Sandy Kory: The Founder Obsessed Investor Behind BillionToOne, BaseTen & Palantir

Summary

• Sandy Kory put roughly half his net worth into Palantir’s bridge round around March 2009 — near the market bottom in hindsight — primarily on a talent-density inference, not a detailed business analysis. A Stanford grad-school friend, whose name varies in the transcript between Shawn/Shyam Sankar, raved about the team; Kory’s read was statistical: “this might be the smartest, most ambitious group of people in the world.” The company could barely explain its business — Kory recalls explaining “the definition of revenue” to Joe Lonsdale in 2008 — and top-tier VCs dismissed it as government consulting.
• His screening method is radically product-light: “90% of the conversation is going to be on the things you’ve done before the startup,” and founders who want to show slides are told to send them afterward. Founders are coached on pitches but not on telling their life stories, so he believes there is more signal in the life before the startup. He also argues the earliest stage offers more alpha and more room for legitimately different approaches.
• Talent magnetism is one of his highest-weight signals — “it might be the most important thing, frankly.” Given a two-year crystal ball, he would ask for the caliber of the engineer who just joined. A recurring reason he passes on otherwise attractive deals is that he is not convinced the founder will recruit great people; hiring quality is his leading explanation for why funded startups slow down.
• He favors missionary founders over mercenary ones and says many VCs do not care about that distinction. He would take the Anthropic founders over founders who left for Meta, while noting that highly competitive winners such as Uber or Facebook may not have been genuinely mission-driven. Mission can attract talent, connecting back to the talent density he saw at Palantir.
• BillionToOne illustrates the approach: Kory found it through a TechCrunch article about YC’s short-lived Fellowship, invested before a later $4 million valuation round, and repeatedly added to the position. The host says he thinks the company is now worth about $4 billion. The public company reported roughly $100 million in quarterly revenue in the discussion — about a $400 million run rate at roughly 100% growth — and was profitable, with most revenue from NIPT prenatal testing and a faster-growing liquid-biopsy line using a “chemical microscope.”
• His competition heuristic is that a real insight will attract funded rivals — Sequoia, Kleiner, and Andreessen-backed competitors — within six months to a year. The key question is how the founder and investor will feel when that happens: with BillionToOne, “100 competitors came to the market — oh, this is going to be fun.” He puts little stock in references, rejects claims that conviction can reliably be formed in 10 minutes, and looks for inconsistencies by asking basic questions without embarrassment.
• SendCutSend shows a different sourcing path: a high-trust M&A relationship with founder Jim Belosic, whom Kory advised not to sell his bootstrapped software business. Kory invested $1.5 million through an SPV in 2021 when the manufacturing business struggled to attract investors. The vertically integrated parts company shifted from mostly tinkerers and SMBs to mostly industrial customers; its customers include people working at Tesla, SpaceX, and Anduril, and demand is largely inbound. Patrick Collison later heard Belosic’s $1 billion valuation target and offered $10 million, then introduced him to Sequoia and Paradigm’s Matt Huang.
• Kory expects the founder formula to remain useful in an AI-heavy future, while acknowledging uncertainty about his own ability to execute it. Invoking a quote he attributes uncertainly to Jeff Bezos or someone else about focusing on what stays the same, he points to exceptional, determined, resourceful, mission-oriented, intellectually honest founders — even if companies have a thousand agents per human — and says the formula should work as long as capitalism runs reasonably well.

Deep dive

1. The Palantir bet: half a net worth on a statistical inference about talent

• Kory’s path in: no VC would hire him out of Stanford — venture seemed like “this thing that the gods of Mount Olympus were doing” — so he joined a boutique M&A firm advising bootstrapped tech companies and funded an angel-investing habit. In graduate school he befriended a classmate identified variously in the transcript as Shawn/Shyam Sankar, who was regarded as likely to succeed, later joined Palantir as employee number 13 in 2007, and raved about the team: “your buddy was friends with LeBron James and saying, these guys are so good at basketball.”
• When Sankar contacted him about a bridge round around March 2009 — Palantir did not need the money, but “the world might be ending” — Kory put in roughly half his net worth. His framing was an inference: “what are the chances that these people are not the most ridiculously talented people in Silicon Valley?” He rationalized the concentration “like NPV”: “I’m going to make money if I need to.”
• The decision was fast: he asked for information, Joe Lonsdale sent a long email, and he invested. Within roughly 15–16 months, Kory had invested in Palantir, been let go, and launched his own M&A firm with partner Mike — raiding his 401(k) to pay rent until the business cash-flowed by 2012.

