gemma Allen
>> Welcome back to theCUBE Studio here at the New York Stock Exchange. I'm Gemma Allen with NYSE Wired's AI Factories and joining me now for a conversation on artificial chip intelligence is Faraj Aalaei, founder and CEO of Cognichip. Welcome Faraj.
Faraj Aalaei
>> Oh, thank you. Thanks for having me.
Gemma Allen
>> So you've had a long and tenured career in this space. You've been in tech a long time, but a lot is changing in this moment, which is some of the catalysts behind founding Cognichip two years ago. Talk to me a little bit about Cognichip, the journey, and what you guys do.
Faraj Aalaei
>> Sure. And that long time, part of it was very painful, of course, building two semiconductor companies and taking them out public. Along the path, obviously the industry was kind of getting more and more complex, but the fundamental problem all those years were really that increasingly, it was just taking longer and longer to be able to design a chip and the cost of designing those chips was just skyrocketing, which incidentally it led to really a decline in how venture capitalists were investing in startups. The number of startups started to go down. It was very unique in the late, let's say 2010s and beyond to actually get funding for semiconductor companies. And again, cost and time had a lot to do with that. And so I was building these companies through that process, through that transition. And then when it came to Cognichip, it was really kind of looking back at all of those experiences and the issues related to chip design in our industry and where our industry was heading and then looking at AI as probably the discontinuity in technology that would allow us to solve some of those problems. So I started to think about, well, what if? What if I could take generative AI and direct it to solving the problems I had been struggling with in building the last two companies? And that's how Cognichip was worn.
Gemma Allen
>> Wow. And we know that no two chips are the same. This is a very an ever growing portfolio, I guess you could say, of chips and their requirements, GPUs for certain things, CPUs for other things, LPUs for other things. Talk a little bit about how you kind of segment this industry and how you think about, I guess, is there low hanging fruit from the design process? Are there challenges that are easier to solve first?
Faraj Aalaei
>> Yeah. So part and parcel of the increasing costs and the complexity of these devices really led the industry to say, we're going to build a lot of really big general purpose devices because we can't do too many of them so we might as well do one that has everything in it. Well, if you build one that has everything in it, it becomes also very inefficient. While it becomes more complex, more risk, but it also becomes inefficient in terms of power and in terms of cost. Where software industry is going through the roof from zero to 100 million users in matter of weeks, it takes us five years to build a chip and bring it to production and that's in a good day. So think about it. When ChatGPT, for example, was announced as a product, the chips that that was running on were probably five, six years old from beginning. They started designing those chips five, six years ago, so they couldn't have comprehended ChatGPT as a workload. And this difference between how fast software can scale and how fast chip can scale creates a major problem in our industry. The hardware industry is far behind beyond where the software industry is and we need to solve that problem. The way you solve that problem is by collapsing the time and collapsing the complexity of these devices, and AI can help you do that. We invented ACI to be able to really look at instead of AGI, which is artificial general intelligence, we focus on artificial chip intelligence, training models that can actually design a chip for you a lot faster and a lot cheaper and get it out to the market. Now, if you do that in process, what you're enabling the industry to do is to build bespoke devices. So if I'm running, for example, if I'm doing a device for a self-driving car, it may be different than if I'm building a supercomputer device and now I can actually build those devices cost efficiently, bring them to market as fast as we can imagine applications being formed in the software side of things.
Gemma Allen
>> When we think about this in terms of what's happening in the industry right now, there is at least a perception that some folks are really nailing this space. Like companies like NVIDIA, like Grace Blackwell's or Vera Rubin chips, or there is a hype factor around certain large incumbents and then somewhat of a gap I think for everybody else. How do you think about that? Do you think there are certain companies that just get this right, their design process right, they use the right level of AI within the design process? Or again, is that somewhat of just a misunderstanding based on what we're told because these titans are as large as they are?
