Neil Patel

I hope you enjoy reading this blog post.

If you want help with your fundraising or acquisition, just book a call click here.

From humble beginnings in Sunnyvale to leading a groundbreaking AI company, Varun Mohan’s journey is a testament to resilience, innovation, and a relentless pursuit of impactful solutions.

In this interview, Varun shares his story, insights on the evolution of technology, and the strategic pivots that led him from autonomous vehicles to the creation of Codeium, an AI code acceleration tool transforming the software development landscape.

Codeium has attracted funding from top-tier investors like Kleiner Perkins, General Catalyst, Greenoaks, and Founder’s Fund.

In this episode, you will learn:

  • Varun Mohan, co-founder of Codeium, emphasizes the importance of resilience, mission orientation, and accountability in building a successful startup team.
  • Codeium pivoted from a GPU virtualization company to an AI code acceleration tool, now serving over 600,000 developers and 700 enterprises.
  • The company’s rapid growth and success are attributed to its focus on providing tangible value and leveraging internal talent to solve problems before hiring specialists.
  • Codeium’s technology aims to significantly reduce the time developers spend on repetitive tasks, thus increasing their overall productivity.
  • Varun stresses the importance of humility and truth-seeking in navigating the uncertainties and challenges of startup life.
  • The company has raised $93M to date, with a strong emphasis on responsible capital management and avoiding unnecessary expenditures.
  • Varun’s transition from engineer to CEO involved learning to effectively communicate business value and building a robust go-to-market strategy alongside technological innovation.

 

SUBSCRIBE ON:

For a winning deck, see the commentary on a pitch deck from an Uber competitor that has raised over $400M (see it here). 

Detail page image

*FREE DOWNLOAD*

The Ultimate Guide To Pitch Decks

Remember to unlock for free the pitch deck template that is being used by founders around the world to raise millions below.

About Varun Mohan:

Varun Mohan is the Co-founder and CEO of Codeium. He was previously a Tech Lead Manager at Nuro from January 2018 to May 2021, working on autonomy infrastructure.

Prior to that, Varun was a software engineering intern at Databricks from June 2017 to August 2017, working on ML systems.

Varun has also interned at Cloudian Inc., Quora, Cloudera, LinkedIn, Samsung Electronics, Stony Brook University, and UC Santa Cruz. His research interests include storage systems and data infrastructure.

Varun Mohan has a Master of Engineering degree in Computer Science from the Massachusetts Institute of Technology (MIT).

He also has a Bachelor’s degree in Electrical Engineering & Computer Science from MIT, with a minor in Mathematics. He attended high school at The Harker School.

See How I Can Help You With Your Fundraising Or Acquisition Efforts

  • Fundraising or Acquisition Process: get guidance from A to Z.
  • Materials: our team creates epic pitch decks and financial models.
  • Investor and Buyer Access: connect with the right investors or buyers for your business and close them.

Book a Call

Connect with Varun Mohan:

Read the Full Transcription of the Interview:

Alejandro Cremades: All right. Hello, everyone, and welcome to the Deal Maker Show. So today we have a very exciting founder, you know a founder you know that has been through through all the different steps that you can think of. Okay. So we’re talking about a founder that you know basically ah has gone with the people thing, with they Basically, we’re going to be talking about the future, the technology that they’re in. We’re going to be talking about to the transition from perhaps you know like autonomous vehicles to like more generative AI and a bunch of other stuff like raising money and and and also growing and and scaling the business that I think you’re all going to be ah very much a you know inspired by and and enjoy. it So without further ado, let’s welcome our guest today, Varun Mohan. Welcome to the show.

Varun Mohan: Hey, Alejandro, glad to be on.

Alejandro Cremades: So originally from Sunnyvale, so give us a walk through memory lane. How was life growing up over there?

Varun Mohan: Yeah, so I was born to immigrant parents, Indian immigrant parents from Sunnyvale, went to school in the Bay Area, did a lot of these math and computing Olympiads in high school. Actually funny story, a lot of us at the company at Codium actually competed against each other in middle and high school. And that’s how we originally knew each other. After I went to MIT, MIT is also where a lot of us at the company sort of met each other as well. um and And that’s where me and my co-founder worked ah together a lot. And then afterwards, joined a company called Neuro, which was an autonomous vehicle company.

