Innovators
4 Curated Dialogues
José
Key Views & Dialogues
Why the Future of Venture Is Hard Tech, Drones & Physical AI | Ian Rountree
Cantos’ edge is deliberate smallness: Fund IV reached $70M, supporting $1.5–4M checks before websites and same-day SAFE execution. Ian favors hard tech, vertical integration and manufacturing scale, with Neros at 1,000 drones weekly and targeting a million annually. The opportunity is cost advantage and defense demand; China’s control of motors, actuators and rare-earth inputs remains a physical-AI risk.
View summary (8)
- José frames Roelof Botha’s “return-free risk” charge around the median fund and notes that Ian voluntarily put a hurdle rate in his fund. José’s caveat is that median venture can underperform public markets or T-bills, while the top 1–2% put up “mind-boggling numbers.” Ian’s answer is persistence: top venture firms repeat across vintages because “the asset picks the investor rather than the other way around” — anyone can buy Micron, but a pedigreed founder shortlists Sequoia — making outlier-fund selection the relevant LP question, not the median.
- Cantos’ answer to the mega-platforms is deliberate smallness: Fund IV grew only from $50M to $70M so it can fly “under the floor of economic rationality for the big funds.” Writing $1.5–4M first checks (“if the company has a website, it’s too late for us”), everyone on the investing team can tell a founder “your company is a meaningful percentage of my net worth” — alignment the multi-stage giants may not offer at inception.
- He now prefers “hard tech” to “deep tech,” having “come to question the myth of the breakthrough.” Most value is “a thousand little things compounding” — engineering and execution rather than primarily science risk — and he skewers fusion’s $10-trillion narratives: “what happens if you have two fusion energy companies succeed? Do you understand that competitive arbitrage is a thing?”
- Howard Marks’ consensus/non-consensus matrix “doesn’t apply to early-stage venture,” and AngelList data from Abe Othman shows price positively correlates with performance. Insolvent startups must become consensus before the next round, so edge is access plus speed — SAFEs can be signed “same day if we meet them before 2:00 p.m. Pacific.” His biggest misses: passing on price, and passing on decks like Castelion’s pre-seed, which Cantos later bought at Series A “a heck of a lot more” expensive.
- The core thesis: don’t sell technology to staid incumbents — vertically integrate and compete with them. Earth AI applies for its own mineral rights rather than selling AI leads to miners; Shinkei sells branded fish, not robots; the lumber startup designed a bigger saw than the industry had built and now runs its own Mississippi mill, selling “a literal commodity” off “a preferential cost structure.”
- Neros went from a pre-seed check into Soren, a drone-racing world champion, in August 2023 to America’s highest-volume drone maker at 1,000 drones a week, targeting a million a year. A Pentagon official’s tell: “we’d give a billion dollars to someone who could make a million of anything” — and Ian’s corollary, “if you have a drone that’s second best but you can make 10,000 more of them… I know which one I’m picking.”
- The strategic risk is China’s grip on the physical-AI supply chain: 50–70% of a humanoid robot’s BOM is actuators and motors, “and we don’t really make cheap motors in the Western world.” Rare earths, processing chemicals, copper and motors are concentrated in China — Ian says Beijing’s leverage already forced the U.S. to back down on sanctions — and if physical AGI arrives on that stack, “they’ve got all the power.” He invests in the U.S. and its allies as a principle, not as a protest; without that principle, “I would find a way to invest in Unitree.”
- On momentum and meme-company dynamics, Ian is torn, while José proposes being “half stubborn and contrarian” and “half” acknowledging “the game on the field.” Path-dependent, unprofitable startups “can’t afford to be non-consensus for more than a few years,” and José argues secondary liquidity now makes venture trade more like crypto — but Ian’s stated endgame remains “something enduring that compounds forever,” not dumping on retail.
Most Venture Funds Are Playing Momentum Games | Michael Dempsey
Venture momentum is uniquely fragile: without scaled liquidity, dilution and preferred stacks can turn a softening bid into a 100% drawdown rather than 10–20%. Compound’s seed data finds that every valuation bucket above the bottom quartile produces similar top-decile outcomes, while Dempsey sees a 2028–2030 AI window and must rebuild the firm’s edge through higher-order analysis.
View summary (8)
- Dempsey’s core warning: a lot of venture funds are playing momentum games, and momentum is uniquely lethal in venture because there is no scaled bid. In public markets a fading momentum trade comes down 10–20% into a live bid; in illiquid, dilution-heavy private markets, “when the bid starts to soften, it doesn’t soften 10 or 20%, it softens 100%.” Investing “looks easy in certain 36-month windows and usually you find out it’s hard in 48-plus months.”
- He rejects the “pay up for legible founders” seed strategy on data: everything but the bottom quartile of seed valuations produces the same distribution of top-decile outcomes. Loss ratios on high-priced seeds aren’t meaningfully better than cheap ones, and Compound’s two largest companies — Runway and Wave — had founders who “even today would not be considered super legible.” With AI lowering the barrier to creation, he finds it strange to expect talent to narrow rather than disperse.