Faraj Aalaei
>> Yeah. There's actually very little AI is used up to this point in chips that today you see are driving the massive market. I think companies like NVIDIA, which has always been an innovative company, invested years developing the technology upon which today we're running all of these AI models. They didn't wake up one morning and say, "Oh, I want to design a chip like a Blackwell," and it just was done. The cement for this thing was poured like six, seven years ago, and then of course they kept making them better and better incrementally as they came up with generations. But whether it's NVIDIA or anybody else, the cost and the time doesn't go away because in the chip design, it's a very precise process to build a chip. You're putting hundreds of billions of transistors on a substrate, on a device that fits in the palm of your hand, and it takes only a few badly place or badly designed transistors to blow up the whole scenario. So precision is there and that's why there is such care. Now, with the advent of AI, of course, we can actually create those designs in a lot more intelligent way at the speed of inference, which is very different than the speed of human beings designing these blocks one by one and verifying them and then making sure that they're working okay before you send the chip out for production. We collapse that time by a factor of 10 and get the device that you want in a short period of time. That allows, by the way, as a side effect, it allows for more innovation. My vision was when I did my first company, it was four of us entrepreneurs and essentially a plan to do a chip. We were first time entrepreneurs and the venture capital industry decided to invest in us to go build those chips. Now, today, that's not possible because the cost is so high that when you actually fund the company, you worry about the amount of money it takes to build those devices. We want to turn the clock back in terms of that ability to a point where the next generation of entrepreneurs can actually build any chip they want for the markets they want and not have to raise hundreds of millions of dollars to develop it.
Gemma Allen
>> And just to benchmark that cost for a second, what are we talking? Are we talking hundreds of millions of dollars per chip design? Is there an industry standard? Just for the audience that might not know, in fact I do not know, what does that cost roughly to build?
Faraj Aalaei
>> Yeah, so it depends on the complexity of device obviously that you're building, but a lot of devices that people talk about, they've almost become household names thanks to NVIDIA. These devices are very complex. So sometimes there's thousands of engineers that, from beginning to an end, are actually working on this device over a certain period of time. That period of time is not in months, it's in years. So not only you have to hire all of these folks with these different expertise, but they have to stay on the project for duration of that whole time until you actually put it in production and know that you can now just manufacture it. And so yeah, a device can easily cost several hundred million dollars even for a medium complexity device and then therefore it's a big bet. And if you look at the trajectory of semiconductor companies being funded, we've had a massive decline since the year 2000. And even today, with all the greatness that has come to semiconductor companies in terms of valuations, the number of new semiconductor companies that are pure play semiconductor companies, meaning they're just building a chip and selling it to their clients, it's in single digits and maybe 10, 12 at best a year. We used to have 200 of those a year.
Gemma Allen
>> That is really interesting. So the TAM for you is quite large, the addressable market.
Faraj Aalaei
>> Absolutely.
Gemma Allen
>> Who are you targeting? Talk me through a typical customer profile, partner profile for Cognichip.
Faraj Aalaei
>> Yeah. This also ties, by the way, in the history. When I started in this thing, there were chip companies and then there were system companies like Cisco and Dell and those folks. You designed chips, you sold it to them. Over time, the need arose with these system companies and now cloud companies, like Google and Amazon and others, they want to build their own bespoke devices. So today, a target customer for our system that allows them to do this through AI can range from small chip companies, pure play chip companies, large ones, all the way to the system companies and cloud companies, because every company out there in tech today, electronic tech, is actually a chip company. So for us, the market is huge. It's a great opportunity and great opportunity that goes along with really kind of providing the industry with a new capability to reinvent itself and proliferate the number of companies that are solving problems. Ultimately for us, our vision is that if you're a scientist, you don't know anything about chip design, you ought to be able to get in front of a terminal, connect to our system and using ACI, design the chip that you want to perform the function that you want. You may be a chemist, you may be a physicist and you need to accelerate the development of your idea and to accelerate it always means you have to put it in hardware. And so our vision is to enable everybody to become a chip designer.
Gemma Allen
>> So it's almost like a form of CAD for chip design. From the perspective of the ultimate buyer profile, who's leaning in first?
