Nov 24, 2025 · 57m · sourcery
Inside Thrive Capital: Investing in OpenAI, Wiz, Cursor, Nudge, Physical Intelligence · Sourcery with Molly O'Shea
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of Sourcery, host Molly O'Shea interviews Philip Clark, Partner at Thrive Capital, exploring the firm's high-conviction investment strategy, high-profile bets on OpenAI, Cursor, Wiz, and Nudge, and the future of agentic AI and physical hardware.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Molly holds 18.2% of the talking time here. How this is scored →
Molly as informed peer 4.3 Guest teaching 3.9 Guest disagreement 0.4 Molly pushing back 1.2
speaking balance: gold is Molly, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 14:14 Philip rejects the premise of competing against other VCs
Philip explicitly pushes back against Molly's framing that flying into a war zone was about deal competition, asserting that Thrive plays against itself and focuses solely on founder commitment.
Hardest push from Molly ▶ 13:58 Molly presses on the wartime deal closing narrativeMolly directly challenges Philip by citing reporting that no other investor would fly to Israel and asks if the extreme action was driven by deal competition.
Biggest teaching moment ▶ 32:10 Philip breaks down read vs. write neurotechnologyPhilip provides a clear technical differentiation between Neuralink's invasive read implants and Nudge's non-invasive ultrasound targeting versus conventional TMS electroshock.
Molly holds their own ▶ 38:26 Molly recites OpenAI's complete valuation escalationMolly demonstrates exceptional preparation by citing past Kushner interviews and precisely detailing each OpenAI valuation round from $29B up to $500B.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Molly as informed peer | Guest teaching | Guest disagreement | Molly pushing back | Why |
|---|---|---|---|---|---|---|
| Philip Clark's Journey from Physics to Venture Capital | 4 | 2 | 0 | 0 | Molly establishes Philip's background and humorously draws a comparison to Jeff Bezos switching out of physics. Philip amicably shares his background, his semiconductor Substack, and how he joined Thrive. | |
| Data Center Infrastructure & Mesh Optical Investment | 4 | 4 | 0 | 0 | Molly prompts Philip on value accrual across the AI stack and his portfolio companies. Philip provides an in-depth breakdown of data center interconnects with Mesh Optical and shares Michael Dell's speed chess analogy. | |
| The Origin and Pivot of Cursor | 5 | 3 | 2 | 3 | Molly cites the Colossus profile and presses Philip on whether flying into Tel Aviv during wartime was necessary to win a hyper-competitive deal. Philip reframes the premise, explaining that Thrive acts out of founder partnership rather than playing against other venture funds. | |
| Wiz Diligence Metrics and Multi-Act Platform Potential | 4 | 5 | 0 | 2 | Molly pushes Philip to define what made Wiz's diligence metrics stand out beyond generic praise. Philip educates on the rare ratio of deal size to sales velocity and how Wiz operated across multiple software stack layers. | |
| Hardware Investment Thesis and Childhood Ambitions | 4 | 4 | 0 | 1 | Molly brings up Philip's childhood aspiration to be the CEO of Lockheed Martin. Philip explains the declining cost of hardware sensors like lidar and how mature startup talent pools make hardware viable today. | |
| Sponsor Segment: Brex Financial Services | 0 | 0 | 0 | 0 | Mid-roll advertisement reads for Brex, Turing, and Carta with no interview interaction. | |
| Economics and Scalability of Hardware vs. Software | 5 | 5 | 1 | 3 | Molly challenges the venture viability of hardware given historical VC underperformance. Philip playfully acknowledges past skepticism before detailing geometric scaling curves, Program of Record defense contracts, and Starlink cash flows. | |
| Hardware Market Cycles and Neurotech Innovation (Nudge) | 4 | 5 | 1 | 2 | Molly asks whether the current hardware rush represents an unsustainable bubble. Philip explains power laws and alpha versus beta dynamics in venture capital, transitioning into Nudge's neurotech approach. | |
| Nudge vs. Neuralink and Traditional Brain Stimulation | 4 | 6 | 0 | 1 | Molly asks for a comparison between Nudge, Neuralink, and conventional brain therapies. Philip explains the difference between read versus write mechanisms, invasive BCI, and non-invasive ultrasound versus transcranial magnetic stimulation. | |
| Concentration Strategy and Fund Philosophy at Thrive | 6 | 4 | 0 | 2 | Molly brings data from Carta researcher Peter Walker comparing concentrated portfolios to spray-and-pray strategies, questioning Thrive's reported $25B AUM approach. Philip details why concentrated ownership aligns with missionary founders. | |