2. Palantir could not explain its own business — and that was the lesson

• Kory’s best specimen of contrarian conviction: he remembers having coffee with Lonsdale in 2008 and explaining “the definition of revenue.” No one there knew the technical definition, but Kory says it did not matter to the business because they were making great progress. He still believes someone should understand GAAP financials if a company is going public, but that talent density and ambition matter more.
• Top-tier VCs at the time passed because Palantir looked like “a consulting business” selling to the government. Kory says the company was “radically contrarian” in many ways; not every contrarian choice created value, but its extreme first-principles thinking and genuine mission orientation were real rather than performative.
• Why did Sankar bring in a new angel? Kory says his friend was doing him a favor: he thought Kory was smart enough to understand the opportunity, and they had done favors for each other. The host’s reaction was: “Sandy, I need better friends.”

3. The double life: M&A by day, angel investing nights and weekends

• His M&A niche taught him an inverted signal: with bootstrapped businesses, “the more profitable the business, the less likely the founders are to know the numbers.” Outsiders saw sloppiness where Kory saw value, and he developed a sense for the many ways founders can mislead investors.
• Meanwhile he was cold-sourcing YC startups and companies in Australia and Africa on weekends while advising bootstrappers in places including Saskatchewan, Ontario, Texas, and Florida. Founders assumed the M&A guy wanted to sell them; his actual view was that companies such as Palantir and Canva were unlikely to sell, and he often gave struggling portfolio companies free advice.
• On whether he discarded a valuable banking skill set, Kory says most VCs do not know much about M&A because they do not need to. His WhatsApp example was Zuckerberg ordering a $10 billion deal to close on Monday — a one-off unicorn transaction from which a VC would not learn a general M&A playbook. M&A knowledge helped on two smaller Fund I exits, but the major wins would be companies such as Palantir and Canva, where it was irrelevant.

4. Horizon was conceived in late 2021 and began in 2022 on trust, not a fundraising grind

• Kory refused the emerging-manager slog: “I just didn’t want to be the guy pounding on doors saying, give me money.” Instead, over the years he had brought people into deals for little or no compensation — including Canva and a BillionToOne bridge round — “kind of like what” his friend had done for him with Palantir.
• In 2021, he ran two SPVs with a client whose bootstrapped technology company returned roughly 1.5–2x capital in six months, with more expected. Combined with years of making M&A clients money, that gave him a high-trust base for Fund I. He jokes that “great markets make us all think we’re smarter and better-looking than we really are,” so he was very smart and very good-looking in 2021.
• He decided in late 2021 to start the fund and was full-time in venture from 2022. He left almost all of the M&A cash flow to Mike, retaining only a limited sharing arrangement for a couple of years. The arc took about 20 years from the Stanford class taught by VCs; his consolation was that Ray Kroc started McDonald’s at 54, leaving Kory eight years to “mess around.”