Varun Mohan: Finally, I joined the company even before knowing what it was going to what it was doing, because they were so in stealth mode.

Alejandro Cremades: you

Varun Mohan: um But I just realized I wanted to work on where the future of robotics and autonomy would ultimately end up going. um I was there one of the first 15 engineers at the company and ended up being a manager of a team running large-scale deep learning infrastructure offline and then after that decided hey I wanted to start something and actually build a product that people used at scale and built a company that wasn’t what Codium is today. It was actually a GPU virtualization company. So we actually built software to make GPU applications run significantly more efficiently.

Alejandro Cremades: you

Varun Mohan: And at peak, we were managing upwards of 10,000 GPUs for a handful of companies, which actually accounted for 20% of GCP Google Cloud Platform’s GPU inference capacity in multiple data centers. And we were doing this with a team of eight people. But in middle of 2022, we saw the advent of the transformer in these generative models, which is where the GPT started taking off. This was pre-chad GPT. And we realized that that would usher in a brand new set of applications, but also that would mean everyone who was going to run transformers.

Varun Mohan: so And if everyone ran this type of model architecture, we didn’t see much value in being a GPU virtualization company. So we actually decided to verticalize and pivot and build a company that we see called Codium, actually.

Alejandro Cremades: Thank And we’ll talk about that in detail. ah But let’s let’s let’s start you know from from the beginnings, from the roots. you know let’s Let’s go back in time you know when you got into math. you know What got you into math to begin with? Because I know that was a big deal. You did competitions, you know whether it was there in in high school or or then going to MIT and still doing you know this kind of competition. So what got you to really develop that love for math?

Varun Mohan: Yeah, so I think the big thing is just the objectivity. and It was one of those things that you just can’t BS. And the fact that you can kind of like wrangle with a problem for a long time is very attractive. I’ll just say like an interesting statement. In seventh grade, I made this competition called the USAML, which is the top 250 kids in the entire nation, including high schoolers. And I took the contest and it’s a contest that is two days. Each day is four and a half hours and you have six problems and the score is out of 42 and I got a one. I got a one out of 42 and that’s the median score. The median score is a one. So I think there’s something so beautiful about basically problem solving for a long period of time and potentially not even solving the problem.

Varun Mohan: And I think that gives you a level of courage of solving hard problems. um And I think that’s what all of my peers really enjoyed. We liked not solving problems where we could instantly solve it in like a minute. We wanted to be able to wrangle with hard problems and Olympiads were sort of at and ah a form factor that that ah that was conducive to that.

Alejandro Cremades: So you eventually you know graduated from MIT, and and it was a really interesting transition that you did there from being an engineer to all of a sudden becoming a manager. So how was that transition for you?

Varun Mohan: Yeah, i think I think at the company, if I can just say something, it was it was pretty natural because I was so early there that I built a lot of the systems and infrastructure at the company that I naturally ended up leading people at the company. I think what I realized is my dream was not to just manage people. It was to to actually like build a product that people use at massive scale. And I think the unfortunate thing is in autonomous vehicles, the timelines to actually do that were quite high. And that’s because the problem is so hard. And I think because of that, I sort of moved, wanted to do something where I could have an impact, but but an impact today rather than a delayed impact where where we potentially build something that would work 10 years from now.

Alejandro Cremades: So you actually were an employee for a bunch of time you know on on other companies. So at what point you know did it become clear that it was your time to to take a stab at it you know on your own and create your own destiny?

Varun Mohan: Yeah, so me and my co-founder, actually, Douglas and I, we’ve we’ve known each other since middle school. um We’ve always wanted to start something together, but it never seemed like the right time. And I think for us, we we decided to, I guess, just jump off the deep end and ah probably irrationally. Irrationally, that the idea that we originally had, we didn’t completely understand why it was the best idea possible. um But I would say, like, you will never rationally come up with the best idea to to work on after quitting your job.

Alejandro Cremades: So then let’s talk about starting what thing obviously was at that at the time AXA function, and and obviously that would develop into a pivot, you know which is what you guys are doing now with coding. But how how did that start? How do you guys get going? And and then what happened?

