The Six Five Insider Edition with Ram Velaga, Broadcom

By Patrick Moorhead - April 5, 2023

On this episode of The Six Five – Insider Edition, hosts Daniel Newman and Patrick Moorhead sit down with Senior Vice President and GM of Broadcom Inc., Ram Velaga.

Their discussion covers:

  • Broadcom’s history of successfully integrating new acquisitions into their portfolio
  • The importance of innovation paired with execution
  • Broadcom’s industry-leading portfolio of new innovations
  • How Broadcom is helping to push the ecosystem
  • The impact Broadcom has had in the market

Be sure to subscribe to The Six Five Webcast so you never miss an episode.

You can watch the full video here:

You can listen to the conversation here:

Disclaimer: The Six Five On the Road is for information and entertainment purposes only. Over the course of this webcast, we may talk about companies that are publicly traded, and we may even reference that fact and their equity share price, but please do not take anything that we say as a recommendation about what you should do with your investment dollars. We are not investment advisors, and we ask that you do not treat us as such.


Patrick Moorhead: Hi. This is Pat Moorhead, and we are back for another Six Five Insider Edition where we chat with the most influential executives at the most important tech companies out there. Dan, I love these Insider segments. They’re absolutely my favorite.

Daniel Newman: Yeah, definitely like talking to these people because they have the insights, they know the industry, what’s going on. Of course, you and I always try to get underneath a little bit of what’s not available out there publicly. Of course, we sometimes succeed, and oftentimes we get the Heisman. But at the same time, I do love this format because it’s getting a little deeper than what typical media news, broadcast, interviews can ever do, and helping the market really understand what these companies are doing. It’s so important.

Patrick Moorhead: It is. One of my favorite topics, not from being in the semiconductor market as an executive for, gosh, 11 years and being kind of a consumer, a buyer of semiconductors for another decade, we no longer have to convince anybody that semiconductors are awesome. People know them, they see new applications at companies, even to the end user, kind of changing the way on the planet. And then that little supply chain blip that we had where we weren’t able to ship cars because of a 12-cent analog piece of equipment, but anyways-

Daniel Newman: Oh, they just didn’t have radios in them.

Patrick Moorhead: That’s right. That’s right. They did not. But without further ado, want to introduce Ram. Ram, how are you?

Ram Velaga: Outstanding. How are you guys?

Patrick Moorhead: Great, great.

Ram Velaga: Thank you for having me.

Patrick Moorhead: First time on the Six Five. We’re so excited to have you on, and you’ve got such a storied background. And I’ve been looking forward to this for a long time and I appreciate you coming on the show.

Ram Velaga: Absolutely. Thank you. You’re being kind.

Daniel Newman: So Ram, before I dig into the very first question, because it’s an Insider Edition and we have a little bit more time with you, just give a quick synopsis of you, your role and how you got to Broadcom.

Ram Velaga: Yeah, so I run the switching and routing business unit inside Broadcom. Almost all my career that is worth anything has always been in networking. I started in 2000 just about the dotcom bubble crash. Cisco is a networking company, been there for about 12 years, and after that moved to Broadcom. When I joined Broadcom, I thought, “Oh, I’ll probably be here for four years. Why would a guy from a systems industry be in the semiconductor industry?” And then the next thing I know, it’s almost 11 plus years at Broadcom. So that’s me.

Daniel Newman: So let’s talk M&A and Broadcom and it’s background. So the company is well known for M&A.

Ram Velaga: Yeah.

Daniel Newman: When you talk to, you see any interview about the company, about Hock Tan, it’s the CEO that makes… He’s a deal maker. And of course you have a pretty big deal that’s going on potentially right now.

Ram Velaga: Yep, yep.

Daniel Newman: Talk a little bit about in 2015, the Avago acquisition, being part of that. What was that like?

Ram Velaga: Yeah, actually 2015, roughly, I think it was about May of 2015, I got a call from my then boss who used to report to Broadcom CEO, and he said, “Hey, dude. Can you go to Europe? I’m supposed to go meet this customer, but I need you to go.” I was like, “The last time also you bailed on this customer, so you want me to go again and cover up for you?” He’s like, “Yeah, yeah, just go.” I’m like, “Okay.” I went, and then I covered for the customer. He was like, “Where’s your boss?” I’m like, “I’m here. Let’s talk about me.”

And then towards the end of the day, we see the news that Broadcom was being acquired by Avago. So for those who don’t know it, we were acquired by a company called Avago. But they kept the name Broadcom for two reasons. One, our chairman, Dan and chairman now, Henry, wanted to keep Broadcom as a name, and what I understood was he offered Hawk a discount if he kept the Broadcom name instead of Avago. And then Hawk being Hawk, he took the discount and he kept the name.

