What's Up with Tech?
Tech Transformation with Evan Kirstel: A podcast exploring the latest trends and innovations in the tech industry, and how businesses can leverage them for growth, diving into the world of B2B, discussing strategies, trends, and sharing insights from industry leaders!
With over three decades in telecom and IT, I've mastered the art of transforming social media into a dynamic platform for audience engagement, community building, and establishing thought leadership. My approach isn't about personal brand promotion but about delivering educational and informative content to cultivate a sustainable, long-term business presence. I am the leading content creator in areas like Enterprise AI, UCaaS, CPaaS, CCaaS, Cloud, Telecom, 5G and more!
What's Up with Tech?
Charting the Tech Frontier: PK Gupta's Odyssey in VoIP Innovation and Startup Culture
Interested in being a guest? Email us at admin@evankirstel.com
Embark on a technicolor journey through time with PK Gupta, an innovator whose fingerprints are smeared across the landscape of modern tech. As we sit down with this luminary, we trace the arcs of his career—from pioneering VoIP gateways to shaping the contours of air-to-ground telephone services. PK's anecdotes offer a rare glimpse into the gritty startup culture at Dialogic and its dramatic absorption by Intel, all seen through the prism of his formidable academic achievements, featuring a PhD in Electrical Engineering and an MBA from one of the most prestigious business schools, Wharton.
Venture into the heartland of startup success as PK Gupta decodes the DNA of breakthrough innovation. He divulges the essential elements that give startups their fighting edge: solving real-world quandaries, penetrating lucrative markets with unique products, and orchestrating teams that resonate with dynamic synergy. As we peel back the layers on edge computing, analytics, and the seismic shift promised by generative AI, PK forecasts a brave new world of entrepreneurship. His current crusade? Channeling his boundless insight into mentoring the tech titans of tomorrow, while weaving tales of a vibrant global innovation scene where creativity knows no borders.
More at https://linktr.ee/EvanKirstel
Hey everybody, really intriguing guest today, PK Gupta, pioneer in so many fields in emerging tech. Pk, how are you? I'm great, Owen, Thanks for having me. Well, great to have you here. I'm fascinated by your personal professional journey. You've gone from developing a number of industry-first technologies back in the day VoIP gateways and YMAX base stations and more. For those tech insiders, they'll know what I'm talking about. Maybe introduce yourself a little bit about your current mission and related topics here for our audience today.
Speaker 2:Yeah, yeah. I've been fortunate to have a great journey in technology. I worked at various leading companies like Hughes and Intel and then did my startup, most recently Mac computing, focused on video analytics, and now I'm getting into investing in advising startups. So it's been a very long journey, very successful, very exciting, and love to talk to you about different aspects of that.
Speaker 1:Yeah, there's a lot to unpack, including going back earlier in both of our careers. 30 plus years ago we intersected at a little known fast hot startup called Dialogic and, yes, in the early 90s there were hot tech startups around. It didn't all happen the last two years. So maybe talk about your background, going back to early in your career, and you have a very interesting educational background. I would love to understand how that has helped shape your career as well.
Speaker 2:Yeah, so education background I did a PhD in electrical engineering and then a few years later, maybe 10 years later, I did my MBA, also from Warden. So I have a background in. I've always been interested in business and technology, so the educational foundation helped me navigate through the different aspects of my career. So I started at Hughes just to kind of give you a quick render and Hughes, which is a very interesting place, and this was Hughes network, right Separate from Hughes aircraft, which was in defense in California. Hughes network was outside of Washington DC and focused on communication technologies and Hughes had a culture of building new technologies in partnership with industries and bringing them to market. Some of them worked and some of them failed, so they took that risk. So one of the technologies that we worked on that was very exciting, was at Hughes was air to ground telephone. Do you remember that?
Speaker 1:back at the service. I remember it well. Yeah, I used to swipe your credit card and fast forward. We're going to have it again with Starlink.