5. BillionToOne: sourced from a canceled YC program, with the $4 million valuation coming later

• The origin was shoe-leather: TechCrunch profiled six companies in YC’s short-lived Fellowship, which allowed companies to participate without much traction and did not itself involve investment. Kory contacted three, and one was BillionToOne.
• The three founders were about to finish PhDs, apparently at Stanford, and were brilliant, energetic, raw, and unusually honest even though they were not polished fundraisers. Kory “slept through biology in ninth grade,” asked basic questions, and was impressed that they could explain a deeply technical concept clearly. In healthcare, he considers the ability to communicate across levels of background knowledge a meaningful signal.
• He later heard stories that the CEO had been the top scorer on Turkey’s high-school science exam and that one founder had brought blood samples from his father’s clinic into the United States in a coat or suitcase. He says these were otherwise rule-following people and that the stories conveyed dedication and determination.
• Kory wrote a small pre-seed check, doubled it while they were raising, and added again when they raised more at a $4 million valuation about a year later. These were small checks, but BillionToOne became the company into which he wrote the most checks and invested the most as an angel.
• The founder who left did so because the opportunity changed, not because of an integrity issue. The original plan targeted beta thalassemia in India and East Asia because they expected U.S. regulation to be harder. When the team realized the U.S. was accessible and a better market, the India-focused co-founder left. Kory says BillionToOne is the only company he has invested in that has overachieved every time, including through consistently professional quarterly updates.

6. The chemical microscope: roughly $400 million run rate, profitable, growing about 100%

• BillionToOne was public at the time of the discussion and was doing roughly $100 million in revenue per quarter — about a $400 million run rate — with roughly 100% growth and meaningful profitability, which Kory called unusual for a diagnostics company. About 90% of revenue came from NIPT, or noninvasive prenatal testing: fetal DNA circulates in the mother’s blood, allowing one maternal blood draw to screen for fetal genetic risks.
• The other roughly 10% was oncology, or liquid biopsy, which was growing faster and might be the larger commercial opportunity. After cancer treatment, a blood test can look for lingering mutant cancer DNA rather than waiting for a tissue biopsy to reveal recurrence.
• Kory describes the technology as a “chemical microscope.” Traditional approaches amplify a small signal but also amplify noise, like magnifying a penny until it becomes blurry. BillionToOne adds known DNA fragments to the sample, amplifies the material by up to about one million times, and uses what it knows to remove the noise, enabling extremely sensitive, single-molecule detection. “It sounded so scientific back then,” he says, “and it worked out.”

7. The lens: 90% of the conversation is life before the startup

• Kory’s core philosophy is that the earliest stage offers more alpha and more room for legitimately different approaches. He acknowledges that founder-focused, market-focused, and even less conventional approaches can all produce persistent success, while later-stage investing converges around a more legible set of analyses. His ambition, in his own flattering comparison, is like BillionToOne’s: finding the needle in the haystack.
• His method is explicit: “90% of the conversation is going to be on the things you’ve done before the startup.” Founders sometimes want to show slides, but he tells them to send the slides afterward. He has passed on founders whose initial ideas were bad but who later pivoted successfully, reinforcing his view that there is more signal in the life before the startup.
• He says founders are often coached to deliver the things investors want to hear, but they are not coached to tell their life stories. He also admits he experimented before settling on this lane, including traction-based investing — “$5 million valuation and $500,000 in revenue? Count me in” — and market-focused investing. Veeva is his example of a market insight that could have produced a great investment even without his current founder-first approach.

8. Founder trauma is overdone — look for absolute outlier signals

• Kory pushes back carefully on the idea that exceptional founders need traumatic childhoods. He notes that people such as Mark Zuckerberg, Bill Gates, and perhaps Larry Page came from relatively good families, while emphasizing that different people experience hardship differently. He therefore looks for “absolute signs of outliers,” which can take many forms.
• His example is smart-glasses founder Tom Suarez, whose interest in augmented reality began after Google Glass appeared while he was in high school. A former schoolmate later confirmed that Suarez was unusually obsessed and wore Google Glass constantly in their Georgia Tech engineering lab. Suarez dropped out, had limited success with a first company in the space, and is now pursuing a second one.
• The signal for Kory is obsession plus polymathy: Suarez is self-taught, can write software backward and forward, and can discuss waveguides and optics. Kory lacks depth in those fields but believes he can recognize legitimacy across them, often with reinforcement from world-class experts.
• The formula is extreme talent in one or two areas plus “an absence of negatives.” The negatives he watches for include bullshitting, a lack of intellectual honesty, and excessive commerciality. He adds the caveat: “I might get things wrong.”

