"The most impressive person I've met": Nvidia's CTO on Jensen Huang, $5T valuation, and the secret formula

On Wednesday, October 29, Michael Kagan did not deviate from his planned schedule. He attended the Nvidia conference in Washington along with ten thousand other people. But here and there, between lectures, meetings, and mingling, he received an update: Nvidia’s value had passed the $5 trillion mark, less than four months after reaching $4 trillion. No company has ever reached such a valuation, and Nvidia has held the title of the world’s most valuable company since June 2024. It achieved these numbers in large part thanks to its connection with Israeli company Mellanox, which means it also owes much of this success to Kagan, one of Mellanox’s founders and now Nvidia’s global chief technology officer (CTO), the most senior Israeli at the company.

“I received the update, and I thought about the phrase we’ve been saying since we reached the highest valuation in the world: ‘Play the game and the score will take care of itself,’” Kagan tells Calcalist in an exclusive interview.

And that’s it? No excitement?

“Every time you cross a trillion, you can’t be indifferent to it. And we’ve crossed it five times. But the company’s valuation, the stock price, it doesn’t really speak to me. It’s not that I don’t care, but it’s not the main thing that excites me.”

This becomes clear when talking to Kagan. The moments of true excitement come when he speaks about his personal life, about his grandchildren, or when he mentions, “I was excited to hear that my childhood teacher was proud of me.” But he is most animated when he talks about his work.

Kagan, an electrical engineer by training, is today one of the world’s leading experts on AI. Market value, trading, and historical milestones interest him less. It’s clear he’s not entirely comfortable discussing the stock market. But when he talks about technology, he lights up. When he describes the revolutions he’s been part of, his eyes sparkle. When he explains AI, he’s as enthusiastic as a child, even at 68. While investors focus on the stock, Kagan talks about the employees. When the world frets or speculates about the AI revolution, Kagan dives into the technical “wires,” as he calls them, breaking down the infrastructure with professional precision and vivid metaphors.

So, just moments after describing his indifference to the $5 trillion milestone, he adds: “If there’s an idea that a product is built on, that excites me much more than valuation.” (For the record, the valuation has since dipped to about $4.7 trillion. Kagan doesn’t seem particularly concerned about that either.)

So what excites you now?

“The latest generation of the GPU, our new processor designed for artificial intelligence models. This is the first time Mellanox and Nvidia technologies have been integrated into a single system. Imagine this processor as a unit with a million and a half components working together. The main elements are Nvidia processors and Mellanox chips, enabling extremely fast and efficient communication. That excites me much more than what happens on the stock market.”

These are the kinds of numbers that thrill him, the tangible ones. He makes no claim to understand market valuations. “A man walks down the street with a dog and says, ‘I’m selling it for a million dollars.’ Everyone thinks: ‘Who would pay a million dollars for a dog?!’ A week later, they see him without the dog and ask, ‘Did you sell it?’ He says, ‘Yes, for a million dollars.’ ‘What, really a million? In cash?’ He answers, ‘No, I got two cats, each worth half a million dollars.’ Maybe some people know how to value a dog compared to a cat. But I don’t know how to value abstract things, I’m an engineer. I can’t say whether a company’s value is real or not. I only know how to talk about a concrete product that solves a real problem.”

And what you do, is it real? Real enough to be worth $5 trillion?

“Yes, what we do is real. In the early 2000s, we needed an entire sports hall to build a computer with this kind of capability, and we did. Today, those capabilities exist in computers the size of your palm. In the age of AI, the basic computing unit is a data center, and it depends on fast, efficient communication, like the nervous system of the human body, and that’s what Mellanox brought to Nvidia.”

You often compare AI to infrastructure. For most people, the whole concept feels much more abstract.

“That’s exactly the revolution. The story of AI is like our dependence on electricity, it will become infrastructure for everything. You’ll use it without thinking, just as you use electricity or water without wondering how it reaches you. When you’re driving and your car keeps you safely in your lane, that’s AI. Or in a smart home. Even with electricity, people didn’t initially understand how far its uses would go; they started with the light bulb, and now look, we have electric cars and smartphones. The same is true of artificial intelligence. We’re only at the beginning.”

And all this infrastructure, it’s yours?

“Almost every place in the cloud connects through Nvidia ‘wires.’ Most AI models run on our infrastructure. When you ask ChatGPT something, the data travels through Nvidia processors and is processed in the cloud on our chips. You don’t feel it, just like you don’t think about electricity or the battery when using your smartphone.”

And that’s what you saw coming? That’s why you didn’t leave after the Mellanox acquisition and chose to stay with Nvidia?

“Yes. I knew. I understood that I was getting on board a train bound for a revolution that would change the world.”

“My pension is a job, and Moore’s Law is already dead.”