Speaker 2:So Starlink is obviously talking to the satellites and then beaming down. The Hughes air to ground was talking down to the ground base stations, right, and trying to hop as the plane was flying hop between ground base stations. So that technology was developed in the mid 90s and a couple of airlines deployed it Alaska, american Airlines. But in a talk about technology, some technology succeeds and don't, based on timing. So the air to ground telephone was a great project, worked well, but it was killed as the cellular phones were introduced in the late 90s right, the first brick phones, right. So instead of trying to make a call on the flight for $5 a minute, it was five dollars a minute. People would wait till the flight landed and then they would make the call. Right, so that killed that technology. The company invested a lot. It also.
Speaker 2:Hughes also did something pioneering in cellular base stations. If you go back to now, we have 5G. It was in the 90s, it was 2G and they pioneered a technology called TDMA, time Division Multiplex, and they lost out to Qualcomm's CDMA. So that was another technology where Hughes made a bet, big bet, and lost out on the cellular side. But some of the other technologies they worked on actually succeeded, including visa terminals, and today you might be familiar with direct TV and direct, and so those kind of technologies worked out for Hughes. So after Hughes, I went to the startup you were referring to, dialogic, and our bats kind of crossed that, I think, for at least a year. Right, you were there also, you were leaving.
Speaker 1:Yeah, I was a sales engineer, you were an engineering, so they didn't let us talk to each other that much. But no, I used to enjoy the startup culture. It goes well back, it's not a new tradition, so the beer in the parking lot and a really great culture fond memories. And they were acquired by Intel and you had quite a long run as an executive at Intel. What did you look after there?
Speaker 2:Yeah, so dialogic. I was there, you guys were there. And one thing I just want to mention before I forget one of the key technologies we developed at dialogic was a voice over IP gateway, and dialogic was in the computer telephony business, which was basically using your computer to make phone calls and voice mail and all that. So the first voice over IP gateway was basically a PC with a voice card in it and some software and it was used to bypass the TARIS network, right. So it was hugely successful. Obviously, everything today is voice over IP based. But that's what got Intel interested, because Intel at that time in, say, late 2000, early 2000, was looking to invest in communication. So they acquired dialogic in 2000 and at Intel we started working on different things.
Speaker 2:One of the first things we worked at Intel in the communication space was building a YMAX based station, right. So YMAX was another one of those big bets, right, made by many companies in the early 2000 timeframe, and the best way to describe YMAX is Wi-Fi on steroid, and so you could go very long distance. Use for connectivity, backhaul, but also for mobility, different than Wi-Fi. It was also supporting mobility and people tried it. You know there was a huge consortium. Intel invested very heavily in YMAX, but in the end it lost out to regular LTE and cellular technologies right, Because they were superior, right. So it just didn't work out. I believe today it might be deployed very selectively, but not much, right. So YMAX was a great technology we worked on.
Speaker 2:The next thing we actually did at Intel that worked out well is we worked at using FPGAs field-programmable gate arrays, which I'm sure many of you listeners are familiar with using FPGAs as a computing accelerator. So that was a very interesting concept because Intel at that time obviously was making Xeons and the Xeon IA architecture and promoting that and they were looking at other accelerator technologies. So we set up a small team within Intel, like the Skunkworks, and we started looking at. This is back 2008, 2009 timeframe, right, Almost 15 years ago. So I'm looking at it using FPGAs as an accelerator, which was a Very innovative concept at that time. Right, Because it's almost like a foreign silicon, an FPGA you wanna put close to a Xeon silicon and make them talk to each other. So we worked on that. We showed how to put it together, how some of the compute could be offloaded to an FPGA and accelerated, and the long story short is. It got good traction. It got good support from other veterans in the industry, like Google, Microsoft. All of them were interested in the technology. So Intel pursued it and it led. So this effort led to Intel's acquisition of Altura back in 2015,. 16, right. So a project that started as a very small project, de-risking it, proving it out and then leading to Intel's acquisition of Altura, which remains, by the way, today, Intel's largest acquisition. Oh wow, Other companies that acquired AI companies, and all that were in less than the 16 billion right. So at that time it was considered to be a huge amount of money to pay for an FPGA company. Of course, till AMD bought Xilinx, maybe 10 years later, for double the price, right. So both the FPGA companies were acquired by the computing companies, which is kind of an interesting development. But Intel at Intel, we stayed back.
