Why People Matter More Than Science in Deep Tech with Shahin Farshchi

People thought of venture as like a club of old white guys that play golf with each other. I used to spend so much time digging into the technology, and what I learned was that I was completely wasting my time and that we are ultimately in the people business. I'm a glorified recruiter, and technology can get obsolete. Technology cannot work. So I could have diligence that technology to death, only to realize that 18 months later, It's completely irrelevant.

Shaheen Farchi, PhD in electrical engineering and ex-chip designer. He didn't break into venture the usual way. As a postdoc, he worked with Lux Capital for 18 months unpaid. 18 years later, Lux manages over $5 billion with early bets on Zoox, Relativity Space, and Mosaic. This is what that journey taught him. All right, everyone. Today we have Shaheen, friend and investor. He's a general partner at Lux. He got his PhD in electrical engineering, started a company, and now he invests in all kinds of companies from consumer software to very deep tech. He's invested in companies like Mosaic, Zoox, Relativity Space. And I think the thing about Shaheen that's really interesting is he has the entrepreneurial mindset, the investor outlook, but also the deep, deep technical understanding of things as complex as chip design, which is very rare. So very excited to have you here today.

Thanks for having me and excited to be here.

Yeah, awesome. Well, can start at the very beginning. Tell us about your journey from PhD to founder. When you decided you want to get a PhD, was it just out of curiosity or did you have this intention of finding something and turning that into a company?

Yeah, I have this tendency to inflict self-harm. So when I saw that, like, so my dad, for example, he failed his PhD entrance exam or his quals. After I was born because naturally he was busy dealing with a newborn. So I figured, okay, if like, if it was really hard for my dad to do, it must be something really hard. So I'll just do it because it's hard. So that's really what the main impetus for doing it. But from a career perspective, my intent was to solve a hard technical problem and then build a company around the solution to that problem, which I did. And it was a great lesson in that I learned the hard way that it's one thing to have a really interesting solution to a product or to a, to a problem or a technology. And it's another thing to have a real business. And I had the, the former, but not the latter. And so the great takeaway there was you can have the best solution to the hardest problem on earth, but it doesn't necessarily imply that anyone's gonna pay for it, that there is a business to be built around it. In fact, it's the exception. Most of the great tools and technologies we have around us, we consume for free. And it's a small fraction of those that we actually pay for. So the lesson, the takeaway for me there was like finding the ones that people pay for is an art in itself versus just coming up with an answer, if that makes sense.

So would you have done something differently with what you worked on? I would love to hear what was it that you built?

Yeah, so my PhD thesis was on building a small kind of radio that gets implanted into a mouse to be able to capture their brain signals and to offer that to scientists. I mean, like, it was, I mean, that technology existed, but it was tethered. It wasn't wireless. It lacked a channel count. Like if you hear Elon talk about Neuralink today, he talks about having many thousands of channels. They had like 2 channels back in those days. And so building something that has many channels that doesn't fry your brain when it's in there. So that's what I set out to prove to the world that you can do this. But like, that's not a commercially viable pro— you can't really build a product around it. It's like a technology that someone else would use to build a product around. So I shut it down and I decided to do it right second time around. So I was kind of thinking about my next move when I kind of stumbled into venture.

Want to go deeper? Talk to Shaheen's digital mind on delphi.ai. Link in the description. Yeah, so how did you stumble into venture and did you have to rethink the way you operate? Like what new mental models did you have to adopt?

