20VC: David Tisch on Why High Ownership and Concentration is BS | Why No Manager Actually Wants to Stay Small | Why Competing with the Mega Platforms is a Losing Game | Why the Lean Startup is a Fake?

Host Harry Stebbings opens by admitting that he was wrong about David Tisch’s BoxGroup model, especially its low ownership targets and highly diversified portfolio. He points to BoxGroup’s early investment in Cursor, which Tisch says returned the fund multiple times, as well as investments in companies including Clay, Base10, and Rogo. Tisch says the two-and-a-half-year gap since his previous appearance has not made the basic job different: meet strong founders, decide whether to invest, and persuade them to take the firm’s money. Tisch rejects comparisons between today’s AI-driven market and 2021. He says every market period—2010, 2015, 2018, 2021, and the current moment—has had different macro and micro conditions. Although some companies are raising huge rounds at aggressive valuations, he says early-stage investors should focus more on whether a specific company is exceptional than on broad market conditions. AI is creating unusually fast mini-cycles: new model releases, rapid revenue growth in vertical AI companies, and major acquisitions can change market sentiment within two or three months. He still believes the technology is in the early innings, but says founders must navigate rapidly changing conditions rather than rely on a simple “bubble or no bubble” diagnosis.

Harry asks how venture economics can survive seed rounds that routinely reach $10–20 million, with some even larger, when an investor may need to commit a large share of its fund before much of the business exists. Tisch’s answer is that investors can either complain or accept the free-market price. If the company matters and the founders are worth backing, the investor has to decide whether to say yes at the available terms. He identifies two changes that can offset higher entry prices. First, the potential outcomes have expanded: a $10 billion company, once an implausible result, is now a valuation investors can reasonably try to envision at the early stage. Second, successful companies are often experiencing less dilution in later rounds. Tisch says Series B, C, and D financings can dilute founders and early investors by 3–8%, compared with 12–20% in the past. BoxGroup’s model is more exposed to entry-price variation because it invests across many companies rather than seeking a fixed ownership percentage. Tisch says a diversified portfolio must therefore blend different entry points, while Harry acknowledges that he previously failed to account for the expansion in outcome sizes and the lower dilution of breakout companies.

Tisch explains why he does not want to compete directly with what he considers the world’s best venture firms for 10–20% ownership in every seed company. Established multi-stage firms have spent decades building brands, relationships, and founder access, and they now compete aggressively at the seed stage. In a post-COVID fundraising market, founders can generate 10–15 competing term sheets, making the market much more efficient than it used to be. He challenges the standard claim from concentrated managers that they win 90% of the deals they want. Tisch calls that “total BS” and estimates the real figure may be closer to 20%. A concentrated fund might offer a large check and demand 10% ownership, but it may have to approach 120–150 companies to fill a portfolio of roughly 30 investments. The companies that make it into that portfolio are not necessarily every company the manager most wanted. BoxGroup’s lower-ownership model lets it invest in more of the founders it finds compelling, including deals where its initial ownership is only 1%. Harry agrees that a lead-only model becomes difficult when larger firms can offer triple the check at triple the price, and Tisch says he would rather not compete in that arena.

Harry worries that oversized or tranched rounds can distract founders, encourage them to spend too quickly, and allow another investor to enter at double the original valuation without any meaningful change in the company. Tisch says that risk exists, but does not apply to every founder. More runway, a higher valuation, and market momentum can help recruit employees, win customers, and signal credibility. Founders who accept more capital must execute differently from founders operating under severe capital constraints. Tisch dislikes tranched rounds in principle because he would prefer to invest at the lower price, but says the decision is ultimately a partnership question. If a founder insists on a higher-priced second tranche, BoxGroup may accept it even while being unhappy. He jokes that he will be “sad and take it.” He says investors must earn the right to structure the deal they want rather than treating founders as obligated to provide it. This leads to his strongest criticism of the Lean Startup: he calls it a “total fake” and an “awful” decade-long form of brainwashing because successful venture-backed companies generally require substantial funding. He distinguishes being well-funded from being undisciplined, arguing that companies seeking the venture flywheel are eventually “fat,” not lean; he acknowledges that bootstrap businesses such as 37signals are an exception.