- Maximal obviousness makes a sector uninvestable for Compound: venture is “the lowest conviction asset class in finance,” so consensus themes see model collapse and identical companies. José suggests AI is now roughly twice as obvious as its “maximally obvious” 2024 reading; Dempsey would rather wait for the dust-settling window he pegs at 2028–2030, when cheap intelligence commoditizes much of the market while frontier intelligence creates new opportunities.
- Compound’s edge is decaying by design and must be rebuilt: the 2016–2022 edge of simply understanding academic research is “largely gone” now that anyone can put a paper into ChatGPT. The replacement is prescriptive first-, second-, and third-order-effects thinking — something AI still can’t do because “nothing shows me that AI is able to think out of distribution” — plus understanding how narratives drive multiples and late-stage flows, which models are “horrible at.”
- The public-markets fund targets venture-like IRRs through dislocations the market misreads: software facing AI-driven margin compression can re-expand as inference costs fall — Pinterest’s fine-tuned open-source models reportedly run at 8% of frontier-API cost. Bio longs are undervalued data assets and left-for-dead small caps; shorts are narrative-driven deep tech where the first sign of invalidation can send stocks down “40 to 60% in 4 to 6 weeks.”
- Mag 7 views: Meta is “probably best positioned to monetize AI regardless of whether they build on the frontier”; Google’s structural advantages — including GCP, models, and unusually cheap debt — make him bullish; Apple is well positioned. He is less high-conviction on Nvidia and memory stocks. His unchanged 2024 “AI obvious basket” of about 30 names (ASML, Micron, hyperscalers, Alibaba) “has just so drastically outperformed,” and he thinks theme baskets à la Citrini may be the future of active management.
- Robotics was the big mistimed call — 2016–18 bets on generalizable deep-learning policies were “8 to 10 years too early” — but “this time is really different,” and he’s backing a stealth vertically integrated frontier lab plus vertical players like Alquist. On the $0-to-$10 trillion humanoid claim that José says he has heard Andrew Kang make, Dempsey urges a more measured view: humanoids make most sense where the relevant cost is human life — military or industrial — and industrial robotics is already the most penetrated area.
- Macro stance is “rolling bubbles,” not a single top: CoreWeave went to 79 and back over 100, while Palo Alto fell more than 40% and recovered toward highs — whiplash is structural because everyone’s stop is the same 20% drawdown. On selling discipline, Compound’s data says exiting hyper-reflexive names at all-time-high NTM revenue multiples beats selling on a 25% retrace — though “you might miss the leg 3 years later.”
Building the SoftBank of Robotics | Andrew Kang
Humanoid robotics may be approaching the pre-ChatGPT phase of AI, with $50,000 robots requiring only 20,000 units for $1B revenue and a projected tens-of-billions industry within two to four years. Robo Strategy (NASDAQ: BOT) seeks to challenge SoftBank through public markets, favoring vertically integrated players with embodiment-specific data; US restrictions on Chinese robots could help, while robotics data businesses face a two-to-five-year durability test.
View summary (7)
- Kang’s core thesis is that humanoid robotics today is where AI sat in 2022 and crypto sat in 2014–15 — a pre-ChatGPT-moment industry the general investment community still misprices. His unit math: at $50K per robot, $1B of revenue is only 20,000 units and $10B is 200,000 — trivial next to phones or cars made “in the hundreds of millions or billions” — and he expects industry revenue in the “tens of billions somewhere between the next two and four years.” His edge claim: “the more difficult an industry is to understand, the more alpha you can generate.”
- The TAM answer to humanoid skeptics is the general-purpose-device argument: “all human physical labor is your TAM,” and scale economics beat special-purpose machines the way smartphones beat cameras and GPUs beat ASICs. Proof by existence: BMW, Volkswagen and BYD still employ hundreds of thousands to million-plus workers despite decades of industrial automation, because most factory work changes too often to justify fixed automation below multi-million-unit SKUs.
- Robo Strategy (NASDAQ: BOT) exists because Tom Lee raised $18B in 8–9 months “for an asset that went down 50%… in a thesis that wasn’t that good” — so an asset class actually going up could raise “maybe an order of magnitude more.” Kang’s ambition is explicit: use public capital markets to reach beyond a16z-scale ($15B last fund series) and “become a new challenger to SoftBank,” applying a similar private-to-public valuation concept to the one TransDigm and Constellation Software use (buying at 3–8x earnings, marking at 15–40x).
- Post-ChatGPT power laws changed the venture math: you can now invest at multi-billion valuations and still expect “a 50x, 100x, 500x investing at a $2 billion valuation” — which is what makes deploying billions viable at growth stage. Portfolio construction is a barbell: 6–10 companies at ~70% of the fund, 20–30% in pre-seed-to-Series-B optionality, and “let winners ride.”