Yeah, so that comes after. So the new mental models came after. So I was still very naive going into this kind of like next chapter of my career. So back in 2006, there was no Paul Graham, there was no Brad Feld, there was no kind of like first round, there was no information, there were no blogs, there were no blogs, no Twitter. Venture was a very opaque business. People thought of venture as like a club of old white guys that play golf with each other and fund each other's companies. That was my view of what venture capital is. And I thought that to register on any of their radars, I need to build and exit multiple companies so they would finally maybe at one point invite me into their club. But then there was a catch-22, which was like, how do I do this without raising venture dollars? And I was told that the best way to, to get the attention of a venture capitalist is to approach them without, without asking them for anything. 'Cause VCs are accustomed to people reaching out to them, asking for advice, asking for jobs, or asking for, for money. So if you reach out to them offering them something and asking nothing in return, that at least will make you stand out. From the many other people that reached out to them. So back in 2006, nanotech, I don't know if anybody remembers, but nanotech was a buzzword. Nanotech was this magical technology, which was really just glorified chemistry that was used by academics to raise, you know, large amounts of money from the government to do research on like building these magical materials with these crazy amazing properties. And so I reached out to a bunch of firms like info@sequoia.com, you know, info@whatever, you know, benchmark.com. And I just said, hey, I'm a postdoc. So back then I was postdocing to pay my bills. I'm a postdoc. I did electrical engineering. I know microfabrication, nanofabrication, semiconductors. If you're remotely interested in these areas, I can send you opportunity, like scouts, like scouts, not even a term back then. I can send you opportunities. I can do diligence for you and I'm not looking for anything in return. I just wanna build a relationship. And a lot of firms said like, either they didn't reply or if they did reply, they said, yeah, yeah, sure. Send deals to this email address. And then I Googled nanotechnology venture capital and Lux was the first Google search results. And I go on the website, luxcapital.com, I see a bunch of guys in their 20s trying to like dress and pose like as if they're in their 40s, based in New York. I thought to myself like, why would a venture capitalist be in New York? Who are these young guys? None of them had technical backgrounds., but they seem like very, very ambitious people. So I said, well, what do I have to lose by reaching out? So I reached out with the same messaging and I got a call the same day and they said, hey, we're raising our first institutional fund. We want someone on the West Coast to go out there, hustle, put together a deal flow so we can show it to potential limited partners, fund investors, and raise this fund. And I said, sure. So I would basically, uh, apply to present at technical conferences that were taking place in New York and have basically the school send me out to present at these conferences in New York. And I would basically hang out with the Lux guys. Long story short, you know, I did this pro bono for a year and a half, and then when they raised that fund, they called me and said, hey, do you wanna join us? And I had no idea what that meant. Like, so what, what does it mean to be like, oh, do you wanna be an associate here? I'm like, what does that mean? What is, what is a venture capital associate? So they said, it's not a whole lot different from what you're doing right now. You're sourcing opportunities, you're helping us with diligence, you're spending time with the founders that we end up funding. It's just basically doing that. I thought that's, that's a pretty amazing job. And I was very impressed by the founding team of the firm and I wanted to be part of whatever they were doing 'cause I knew they're gonna do great things. And so here I am 18 years later, we've since raised 10 funds totaling $5 billion, current funds a billion. We cut checks from $50,000, we cut checks at $100 million, and we can invest from de novo kind of idea stage. I met Dara, you know, shortly after he'd started the company, and that was kind of like on the, on the earlier side for us. Um, but our sweet spot is kind of seed and Series A, and then we have the ability to be a long-term capital partner., with our founders. So there's an example where I met a couple of guys at Apple. They were thinking about starting a company. We ideated on business plans together. 6 months later, they start a company. We're a, a co-founding investor. Company went public 6 years later. I stayed on their board and, and they, they thankfully, uh, asked me to step down to bring on a real professional public company board member. Uh, but that's an example of how, you know, you, we tend to work with people very early on. All the way through being a public large company. Now they're like 1,500 people and doing great things. So that's the high level of the journey.

And Lux was one of the earliest to have an opinionated stance on we're investing in deep tech in a time where everyone thought it was too risky and too capital intensive. Was it ever tempting during the craze of software with Airbnb and Uber, the shiny object syndrome of like, hey, maybe we should not do that and, and focus on something that's more sexy. But you, you stuck through it.

We stuck through it outta necessity because, uh, Bill Conway, who basically put up the money for what was like Lux One, which was basically just his checkbook and Lux Two, he anchored. And that's when I joined. Um, Lux Two was our first institutional fund. He had this very, um, uh, insightful kind of advice, which was the world doesn't need a venture cap, another venture capital firm. This is back in 2000. The world doesn't need another venture capital firm. And so if you guys want to build a new institution, you need to focus on a niche where you feel like you're differentiated and you have to build an institution that would differentiate you from other investors. And so that was sort of our initial mandate that we stuck to. There was a lot of, uh, you know, there was a lot of, uh, of temptation to go after the shiny objects and ad tech back in the day to doing, you know, cloud, which emerged like in the late 2000s. And then it was, you know, SaaS was starting to become a thing and then big data and analytics. And then there was the first wave of crypto, then there was the first wave of AI, then another wave of crypto. And now we're in the second wave of AI post-Transformers. And, you know, we, we did invest in those areas where we felt like there was a strong kind of competitive advantage generated by some kind of innovation and technology. So we've stuck to our, our knittings in that sense, but we have admittedly kind of, you know, broadened our mandate to include all kinds of companies now. But again, trying to focus on companies that are solving hard problems, 'cause we feel like as if that's where our edge is.