Varun Mohan: Yeah, so this is um this is the the sort of interesting story here. As I was saying, we ended up building GPU virtualization software. So the idea was you could take applications that would previously run on GPUs and you could run them on computers, normal s CPUs, and we would transparently offload the computations to a remote machines. And we ended up being able to, I guess, optimize some workloads by five to 10x. So it was a big, big improvement, especially in 2021 when GPU supply was under shortage. There was shortage for many reasons during COVID. And for us, though, what we felt is, and this was the hard part, we were a team of eight people. We were making between two and three million in revenue.

Varun Mohan: And which was good, we were free casual positive as a company. But for me, I did not know how could we scale this thing to be 50 times bigger. And this is maybe a hallmark of the company, which is that if I or someone at the company cannot figure out how to increase the revenue by $1, we don’t just hire someone to increase it by 10. which is to say we make really, really strong generalists try to solve problems. And if they can make no progress, the solution is not to hire a specialist and see if they can succeed. So this is a big thing where we never hired a sales rep, even though the business was doing millions of dollars in revenue, we never hired a sales rep internally, ah which was a good decision in retrospect because it enabled us to pivot the business significantly with significantly more ease. Right.

Alejandro Cremades: and In your guys’ case, i mean you you actually raise money, you know quite a bit of money, i mean over $20 million dollars for this, and then all of a sudden you realize that it’s time to pull the plug and and and and and orient you know kind of like the path you know towards a different day site, a different type of future that you would you know be living into. a show What triggered that? you know and and and Also, how was that you know type of conversation with some of those existing investors too?

Varun Mohan: Yeah, so i think I think at the time we were probably doing more revenue than a lot of companies that had raised more money, which is more of an indication of the zero interest rate phenomena environment ah in 2021. But I think what really flipped the switch in my head was I was looking at some of these other generative AI companies or application companies and seeing them with small teams getting to substantial amounts of revenue very quickly. And what it actually showed showed to me was the right comparison for us to do wasn’t to compare ourselves to other ML infra or ML ops companies, but to what the best company we could be was. And I just realized we were not operating at our potential as a company. And then at that point for us to be a worthwhile or or useful sort of venture investment, we couldn’t just be a company that made tens of millions in revenue. That is not a venture return.

Varun Mohan: ah right And I think for us, we had much larger aspirations for the company. I want this company to to be making billions of dollars a year. And once that that sort of switch triggered in our head, we had an option. right We either die die a slow but certain death, or we die with a non-trivial chance or we succeed for a massive reward. And yes, the thing about taking the second option is we would die significantly faster in the case where we died, but we decided to go about that approach because it is never good to do the wrong thing for longer. And I think right now, if you were to ask any of our investors, did we make the right decision? I think all of them would say uniformly, this was the right call.

Alejandro Cremades: so then So then walk us through how was that you know communication that you guys endured with the investors to make sure that everyone was up to date, that everyone was OK with the decision. And in that moment where you were like, OK, now he is saying it’s time to to get going.

Varun Mohan: Yeah, i think I think we were able to convince ourselves, both the founders, at the time like we have we had control of the board. We obviously have like a great conversation with our investors to make sure, like hey, this is what we will be pursuing, and here’s our thought process. ah But ultimately, like for for us, you know if If we felt in our gut this was the wrong call, like we could if the investors told us like we have to do the ah thing that we believe is wrong, like the company was going to fail anyways. So I think our investors, largely speaking, ah were like did not oppose our decision to to move forward in this direction.

Alejandro Cremades: So, for the people that are listening to get it, what ended up being the business model of Codium? How are you guys making money today?

Varun Mohan: Yeah, so a little bit about Codeium, we’re an AI code acceleration tool. our Our purpose is to provide the most amount of leverage to developers in the world. um A little bit of a backstory on the company. In the beginning of 2023, we had less than a thousand users on the product. And now we have over 600,000 developers that use the product. We process over 100 billion tokens of code every day, which is over 10 billion lines of code every day. We’re one of the top five largest Intervia apps in the world. And we actually train our own models and run them at massive scale ourselves. And this is where one of our massive advantages is.

Varun Mohan: And currently, over 700 enterprises use the product with over 34 to 500 using the product internally. And the big reason why enterprises want to use the product is for two reasons. First of all, we support all the places where their developers are on every ID, on every single language that the that the users can write. We provide self-hosting capabilities when necessary. So a lot of companies don’t want to send their code outside of the company. And because we train our own models, we can deploy it entirely within their own environment so the code never leaks, right?