- He concentrates in vertically integrated players because embodiment-specific, on-robot data is “absolutely necessary” — a model trained in the wrong body is like learning basketball at 7 feet then playing at 5'10": “all wobbly.” The “Scale AI of robotics” cohort is the Buzz Lightyear meme — a thousand identical companies — and a “short-term cash cow” whose revenues might plummet “two, three, four, five years down the line” once robots self-improve; “it almost feels like trading as opposed to investing.”
- On US vs China he rejects the who-wins framing: both build great products, but American companies win for shareholders on 20–50%+ margins and stronger capital markets — “you can sell 10 times less product but be a bigger business.” Plus bipartisan bills restricting Chinese robots in the US are “very high probability,” potentially ring-fencing the biggest market (30–60% of any major product): “America’s the place you want to invest.”
- His philosophy is pricing risk/reward, not contrarianism: citing Coatue data that bigger public companies 10x more often than small ones, he backs the strongest founders at higher valuations — “winners win” — versus the Character AI/Inflection AI trap of buying “grittier and cheaper.” One admitted change of mind: more bearish on model-only companies, more bullish on open source — “Kimi getting as good as Claude on some benchmarks” will repeat in robotics — which he reads as bullish for hardware companies with secured supply chains, not for restricted Chinese robots.
Identifying Generational Managers with Jordan Nel
Jordan Nel argues venture’s bottom-up opportunity can survive weak venture beta because outlier funds capture multiples from overlooked companies and signal larger funds before they can deploy $30–100M tickets. The investable edge lies in pre-consensus access and GP calibration, but frontier AI is a winner’s market, advantages decay, and first funds need a prior angel record.
View summary (7)
- Jordan Nel, ex-Hummingbird/Nomads, refuses to defend venture as an asset class: “I’m not sure that venture beta justifies the illiquidity premium.” His case is strictly bottom-up — “for every outlier company, there are at least three outlier funds”: every Coinbase has a Ribbit, a USV, and a dozen funds that return multiples off one name. Even in 2021 there were pockets of arbitrage (“if you were fishing in Indian life sciences, probably you’d be the only one”), and the LP’s job is finding those pockets, not timing beta.
- Emerging managers shouldn’t try to beat Andreessen or Sequoia — they should see companies first and “act as signals to the bigger funds such that they will eventually follow on from you.” Big-fund GPs on 20-person teams aren’t incentivized to fish for a $5M post-money-cap founder in the backwaters; they need to deploy $30–100M tickets to drive carry. It’s not either/or: big funds run fund-of-funds programs that seed emerging managers and use them as deal flow.
- Nel rejects “contrarian” for “pre-consensus” — the art is making founders “legible to capital” and sequencing the cap table. Bridget Mendler (Disney child star turned PhD, founder of Northwood Space) could have raised from megafunds but deliberately built her pre-seed with signal angels; a founder from Abuja, Nigeria couldn’t get a term sheet for six months until one risk-on pre-seed GP unlocked the San Francisco hard-tech ecosystem. On a founder with direct access to top funds, Nel says he would do it for “Sam Bradberry Taylor,” but adds, “I’m not Bret Taylor,” and most founders can’t.
- After four years and meeting 5,000-plus GPs, Nel gave up on a precise GP archetype; the key lenses are “how do they make decisions and who do they hang around?” Indexing on hyper-articulate theses was “a mistake” — sometimes “this is the best founder I’ve met this year” is a back-up-the-truck signal, and the killer calibration question is “Okay, well, who else did you meet this year?” Compound’s Michael is the case study: the theses are visible, but the real edge is “a very, very good nose for who to listen to.”
- Edge is cycle- and sector-dependent: in hot Bay Area frontier AI “GPs who win will win,” and if you can’t outbid Sequoia, Conviction, and Benchmark, “you probably shouldn’t play.” In colder sectors, picking wins — which is why institutional LPs were interested in Dimension, the Lux spinout focused on tech bio, as “in some sense… the only game in town.” Forward-looking opportunities include Indian life sciences, Brazil pre-seed (“so few VCs actually writing pre-seed checks”), and capital-starved China — as long as you keep the bar at “global maximum” with San Francisco benchmarks.
- Concentration is “downstream of convexity,” not a rule — 40 checks at $2M post are “basically free call options,” but almost nobody besides Silicon Valley Angel has the network to do it without adverse selection. Price discipline swung from lax to overtight before settling on mentality over absolutes: what matters is getting “the best terms at the table,” because “the returns follow the mentality of the GP.” First-fund managers are backable only with a prior angel record: “if you’ve never written a check before, I have no data points to move on.”
- The process lesson is constant WhatsApp debate rather than a scheduled IC. Nel’s formative scar tissue is a South African airline he bought going into COVID as deep value, only for the government counterparty not to pay and the business to go bankrupt: “I think probably there is no price.” The venture translation is emotional, not tactical — “I just don’t want to lose the money.”