Yeah. And how do you know if something is too hard or not? When, when is the line where it's like, let's double down on this?

So there's, there's two types of hard. So there's, there's, there's technical difficulty and there is, there is kind of scientific risk. So, so when, when there's scientific risk, that means there are fundamental discoveries that need to be made before a group of very talented engineers like you could turn this into a product. We feel, or at least I as an N of 1, I feel uncomfortable taking on that kind of burden of having to solve a fundamental science problem. I feel like the best, um, environments for those kinds of problems would be kind of like in the more academic institutions. Now, there are exceptions to that. Like we invested in Evolutionary Scale. Which has a big science component to it. And so there are exceptions where you can build a startup company that attracts the kind of, the kind of patient capital that you need to be able to do those fundamental scientific discoveries. But my personal preference has been to invest in teams that are taking a basic scientific, like, advancements and engineering that into a product that would underpin a monopolistic business. I feel like that's where my personal strength is. But as a firm, we have mostly done the companies that are, that are, that are doing more in engineering than they are in basic science. We've done a bunch of companies in the life sciences space that focus more on, on basic science, and that's a very different risk profile because there are clear rewards for de-risking the science by getting passed various stages of, of clinical approvals, of regulatory approvals, you build value in the company. That's not the case usually outside of the life sciences where you have to actually build a product, take it to market, generate some kind of metrics around your business, around margins and growth, which takes a very, very long time if you're starting from basic science, if that makes sense.

And speaking of patients, I, I, a lot of VCs and founders will say speed is the most important thing in, in a startup. In areas where you're spending a lot of money and you can't really afford multiple iterations, like you really maybe like have like a couple shots at goal.

Yeah.

How do you think about speed versus like, yeah, care?

100%. Like I have one company that doesn't have a couple shots on goal. They have one shot on goal. Like they have to raise $100 million to put out a product. And if that product is not a smashing success, then they're going to be in trouble. We're going to be in trouble. And so that is one of the challenges with, with deep tech where you do not have the, the luxury of being able to iterate on, on the product the way most software companies can. And that's where you're really betting on the company's ability to predict where the market is going. Although ultimately nobody knows their ability to execute and get it done, get it right the first time, because there's not going to be a second time and the level of confidence that there's going to be capital out there to continue funding the company when they deliver on these milestones. And a lot of times, like if you look at driverless cars back in 2016, there was billions of dollars going after these companies, but then 2021, that money kind of dried up. And so, um, it's just being, being mindful that if there's a feeding frenzy on behalf of investors today, that may not exist. Tomorrow. So if you're gonna be burning millions of dollars a day at some point in the company's future, like you have to feel really comfortable that that money's gonna be there to keep it going.

So you would advise taking advantage of the frenzy?

When you can get ahead and then make sure you have other kind of partners around the table that have a vested interest in your survival that have deep pockets. So like what a lot of these driverless car companies did was kind of marry themselves to the car companies to Uber to the hyperscalers in the case of Zoox, they partnered and ultimately got acquired by Amazon. And so it's kind of that kind of strategic thinking that really helps kind of hedge your risk. Yeah.

While a part of your job is evaluating technological feasibility, most, a lot of VCs also evaluating people.

Yeah.

And so I'd love to learn common patterns in both founders and their teams.

Yeah.

That you have seen be most directly correlated with success.