Kagan is married to his second wife, Iris, who worked with him at Intel (“Without her, I wouldn’t have gotten where I am. I envy myself for having a wife like her”), and together they have six children, three from his first marriage, two from hers, and one daughter together, and 14 grandchildren. Five of the six children, aged 26 to 46, work in high-tech; one is a teacher. He likes to travel with Iris, ride bikes, and take photos, but mostly, he works a lot. “A third of the time I’m in the United States, mainly in Santa Clara, California. I travel there once a quarter for a month. It’s not easy, but at Mellanox it was harder, more trips, and shorter ones. In Israel, I have a house in Zichron Yaacov and an apartment in Tel Aviv. I work from Tel Aviv and Yokneam, and I work a lot with the United States. I have meetings in the evening, even on Fridays. I live in two time zones. I finish meetings in the Tel Aviv office late in the evening and go downstairs to walk around a bit. That’s why I prefer to work from the office and not from home, in Zichron Yaacov there’s nowhere to walk around.”

Don’t you ever feel the need to rest a little? To make more time for family and hobbies? You’ve done a lot, you’ve earned a lot.

“It’s not that without the job I’d be hungry for bread. I do it because I enjoy it. If retirement means doing whatever you want and getting paid for it, then I’ve been retired since I was 25. I enjoy it, I have an impact, and I help people succeed in doing their life’s work. And if I’m having fun, I stay. I’m not on any exit path.”

Do you feel your age at work?

“I have experience, and experience is good. I can’t program like I did when I was young, but now I have AI. It does everything faster. It democratizes capabilities, allowing billions of people to use intelligent tools.”

But everything is moving so fast. You’re at the center of the fastest-growing field in the world.

“That’s right. Until 15-20 years ago, things worked according to Moore’s Law, every two years they built a faster processor. In the world of AI, the model of doubling computing power every two years is no longer relevant. Now it doubles every quarter.”

How do you deal with that?

“We have a program that supports 1,200 artificial intelligence startups in Israel, guiding them with an emphasis on finding technological solutions to real problems. And the connection between Nvidia and Mellanox especially allows us to keep up. Nvidia without Mellanox couldn’t have done it, it didn’t have the communication technology.”

“I’m recruiting 5,000 employees every day, and AI will change Israel.”

The American Nvidia was founded in 1993 and initially stood out for its graphics processors. The Israeli Mellanox was founded in 1999 and developed products that accelerated and optimized communication between various computing components. In March 2019, when Nvidia was worth a little less than $100 billion, it announced the purchase of Mellanox for $6.9 billion. The deal was completed in April 2020. It was born from years of collaboration and mutual respect between Eyal Waldman, one of Mellanox’s founders and its chairman and CEO, and Jensen Huang, Nvidia’s CEO, president, and founder. “Jensen always told him: ‘I don’t initiate purchases, but if there’s a sale, I’ll make the last offer, and at the highest price,’” says Kagan. “The connection between the companies enabled the integration of all the body’s organs, the ones Nvidia produced, with the nerves and arteries that connect them, which is what Mellanox produced.”

You built an empire here and then sold it, when it was a leader in both the Israeli and global industries.

“We didn’t want to sell the company, but offers came in and we had no choice, we had obligations to investors. And we were sad because we wanted to build an Israeli company. But now everything is on a different scale. We wanted to build a big Israeli company, and we did, just not in the way it could have been under Nvidia. It’s hard to imagine a bigger impact than that on Israel. We didn’t make an exit; we made a sale. And clearly, this sale was the best thing that could have happened to the company.”

Why?

“Because we brought here a company that is changing the world, and we are at the center of it, Nvidia Israel is at the center of Nvidia.”

What does that mean in numbers?

“Our operations in Israel are responsible for a little more than 20% of Nvidia’s revenue. Their revenue from us in one quarter is more than what they paid for us.”

You put Yokneam on the map.

“It’s exciting. We founded the company here in 1999, on a site that looked like a toll booth near Soltam. We wanted to use the money we got from investors in the most efficient way. Back then there was no cloud, we bought computers that came in cartons. I remember the pile of cartons on the floor, the door we took off its hinges to use as a desk. When my wife Iris came to visit with our daughter, she was looking for a place to breastfeed, so I put up a large cardboard box for her to use. After a year, we also opened offices in Tel Aviv, but we started in Yokneam.”

And now you’re on your way to huge offices in Beersheba, about 3,000 square meters, with hundreds of new employees, and at the same time you’re looking for a site for a new campus in the Haifa area, around 100 dunams, with an investment of about 2 billion shekels. Local authorities in the area are competing for the right to host your operations.

“From the moment it was announced that we were looking, I haven’t stopped receiving inquiries. But I don’t deal with the location, I passed the inquiries on. The interest of the local leaders is clear, but from our point of view, the main consideration is accessibility for as many employees as possible.”