Varun Mohan: You can imagine for some of the largest companies, code is their most important IP.

Alejandro Cremades: you

Varun Mohan: The second piece is we actually work on every source code management tool. The big competitor in this space is this product called GitHub Copilot, which is owned by Microsoft. And it’s a massive product. But we don’t care about if you’re running things on GitHub, GitLab, Bitbucket. And there are many, many different source code management tools. And some of the largest companies actually have many source code management tools used internally. on And finally, we actually give you a personalized experience. which means that we actually tailor make the suggestions to the private code that lives inside the company. And we found that actually for a lot of companies, they have noticed that their developers, their new developers onboard onto codebases in four to six weeks instead of four to six months now with Codeium because our system actually deeply understands the software that exists inside the company. And they’re also noticing that over 45% of all software that is getting committed is generated by Codeium.

Alejandro Cremades: So in total, how much capital have you guys raised too late for the company?

Varun Mohan: $93 million. dollars

Alejandro Cremades: So at what point do you guys realize you’re turning a corner here with this new pivot? And also how did that look like from a you know going from life cycle to life cycle and putting in parallel a financing cycle with new reinvestment to really keep things you know pushing along?

Varun Mohan: Yeah, so interesting fact, by the time we raised our Series B, we had barely spent our seed round. So we are not a company that I guess goes in and Series B was $65 million. dollars We’re not the kind of company that goes out and spends a lot of capital unless it’s actually making the business work in a material way. And we didn’t try and go out, once we did the pivot, we didn’t try and go out and raise capital. We actually just went out and built a product. and scaled the product to, I guess, a massive number of users. And a lot of large enterprises started to use the product. Companies like Dell that currently have 40,000 developers inside the company um were starting to use the product.

Varun Mohan: And I think the the key aspect of this was we were never in a position where the only thing that would make us go from not exist or existing to not existing was the lack of existence of capital.

Alejandro Cremades: you

Varun Mohan: um And because of that, actually, we were we ended up being able to raise around fairly easily ah from the folks at Kleiner Perkins, who it’s a very well-repeated sort of venture capital firm.

Alejandro Cremades: So what did you ah guys learn about? Obviously, you know when you’re doing well like that, you know you have a little bit more leverage when it comes to the capital raising um efforts. So I guess from you know a perspective of really developing a very streamlined process when it comes to fundraising, what have you learned?

Varun Mohan: Yeah, so this is maybe a little bit of an anti pattern, but in terms of our fundraise process, we have never we’ve never actually done a very, very proper sort of fundraising process, largely because for different points of the business, we have been getting we ended up getting preempted ah by folks that were particularly interested in the progress that we made and the proof points of the company. And I guess for us. The reason to fund to really go out and want to fundraise really badly is if if you believe that you will you will not be able to make it unless you get that amount of capital today. right And I think for us, we have largely been good stewards of capital in a way where we are willing to invest a good deal of money to see the company work, but we don’t invest millions of dollars into different aspects of the company if we can’t see a clear ah ROI. right um And because of that, we always operate in a way where we could be cashflow positive if necessary.

Alejandro Cremades: and And I guess you know here when we’re talking about building the company too, if you were to go to sleep tonight, Varun, and you wake up in a world where the vision of Codeum is fully realized, what does that world look like?

Varun Mohan: I think in a world in which the vision of Codeium has been realized, developers are now have 10 times the amount of leverage. That means that they can generate code 10 times faster, deploy code 10 times faster, or debug, test, review, um and finally finally actually navigate code 10 times faster. And in that world, the best thing that I envision happening is the next time Nvidia goes out and does something very hard, they’re building the X100, which is their newest chip, let’s say. And that chip takes 12 months. I want the time it takes for that chip to get built to go down from 12 months to one month with Codium. And I think if we are able to do that, we will capture a lot of value at that point, but the sheer amount of value we will be able to generate is going to be so massive um that i think I think that would be the the true north star for us as a company.

Alejandro Cremades: So when you put the technology in parallel with that, you know on where we are today and where it’s heading, you know what where where can you tell the listeners that the technology is going?