Faraj Aalaei
>> Chip companies. So we are engaged with over 30 customers today, and these are some of the largest companies on the planet, chip companies on the planet that are using actually this system. Currently, it's in trial period for us because we just made it available to this set of customers to start using. But the results that they're reporting back to us, it absolutely blows my mind even though I'm the one pushing this idea that we ought to be using AI for chip design, but also customers. I mean, how we arrived at this point in the semiconductor industry was by necessity. The more complex the devices became, the more people we needed on the teams. The more people we needed on the teams meant that we have to kind of create silos of expertise. And then you stick the silos together to get the chip from an idea into an actual piece of hardware. And there was a whole EDA industry that was developed, companies like Synopsys and Cadence, and so on. And those companies provided really the ability for engineers to be able to see the result of their work. And so we got used to this silo, but this silo is the one that creates time and it creates expense, and what we do with AI is actually get out of that silo and be able to do a design end to end without having to stay within those confinements. And you can teach AI models to do that. You can teach them to be experts in anything you really want them to be experts in. But you also have to kind of create the workflows to make sure you contain the integrity of the process, which in this case would be the chip design.
Gemma Allen
>> And your own tech, your own development, is that happening in partnership or in conjunction with some of these large frontier models? Is a lot of this proprietary? What does it look like underneath the hood?
Faraj Aalaei
>> Yeah. So to do what we've done, I had to bring together three different disciplines in tech. So we had to bring in scientists, math and physics scientists, that knew how to create AI models and knew how to train models. Then we had to put those in the same room with highly experienced chip designers with different disciplines of chip design. And our job was to really put these two teams together to meld mind. Then a third team is a software team that actually builds the product that the engineers will use to design their own chips. So that's been sort of like really itself a learning process, because AI scientists don't really know much about chips and chip guys typically don't know a lot about AI. And by putting them together for the last two years, we've created basically cyber people out of the whole team. And what we do is there are some models that we develop on our own and there are some models that we use open source models and train those open source models.
Gemma Allen
>> Interesting.
Faraj Aalaei
>> And our approach is actually a mixture of models because what we want to do is we want to make sure that models are small enough, fast enough for the task at hand and they are purposefully trained for that particular task. And you put them all together and you create essentially what should translate into the best chip design mind on the planet.
Gemma Allen
>> And a mesh of some brilliant minds. So you've raised some money because obviously you mentioned that the semiconductor industry hasn't really gotten the love it maybe could have. It's CapEx heavy. It's seen as somewhat of a different play to software. But there are some folks who certainly believe in this vision. Talk to me about the financials and what's ahead.
Faraj Aalaei
>> Yeah. So we've been lucky to be able to get tier one VCs to believe in this idea. And when I went out to raise money for this, I purposefully went to folks that I knew that understand both AI and understand the chip. They've been veterans in the chip business, so they understood the problem. So we went out to Lux Capital and Mayfield as the two leading guys in that sector. And lucky for me, they both agreed that they provide the seed funding for the company. And then what happened was we recently raised ... So we raised 33 million there and then recently we did a series A of 60 million, and with that we brought in a new set of investors to come in and help us grow the company into commercialization. And Seligman Ventures was the leading investor in this round. And yeah, SBI from Japan was another, it's the largest VC in Japan. Also very interested in the semi space both as a country and also as a firm. So we put together really a good coalition of investors that understand the market, understand our vision, and are willing to support us. So we're thankful to them.
Gemma Allen
>> Wow. Well, Faraj, you've rang the bell at Nasdaq. You've rang the bell at the NYSE. Where are you ringing the bell for Cognichip?
Faraj Aalaei
>> It's going to be right here, NYSE. I can tell you that.
Gemma Allen
>> I hope so. Please invite me to the party.
Faraj Aalaei
>> We will.
Gemma Allen
>> Thank you so much for coming on theCUBE.
Faraj Aalaei
>> My pleasure. Thank you.
Gemma Allen
>> I'm Gemma Allen here at theCUBE Studio at the New York Stock Exchange. This is AI Factories, part of our program with NYSE Wired. Thanks for watching.
>> Palo Alto studio connecting Silicon Valley and Wall Street. I'm John Furrier, host of theCUBE, here with Dave Vallante, my co-host.