How involved are you in the decision to expand operations in Israel? Did you push for it to happen?

“I wasn’t involved in the decision itself, but operations in Israel are important to me. The expansion here stems from the company’s trust in local human capital. There’s an extraordinary concentration of excellent, creative engineers here who understand systems in depth and operate at a pace that fits the world of artificial intelligence. That’s why Israel continues to be one of Nvidia’s most important growth centers globally.”

And why in the north and south?

“Because in Israel, talent isn’t only found in the central region. We want to give talented people with a passion for building great things the opportunity to work here, and ensure that where they live won’t be a barrier. That was true in the past too, when we were told at Mellanox to outsource to India, we said no. Instead, we went to Tel Hai and to the Palestinian Authority.”

And do you still employ Palestinians?

“Yes, we have more than 100 employees in Rawabi, Nablus, and Ramallah, and they’re an integral part of our operations. We see great importance in this bridge, and we continue to cultivate it today.”

When you joined Nvidia, you had about 2,000 employees in Israel (out of about 3,000 globally). Today, you have around 5,000 employees out of Nvidia’s 36,000. How many more will you recruit?

“A few hundred this year, and more next year. It also depends on market conditions, but from our perspective, we’ll continue recruiting at the same pace.”

Despite being a technology man, it seems that Kagan’s excitement for his employees equals his enthusiasm for processors and computing power. When asked, for example, whether it isn’t thrilling to surpass Apple (nearly $4 trillion) and Microsoft (around $3.7 trillion) in value, he instead talks about people. “The thrill is in running a company that people think is a good place to work. Getting to number one is hard, maintaining it is even harder. Every day I ask myself how to recruit 5,000 employees.”

What do you mean? You already have 5,000 people.

“That’s exactly what I mean. When people ask how many we hire, I feel like I have to recruit 5,000 people every day. At the end of the day, they go home, and I have to get them to come back the next day.”

You reward them well. Many of them, at least 2,000 of the original Mellanox employees, it is estimated, have become millionaires as the stock price soared.

“Exactly. Some of them don’t have to keep working; they don’t work for the money. Our job is to create an environment where they feel good, where they do something meaningful, where they can fulfill themselves through a technological challenge. That’s the greatest reward in my opinion, more challenge, more interest, more fun. Allowing them to create, to invent, to develop a new dimension in the universe, to be part of the revolution.”

Won’t this revolution make some of them redundant?

“This is nonsense. Some professions will change, and some will disappear, but many more will be created. Look at the smartphone, how many people make a living from developing mobile apps? That’s a new profession. If AI increases a programmer's productivity so they can do ten times more than before, you might think ten times fewer programmers will be needed, but it’s exactly the opposite. More software applications are being born that didn’t exist before, creating entirely new worlds of employment for software professionals.

“The coachman became irrelevant when the car arrived only if he didn’t realize he wasn’t in the horse business, he was in the transportation business. That’s why AI has tremendous social potential: it can create jobs for people who never studied formally but know how to ask the right questions. That’s what matters today. Take, for example, the ultra-Orthodox community, even those who didn’t study core subjects can integrate. AI can narrow gaps, change society, and change the world."

And also create many new problems, like the flood of fakes.

“This is a social problem, not a technological one. Every society faces challenges that need to be addressed. There already are, and will be more, tools to detect fakes and identify AI-generated content. In the United States, for instance, some teachers already use technological tools to check whether papers were written with the help of AI."

And technologically, where will the challenges come from? Competition from the East, for example? China is trying to close the gap and develop its own AI infrastructure.

“There are geopolitical issues I don’t want to get into. I once had a teacher who told me that to win a marathon you have to run as fast as possible, and constantly accelerate. That’s true: you have to keep running. You’re always being chased. To stay relevant, you must do better. Even if you’ve baked a good cake, you have to constantly think about how to bake a new one, how to open up new dimensions. Taking a GPU and transforming it from a graphics processor into an AI processor, that’s opening a new dimension."

“I worked on building miracles, and Intel wasn’t moving fast enough.”

Kagan started his race early. He grew up in St. Petersburg, “a completely Russian family, no bar mitzvahs, no holidays, I never went to a synagogue. The family’s view of Judaism was: ‘A Jew - to be like everyone else, you have to be better than everyone else.’ When I got a 90 on a test and was so proud, my father asked where the other 10 points were. I explained that almost everyone else failed and that it was the highest grade in the class, but for him it didn’t matter.”

As a “completely Russian” family, other things loomed larger. “My grandfather was executed in the name of various causes, my father was imprisoned after saying something disrespectful about Stalin, and a friend informed on him. Only later did I understand what it meant to grow up under such terror.”