Speaker 2:I led the effort to integrate the FPGA core. So we took an FPGA core and we took a Xeon core and we put it in the same package. So imagine now on a regular motherboard, you go to your Xeon CPU in one socket and the other socket you have an FPGA. They're talking to each other over the bus the core-end bus between them, and it opened up a whole slew of new use cases. Right, Because the prevailing way to do this was using a PCI card.
Speaker 2:Right, the typical way to attach something to a CPU you put it on a PCI card and over the PCI bus, you do your data in and outs. Right, and you offload With the FPGA attached to the CPU. You could get access to the Xeon's memory currently, and so it opened up a whole bunch of new use cases, which led to some very aggressive forecast about the deployment of FPGAs in the data center. Right, and you might have heard about Microsoft deploying FPGAs in that data center, so they use FPGAs. Now Every server in Microsoft Azure actually has an FPGA card in it. Right, For networking, so it's a smart mic, so they deploy that as a smart mic.
Speaker 2:Aws has FPGA instances, so does other cloud service providers, and so FPGA acceleration itself was supposed to take off. It did not reach its initial forecast, but it is being used in a wide variety of use cases and we saw the opportunity to deploy FPGA accelerators for enterprises right For accelerating enterprise workloads, like big data analytics, which at that time, was taking off. So that was the motivation for us to start Mac computing right. So the next part of my journey was leaving Intel in 2017 and starting Mac computing. So we started Mac computing with a few folks from Intel who were part of my team. We started in Portland, oregon, because that's where we were located with Intel In 2017, we started, and I just personally exited just last year, in December.
Speaker 2:But over this five, six year we've had a great run. We introduced some very new technology on edge computing, which I love to talk about more about edge computing and doing real-time video analytics. So that was a very successful run that led to Mac computing, and now I'm looking at investing and advising other startups right, hoping to help them leverage the learnings from my lessons over the years.
Speaker 1:Wow, that's a great synopsis and lessons indeed. You've seen so much in the technology space the successes, the failures, the challenges, the creative disruption that hits across technology. Has that made you more of an optimist or a pessimist when it comes to new ventures, new startups, new investments? Given the challenges you've seen, it's not for the faint of heart.
Speaker 2:Right now, I'm always an optimist in terms of introducing new technologies, but also being pragmatic that you cannot just develop technology for technology's sake, right. There are other factors, right, that go in in making the technology adoption successful, right? So the first thing is timing. Obviously, timing is critical. When you take money, it's developed, and if the timing is off, then no matter how good it is or how well funded it is. And there have been many cases documented where people try to bring a technology to market a bit before time and then it just didn't pan out. So there are a lot of examples of that. So timing is critical. So is execution.
Speaker 2:You might have a great idea, but if you don't execute well and if there's other technology, the implementation and the execution of the technology, that can also lead to failure. But besides these two things, the other things obviously is this you have to be solving a problem. You cannot just be developing technology because it's sexy or interest you in a certain way. So as long as there's a problem you're solving and which will then lead to adoption of the technology, which is the ultimate final step, the technology has to be adopted by the market, it has to get deployed, it has to generate revenue, to be successful. So there's all these phases you go through and in my personal journey and I'm sure other people, we can cite different examples. There are always cases where one or the other of these levers is missing and that leads to a technology not getting to market. Other cases where everything just clicked perfectly, where then things led to wide adoption and success of the technology.
Speaker 1:Yeah, that's an amazing insight. And you talk to entrepreneurs and founders and private equity all kinds of collaborators. What are some of the common challenges you see with those organizations, individuals, what's your high level advice like to them? Yeah, maybe shed some light on your advisory work.