Gemma Allen
>> Welcome back to theCUBE Studio here at the New York Stock Exchange. I'm Gemma Allen with NYSE Wired's AI Factories and joining me now for a conversation on artificial chip intelligence is Faraj Aalaei, founder and CEO of Cognichip. Welcome Faraj.
Faraj Aalaei
>> Oh, thank you. Thanks for having me.
Gemma Allen
>> So you've had a long and tenured career in this space. You've been in tech a long time, but a lot is changing in this moment, which is some of the catalysts behind founding Cognichip two years ago. Talk to me a little bit about Cognichip, the journey, and what you guys do.
Faraj Aalaei
>> Sure. And that long time, part of it was very painful, of course, building two semiconductor companies and taking them out public. Along the path, obviously the industry was kind of getting more and more complex, but the fundamental problem all those years were really that increasingly, it was just taking longer and longer to be able to design a chip and the cost of designing those chips was just skyrocketing, which incidentally it led to really a decline in how venture capitalists were investing in startups. The number of startups started to go down. It was very unique in the late, let's say 2010s and beyond to actually get funding for semiconductor companies. And again, cost and time had a lot to do with that. And so I was building these companies through that process, through that transition. And then when it came to Cognichip, it was really kind of looking back at all of those experiences and the issues related to chip design in our industry and where our industry was heading and then looking at AI as probably the discontinuity in technology that would allow us to solve some of those problems. So I started to think about, well, what if? What if I could take generative AI and direct it to solving the problems I had been struggling with in building the last two companies? And that's how Cognichip was worn.
Gemma Allen
>> Wow. And we know that no two chips are the same. This is a very an ever growing portfolio, I guess you could say, of chips and their requirements, GPUs for certain things, CPUs for other things, LPUs for other things. Talk a little bit about how you kind of segment this industry and how you think about, I guess, is there low hanging fruit from the design process? Are there challenges that are easier to solve first?
Faraj Aalaei
>> Yeah. So part and parcel of the increasing costs and the complexity of these devices really led the industry to say, we're going to build a lot of really big general purpose devices because we can't do too many of them so we might as well do one that has everything in it. Well, if you build one that has everything in it, it becomes also very inefficient. While it becomes more complex, more risk, but it also becomes inefficient in terms of power and in terms of cost. Where software industry is going through the roof from zero to 100 million users in matter of weeks, it takes us five years to build a chip and bring it to production and that's in a good day. So think about it. When ChatGPT, for example, was announced as a product, the chips that that was running on were probably five, six years old from beginning. They started designing those chips five, six years ago, so they couldn't have comprehended ChatGPT as a workload. And this difference between how fast software can scale and how fast chip can scale creates a major problem in our industry. The hardware industry is far behind beyond where the software industry is and we need to solve that problem. The way you solve that problem is by collapsing the time and collapsing the complexity of these devices, and AI can help you do that. We invented ACI to be able to really look at instead of AGI, which is artificial general intelligence, we focus on artificial chip intelligence, training models that can actually design a chip for you a lot faster and a lot cheaper and get it out to the market. Now, if you do that in process, what you're enabling the industry to do is to build bespoke devices. So if I'm running, for example, if I'm doing a device for a self-driving car, it may be different than if I'm building a supercomputer device and now I can actually build those devices cost efficiently, bring them to market as fast as we can imagine applications being formed in the software side of things.
Gemma Allen
>> When we think about this in terms of what's happening in the industry right now, there is at least a perception that some folks are really nailing this space. Like companies like NVIDIA, like Grace Blackwell's or Vera Rubin chips, or there is a hype factor around certain large incumbents and then somewhat of a gap I think for everybody else. How do you think about that? Do you think there are certain companies that just get this right, their design process right, they use the right level of AI within the design process? Or again, is that somewhat of just a misunderstanding based on what we're told because these titans are as large as they are?