Daniel Newman: What a story.

Patrick Moorhead: I know, this is great. I have never heard this before.

Ram Velaga: Oh, yeah.

Patrick Moorhead: This is so good.

Daniel Newman: There’s that behind the scenes we were looking for.

Patrick Moorhead: I know.

Ram Velaga: Exactly. So everything has a price. So I was in Europe and then I started to get texts from my team members saying, “What’s going on?” With other more colorful words in there. Like, “Who’s Avago? Why is Avago buying Broadcom?” Because Broadcom was known to be a technology leader at that point. We were in the set-top box business, we were in networking business, we were in WiFi.

But every company that’s in the public market, at some point in time, there’s a bunch of products which are extremely successful, and Wall Street’s looking for you to continue to grow. And then you put a bunch of bets into new markets that may not necessarily materialize, or the amount of money you’re spending on these new markets might actually not make economic sense. And so Broadcom was doing some of those. We were trying to go after the mobile application processors. We were trying to get into the 4G, 5G modem space and so on and so forth. And we were pouring a lot of money into it and really wasn’t getting much returns on it.

And the other thing that was happening because of that was some of these businesses, which otherwise were very successful, what you would call the core businesses, were getting starved for resources because you’re all investing in what is going to be the new thing going forward, and you’re getting starved in the core businesses. So that sets you up for a perfect scenario where somebody comes in and says, “Look, you’re going to get acquired, we can run this place far more efficiently.” And that was kind of the thesis that was presented to us by Avago, which is led by Hawk Tan.

And so first thing when something like that happens is there’s a lot of uncertainty. I was like, “Okay, will I have a job? I’ve been there for three years now.” And I like, “Anyway, if they fired me, I was going to leave the fourth year. So no big deal.” But you start to think through all of these things. But what I would tell you is what people don’t know is when people think about Hawk or what now Broadcom does as acquisitions is they think of it as you’ve reduced R&D.

The reality of what we end up doing is we look at businesses inside of a company and say, “Some are worth investing in, investing a lot more in. Some we will never be successful in this space, and it’s probably best to divest out of it so that we can actually invest in the businesses that are worth investing in.” So back then when we got acquired, my team was about 1300 people and we were investing a couple of hundred million dollars in R&D. I would say now you look forward in the seven years later, roughly the same number of people, but we’ve doubled the amount of R&D. So then you say, “Okay, same number of people, double the amount of R&D, where’s the money going?”

The money’s actually going back to the team members, but make sure that they participate in the overall success of the organization and the products that we put out there, and hire the best out there. So I would say more likely than not, if not for the acquisition, I probably would not still be at Broadcom. But in terms of what the acquisition did for us in being able to invest in the team and actually invest in the products that we are actually doing well with has been an outstanding run.

Patrick Moorhead: It’s interesting, there’s always lore about a company, and I’d heard the lore about Broadcom and how you pay folks. So it’s great to hear that directly from somebody who has worked there forever. So you obviously have your own organic innovation, you acquire, and like you said, you go in and you look at, “Hey, what are some of the best bets that we can double down on, we can triple down on?” Can you talk about subsequent of when Avago was acquired? What are some of the acquisitions that you did and how did those integrations go? What areas did you double down on?

Ram Velaga: Yeah. Post Avago acquiring in Broadcom at the company level, we obviously acquired a few other companies. One was Brocade, right? Brocade was another classic example of… Brocade was known for fiber channel business, but then Brocade had itself acquired Foundry to try to get into the IP ethernet side of the business. And then they acquired, I believe, Ruckus to try and get into the wireless business. And their fiber channel business was starving. When you look at what happened with Brocade, we’ve really focused on how you make the fiber channel far better. But within my organization too, we’ve made acquisitions. These are much smaller, and you can either acquire companies for revenue or you acquire companies for an IP or you acquire companies for people.

Generally, what we have done inside my team is we’ve acquired companies for people, great talent, and actually both of them were in Europe. One of them was two professors and six PhD students. They were doing very, very deep research into how do you make networks far more higher performance, especially in the context of machine learning and AI workloads, which have a very different traffic patterns. Another was a company that we acquired in Romania, about 900 engineers. It was a great talent. So when you acquire these companies, the first thing you think about is what makes this acquisition successful?

And for us, it was about four years later, five years later, can you look back and say, “The team is still there.” Because you’re really acquiring people, and if you lose all the people in two years, what have you done? So I can say around now, the team that we acquired about five years ago, the leaders of the team and almost everybody from the team is still with us. And actually even the leaders have grown in the organization over the last five years. And the other one that I did with a couple of professors was about six months ago, time will tell how things will look.