Speaker 2:Yeah, absolutely so, I think. First of all, let me just say this is a great time to be an entrepreneur. There are a lot of challenges, but it's a great time. Technology advancing, though, the VC market was a bit frozen last couple of years. Funding was a bit difficult, but there's always ways to overcome that and things are loosening up a little bit right now, so I see things going forward as very positive.
Speaker 2:So when I talk to new CEOs or teams that are starting a new venture, a small startup, I think there's four things that I'm looking for personally and I advise them accordingly. So one is there has to be a problem that you're solving. So you start with that. What problem are you solving? And if the startup founders have personnel experience in the problem so they are familiarity with the problem the better the chances of success. So it cannot be, and there are cases where a few very smart people will get together and say let's go and build something, and they'll try to build something because they are familiar with the technology. They start from the wrong side and those typically don't end well. So starting with the problem, understanding the problem, is key.
Speaker 2:The next thing is the problem in solving the market has to be attractive. Eventually you want to get funding or go to market and scale, so the market has to be attractive. The share of market or whatever metric you use the product revenue potential has to be large and attractive enough to attract investors. So that's the second thing. The third thing, and the product that you are defining based on this, has to have certain differentiation. It cannot be a me too. It cannot just be okay, I'm building this new product and it's got these neat features which are 10% better or even 20% better than an existing product. That will not kick in. You've got to focus on maybe one feature that's 10x better and that will differentiate the product from the computer. There has to be something like that that's highly differentiated. So the product definition has to be well defined and executed.
Speaker 2:And then, finally, but not the least one, is the team. So the team that's coming together. There should be a team like if they're two co-founders. There should be complementary skill sets, not just identical skill sets, so people who bring a different perspective, a team that can execute well. The questions that investors will ask is why you, why this team, what is different about you, how you're going to execute? So some evidence of that it helps, but doesn't have to be always successful entrepreneurs. On the second venture it can be your first venture, but you have to demonstrate that the team can work well together. So those are the four things that I advise I look for, I think other investors also look for, and those are the metrics that I think would make startups more successful.
Speaker 1:A wonderful advice. So, across the tech landscape, what are the technologies, the niches that you are most excited about and why? Iot, ai, obviously cloud edge. Where are you exploring exactly these days?
Speaker 2:So if you look at the hype cycle that Gartner publishes, some of the technologies are in the hype cycle someday, and if you look at some of the technologies that are getting deployed today, so I would say edge computing, edge analytics, ai at the edge is at a stage right now where it's getting widely deployed, right Everything has come together the hardware, the edge platform, the software, hardware architecture required for edge analytics, for edge computing, is well established now and you are seeing, and it's supported by the market trends, right, like 50% of the data is going to be generated at the edge right this year and next year. Right Over a billion cameras are deployed. They're billion, tens of billions of sensors deployed. So a lot of data being generated at the edge and the data needs to be processed at the edge. So edge computing in general, edge analytics these are the kind of things we were doing at Mac computing, right, my startup right, which is very, very attractive. So that will continue right, and so companies coming in with some new technologies there are getting very successful.
Speaker 2:The trend now that everybody talks about is gen AI right now. So obviously AI. So AI we used at edge computing was more you could call it the traditional AI where we used it to detect objects, classify objects, right those kind of things, and with all based on deep learning models that we developed and we did some innovative things like being able to train the models continuously for better performance. So all those were kind of differentiation at the edge. But with gen AI now getting widely deployed, it's hugely exciting, right.
Speaker 2:If you just look at from a startup perspective, what gen AI is enabling besides everything else that you hear about, it right, is actually the cost of bringing a new product to market. You know it's coming down substantially, right. The change is similar to, if you go way back, when AWS introduced the infrastructure. Right Before that, if you recall, you know, a startup had to buy the sun servers and set up their own infrastructure and everything right. Then AWS came along with the shared infrastructure and that cut down the infrastructure cost a lot and drove a whole new wave of startups who could leverage that and the cost of a startup may be decreased to a few million dollars.