| Investing in OpenAI: Early Skepticism to $500B Valuation | 6 | 4 | 0 | 1 | Molly displays research by quoting Josh Kushner on Patrick O'Shaughnessy's podcast and listing exact valuation steps for OpenAI from $29B up to $500B. Philip explains the early skepticism surrounding OpenAI and their preview of GPT-4. | |
| OpenAI's Talent Density and Youth-Driven Research | 5 | 4 | 0 | 0 | Molly references an earlier discussion with Altimeter about speed as a moat and OpenAI's release cadence. Philip explains why OpenAI cultivates young researchers who intuitively understand model architectures. | |
| Evolving Paradigms in AI Research and Post-Training | 5 | 5 | 0 | 0 | Molly brings in technical insights from AI research accelerator Turing on synthetic data and reinforcement learning. Philip draws a parallel between AI training shifts and historical semiconductor core scaling. | |
| What Comes After SaaS? The Shift to Agentic Software | 5 | 4 | 2 | 2 | Molly asks what comes after OpenAI and cites Kalshi prediction markets forecasting tech layoffs. Philip reframes the question from what comes after OpenAI to what comes after SaaS, arguing AI will augment workers and unlock unaddressed problems. | |
| Navigating AI Market Cycles and Public Exit Outlook | 4 | 4 | 1 | 2 | Molly inquires about AI bubble risks and public market exit timelines for high-valuation startups. Philip cites Ray Dalio's Bridgewater philosophy of focusing on the decision process rather than quarterly market swings. | |
| Future Horizons: RL Production, Hardware, and AI for Science | 3 | 4 | 0 | 0 | Molly asks what Philip anticipates over the coming year. Philip highlights practical reinforcement learning deployments, physical intelligence robotics, and AI applied to science like Isomorphic Labs and Periodic. |
Statements from this episode (32)
Clark: US semiconductor manufacturing dropped from about 40% to 10%
“Back in the 19 nineties, about 40% of semiconductor manufacturing happened in the US. Today it's something like 10%. So drastic decrease.”
Clark: Thrive Capital led the seed round for Mesh Optical
“Mesh is a pair of founders from SpaceX named Travis and Cameron, who I actually knew back from college and then went to work on the Starlink photonics team. So the team that uses lasers in space to pass data from one satellite to another to give us internet. A…”
Clark: Thrive Capital has not invested in chip companies
“So that's where we've invested in data centers thus far. No chip companies thus far.”
Clark: Cursor reached hundreds of millions in run-rate revenue in 18 months
“I think when we invested, there were a couple tens of thousands of users and you know, low single digits, million air are fast forward to today. I don't think the company publicly publishes revenue, but call it on the Many, many hundreds of millions of run rat…”
Clark: Thrive Capital invested in Cursor after a yearlong pursuit
“Myself and my partner, Miles spent the next year trying to see how we could ultimately partner with them. And fast forward to May of 20, 24, we finally convinced them that we were the right sorts of people to bring onto the cap table.”
Clark: 50% of Wiz's users in 2022 were developers, not security staff
“50% of Wiz's users at the time were software developers, not security people.”
Clark: Wiz had churned only one customer before Thrive's round
“I think they'd almost never, they turned one customer in the course of all the time between when they started the company and when we were looking at this round.”
Clark: Wiz closed six-figure enterprise deals in weeks or months
“And so Wiz was Closing on these like very healthy six figure deals in a matter of weeks or months, which was just unprecedented. You know, normally you have small deals and fast cycles. You can have big deals and slow cycles, which had more or less figured out…”
Clark: LiDAR sensor costs fell from $80,000 to hundreds of dollars
“And these days, lighter costs are a couple hundred dollars a sensor back in 2014. So barely 10 years ago, there was something like 75,080 thousand dollars a sensor. So wild decrease.”
Clark: Hardware startup opportunities are at a 20-year high
“And so I think as an investor and someone who's really excited about companies that can counterfactually impact the world, that can make the world look physically different in some meaningful way, the number of opportunities now in hardware are far greater tha…”
Clark: SpaceX, Tesla, and Anduril accelerated growth at multi-billion scale
“I was looking at the revenue curves actually of and roll and SpaceX and Tesla the other day, and they are some of the only companies that have actually accelerated growth at scale, which is to say they are growing at a faster rate at multi-billion dollar scale…”
Clark: Anduril won a defense program of record within two years
“They, I think, won a program of record, which are the very big budget line item contracts in the National Defense Authorization Act within two years. Fastest timeline since the Korean War.”