Varun Mohan: Yeah, so I think, you know, I previously worked in autonomous vehicles where every year we said the next year was the year autonomous vehicles was going to ship. ah Right. I don’t know if you’ve you’ve seen this but in 2015 TechCrunch said ah this is going to be the year of AV and in 2024 they are now saying is this the year of AV. right So that’s largely because we underestimate what is possible over a long time horizon, but overestimate what is possible over a short time horizon. So I believe we will still be, developers will still be writing code inside applications in the next couple of years. But I think what is going to happen is the amount of work it takes for them to write boilerplate applications to do tasks like migrations, to do tasks like refactors is going to go down tremendously.

Varun Mohan: But then for every piece of the pie, you know, what I like to quote is there’s this principle in computer science called Andas law. And what that says is if you, if you take many steps of a process and you make one step faster, there’s a limit to how much that makes everything faster. And I’ll give you an example. in a world in which there are 100 units of time and 30 units of time is being spent writing software. If all we do is make the 30 units of time writing software 10 times faster, we only reduce the total amount of time spent from 100 to 73. This isn’t a complete game changer, right?

Varun Mohan: The game changer is how do you reduce the remaining 70 units by a factor of 10?

Alejandro Cremades: you

Varun Mohan: And I think the way I would like to think about it is it will be a gradual process or incremental process to continue to shave down time from every single aspect of this process, right? Because even today, if you could write code instantly, the time it takes to review code will become a bottleneck. The time it takes to deploy code will become a bottleneck, right? And people are not willing to completely deploy arbitrary software than and that an ML or AI system generates into production without any review. We are far from that right now.

Alejandro Cremades: I mean, in your guys’ case, you know you have um about 50 employees. And for a company like this, I mean, obviously, it’s very heavy on the technical side of it. So how do you go about really getting incredible engineers and not only getting them, which is tough, but even tougher is retaining them. you know What does that look like in an organization like this?

Varun Mohan: Yeah. So I think, I think first of all, as a company, we want to hire people for resilience. That’s ah a big one. And I think us having gone through the pivot is an important piece where we realize right now things are going well, but there will be, there will be low points for the company. And we don’t want people that are here just because of some, some great vision that we promised them that the company will always be having a great time. Right. And I think what that means is. We want people that are incredibly mission-oriented. That means that for them, hell or high water, they want to see this thing succeed. right And even if that means that we have some some struggles along the way, they will be around. right Because ultimately, what drives them is the vision of the company, which is providing as much leverage as possible.

Varun Mohan: And I think the reason why that’s really, really critical is, you know, and and it’s important to me also is let’s say we were building software to make tax accountants more efficient. ah Right. That is not a vision that particularly excites me. And if the company gets hard, I’m going to lose motivation to work. Right. So we look for folks that are incredibly mission oriented, but also we look for folks with, with just very high sense of accountability. right and And aptitude is ah is a non-starter, right? We have a very high technical bar at the company. It’s one of the highest technical bars in all of Silicon Valley, I would say. but But accountability is very important. We want people who ultimately, if something doesn’t go right, their goal isn’t to point the finger at someone else.

Varun Mohan: And I think a lot of people misjudge high-trust environments as very nice environments. But that’s actually not true. You have a high-trust environment if you have people that are very capable that are high accountability. When you tell someone, hey, are you going to do this by X? They not only say yes, but they deliver. right And I think we look for these characteristics. right People with inbuilt resilience, people who are mission-oriented, and people who are high accountability. And if we see those aspects in someone, we’re very excited to move forward with them.

Alejandro Cremades: So I guess, say as we’ve been talking about the the future earlier, but I want to talk about the past and in doing so with a lens of reflection. If I was to put you into a time machine and I was to bring you back in time, you know, let’s say, you know, a few years when you guys were thinking about doing something here, and let’s say you’re able to stop that younger self, that younger balloon and give yourself one piece of a advice before launching a company. What would that be and why, given what you know now?

Varun Mohan: i think I think the only thing I would i would tell that younger Varun is ah be significantly more humble. And I think that is largely because having a company is a very humbling experience. You may be able to convince investors that your vision is good. You may be able to convince employees that your vision is good. But ultimately, the water finally settles. The company really needs to work. And there’s a level of humility you need to have that milestones like fundraising don’t really mean much. Milestones like like building a technical product don’t really mean much.