Tisch describes BoxGroup as operating two $275 million funds: a core seed fund that will invest in about 150 companies and a follow-on fund that makes roughly 20–30 larger “lean-in” investments, usually in existing portfolio companies. Those follow-on checks range from $5 million to $30 million, and the largest aggregate investment in one company has been about $40 million. He names Ramp, Clay, Base10, Rogo, and Tandem as companies where BoxGroup has written substantial checks. Cursor is the clearest fund-returning example. BoxGroup met the company before it had scaled into the business it became, and Tisch says the investment returned the fund multiple times. He emphasizes that Cursor’s outcome arrived in about four years, but that other fund companies can take longer. Clay took a bumpy four-year path before its explosive growth, while Base10 needed inference to become a meaningful category before turning into an outlier. Tisch says early-stage investors should look for founders running a world-class experiment: hiring unusually strong people, winning inaccessible customers, and making product and go-to-market progress. The outcome may not be immediately obvious, and founders must be able to navigate the messy middle rather than merely display an impressive initial story.

Tisch tells the story of Rogo, an AI-for-finance company led by Gabe. After BoxGroup invested at seed, Tisch helped introduce the company to approximately 40 Series A investors. Thirty-nine passed, sometimes quickly, citing insufficient traction, uncertainty about the market, or doubts that the company could become large. Keith Rabois ultimately led the round, and the company later attracted Sequoia and Kleiner Perkins. Tisch uses the example to argue that companies can be misclassified by the market at a particular moment and still compound into a major business. The conversation then turns to follow-on rights. Tisch rejects the idea that pro rata is simply unnecessary marketing, but says the right to invest in subsequent rounds is earned whether it appears in the documents or not. A founder may ask an early investor to step back, and refusing can create either financial risk—losing the allocation—or reputational risk by damaging the founder relationship. He nonetheless believes contractual pro rata rights help founders negotiate later rounds: if a new investor wants 22% but existing investors have reserved allocation, the founder can use that cap-table constraint to resist excessive dilution. Tisch says large firms’ ownership demands vary by partner and situation; when a firm says it “needs” 20%, he often interprets that as “would like” rather than a true necessity.

Tisch says his primary responsibility is to BoxGroup’s LPs, while founders are his customers. He generally does not sell early because a founder may view an early secondary sale as a breach of trust, and because the greatest compounding often happens late: moving from $2 billion to $4 billion and then $8 billion can be easier than reaching the first billion. He distinguishes a founder’s liquidity needs from a VC’s fund-management incentives. A founder may use secondary proceeds to reduce personal stress, take greater business risks, or retain employees through tender offers; those reasons do not automatically justify an investor selling. He believes the venture market is consolidating, with fewer genuinely active new firms even though old names remain technically alive. Many new managers begin with a small seed fund but ultimately aspire to become multi-stage firms, so Tisch asks them what “fund four” will look like. BoxGroup, by contrast, has expanded from its original scale while keeping the same early-stage identity. Its ten-investor team sees roughly 50–100 companies per week, meets twice weekly for investment discussions, and does not vote. One person must be genuinely excited about a deal. The firm also prohibits negativity from people who have not spent time with the founders and does not allow “grenades” to be thrown into another partner’s deal. Tisch says the magic of early-stage investing is not explaining why a company will fail, which is easy, but squinting to see why it can work.

In the rapid-fire section, Tisch names Y Combinator as the strongest business model in venture. He says YC created a distinctive product: a world-class brand and a powerful attractor of talent. For later stages, he lists several impressive firms rather than selecting one, including Benchmark, Sequoia, Andreessen Horowitz, Thrive, Addition, Ribbit, Founders Fund, and Khosla Ventures. For growth, he mentions Spark Capital’s growth strategy, Meritech, Menlo, Coatue, Dragoneer, and Green Oaks. Harry chooses YC for seed, Benchmark for Series A, and Sequoia for Series B and beyond. Tisch says today’s strongest founders are more technical, and there is less room for mediocre execution. He sees fewer sales-led founders than in previous eras but expects go-to-market and sales to become differentiators again as current products mature. He says consumer brands can still become large businesses, but BoxGroup no longer funds them at seed because there is not enough early data on customer retention, lifetime value, and acquisition economics; a specialized Series A investor may be better suited. Looking ahead, Tisch is excited by the uncertainty surrounding AI, autonomous driving, science, and physical-world applications. He believes autonomous driving is sufficiently safe to prevent many deaths, but society’s attachment to existing freedoms and institutions slows adoption. He hopes technology, especially AI, becomes an equalizer whose benefits reach more than a small group.