Both of his parents were physicists. “And ever since I can remember, I wanted to be an engineer or a scientist. I was always curious about how things worked. Even as a child, when I got a gift, I would first take it apart to understand how it worked, and then put it back together.”

So naturally, he tried to get accepted into academic studies in the field, but failed. Judaism, which had never been an issue before, suddenly became one. “I got a very low score on the entrance exam and wasn’t accepted. Then someone who worked at that institution and knew my family said to us: ‘With your last name, what exactly did you think would happen?’”

Was the dream shattered?

"The sky fell on me. Everything I had built in life collapsed."

What did you do?

"My father immediately said: 'We're leaving.' Everything, just so I could study. We immigrated to Israel. If I had been accepted, I’d probably still be stuck in Russia."

In 1975, at the age of 18, he arrived in Israel with his family. “I didn’t know anything about the country. Only two days before we arrived did I discover that people write from right to left here.” He joined an ulpan in Jerusalem (“I always thought of it as a city from fairy tales, like Atlantis, and I discovered a real city”), got married at 20, and was accepted to study electrical engineering at the Technion.

If he were starting his studies today, he says, he would have chosen a different field. “Because I’ve always been interested in how things work, and man is the marvel of creation, I would have wanted to understand how humans work. I would have studied computational biology.”

“But in those years, they didn’t teach computational biology. So I studied electrical engineering. The first year at the Technion was the hardest year of my life. I barely knew Hebrew or English, I was alone after being surrounded by friends all my life, and I washed staircases and dishes in a restaurant to make a living. By the time I graduated, I was already married with two children.”

Then he served in the Israeli army, and in 1983 he joined Intel, where he worked for 16 years, rising to senior management and eventually becoming the company’s chief architect. It was a stable, secure place that allowed him to grow, but eventually, it disappointed him. “The last and best-known project I led was the Pentium MMX. It was a great success. I also led development of the 860XP processor, which was 50 times faster than others, but Intel didn’t develop the software it needed.”

Is that why you left?

“I wasn’t looking to leave. I’d been getting offers for years, startups were popping up all the time. But when I felt like I was ready for something new, the offer from Mellanox came."

It seems Intel hasn’t been moving fast enough in recent years.

“I don’t need to grade them. We work with the company, Nvidia is currently investing $5 billion in Intel and we’re developing chips together. It has pipelines to markets that are less accessible to Nvidia, and vice versa. But yes, companies always need to think about how to move forward.

“Nvidia, for example, used to be a chip company, then it started making graphics cards, then computers, and now data centers. We’re building an infrastructure company. If you don’t move in that direction, at some point, you’ll be left behind. If everyone is talking about computers and you’re still focused on chips, you’re less relevant."

At Intel, Kagan met not only his wife, Iris, but also Eyal Waldman, who in 1999 invited him to co-found Mellanox. “I was 42 years old, remarried, with five children and a baby girl, when Eyal called: ‘Let’s start a company.’ He told me who else from Intel was joining, and promised that ‘together, we’ll change the world.’”

You like being promised revolutions.

"I was, and still am, in the eye of the storm of the three biggest revolutions ever, at Intel it was the personal computer revolution, then at Mellanox the cloud revolution, when we realized we needed to build the infrastructure for the cloud, and now AI."

Is it a matter of luck, being part of all of this?

"I don’t believe in miracles, I build them. Miracles happen when you work hard. You move forward with the tailwind of reality, which shows your direction is right, even if the path isn’t what you planned."

Did you consider leaving after the Mellanox sale?

"If it weren’t for Nvidia, I’m not sure I would have stayed. But Jensen offered me the position of CTO. I was parachuted into a technology center of people who had worked together for decades, and I was welcomed thanks to his leadership and spirit."

What is he like as a manager?

"He doesn’t do personal conversations, interactions with him are always in meetings. The intensity varies, and usually if he doesn’t approach you, it’s a sign you’re doing things right. He’s very hands-on, dives into details impressively, highly intelligent, articulate, and has an extraordinary ability to distinguish between the essential and the trivial. You can’t tell him stories, he sees right to the core of the issue. He’s the most impressive person I’ve met."

During the war, he emerged as a clear supporter of Israel, and personally followed the struggle for the release of Avinatan Or, your employee who was freed only about a month ago from Hamas captivity. Will he return to the company?

"His workplace isn’t going anywhere. I hope he recovers, and when he’s ready to return, we’ll be waiting for him."

You praise Jensen Huang’s management, but don’t you want to be number one yourself? You’ve never been until now.

"No. I’m not Eyal Waldman. If being number one is important to him, he’s said so many times. As far as I’m concerned, my wife is number one and I’m number two. I’m not suited to being CEO, there are people who do it better than me. And in general, life is more complex than a linear scale. I have my own world where I’m both number one and one hundred."