The 50,000x Problem

You Can Watch the Full Episode Here 👇

But this newsletter is not really about lasers, and neither is my show. It is about what happens to a person between two worlds, and Vivek Raghunathan is one of the clearest cases I have sat across from. We go back to our MIT days, which is why this felt less like an interview and more like two people comparing notes on the same crossing. The boy who boarded his first international flight at 22, having never left India, is now the CEO of Xscape Photonics, an NVIDIA-backed company built with four Columbia academics, three of them professors, one a MacArthur Fellow. The science came out of their labs. He is the industry outsider among academic royalty. And he runs it. How a first-generation immigrant ends up leading the very people who taught the field, and what from India he used to do it, is the real story of this episode.

Everyone in India knows the IIT-to-America pipeline. Almost nobody talks about what it feels like from inside. Vivek's first surprise was not culture shock. It was welcome: MIT had assigned him a buddy who picked him up at the airport and walked him into his new life. His second surprise was that he did not feel foreign at all.

What he carried from India: the mathematical foundation and competitive fire that clearing the IIT entrance exam demands, and an independence learned in the hostels, where friends become your family because they have to. What he had to set down was subtler, and it is the most Indian part of his arrival story. Everything in his education until then had been structured: a defined syllabus, a known path, success by conquering the curriculum. MIT, he told me, was the first place he learned how to learn, to walk into uncharted territory, define his own problem, and solve it. He calls that his first teaser of what building a company would feel like. I heard something else in it too: the quiet unlearning every one of us from structured systems goes through here, where the rules stop being handed to you and the blank page becomes the test.

One more arrival detail stayed with me. His resume had no photonics on it, none, when he knocked on the door of Professor Kimerling's lab at MIT. He was a metallurgy undergrad with a physics minor and an obsession. Kimerling bet on the passion instead of the paper. Eighteen years later, Vivek has never left the field. Somebody opened a door your credentials could not have opened. If you have made a crossing of any kind, you have a version of that story, and you remember the name.

The company matters, so here it is through the lens that actually explains it. Vivek spent fifteen years as a translator: at Intel, Rockley, and Broadcom, turning research into products that ship, including hundreds of millions of dollars of transceivers for hyperscalers. The bet he eventually made rests on an argument worth hearing in full, because once you hear it, you cannot unsee it.

The AI industry has spent years obsessing over faster processors. But your brain is not fast because its neurons are fast. It is fast because it is communication-heavy: by Vivek's telling, the brain spends roughly 35 times more energy on communication than on processing. The AI clusters we have built are the opposite, processing-heavy, thousands of GPUs that burn enormous time and power just moving data between one another. When a processor and its memory sit side by side, they talk at full speed. Send that same data across a cluster and the speed collapses, by orders of magnitude, because it has to squeeze through copper interconnects that physics has pushed to their limit. The bandwidth that can physically get out of a chip into the wider cluster is called escape bandwidth, and it is the wall the whole industry is now hitting.

Xscape's answer is light. A laser that can emit up to a hundred colors down a single fiber, each color carrying its own stream of data, where the industry standard tops out around four colors because every color traditionally needs its own laser. One programmable multicolor source instead of racks of discrete ones. The vision: a thousand GPUs that behave like one giant brain, because the distance between them stops mattering. And here is where the translator part matters: Vivek had already published on this exact bottleneck at Broadcom, so when three Columbia professors showed him the laser, he was maybe the one person positioned to see both what it was and what it was worth.

And the whole industry knows it. Ayar Labs, valued near $4 billion with NVIDIA and AMD on its cap table, builds optical I/O chiplets. Lightmatter, valued higher, builds Passage, a photonic layer chips sit on. Celestial AI was just acquired by Marvell. Broadcom and NVIDIA are already shipping optical switches, and NVIDIA has poured billions into photonics this year alone. So does Xscape's edge survive that crowd? I think it does, for a reason I find genuinely clever: almost everyone else is building the engine. Vivek is selling the light. Every optical interconnect in that list needs a laser source, and programmable multicolor sources that work in a standard foundry are the scarce ingredient. Investors initially dismissed selling lasers as a commodity race to the bottom. The same plan is now viewed as one of the most scalable models in the space. If multicolor wins, Xscape is less a competitor to those companies and more the arms dealer to all of them. He was honest about the caveats on air, at least one other startup is chasing hundred-color lasers, and incumbents move fast with fresh billions, so the moat is real today and needs defending every quarter. But the deeper pattern is the immigrant one: the translator saw value where the specialists saw a commodity.

Now the question I most wanted answered: how does a first-time CEO lead co-founders who are literally the professors the industry learns from? I expected an American answer about flat hierarchies. I got an Indian one.

Vivek took me to a government office back home, where a 35-year-old district collector commands the respect of a 60-year-old secretary who has held his post for decades. Not because the collector is smarter or more experienced, but because the role carries the authority, and everyone in the room honors the role. He watched that as a boy and carries it as a CEO: respect the role, not the resume. His American translation was football, a more decorated lineman still runs the quarterback's play, and nobody calls it an insult. Titles, trophies, and age stay outside the huddle. What moved him most in his first meeting with his co-founders was their side of the same bargain: world-famous scientists humble enough to embrace someone junior as their potential leader.