Yeah. So Dara, that hits close to home. So in my, in the earlier part of my career, I used to spend so much time digging into the technology and understanding like, you know, the physics and like the, the feasibility and doing all these reference calls with technical people, understanding like what is the actual risk here in, in reducing this to practice. And, and what I learned was that I was completely wasting my time doing that, which is why like you see a lot of non-technical people being amazing technology investors and that we are ultimately in the people business. As Dara said, we are investing in people who understand the problem, know who has the answers, and have the ability to recruit those people that have the answers and be able to punch above their weight with those people that they've recruited to be able to raise more money and hire more great people and create this virtuous cycle. That the kernel of that kind of virtuous cycle is what I'm looking for when I'm investing. So this is ultimately, uh, people ask me like, so what do you really do on a day-to-day basis? And I say, I'm a, I'm a, I'm a glorified recruiter. My job is to attract folks like you to take money from us and, and build and hire other great people like yourselves and punch above your weight and raise more money and hire more great people. And ultimately, eventually build a, a giant, successful, profitable business with, with, with, with massive enterprise value. That is what we're ultimately solving for. So it's all about people. And, and technology can get obsolete. Technology cannot work. Technology can be completely discarded and, and, and rebuilt from scratch. I can't think of a single one of my companies that did not go through the cycle of throwing their company, their technology in the garbage. And starting all over again. So I could have diligenced that technology to death only to realize that 18 months later it's completely irrelevant. And so ultimately we are also in the people business.

Love that. Has there been an investment that you've made that was very controversial? Most people were like, Cheyenne, what are you doing? And you decided to stick with it and then ended up being amazing and would love to learn like what kept you confident even though you had all this external doubt.

Yeah, I mean, Mosaic is, is one of— Mosaic up until the very end, in fact, up until like literally days before Ali offered $1.5 billion to buy it, um, was a, was a controversial investment. That company, um, what, what, what were they doing back in 2000? Their hypothesis was that there could be a business built around algorithms that make it cheaper and easier to deploy AI models. And, you know, AI infrastructure was not even a thing back then. It wasn't a term that was thrown around. And the name of the company was thesis.ai because it was basically being built around the PhD thesis of Jonathan Frankel, who was one of the technical co-founders of the company. And so like we couldn't even come up with a name for this company. And when I presented this to my partners as thesis.ai, like my partner's reaction was like, are you kidding? Are you joking? Is this a company or is this a science project? And the company kind of meandered for a good amount of time. Like they were building these algorithms that reduce the cost of, they were optimizing models for training. And they were reducing the cost of doing inference. And like, it just was, there was really no business until they found one company called Avalara, which I'd never even heard of before, that said, hey, like, you know, we will buy these like inference, you know, fine-tuning kind of products from you. And that's when the company became a business. But at that point, still it wasn't clear as to like, Why the hell should this company exist? And just over time they were able to identify like what problems the market was having as it related to being able to deploy AI models. And they started building these products around inference, fine-tuning, you know, training. And that the business suddenly went from like almost nothing to like the high double digits of millions of dollars in annualized revenue. And it became a prime target for a couple of deep-pocketed companies that wanted to bring that capability onto their platforms. And so, you know, it was, it was never obvious, but it's a testament to a, a strong founding team who had high conviction in what they were building being important and then being the best at what they were doing and being able to identify an opportunity and jumping on it. They say that being lucky means identifying an opportunity and positioning yourself to be able to capitalize on it. That was a great example of where a lot of people say, well, okay, well they got lucky. Well, well, yes and no, because they built something, an opportunity presented them, presented itself, and they were able to capitalize on it quickly, which what led, which is what led to that outcome there.

Probably felt pretty good. Yeah, it did.

Felt pretty good for everybody.

Yeah. On the other side, what has been the most difficult or darkest moment of your investing career and what, what did you learn from it?

Yeah. Yeah. The darkest moment was actually relatively recently. It was shortly after my, my first baby was born and I had a company that was— they had raised a large amount of money and they were concentrated on one customer. And that one customer churned and they were, they were burning a lot of cash in anticipation of serving this, this one customer. It was an infrastructure-heavy business. And there was this decision that we had to make. Either we basically cut our losses and scale down to basically a company this size and try to, and, and, and basically, you know, husband our cash and try to figure out a different product to pursue, or we just, I mean, we were burning like $5, $6 million a month, or we just keep on, you know, we had 100+ people burning $5, $6 million a month, or we just keep going, risking just, you know, running over a cliff. And there were different views on the board. There were different views in the management team. Some people on the management team wanted to go down that path of leaning down, fighting another day. Others were saying like, no, we gotta keep going. And, you know, we don't divert from this path. There were investors that said, no, we can't scale this back. It'll all go to zero, but it ultimately ended up going to zero. So that was a very, and this company was like a baby to me because I was part of it before its founding. And unfortunately, like, you know, the worst happened. It ran out of money and And right, and shortly after there were a bunch of companies that popped up that, you know, had we perhaps like given ourselves more time, we could have gone after those opportunities. But that was a very challenging time for me.

