Venture Insights: Why We Are Investing Early and Often in AI-Robotics

Frequently Asked Questions

FAQ

• Where is the transcript for this webinar?

  Speaker 1:

  Hi everyone, and welcome to our webinar, Betting Early on AI Robotics: The Rise of Physical AI. I’m Rupika Brito here with my colleague Naren Ramaswamy, and we’re part of the AV Deep Tech Fund. Thank you so much for joining us today.

  Before we get started, I actually wanted to pause and remind everyone that we’re speaking today about Alumni Ventures and our views of the associated investing landscape. This presentation is for informational purposes only, and it’s not an offer to buy or sell securities, which are only made pursuant to the formal offering documents for the fund.

  So as a quick starter, I wanted to first give a brief background on myself. I’m a principal at Alumni Ventures where I focus on investing in promising startups in the deep tech space. I earned my BA from UCSD and my MBA from Wharton. My professional background includes scaling a search fund, private equity investment banking at Bank of America, and leading corporate venture initiatives at a Bitcoin infrastructure unicorn. Currently at Alumni Ventures, I continue to seek innovative investment opportunities within the vibrant startup ecosystem here in the Bay Area.

  Speaker 2:

  Thanks, Rupika. Hi everyone. I’m Naren. I’m a partner at Alumni Ventures working with the Deep Tech Fund as well as the Stanford Fund, both of which invest in groundbreaking technologies.

  I studied mechanical engineering at Stanford and did my MBA there as well. My professional background was with the iPhone engineering teams at Apple, as well as at startups, including an internship at Tesla. I’ve invested in deep tech companies at DCVC, which is a prominent deep tech fund, and I continue to explore investments in software and hardware at Alumni Ventures.

  So before we get started, let me briefly share who we are at Alumni Ventures. Alumni Ventures is among the most active venture firms globally, doing over a hundred deals a year. We pool capital from accredited investors into diversified venture portfolios. We’ve raised more than $1.4 billion from over 10,000 individual investors. We have over a hundred employees with offices in Menlo Park, Boston, New York, Chicago, and our headquarters in Manchester, New Hampshire.

  We were founded to solve an important pain point. Ordinary individual investors rarely get access to prime venture capital deals. If you think about it, companies like Apple, Nvidia, and Uber were all funded initially by venture capital. Venture-backed companies like these shape the future of industries around us, and some of these can generate outsized financial returns. So we believe that venture belongs in many personal portfolios, but still, accessing deals like this remains hard. In 2024, we were ranked a top 20 VC firm and also listed among the most founder-friendly investors.

  Our Deep Tech Apex 20 list of companies represents some standout companies, which our investors have received access to through Alumni Ventures, such as quantum computing company Retti, AI companies like Lambda and Cohere, as well as the nuclear company Radiant.

  Rupika and I are VCs on the Deep Tech Fund at Alumni Ventures, and we have a number of areas of interest. This webinar is about our thesis on physical AI, but we’ve been active across frontier computing paradigms, AI for biology, space tech, energy, clean tech, as well as networking and infrastructure. So if you’d like access to these sectors in your personal portfolio, we invite you to consider investing in Alumni Ventures.

  We have relationships with hundreds of VC firms globally, and some of our frequent co-investors are Andreessen Horowitz, Kleiner Perkins, KLA Ventures, and more. We co-invest alongside these top-tier VCs, aiming for a well-rounded mix of seed, early, and growth-stage private technology companies. The Deep Tech Fund invests in a portfolio of 15 to 20 deals over a 12 to 18-month period across groundbreaking startups in AI, robotics, semiconductors, quantum computing, etc.

  Like I mentioned, the portfolio is diversified by sector, stage, and geography, and the minimum investment for you to become an investor is just $10,000 for you to get access to venture capital in your personal portfolio. You’re welcome to book a call with us if you’d like to learn more.

  Now over to Rupika  to talk about our physical AI thesis.

  Speaker 1:

  Amazing. So we’re at a unique inflection point where AI is no longer confined to screens or servers. It’s increasingly embedded in robotics, autonomous vehicles, drones, and devices that can perceive and act in the real world.