Clark: Starlink generates large majority of SpaceX revenue and cash flow
“Created the opportunity for them to build a Starlink, which is now one of the most powerful internet constellations in the world. And it's actually driving a very, very large majority Of overall free cashflow and revenue for the business.”
Clark: Some of the most valuable companies will be in hardware
“I am fairly confident that We are going to produce some of the world's most valuable companies in hardware over the next decade.”
Clark: Five of the ten most valuable companies are hardware companies
“Even right now, five to 10 most valuable companies in the world are hardware companies. Apple, Nvidia, Broadcom, Tesla, TSMC.”
Clark: Thrive Capital backs non-invasive neurotech startup Nudge
“Also investors in Nudge, like you mentioned, Jeremy, who used to help run engineering at Neuralink and Fred Ursum, who was one of the co-founders of Coinbase, who are working on these very unique missions.”
Clark: Neuralink will eventually restore sight and motor function
“I think they'll actually do a number of things that are really helpful on the right side too. You know, things from there roadmap like blindsight, which they're gonna try to restore sight to the blind. You can imagine things that are actually gonna try to rest…”
Clark: Ultrasound brain stimulation can outperform TMS and improve compliance
“And so not only do we think that Ultrasound can be more effective than TMS, but you know, wearing a pair of headphones in your head, way easier than going to a clinic to get zapped.”
O'Shea: Carta data shows concentrated venture funds outperform broad portfolios
“Turns out highly concentrated funds, typically like 20 companies per portfolio. Outperforms spray and pray, but spray and pray looks pretty good.”
Clark: Thrive received a GPT-4 demo in November 2022
“We actually got a demo of GBT-IV in November of twenty-twenty-two before it was publicly released in March, and it was one of the most astounding pieces of technology we'd ever seen.”
Clark: Only two venture firms gave OpenAI a term sheet in early 2023
“And so when we invested in open AI in that round, I think we were one of two firms that gave the company a term sheet.”
Clark: Thrive's original OpenAI memo made no mention of ChatGPT
“We looked at our original investment memo. There was no mention of ChatGT in there.”
Clark: AI-native youth excel at frontier model experimentation
“The people who almost hold the model weights in their brains because they've been brought up thinking about AI natively are really, really good at the creative thinking and experiments and hypothesis testing that is necessary to push the performance of these m…”
Clark: OpenAI's Sora team has an average age in the low twenties
“Sora
You know, recent OpenAI product launch.
I think if you look at the average or median age of that team, I think it's in the low twenties.”
Clark: AI Labs' Key Edge Is Finding Next Paradigms While Productizing
“The ultimate competitive advantage for the labs and for OpenAI in particular is actually being able to constantly figure out what the next paradigm is, even as you're productizing and implementing the last paradigm.”
Clark: Every major large model paradigm has originated at OpenAI
“Every single major paradigm In large models at this sort of macro level that has been very successful to my knowledge has come from open AI or at least been originated at open AI.”
Clark: OpenAI will endure in the near term like Meta or Google
“I think the beauty is one, hopefully open AI is going to be an enduring technology. And so I'd be very surprised if in the near term we were talking about an after open AI, you know, is there an after meta or an after Google?”
Clark: Productized intelligence will create a post-SaaS era of software companies
“Now there's a new substrate for technology, which is what if you could productize intelligence and call it via API or fine tune a model with it? And I think that it engenders an entirely new breed of software companies. So I don't think it's an after open AI. …”
Clark: No Thrive portfolio company has laid off engineers due to AI
“My bottoms up answer is I'm an investor in a lot of companies that use AI tools. I cannot think, especially on sort of the coding side and engineering side of a single one that has laid off engineers because of these tools.”
Clark: Software companies will productize reinforcement learning to automate high-drudgery work
“I think we're in this moment where reinforcement learning, as you mentioned, is this new lever, but we've seen it make its way into very few new software products. I think few products in the labs rumor has it some of the coding companies like cursor have done…”
Clark: A new wave of hardware companies will tackle physical GDP
“I think we're gonna see a real renaissance of, you know, problems that don't just have to be solved with bits. They can also be solved with atoms as well. You know, sometimes to make a real impact on the world, you actually need to move something in reality. A…”