Varun Mohan: If the end result is not you providing value to someone. And focus ruthlessly on providing value to other people and and and the rest will figure its way out.

Alejandro Cremades: you

Varun Mohan: But that requires a deep amount of humility because the first idea you’re going to have when you start a company is very unlikely to be the right idea. It is incredibly unlikely that it’s the right idea. and actually needing to constantly sort of change directions and assume you were wrong, right? The hardest part about a startup is you need to operate with with two modes in your head. You need to be irrationally optimistic because if you aren’t irrationally optimistic, then you will do nothing, right? You will actually just say a big company is going to be able to do it, right? Because why a big company has more people, more capital and more distribution.

Varun Mohan: So you need to be irrationally optimistic that you will be able to execute faster on on an important vision. But you also need to be uncompromisingly realistic. And that requires a tremendous amount of hum humility and and a truth-seeking nature for the for the company. And I think i would just I would just give that piece of advice to myself.

Alejandro Cremades: I guess the the other thing that comes to mind too is, I mean, obviously say you are a trained engineer and you’ve you’ve been very heavy on the on the technical side of things. and And now is the first time, you know, that you’re encountering having a CEO role and and more on the business side of the equation. I know that transition from engineer to business, you know, ah kind of like mindset, it’s not very easy. And it’s a tough bridge to cross. How have you gone about that transformation to make sure that you know you were able to be as effective as possible as a leader?

Varun Mohan: Yeah, I think the biggest thing is I actually did sell the software even before we hired our first sales rep at the company. I don’t know. Once again, for Codium, um we didn’t hire a sales rep until 10 Fortune 500 companies reached out to us. So we are not the kind of company that goes out and asks other people to solve our problems for us until we have a rough sense of what success looks like internally. So I had to go out and do that. I had to understand, hey, like we’re running a bunch of these sales processes. How do we actually convey a message that is compelling? And what is compelling is not just talking to them about technology. right Technology is not business value.

Varun Mohan: right How do we actually properly build champions in inside companies? How do we actually, in ah in a lot of ways, prove business value in a way that is not just hand waving, right? In a way that we can actually go to an executive and and convincingly give them a vision that, hey, if we did deploy Codium across the organization, it would give you this kind of a return, right? And you would then see this kind of productivity improvement inside the company. And I think for us, It is, in a weird way, go-to-market, as you said. Our company is not a product-led growth company in a conventional sense. We are actually an enterprise sales motion company. And this made this is like ah ah like ah a decision we made consciously in the early innings of the company. And we’ve embraced go strong sort of go-to-market team internally at the company. We’ve hired an amazing, amazing go-to-market team. And the technology is is as important as go-to-market, right? Because when you look at the success of a company,

Varun Mohan: It’s three things that ultimately matter. right The first is distribution. but And a startup naturally, when it starts, is not going to have amazing distribution compared to a large company. The second is technology. How great a product can you build? And the third is partnership. How great a partner and a go-to-market organization can you actually build? And these three pillars are what make a company succeed. And if a startup can’t do the last two successfully, it will not get distribution.

Alejandro Cremades: I love it. Varun, for the people that are listening, I would love to reach out and say hi. What is the best way for them to do so?

Varun Mohan: Yeah, just reach out at verun at codium dot.com. That’s C O D E I U M dot com. We’re hiring across the board on both engineering and, and the go to market team. Um, and if any of this sounds interesting, please apply.

Alejandro Cremades: Amazing. well Hey Varun, thank you so much for being on the Dealmaker Show today today. It has been an honor to have you with us.

Varun Mohan: Thanks a lot, Alejandro.

*****

If you like the show, make sure that you hit that subscribe button. If you can leave a review as well, that would be fantastic. And if you got any value either from this episode or from the show itself, share it with a friend. Perhaps they will also appreciate it. Also, remember, if you need any help, whether it is with your fundraising efforts or with selling your business, you can reach me at [email protected]

 

Facebook Comments

Neil Patel

I hope you enjoy reading this blog post.

If you want help with your fundraising or acquisition, just book a call

Book a Call

Swipe Up To Get More Funding!

X

Want To Raise Millions?

Get the FREE bundle used by over 160,000 entrepreneurs showing you exactly what you need to do to get more funding.

We will address your fundraising challenges, investor appeal, and market opportunities.