Speaker 2:With a few million dollars, a startup with small team could go and build something and bring it to market. Now that has decreased by 10x again with Gen AI, right. So now the scenario is that just one person or even a couple of team, right, small team with maybe some key skill sets which they can offshore, can take a new concept, build a new product, you know, and bring it to market very quickly with Gen AI, right With these vertical language models and other technologies, at a fraction of the cost of previous startups. So what you'll see is the cost of startups is gonna change, is gonna come down. The financing required for startups to go to market is gonna come down substantially, right, and that's gonna drive a whole new wave of innovation right as Gen AI technology and the other associate technologies get more widely deployed, right.
Speaker 2:So that's what's really exciting right and that we see happening this year and next year, right as Gen AI then starts getting widely deployed and solving real problems.
Speaker 1:I love that, and it's such a fascinating global landscape as well. Tons of innovation coming from everywhere around the world. You must have your eye on opportunities around the world. What excites you the most, whether it's what's happening in Europe or in Asia? What's new teams, you know innovating, kind of popping out of the woodwork now.
Speaker 2:Yeah, it's a huge opportunity everywhere. You heard all the cases of startup with AI. You know, getting 100 million funding and very quickly in a short term, and then billion dollar unicorn evaluation. So those are the well-publicized cases right that we hear a lot about, but behind that there's a whole in a slew of startups right that are just starting slowly getting small amount of angel funding, small amount of investing, and many of these are gonna break through, right.
Speaker 2:So, the challenges, and these are, as you said is happening globally, right? You don't have to be in Silicon Valley now only to be successful. You could potentially be anywhere in any place, right, and have access to this technology, and as long as you have the right idea the right team, right, you think you can go and make something happen.
Speaker 2:So that's what's really exciting, right, and our opportunity for us as investors' advisors is trying to find that diamond in the rough right, trying to find that opportunity, the team, that idea that they're working on very early right. So our goal is to identify as an investor advisor, we are trying to identify opportunities at a very, very early stage, early seed stage, so we can make a bet on that team, that small team, give them the amount of funding they need, which is actually right now manageable it's a small amount of funding and then enable them to bring it to market and be successful. So, yeah, I think it's across different verticals right. But generally, if you ask me, it's deep tech, right. So deep tech enabled solutions for different verticals relying on a deep tech stack.
Speaker 2:Obviously AI part of it, but also distributed computing, edge computing, deploy and hybrid between edge and cloud. So all these common technology stack is coming together. The challenge is gonna be okay, given this technology stack everybody, by the way, has access to right, so it is kind of leveling the playing field right now, just a high stack, and so the challenge is gonna be okay, since everybody has access to this. Ideas are now the ones that you want to bet on. So who has the right idea to solve a problem that they know about and how they're gonna solve it right? So those are the things you're looking for, which makes it actually quite interesting.
Speaker 1:Quite interesting, to say the least. Really great insights and appreciate your vision. So what are you excited to travel-wise, personally, professionally, any events or tips coming up or calling out?
Speaker 2:Yeah, so I'll be, I've been invited to do a keynote at a FPGA conference in Monterrey in March, so this is very exciting. This is my FPGA journey that I talked about starting in Intel in 2008. So I'll be. The title of my talk is my 15 years journey in FPGA platforms, or something like that I'm gonna be talking about. So basically talking about my journey from Intel to Meg to now, to where I see the FPGA industry going right. So I've been, I've been presenting that in March and that I'm looking forward to some other few engagements after that that are still being discussed Fantastic.
Speaker 1:Well, we'll keep an eye out to everyone you know. Follow PK on LinkedIn. He puts out some great insights and content and thanks so much for watching. Feel free to comment like share always welcome. Thanks, pk, thanks Evan, Take care everyone. Bye, bye.