Is there anything you take from that that is actionable or changing the way you do things?

Oh yeah.

Or is it more situational?

Oh yeah, 100%, 100%. That for me was a learning experience that ultimately you are betting on a management team, and in that scenario, the management team had a certain path they wanted to go down. And you, you know, I, I had already made my bet on them, and so I can't like change course down the road because I disagree with them. Um, I, I, I should've just come to grips sooner rather than later, like instead of chafing. I should have just said, okay, I believe in you, which is why I backed you, and I'm going to support you down whatever path you choose to take. That should have been what I did instead of fighting and fighting and convincing and begging and pleading. Literally, my child was being born and the nurse was saying, hey, do you want to help out here a little bit? And so it was that much stress and involvement. But like now I look back and I think like, hey, you know, as a venture capitalist, like I'm not an executive, you know, you guys are the operators here and I'm, and I'm backing you guys. And so if you choose to do something, the best I can do is support you in whatever choice you make. So that was the big, big learning for me as opposed to thinking that I was doing the right thing and applying prior experiences and trying to apply it here to avoid what I thought was imminent, but there's a small chance that they would've been right and the company would've still been around right now and I would've been wrong and I potentially could've gotten in the way of that. So, big takeaway for me is that you back founders, you back management teams, you back companies, and you let them do what they wanna do.

Thank you for sharing that. Speaking of thinking about whether you should apply prior experiences, AI is changing a lot. About a lot of things, venture, how much money a company realistically needs. And then also, you know, Elon Musk famously was not like a rocket scientist and he learned about rockets and started SpaceX. AI allows anyone to pretty much learn an expertise very fast. Does AI change the way you think about who to invest in and what to invest in? Because maybe you would invest in a PhD before, but someone could potentially upskill very fast.

Yeah. I'm, I feel like we've made a bunch of progress. Like all the dumb questions that I would've asked at board meetings are now being answered by ChatGPT. So I feel like it's really helping me there just kind of like get up to speed on things. But I feel like there's still a ways to go for someone who is not an expert in the space, in a space that can help you solve a problem. Can now suddenly use AI to solve that problem. I think we'll get there eventually where the machine can do the task, and we're already seeing that. And for more rote tasks where like a machine can do the rote task a lot better than a human can, like preparing documents and stuff like that, they're pretty standard of which there's like millions of examples. But for the stuff that we do at startups, I feel like it's just still you still need a lot of institutional knowledge and a lot of experience. And so I feel like AI still has a ways to go, but maybe Delphi will get us there.

Maybe. Scott Belsky has this book, The Messy Middle. I don't know if you've read it.

I haven't read it. I'll put it on my list.

The middle areas of a company, pretty much where you grow from a startup to a company, the inflection point. And we are currently in that messy middle. And obviously a startup could go from 0 to 1 and fail miserably. I'm curious, I don't know how much of an inside look you've had at some of the companies you've worked at, but at that point where a startup goes to 0 to 1, what are the biggest mistakes you've seen that have prevented them from going from 1 to infinity?

Oh, how much time do you have? And so, so, so one big mistake is assuming that what got you to 1 will get you to N. What got you to 1 will, will not. Gets you to end. So one big challenge is product. So there's guys like Elon and Steve Jobs that had their fingers in everything and the company was extremely successful. I'm not sure if that's the rule. I think every company, just like how every person is different, every company is different. So in some companies that works, in other companies, You, you bring on the pros, you bring on the experts that are going to help kind of uplevel the company, whether it relates to product, whether it relates to engineering, whether it relates to, to sales, whether it relates to marketing, uh, and so on. So your job as a founder goes from doing everything and overseeing everything to recruiting the people and motivating the people that end up taking the company to the next level. So there are some companies that go through this transition well. They go to 1 and they realize, okay, we're at 1, now we need to institutionalize and to be able to kick it into the next gear. Other companies aren't as successful at doing that. And I think it's not something that's particularly specific to a management team. I think it's also specific to the product that you're trying to sell and the market that you're trying to sell into. There are, there are some markets that, you know, that are, that are more conducive to being institutionalized internally. There are other markets that are more haphazard and emerging where you be able, need to be able to pivot quickly. And, and the, and the kind of the zero-to-one mentality still applies as you're scaling to end. So the reason why endless business books exist for this and not just one book. Like, like there's one physics book. Yeah, everyone can learn physics by reading one book, but there's like 100 or 1,000 management books. And the reason why is because there's no answer to this and every single scenario is different. So I think just being open-minded to how every business is different, the same rules don't necessarily apply, and being open-minded and being able to try and being open to failing and then trying something else. Like we We look at Elon and we think, oh yeah, whatever this guy touches is gold. And we don't see the failures that happen behind the scenes. And I think what he's particularly good at is being able to attract great people, put 'em on the job, quickly assess whether or not they're the right person for the task, or if the task or the problem is shifting, and then perhaps take 'em out and put in somebody else to do the task and quickly iterate and revisit assumptions.