But the key is always focus on what their motivations are. Let them do what they’re good at doing, don’t come in their way. And our job is really trying to make sure that they have what it takes to be successful. And I think time has shown over a period of time that these people enjoy what they’re doing. They’ve contributed meaningfully to the organization. And sometimes I go back and ask them, “Hey, would you have done this again?” And generally the answer is yes.

Daniel Newman: That’s pretty good. But you of took us a little through chronology. We talked a little bit about Brocade. We’ve now talked about Avago. I know CA was in there somewhere. There’s been a number of fairly large acquisitions. Of course everybody’s watching to see what happens with VMware. The portfolio’s rounding out. Broadcom’s definitely… Its roots as a chip company is clearly changing. But what are some of the recent innovations that continue to build on that story that it’s not about spending less on R&D and it’s not about slowing innovation, but that Broadcom really does have DNA of continuous innovation?

Ram Velaga: One of the gentlemen I used to work for, he had this saying, he said, “Innovation has shelf life.” And so what he meant is you can think a lot about innovation, but if you don’t execute, it’s worthless. And so I would say in Broadcom, especially in my team, we really, really focus on execution. And when you think about execution, execution as in the context of eventually you have to get a product out. And when you’re thinking about getting a product out it is against a lot of the laws of physics when you start to think about it. Moore’s Law is coming to an end as you go from 40 nanometer, 28 and 28 to 16. And now when you actually look from 16 to seven to five, you’re not getting the transistor performance improvements and the costs are going up, and you’re not getting those power reductions right, right?

With Moore’s Law actually starting to taper out, at the same time on the other side, the amount of data that is being transferred, whether it’s because of video being created or artificial intelligence actually shuffling a lot of data, that is not coming down. If you actually look at the petabytes of data that’s either being created, going up.

Daniel Newman: That’s right.

Ram Velaga: It’s going up. So you look at this and say, “On one side there’s diminishing returns on Moore’s Law. On the other side, there is an algorithmic growth, an exponential growth, rather, on the demand side. So supply side’s actually coming down, the demand side’s kind of going up.” So what do you do as somebody in between? You have to innovate because you’re being really, really pushed against two opposing forces. So a lot of what we focus on this innovation is, there’s something that really drives my team, which is we believe in democratizing networking, and then the democratizing networking is, “Can we keep doubling the bandwidth every 18 to 24 months and lower the price?”

So a lot of the innovations that we do are in that context. So for example, predictably we come out doubling the bandwidth of our chips going from three terabytes to six terabytes to 12, 25, 50 terabytes. And the most recent chip that we’ve released is 50 terabytes. I usually don’t walk around. People walk around like they say with a chip on their shoulder. I’m like, “Yeah, this is a chip in my pocket.”

Daniel Newman: Only chip executives pull chips out of their pockets in the middle of videos. I

Patrick Moorhead: I love it.

Ram Velaga: So if you think about it, this chip has 50 terabytes of bandwidth. And I want to put this in context. 10 years ago, if you were trying to get one fourth of this bandwidth, about 12 terabytes, you will be paying a million dollars and that thing would be consuming-

Patrick Moorhead: Well, and it would be an appliance, not a chip, right?

Ram Velaga: Actually, no. It’s not an appliance. It’s probably a box this big, half of a rack height. And it would consume 20,000 watts of power. 20,000 watts of power, a million dollars, at least less price. You guys will probably get an 80% discount and you probably pay $200,000 for it. This now consumes less than 500 watts of power. And you can probably buy this with a box surrounding it for less than $10,000. So talk about anti-inflation, going from a million dollars list price to $10,000 list price, 20,000 watts of power to 500 watts of power. It’s like a 95, 99% reduction. So that’s how we think about innovation. And this is a classic example of one that we just came out with a couple of months ago.

Daniel Newman: Hey, Pat…

Ram Velaga: You probably want to touch it.

Patrick Moorhead: I do. You know I want to touch it.

Ram Velaga: I want to make sure. I said-

Patrick Moorhead: Yeah.

Daniel Newman: Yeah. We’ll have to get a couple of still shots on that. But Pat, what do I always say about the deflationary value of technology?

Patrick Moorhead: No, mean particularly in this time. I mean, when people are looking at where do they invest their dollars, it’s unlike 2008. Now technology is deflationary, and if you can grow the top line or minimize the bottom as much to get where… Or make a customer experience stickier-

Daniel Newman: AI, automation, analytics, I mean, how much more can we do with so much less?