  Today we’ll walk through why we believe physical AI represents one of the most exciting investment frontiers of the next decade. So here’s our roadmap for today’s session. We will define what physical AI is, explore why now is the right time, break down where value is being created, and look at real-world portfolio examples. Then we’ll discuss macro forces shaping the landscape, and finally how we’re investing and how you can get involved. In the end, we’ll save some time for Q&A as well.

  So what is physical AI? Let’s start by defining it. At its core, it’s AI with a body. AI models are integrated into robots or devices that have sensors and actuators to perceive the environment and physically interact with it, whether it’s a drone, a surgical device, or a warehouse robot.

  We’re also seeing a rise in generative physical AI, which is machines that can learn or adapt beyond preset rules. They don’t just see obstacles and avoid them; they can actually develop their own insights and make decisions in real time.

  Moreover, foundational models like Google’s Palm-E or RT-2 are crucial here because they provide a broad knowledge base so robots can understand novel tasks, zero-shot commands, or new objects they’ve never seen before. And all of this is a really big leap from legacy, pre-programmed and rigid robotics.

  And because physical AI is deeply interdisciplinary, you’ll see breakthroughs in mechanical engineering, AI algorithms, material sciences, and even biology-inspired design. That’s what makes this frontier so dynamic.

  Speaker 2:

  So why now? Why is physical AI suddenly so central? Our opinion is that there are three major factors that have converged.

  One is around AI breakthroughs. Thanks to large foundation models, robots can now interpret complex instructions.

  The second is around cost reductions. Critical hardware like lidar is around 90% cheaper now than it was in 2015.

  And thirdly, market demand. Acute labor shortages and pandemic-exposed supply chain vulnerabilities have forced industries to automate.

  This perfect storm means that for robotics, which was long considered too costly or too limited, now suddenly looks viable and essential. Companies are realizing that a well-deployed robotic solution can offset labor gaps, improve resilience, and keep operations running smoothly.

  Speaker 1:

  And as you can see, the market’s responding fast to this. Let’s quantify it.

  Projections say that global robotics markets worth $74 billion in 2024 could exceed $286 billion by 2032. That’s nearly quadrupling in under a decade. Healthcare robotics alone could jump from $12.5 billion to $46 billion by 2034, driven mostly by surgical robotics, diagnostic devices, and patient care automation. While agriculture robots might triple from $16–17 billion to $51 billion by 2029.

  And in logistics, Amazon already deploys over 500,000 robots in warehouses. We’ve seen VC funding for robotics and AV startups hit $7 billion in just one month last year.

  These numbers show that real money and real traction are behind physical AI and will just exponentially explode over the coming years.

  Speaker 2:

  Yeah, so our job as VCs is to understand technology trends, and we’ve seen that when platforms shift, early movers tend to dominate. And that’s what we’re seeing in physical AI.

  Early entrants accumulate proprietary data, integrate deeply into operations, and attract scarce talent—and that creates a significant moat. This isn’t a market where you can find your way into it later. So it’s best to have some sort of conviction early and dive in.

  That adoption is accelerating now. We’re seeing a global labor crisis, we’re seeing rising enterprise trust in robotics, massive AI leaps, as well as government support for robotics. So all signs are pointing to not a gradual but actually an exponential deployment.

  Speaker 1:

  And we see this opportunity in three distinct layers where value is being created.

  The first is the foundational infrastructure—so think sensors, edge computing, middleware, and specialized AI hardware.

  Next is intelligence platforms—operating systems for robots, simulation tools, domain-specific AI models.

  And then at the very top you have vertical applications, which are real-world use cases in healthcare, agriculture, logistics, manufacturing, etc.

  This layered approach really helps us identify investment opportunities at each stage. For example, if a company’s building picks and shovels for the entire robotic sector like advanced sensors, you’re in foundational infrastructure.

  If you’re developing a universal operating system or specialized AI for controlling robots, that’s an intelligence platform.

  Lastly, if you’re delivering a finished robotic solution like a surgical robot, that’s a vertical.

  And at AV, we’re investing across all three layers—from building blocks to full-stack systems.

  Speaker 2:

  So how does this thesis actually show up in the real world? We can show you some examples of companies in our portfolio that are building across these three layers—from radar and robotics to surgery as well as drones.