Faraj Aalaei
>> Yeah. There's actually very little AI is used up to this point in chips that today you see are driving the massive market. I think companies like NVIDIA, which has always been an innovative company, invested years developing the technology upon which today we're running all of these AI models. They didn't wake up one morning and say, "Oh, I want to design a chip like a Blackwell," and it just was done. The cement for this thing was poured like six, seven years ago, and then of course they kept making them better and better incrementally as they came up with generations. But whether it's NVIDIA or anybody else, the cost and the time doesn't go away because in the chip design, it's a very precise process to build a chip. You're putting hundreds of billions of transistors on a substrate, on a device that fits in the palm of your hand, and it takes only a few badly place or badly designed transistors to blow up the whole scenario. So precision is there and that's why there is such care. Now, with the advent of AI, of course, we can actually create those designs in a lot more intelligent way at the speed of inference, which is very different than the speed of human beings designing these blocks one by one and verifying them and then making sure that they're working okay before you send the chip out for production. We collapse that time by a factor of 10 and get the device that you want in a short period of time. That allows, by the way, as a side effect, it allows for more innovation. My vision was when I did my first company, it was four of us entrepreneurs and essentially a plan to do a chip. We were first time entrepreneurs and the venture capital industry decided to invest in us to go build those chips. Now, today, that's not possible because the cost is so high that when you actually fund the company, you worry about the amount of money it takes to build those devices. We want to turn the clock back in terms of that ability to a point where the next generation of entrepreneurs can actually build any chip they want for the markets they want and not have to raise hundreds of millions of dollars to develop it.
Gemma Allen
>> And just to benchmark that cost for a second, what are we talking? Are we talking hundreds of millions of dollars per chip design? Is there an industry standard? Just for the audience that might not know, in fact I do not know, what does that cost roughly to build?
Faraj Aalaei
>> Yeah, so it depends on the complexity of device obviously that you're building, but a lot of devices that people talk about, they've almost become household names thanks to NVIDIA. These devices are very complex. So sometimes there's thousands of engineers that, from beginning to an end, are actually working on this device over a certain period of time. That period of time is not in months, it's in years. So not only you have to hire all of these folks with these different expertise, but they have to stay on the project for duration of that whole time until you actually put it in production and know that you can now just manufacture it. And so yeah, a device can easily cost several hundred million dollars even for a medium complexity device and then therefore it's a big bet. And if you look at the trajectory of semiconductor companies being funded, we've had a massive decline since the year 2000. And even today, with all the greatness that has come to semiconductor companies in terms of valuations, the number of new semiconductor companies that are pure play semiconductor companies, meaning they're just building a chip and selling it to their clients, it's in single digits and maybe 10, 12 at best a year. We used to have 200 of those a year.
Gemma Allen
>> That is really interesting. So the TAM for you is quite large, the addressable market.
Faraj Aalaei
>> Absolutely.
Gemma Allen
>> Who are you targeting? Talk me through a typical customer profile, partner profile for Cognichip.
Faraj Aalaei
>> Yeah. This also ties, by the way, in the history. When I started in this thing, there were chip companies and then there were system companies like Cisco and Dell and those folks. You designed chips, you sold it to them. Over time, the need arose with these system companies and now cloud companies, like Google and Amazon and others, they want to build their own bespoke devices. So today, a target customer for our system that allows them to do this through AI can range from small chip companies, pure play chip companies, large ones, all the way to the system companies and cloud companies, because every company out there in tech today, electronic tech, is actually a chip company. So for us, the market is huge. It's a great opportunity and great opportunity that goes along with really kind of providing the industry with a new capability to reinvent itself and proliferate the number of companies that are solving problems. Ultimately for us, our vision is that if you're a scientist, you don't know anything about chip design, you ought to be able to get in front of a terminal, connect to our system and using ACI, design the chip that you want to perform the function that you want. You may be a chemist, you may be a physicist and you need to accelerate the development of your idea and to accelerate it always means you have to put it in hardware. And so our vision is to enable everybody to become a chip designer.
Gemma Allen
>> So it's almost like a form of CAD for chip design. From the perspective of the ultimate buyer profile, who's leaning in first?