Patrick Moorhead: Ram, I think that’s a great example. Going all the way from a rack system down to… Now, is that an ASIC?

Ram Velaga: Yeah, this is a standard product, a chip. And so basically what ends up happening is you can go to someplace like Taiwan or someplace else, and essentially you’ll get a sheet metal and a power supply and a CPU that sits along with this, and that’s about it. That’s all you need. And then you need a network operating system that sits on top of it.

Patrick Moorhead: In fact, yeah, it does take a village. And one of the good things about getting older is you learn, and hopefully you learn, and you gain perspective. And one thing that I appreciate about the semiconductor business is it can happen without an ecosystem. You are rattling off couple elements of it, high level operating system, low level operating system, infrastructure partners, everything has to connect. So I’m curious, how are you pushing the ecosystem to operate at or near your speed? Because you might be able to operate at a certain speed, but if your partners aren’t moving and swimming in the same direction at the same speed, it makes growth of innovation harder.

Ram Velaga: No, actually you said very well, right? It takes in village to make things happen. And especially so in the semiconductor space because, you could go buy a television at Best Buy, but you don’t go buy this chip at Best Buy. The chip by itself is worthless to most consumers. There is a whole product that needs to be built around this, and that requires an ecosystem. Another interesting data point I would put out there is the rough math that I do just tells me for every dollar that we invest in R&D, there’s probably about six to $8 that our partners would invest in this to eventually have a finished product that they take to the market.

So it actually puts a burden on us when we think about it that way because if you think about it, for every engineer that we have, that means roughly they have six to eight engineers there. So if I’m late to the product by a month on an unpredictable schedule, think about the inefficiency that I multiply out there in the marketplace, because there’s a whole bunch of engineers. It’s like you’re waiting for your plumber to show up and you decide that instead of driving to work, you’re waiting at home and he doesn’t show up. You’ve lost your entire day. And think about that frustration times everybody that’s working in that ecosystem.

So before we push the ecosystem, we are extremely clear that having a predictable schedule is important and we drive to a predictable schedule. But two, when we actually start to look at what we need to build based on what our customers are telling us, and sometimes what we anticipate the customers will need, we go back and say, “Look, hey,” whether it’s TSMC or somebody else who’s a fab, “Your process node is not scaling that much anymore. So how do I go about and build it?”

Not long ago when we built chips, they were half the size of this. And if you built a chip that was over 400 millimeter square, people would start to look at you saying, “Hey, something’s wrong with you.” But then now, unless I build a chip that is 800 millimeter square, that is almost theoretical size of what a large fab can support, we can’t build the chips that we are building. Which means we work very closely with our fab foundries to say, “Can you optimize your process so that even when you’re building an 800 millimeter square chip, you actually have the yield that you need?”

Then you have to go to a packaging in a company and say, “Can you build a package that is the size of this chip?” And packaging companies, as you know, they’ve never really had enough money to be profitable, so they never really invested. And if they don’t invest over a period of time, what happens is you go and you say, “Look, I need to build a big package, I don’t have a product.”

Patrick Moorhead: It always used to be a downstream thing too. You do the design, you get the IP and throw it over the wall to the packaging folks.

Ram Velaga: Exactly. So you have to think about this whole thing and say, “How do you go through every one of the supply chain partners and make sure the whole thing comes together?” And I’ll give you one other example in this domain. We put a ship out there, but if you don’t have an operating system that runs on top of this, you eventually don’t have a networking product that a web or some a web-scale or somebody else is able to consume. So we are actually now the largest contributor to an open source software called SONiC, which is actually started by Microsoft.

Because Microsoft said, “Look, I don’t want to necessarily have a closed operating system which if it breaks while my network is running, I have to call somebody else and they won’t tell me where the problem is. Why don’t I start my own operating system that’s open source?” We are the largest contributor to it. So then now an end customer can take a chip, go to Taiwan or some other place, get a white box built, and they can actually run SONiC on top of it and actually have a fully functional box. So that’s how we just think about the whole ecosystem.

Daniel Newman: Yeah, I like the metaphor about the drive and the, I think we call it the snowball effect. There was a movie called Sliding Doors, I still remember too how each time that we get thrown off in your life, how that chain of impacts of different things that happened down the street. And in semiconductors, I think we’ve noticed that again, the supply chain shortage probably was the first time that the whole world really saw what happens.

But of course there’s vendor level issues too, where the supply chain could be okay, but someone along the way committed something at a certain point and then all of a sudden the OEMs don’t have the product and then their enterprises don’t have the product and then the applications they’re trying to build. So let’s flip to the other side of this though. When Broadcom gets it right you tend to be far up the stream, you tend to be building things that are needed for companies. What is the impact of that innovation? How does a company like Broadcom measure the impact as well?