  The first example here is Zenda, which is building advanced radar for autonomous vehicles and ADAS systems. They have a 10x higher resolution at lower power, tapping a $75 billion industry by 2030. Zenda is backed by investors like KLA Ventures, NXP, as well as Hyundai.

  The second example is Knot Labs. They’ve built microscale bio-knots for targeted drug delivery in neurology. They’ve received two accelerated FDA designations and have raised $43 million, also led by Khosla Ventures

  The third example here is Ghost Robotics, known for their four-legged Q-UGVs that can handle rough terrains. A major defense player acquired 60% of Ghost for $240 million.

  Speaker 1:

  Another couple examples: Andromeda Surgical, which focuses on AI-driven robotic systems for minimally invasive procedures. They just completed their first patient treatment at record speed. Another is Swarm Aero, which is a large-scale fleets provider of low-cost AI-driven UAVs, tackling everything from cargo logistics to defense missions. They’re backed by Founders Fund, Khosla Ventures, Andreessen, and many more.

  These examples show how physical AI is not just a single niche. It’s empowering farmland robotics, surgical breakthroughs, advanced radar—even swarm drones for disaster responses or national security. These are real companies that are solving real-world problems and gaining serious traction.

  In fact, one of these portcos, Andromeda Surgical, joined us at an event that we recently had where we brought this thesis to life at our event called Atoms and Algorithms. It was really a powerful moment where founders were demoing next-gen robots and investors were engaging with the frontier up close. If there was any doubt that physical AI is real and investable, that room completely erased that.

  And now we’ve seen the micro. Let’s zoom out to the bigger forces shaping this category globally.

  Speaker 2:

  No conversation about physical AI is complete without talking about the geopolitical angle. China has been filling up a significant chunk of global robotics patents, making this a strategic priority for the U.S., and the U.S. is responding with calls for a national robotics strategy, especially regarding defense and supply chain resilience.

  On the regulatory side, it’s a patchwork. In the U.S., the FAA handles drones, states handle autonomous vehicles, and the FDA oversees medical robots. Meanwhile, the EU AI Act enforces risk-based standards across all AI. For forward-looking companies that bake compliance in from day one, this can actually be a competitive moat. Customers and governments are more likely to trust robots built with safety, ethics, and oversight in mind from day one.

  So regulation can be an obstacle, but it can also help high-integrity players stand out.

  Speaker 1:

  So now, how do you make money in physical AI?

  The most successful physical AI companies borrow from the best of SaaS. We see a few models emerging. One is robotics-as-a-service, or RaaS. Instead of buying robots outright, customers here subscribe or pay per use. This lowers adoption barriers and creates recurring revenue.

  Another business model is fleet learning or data-driven moats. Every robot deployed gathers operational data and improves the AI for all robots in the fleet. So it’s basically network effects at scale—more robots equals better models equals stronger moats.

  And then finally, there’s ecosystem lock-in, where you want to kind of think about it as app stores for robots, where third parties build on your platform. Basically, the takeaway is: the bigger the deployed fleet, the faster the AI learns, and the wider the moat is. Think of it like Tesla’s car data or Amazon’s warehouse robots, where it’s hard for smaller players to catch up.

  And we’re seeing robot app stores in development where third-party devs build specialized solutions on top of main robotic platforms. That fosters ecosystem lock-in, as customers prefer the platform with the richest set of apps. This turns robotics into platform businesses with software margins, which is an absolute game changer.

  Speaker 2:

  Just to summarize a few key points here—we’re going from bits in software to atoms in the real world.

  For decades, technology has focused on software. Now, physical AI means we’re optimizing the physical world, including manufacturing, logistics, healthcare, etc. And this is a massive shift.

  I think the second thing to highlight is the near-term catalysts. Labor shortages unfortunately won’t disappear anytime soon, and foundation model AI keeps advancing. And governments are backing robotics for economic and strategic reasons.

  And so we think these are all catalysts in the physical AI domain.

  Speaker 1:

  Exactly. It’s truly a generational platform shift—almost comparable in our minds to the internet or the mobile revolutions.

  Early entrants with, as I mentioned before, robust data moats, strong compliance, and deep customer integration could become the next wave of industry giants. Entire industries—from defense to logistics to healthcare—are being reshaped really before our eyes.