9. Missionary over mercenary, and talent magnetism as a master signal

• Asked whether he would take the Anthropic founders over founders who moved to Meta, Kory says yes. He argues that many VCs do not care enough about missionary versus mercenary motivation, while noting that companies such as Uber and Facebook may have been highly competitive and intelligent without being genuinely mission-oriented.
• His mechanism is that believing there is something good in the work attracts talent. “Talent density and talent magnetism are really important,” he says; talent magnetism “might be the most important thing, frankly.” Given one datapoint about a company two years after investing, he would ask about the caliber of the engineer who had just joined.
• On scoring all this, Kory says that during his first five years of angel investing he ranked founders, markets, and other factors, but the later data was noisy. He rejects “gut” and “art and science” as pseudo-scientific clichés while conceding that his science is not very good. The Peter Thiel heuristic he retains is that if he likes something half as much, he should not write a check half as large; he should not invest.

10. The competition heuristic: funded rivals are inevitable — “bring it on”

• The reasoning chain is that a good market insight will attract funded competitors within six months to a year: Sequoia may back one, Kleiner another, and Andreessen another. Kory says the question is not whether competition comes but how the investor will feel when it does. “If it doesn’t happen, it means the market wasn’t there,” he says.
• Some investments leave him thinking, “Oh boy, I’m worried,” when rivals arrive. With a company like BillionToOne, the reaction was: “100 competitors came to the market. Oh, this is going to be fun.” He imagines early Ramp investors having a similar reaction while watching the Brex competition: “Bring it on. We’re going to be on top in the end.”
• The test resolves into execution. Kory describes BillionToOne’s CEO Ozan as a “maniac at execution” and points to the teams at BillionToOne and Ramp as examples of the execution quality his founder lens is trying to identify early. The transcript’s spelling of some names is uncertain.

11. Where he passes: hiring conviction — and why references and speed reads are limited

• A recurring near-miss profile has a quality insight, inbound traction, and a real market, but Kory is not convinced the founder will recruit great people. He calls hiring quality a major reason startups slow down and fail to scale. Growing from 20 to 60 people in a year while maintaining a very high hiring bar and strong mission orientation is theoretically possible but unlikely.
• YC’s advice to “hire your smart friends” may work for some of its hundreds of companies, but Kory wants founders who will try to hire the best people they can. When he is on the fence, recruiting conviction is a key tiebreaker.
• On diligence mechanics, “I don’t put much stock in references.” References on his most recent investment were good, but he notes that of course they were good. The useful extra signal came from having talked to the founder twice before the fundraising and again months later.
• He says it takes roughly 20 minutes to an hour to get a good read. He rejects the idea that investors can reliably know everything they need in 10 minutes, though he agrees that major negative signals can appear quickly.
• To guard against founders reverse-engineering his process, he looks for inconsistencies and exaggerations and asks basic questions without embarrassment. How a founder handles an investor who knows little about the subject is itself a signal. He doubts that genuinely large-outcome founders leave thinking the investor was stupid; they are more likely to think he was a decent person.

12. Red flags calibrate by age; intellectual honesty is non-negotiable

• His tolerance is behavioral rather than purely biographical: if someone says they were stealing cars at 13, that is not automatically a red flag if they are honest about it and have changed. He evaluates a 19-year-old differently from a 29-year-old and says he would consider investing in compelling founders who are 16, 18, or 19, while recognizing that he may miss some edge cases.
• What he will not compromise on is intellectual honesty. Across more than 100 angel investments, including at least one fraudulent company, he wants founders who will tell him the truth about the company and remain willing to say, “This thing isn’t working out.” The best founders are learning machines; bullshitting others or themselves obstructs learning.
• A possible workaround for performative young founders is talent magnetism. If a 19-year-old has an impressive co-founder or has persuaded strong people to join, that is evidence that others trust the founder. Kory can accept different styles if the underlying honesty remains.