Faraj Aalaei
>> Chip companies. So we are engaged with over 30 customers today, and these are some of the largest companies on the planet, chip companies on the planet that are using actually this system. Currently, it's in trial period for us because we just made it available to this set of customers to start using. But the results that they're reporting back to us, it absolutely blows my mind even though I'm the one pushing this idea that we ought to be using AI for chip design, but also customers. I mean, how we arrived at this point in the semiconductor industry was by necessity. The more complex the devices became, the more people we needed on the teams. The more people we needed on the teams meant that we have to kind of create silos of expertise. And then you stick the silos together to get the chip from an idea into an actual piece of hardware. And there was a whole EDA industry that was developed, companies like Synopsys and Cadence, and so on. And those companies provided really the ability for engineers to be able to see the result of their work. And so we got used to this silo, but this silo is the one that creates time and it creates expense, and what we do with AI is actually get out of that silo and be able to do a design end to end without having to stay within those confinements. And you can teach AI models to do that. You can teach them to be experts in anything you really want them to be experts in. But you also have to kind of create the workflows to make sure you contain the integrity of the process, which in this case would be the chip design.
Gemma Allen
>> And your own tech, your own development, is that happening in partnership or in conjunction with some of these large frontier models? Is a lot of this proprietary? What does it look like underneath the hood?
Faraj Aalaei
>> Yeah. So to do what we've done, I had to bring together three different disciplines in tech. So we had to bring in scientists, math and physics scientists, that knew how to create AI models and knew how to train models. Then we had to put those in the same room with highly experienced chip designers with different disciplines of chip design. And our job was to really put these two teams together to meld mind. Then a third team is a software team that actually builds the product that the engineers will use to design their own chips. So that's been sort of like really itself a learning process, because AI scientists don't really know much about chips and chip guys typically don't know a lot about AI. And by putting them together for the last two years, we've created basically cyber people out of the whole team. And what we do is there are some models that we develop on our own and there are some models that we use open source models and train those open source models.
Gemma Allen
>> Interesting.
Faraj Aalaei
>> And our approach is actually a mixture of models because what we want to do is we want to make sure that models are small enough, fast enough for the task at hand and they are purposefully trained for that particular task. And you put them all together and you create essentially what should translate into the best chip design mind on the planet.
Gemma Allen
>> And a mesh of some brilliant minds. So you've raised some money because obviously you mentioned that the semiconductor industry hasn't really gotten the love it maybe could have. It's CapEx heavy. It's seen as somewhat of a different play to software. But there are some folks who certainly believe in this vision. Talk to me about the financials and what's ahead.
Faraj Aalaei
>> Yeah. So we've been lucky to be able to get tier one VCs to believe in this idea. And when I went out to raise money for this, I purposefully went to folks that I knew that understand both AI and understand the chip. They've been veterans in the chip business, so they understood the problem. So we went out to Lux Capital and Mayfield as the two leading guys in that sector. And lucky for me, they both agreed that they provide the seed funding for the company. And then what happened was we recently raised ... So we raised 33 million there and then recently we did a series A of 60 million, and with that we brought in a new set of investors to come in and help us grow the company into commercialization. And Seligman Ventures was the leading investor in this round. And yeah, SBI from Japan was another, it's the largest VC in Japan. Also very interested in the semi space both as a country and also as a firm. So we put together really a good coalition of investors that understand the market, understand our vision, and are willing to support us. So we're thankful to them.
Gemma Allen
>> Wow. Well, Faraj, you've rang the bell at Nasdaq. You've rang the bell at the NYSE. Where are you ringing the bell for Cognichip?
Faraj Aalaei
>> It's going to be right here, NYSE. I can tell you that.
Gemma Allen
>> I hope so. Please invite me to the party.
Faraj Aalaei
>> We will.
Gemma Allen
>> Thank you so much for coming on theCUBE.
Faraj Aalaei
>> My pleasure. Thank you.
Gemma Allen
>> I'm Gemma Allen here at theCUBE Studio at the New York Stock Exchange. This is AI Factories, part of our program with NYSE Wired. Thanks for watching.
John Furrier
Gemma Allen