  So the window is open, but it won’t stay open for long. And so for investors, founders, and operators, this really is the moment to move.

  Speaker 2:

  So just tying this back to Alumni Ventures: we think Alumni Ventures is uniquely positioned to invest in some of the most category-defining companies in physical AI.

  We’ve built the sourcing engine, the co-investor network, the pattern recognition to identify these kinds of companies early. And that’s why we believe that a whole class of generational companies will be built in physical AI over the next few years. Many of them will emerge from deep tech, hardware-heavy, high-impact fields. And we’ve built up our networks and our processes to take advantage of that moment—with firms like CSLA, Sequoia, Founders Fund, etc.

  And our own team blends technical depth with real VC experience. And with the alumni network that we offer to our founders, we find ourselves joining competitive rounds alongside leading VC firms.

  So if you’re interested, we hope you consider Alumni Ventures as you think about your personal portfolio. As a next step, there are a few ways to engage. You can either visit our website to read more about the fund, or you can schedule a call with one of our senior partners using the QR code on the screen.

  And if you are one of the 10,000+ investors in Alumni Ventures already, your account managers are here to help you and to be your primary contact.

  Speaker 1:

  Awesome. So we’d love to hear your thoughts on what we’ve just covered. What questions do you guys have—whether it’s about the thesis, our portfolio, or how you think our returns are in this category? We’d love to hear from you.

  Speaker 2:

  Okay. We’re starting to see some questions come into the chat. This one’s a good one:

  “How do you assess the defensibility of physical AI startups when hardware or software tends to commoditize over time?”

  That’s a good question. So we agree that pure hardware plays are at risk of margin erosion over time. And that’s why we look for startups building some sort of a proprietary layer of software, or some sort of data moat—recurring revenues on top of their hardware.

  And robotics-as-a-service, like Ru mentioned, is a good example where the robot is the platform and the real defensibility comes from fleet-wide learning, customer lock-in, and data-driven optimizations.

  So we assess defensibility not just on the hardware spec sheet, but on the company’s ability to create a compounding advantage. And usually, data is a big piece of that.

  Speaker 1:

  Awesome. Oh, here’s another great question:

  “Can you give an example of how the Alumni Ventures alumni network has helped one of your portfolio companies in the physical AI space?”

  Absolutely. That’s one of the unique aspects of investing through Alumni Ventures.

  For instance, one of our portfolio companies in AI was looking for introductions to Fortune 500 execs, and we were actually able to mine and tap into our alumni community to connect them with senior executives at two Fortune 500 firms. Really, these early conversations helped accelerate their first enterprise contracts. And the alumni network often acts as an extension of the founding team’s business development efforts, which can be especially valuable in frontier spaces like this.

  I see another question that’s just popped up:

  “What’s the minimum investment amount to become an investor in the Deep Tech Fund, and what’s the timeline?”

  The minimum investment amount is $10K, and I believe our first close is September 30th, 2025. You guys get a fee reduction, I believe, if you join the first close. So we encourage you to start the process early to confirm a spot.

  Speaker 2:

  Great. Thanks again for joining us today. If you are excited about this space and if you want to learn more, there are a couple of easy ways to take the next step.

  You can visit our website for a deeper dive into the fund materials. We’ve put together lots of resources about how we invest, our track record, where we’re seeing opportunity in physical AI.

  And if you prefer a more personal touch, you’re welcome to schedule a call with one of our senior partners. They can walk you through how the fund works and help you figure out if it’s the right fit for your goals.

  Speaker 1:

  Yeah, and as Naren mentioned earlier, if you’re already an investor with Alumni Ventures, we’re of course so glad to have you with us.

  Your account manager is here as your first call—whether you have a portfolio question, or you want help understanding performance, or just want to stay plugged in on what’s happening in this fund or others—they’re there to help.

  And we hope that today kind of gave you a sense of why we’re excited about physical AI and why we believe now is the right time to be backing these category-defining companies in this space.

  Speaker 2:

  Yeah, I mean, we’re all quite excited by physical AI and the future, and we’re here to help investors—individuals just like you—to access these companies early and be part of that journey. So thanks again for being here with us, and thanks for your time.