13. AI may not break the formula — the uncertainly attributed “what stays the same” answer

• Asked whether AI will make founder diligence harder, Kory flags his own bias — he hopes the process will not change much — but refuses to be precious about the parts of the job he considers special. AI may automate everything, and he wants to use it rather than assume his work is exempt. His explicit concern is that AI may eventually outperform him at the pattern-recognition game, like a stronger chess player.
• He invokes a quote attributed to Jeff Bezos, or possibly someone else, about focusing on what will stay the same rather than predicting every change. Kory’s constants are exceptional, determined, resourceful, mission-oriented, intellectually honest founders. Even if companies have a thousand agents per human, he expects the most successful founders to remain broadly similar.
• His hedge is precise: “I’m not 100% sure I’m going to be able to execute on it. I think I will. But I’m very confident this formula will be working” as long as capitalism runs reasonably well.

14. Solo operating system: microwave optimization, inbox three, no significant Horizon setbacks yet

• Kory’s fund partner Mike is full-time in M&A, while Kory previously worked with an associate for roughly half of the fund’s history. He constantly reviews his own decisions and process, using a sports-GM analogy: he is always thinking about founders, signals, and how to improve. He tries not to over-contact founders and does not want a schedule of ten founder calls every day.
• The comic but genuine texture is his microwave optimization: if something has 40 seconds left, he thinks about whether to take vitamins or tie his shoes. He writes things down, keeps a task list, and usually aims for “Inbox 3” rather than literal inbox zero. He also criticizes productivity theater, such as driving an hour to save a nickel.
• He says he has had no significant setback in Horizon so far, though he has been unable to get into a deal and expects larger setbacks eventually. Fundraising has not been a real constraint; he has never wanted to spend more than 10% of his time on it and is not trying to raise the biggest fund.
• His angel-era pain includes companies that once appeared to be 20x outcomes and later became worth little. He considers that part of venture investing. His conduct rule is to follow up with every founder and give a reason for passing, because he cannot believe how often VCs fail to do so.

15. SendCutSend: from “no one wants to dance with you” to Collison’s $10 million

• Kory met Jim Belosic through M&A work on Belosic’s bootstrapped software business. Rather than push him to sell, Kory said that a business with good cash flow might not merit a sale at only three or four times cash flow. That advice built trust. Kory later remembered that Belosic was more animated about building cars and other projects, including an expensive machine in his garage, than about the software business.
• Years later, Belosic — who disliked VCs — came back to the finance person he trusted. SendCutSend is a vertically integrated manufacturing company that lets tinkerers order one or two custom metal parts instead of the thousand-unit minimum often imposed by local job shops.
• Unlike many manufacturing startups, it does not build its own machines. It buys capacity from expensive, top-of-the-line equipment suppliers such as Amada and sells the resulting parts. In 2021, about 80% of its business came from SMBs and tinkerers; by the discussion, that mix had flipped toward industrial and larger customers amid reindustrialization, defense technology, and robotics. The tinkerers often work at companies such as Tesla, SpaceX, and Anduril.
• Kory compares the company to AWS for hardware startups: little marketing, mostly inbound demand, and a useful service for companies that need parts quickly. China can often offer cheaper parts if customers wait two or three weeks, but SendCutSend competes on speed and on cases where sourcing from China is not permitted or practical.
• In the summer of 2021, when the market was focused on SaaS, Kory says no investor wanted to invest in the business. He invested $1.5 million through an SPV alongside a $2.5 million check from a VC at MHS Capital, then had the fund invest the following year. The MHS investor is identified in the transcript as “Mark Zuckerberg,” but that name is ambiguous.
• Kory did not initially apply his normal founder process. Belosic had traction, execution, and a category that was out of fashion, while lacking many prestige credentials VCs typically seek. Kory says that gave him an opportunity to use traction as part of the assessment; ultimately, “what you’re looking for is execution.”
• In a later first investor conversation, Belosic opened by saying he was considering a $1 billion valuation. At the end of the hour, Patrick Collison offered $10 million and asked to introduce him to friends, including Sequoia and Paradigm’s Matt Huang. Kory summarizes the reversal as: “no one wants to dance with you, and then everyone wants to dance with you.” He considers it especially satisfying to have been an early believer in an outsider founder and company.