Ram Velaga: So I would actually go back to the impact of our innovation and the example that I gave you before. What used to cost a million dollars and 20,000 watts of power now is less than $10,000 and 500 watts of power. And then you say, “Okay, fine, but what does it all mean?” If you think about it today, and if you actually think about the evolution of computing, we went from clients to mainframes with your friend Greg. All right. And then over-

Daniel Newman: He’d be so happy, by the way, you’ve mentioned him.

Ram Velaga: Absolutely. And still some of the most important things in the world get done by mainframes and whatever Greg does to help them out. But then we went from there to PCs and then eventually the whole idea of distributed computing started to take hold. And the people who pioneered those are people like Google and the others built very, very large data centers and the applications were written. So then they took advantage of this compute being spread out. And then Sun, if you think about it, 30 years ago, coined the phrase The Network is the Computer. And I actually didn’t think about it for a long time what it meant.

Patrick Moorhead: It’s digital equipment, yes.

Ram Velaga: Yeah. I never actually understood what it meant. But now when you think about it, all the applications you experienced today, whether it is an Uber or a Google search or anything that comes from Meta’s platforms or anything else, all of that is based on distributed computing. And the distributed computing essence of it is that any one computer cannot be big enough. So you have a whole bunch of computers, hundreds of thousands of those, and they’re all connected by the network.

And so we are the ones that are actually enabling this network happen because it’s like plumbing. You don’t think about it until your plumbing stops working and then everything breaks loose. So that is what we do, which is we have to continuously keep coming out with the bandwidth at a lower prices and lower power so that the distributed computing as you’re experiencing in your digital life, every day is seamless. And that’s the impact. And that’s why I say Broadcom, most people don’t know it, but we connect everything.

You think about your phone, your WiFi is a Broadcom chip in there. It connects to your access point, which is a WiFi access point, which is very likely a Broadcom chip in there. Then it connects to an ethernet port that leaves your home through your CPE device that a service provider drops off, likely a Broadcom device in there. From there it goes to service provider, metro network, eventually to the data center. We’re connecting everything.

Patrick Moorhead: Yeah. I tell you, I look back and I was a major supplier of a hyperscaler when I was at that unnamed semiconductor company. And I remember, you used the word democratize and what you did to democratize networking at the hyperscalers was just amazing. And it used to be you would go in and quite frankly go through a very expensive network provider. And you shifted the model to an ODM model that literally changed everything out there. And networking is the only way to do distributed computing, otherwise everything would have to be in the same place.

And the great news about the future is, you had mentioned this and we’ve all sat through presentations of the explosion in data, but it is true. And it’s data, not only that’s being created by businesses and consumers, but now by machines. And the requirement too, for the low latency and high performance puts that extra added challenge. So, Ram, you’re going to be busy, you and your team for a long, long time.

Ram Velaga: We’d like that.

Patrick Moorhead: Yeah, for sure.

Daniel Newman: It’s all about going fast like a 911 or a Cayman.

Ram Velaga: Exactly.

Patrick Moorhead: Yeah.

Daniel Newman: Ram, thank you so much for joining us.

Ram Velaga: Yeah, thank you.

Daniel Newman: Really appreciate having you.

Ram Velaga: Appreciate it.

Daniel Newman: All right, everyone, there you have it. The Six Five Insider Edition is all wrapped up. We had Ram here from Broadcom, we talked a lot about M and A, we talked about innovation, we talked about the challenges of when one part of a design gets broken< and we talked about how companies have to work through it. But Pat, I think for this show it’s time to say goodbye.

Patrick Moorhead: It’s great. I love Insider Editions, especially when they’re about semiconductors. So-

Daniel Newman: Anything about chips.

Patrick Moorhead: Anyways, hit that subscribe button, and we really appreciate you coming in. Thanks, and have a great morning, afternoon, night, wherever you are on the planet.

Patrick Moorhead
+ posts

Patrick founded the firm based on his real-world world technology experiences with the understanding of what he wasn’t getting from analysts and consultants. Ten years later, Patrick is ranked #1 among technology industry analysts in terms of “power” (ARInsights)  in “press citations” (Apollo Research). Moorhead is a contributor at Forbes and frequently appears on CNBC. He is a broad-based analyst covering a wide variety of topics including the cloud, enterprise SaaS, collaboration, client computing, and semiconductors. He has 30 years of experience including 15 years of executive experience at high tech companies (NCR, AT&T, Compaq, now HP, and AMD) leading strategy, product management, product marketing, and corporate marketing, including three industry board appointments.