Underneath that framework sits a temperament, and I want to describe it precisely because it is not the usual CEO humility performance. Vivek has simply never considered himself the smartest person in the room, going back to MIT, where in his eyes that title always belonged to his professor. The feeling never left, and instead of fighting it, he built with it, deliberately surrounding himself with people he can learn from, because learning is the thing that actually drives him.

I have sat in enough boardrooms, banks, and fund meetings to know how rare this combination is. Most organizational politics is people relitigating the org chart in every meeting. Vivek's team settled the question once, and it freed them to argue about the work instead of the pecking order. A childhood spent watching how respect works in Indian institutions turned out to be the exact operating system for managing a MacArthur Fellow.

Written into Xscape’s company language is the value “show up for each other like family.” When I asked where it came from, Vivek gave one of the warmest answers of the episode: his father and grandfather back in India, men who would drop everything and arrive at a relative’s door the moment trouble struck. No questions, no ledger. He has seen the same instinct in his wife’s family. He wants the people at Xscape to bring some of that care and responsiveness to one another when the company faces a difficult moment.

As an aspiration for how colleagues treat colleagues, I find it beautiful, deeply Indian, and immediately recognizable. I saw my own family in it.

Another story from the conversation added nuance to that idea. Vivek spoke about hiring exceptionally talented people who ultimately did not share the team’s values, and making the difficult decision to part ways with them, even when doing so risked delaying the product. As an operator, I understand the choice. Culture has to be protected, particularly when a company is moving quickly and the stakes are high.

It also made me reflect on what founders are really trying to communicate when they describe a company as a family. A workplace can offer care, loyalty, trust, and a willingness to show up for one another. But it also carries responsibilities and standards that are different from those inside a family. Perhaps the healthiest version of the metaphor is not unconditional belonging, but reciprocal humanity: asking people to care deeply about the mission while ensuring that the institution treats them with the same honesty, dignity, and consideration in return.

What Vivek appears to be building is a team with unusually strong personal commitment, one where people are expected to support one another while remaining accountable to the work. The closest comparison may be a championship team: shared sacrifice, mutual trust, and clear standards, held together by the belief that nobody succeeds alone.

The phrase “like family” can mean many things inside a company. In Vivek’s telling, it seems less about blurring personal and professional boundaries and more about carrying a specific instinct from India into American business -when someone beside you is in difficulty, you do not look away.

Every episode sits with money, never what a guest has, always how they think. Vivek's relationship with it may be the most settled I have encountered on this show. Leaving Broadcom, he knew the golden handcuffs would only get more expensive with every year he stayed, so he ran a simple test: at 80, grandkids around him, which choice would he regret? Staying was the only answer that scared him. And when I asked what 'enough' looks like now, he refused to give me a number. Enough, he said, is waking up at 80 without a backache, still adding value, needing no medication. Then he laughed and added that he does not even get paid that much.

By “family of 200,” Vivek was referring to the wider circle of employees, contractors and households whose livelihoods are connected to Xscape’s success, not to an employee headcount of 200.

Notice what roughly $95 million of other people's money changed in him: not his risk appetite, his sense of obligation. That is the immigrant relationship with money I keep finding on this show, money as responsibility carried, not status displayed. And notice, too, that when he reached for the weight of it, the word he picked was family. The metaphor I questioned above is also the truest thing about how he carries the job.

We closed where this show always closes. What of India will he never let go? Food, instantly, and when I pressed for one: idli. What piece of America has he made completely his own? Capitalism, no hesitation, no apology. And the thing about living between two worlds that only he understands: that beneath the apparent differences, human needs are always the same. Indians, Americans, everyone, just wanting to be appreciated, loved, and respected.

He thinks in English now, swears in English too, after two decades surrounded by people who do not speak Tamil. But Tamil lives at home, and the last hours of every day belong to his two-year-old daughter, where he and his wife are deliberately planting Indian values in an American childhood. I have a four-year-old at home, and I felt that answer in my chest. This is the quiet work of every crossing: you optimize your days in the language of the country you chose, and you protect your evenings in the language of the country that made you.

Clear lessons readers can carry into their own works

Vivek learned it watching Indian institutions and applies it to a MacArthur Fellow. Settle the authority question once, explicitly, and stop relitigating the org chart in every meeting.

Vivek does not perform humility; he has genuinely never felt like the smartest person present, and instead of fighting the feeling he built with it, surrounding himself with people he can learn from. The feeling keeps him curious. The choice built his founding team.

Culture is what people do when you are not around, and talent without shared values creates friction faster than progress.

Foundries serve their billion-dollar customers first. Plan your timeline, and your fundraise, around their speed, not your ambition.

For irreversible decisions, ask which choice you would regret at 80. Vivek's answer took him out of Broadcom and into the bet of his life.

The full conversation goes further into how five very different founders found one another, what Vivek had to unlearn moving from engineer to CEO, why deep-tech startups can only move as fast as their partners, and the investor advice he deliberately chose to ignore.

If you resonate, subscribe there, and forward this to one builder living between two worlds. That is how The Second Whistle gets heard.

Long-form conversations about what people built, and what built them