I think.

That is where his superpowers lie. And I think there's a lot to learn from that.

What's one question you wished people asked you more often?

People assume, since I'm a venture capitalist, people assume that like I'm all about the money, you know, it's all about like, you know, generating financial returns. And I wish people would ask me like, hey, Shain, like what drives you to do what you do? And for me, it's all about relationships. It's all about the relationships that are built. Like even when the companies fail, like that example that I gave you, That founder and I are friends. Like we are, we are on very good, we had a giant disagreement, but we are still close friends. And so for me, it's the, I'm doing this in hopes of going into the trenches with amazing people and coming out having achieved something that we would've never even dreamed of going into. Going into it. So that's really what drives me.

Do you have any really hot takes that are contrary to maybe the polygrams or the VCs on Twitter?

VCs on Twitter, I think, I feel like there is very little modesty in our business. I feel like the VCs on Twitter, and I don't want to point fingers, but I feel like there is, I feel like VCs take too much credit. For themselves. I think a lot of it comes from extreme insecurity, in my opinion. I'm, I'm, I, you know, I have come to terms with the reality that I'm not building anything. You are. That money is the ultimate commodity. There's nothing more commoditized than money. And, and the best thing I can do for you is hear where, where, what's, what's on your mind. And try to connect the dots, uh, for the company, which is far more important than just having encyclopedic knowledge about what you're doing. Some knowledge is necessary, but having a PhD in that field is absolutely unnecessary.

This might not be exactly your lane, but it feels like a lot of startups now are over-optimizing for distribution over product.

Uh, do you feel in a world of AI that's actually a smarter play?

Or do you think that—

Absolutely. Absolutely. At the end of the day, you can have the greatest product on earth. You could, you could have a perpetual motion machine. You can have the cure to all cancers, but if you don't have distribution, you don't have a business. And so if I had to choose between product and distribution, I would, I would take the distribution. And then figure out the product. Because if you have the product, building the distribution for it is not— unless it's a drug, I guess. But like building distribution is extremely challenging, which is why a lot of companies, to your point, are optimizing for distribution to reach the customers and then kind of figure out the product while they have that, that reach, if that makes sense. And so I've actually seen a lot of companies in my career build amazing products, but then not execute on distribution and ultimately not get rewarded for their hard work. So you need both ultimately, and you should never over-index on, on one or the other. But a lot of people who are hyper technical and product focused tend to be under the naive assumption that the product will fly off the shelves, it'll sell itself, everybody will want this. And that's never the case. Like there always needs to be some deliberate creativity and hard work that goes into building that distribution, even with a world-beating product. And I, I feel like distribution is kinda like the Olympics. You know how you see these gymnasts and they make it look so easy? Oh yeah. Like that. They make it look so— they're so effortless. And we just do one pull-up and we're dead. And so that's, I think, also like this. You see, when companies get distribution right, others look at those companies, say, oh yeah, that's so easy. Whereas it's not like building that distribution is actually really, really hard. Just like all the thousands of hours of training those gymnasts have had for that 30-second performance. It's the same thing.

Thank you so much.

Thank you. Thank you, Dara.

Thank you.

Great to see you guys. Maybe next time we do this, it'll be like 100 people in the audience. Yeah, that's the plan.

Thanks for tuning in. For personalized advice from Shaheen, head to delphi.ai and ask his digital mind.