Keith Rabois on Returning to Khosla Ventures and Investing at Seed

Keith Rabois joins Harry Stebbings after announcing his return to Khosla Ventures from Founders Fund. He discusses partner debate, seed pricing, reserves, founder relationships and liquidity. His deal accounts, market forecasts and political opinions reflect the January 2024 interview; he has financial interests in many companies and funds discussed.

Rabois previously spent six years at Khosla Ventures and says he remained close to several partners after leaving, frequently co-investing and working with them. He cites shared investments and board work, including companies in which he had personal or professional interests. The renewed collaboration made returning feel natural to him.

He missed the extensive Monday partnership meetings, where colleagues debated new deals and existing holdings. Although the sponsoring investor retained the decision, he says hearing different perspectives sharpened his thinking. He also enjoyed learning about fields such as batteries, robotics, AI and medical technologies outside his own primary investment specialty.

Rabois contrasts Khosla’s lengthy, preparation-heavy partner discussions with Founders Fund’s more individual portfolio-manager model, while noting that practices changed and the difference was not absolute. The former demanded substantial Sunday preparation. He valued the debate even when it challenged his initial view.

He describes investment discretion in terms of social capital: pushing an unusual deal against colleagues’ objections uses trust, and a good result can replenish it. Feedback can also increase conviction. He recalls partners encouraging him to pursue Faire and Opendoor rather than always serving as a brake.

He regrets not increasing a seed offer for Rippling after another investor offered a higher valuation. His memory is approximate: he recalls offering around $5 million at roughly $25 million post-money, with a competing valuation near $35 million. He attributes the missed opportunity to his own decision, including concern about internal social capital.

He also recalls declining to join Robinhood’s board at seed after colleagues cautioned that too many early board seats would consume time. He considers the general advice reasonable but wrong for that particular opportunity. Later, he accepted a board role at Faire, whose founders he already knew from Square; he has an investment interest in Faire.

Rabois says Khosla historically was more price-disciplined than Founders Fund, though he found Founders Fund more disciplined than its reputation suggested. In his view the distance between firms has narrowed. These are an insider’s comparisons of organizations with which he has professional and financial ties.

He invokes Peter Fenton’s argument that an early-stage investor’s objection to price can sometimes conceal lack of conviction. For a truly exceptional seed or Series A company, he thinks the right selection can outweigh a higher entry price. By Series B, he says, paying too much can make even a good company an unattractive investment.

Stebbings points out that doubling average entry valuations can materially reduce returns. Rabois agrees that an investor cannot apply the same strategy to a different price environment and expect identical results. The question is when an exceptional opportunity justifies departing from ordinary discipline, not whether price is irrelevant.

He recalls investing in Ramp’s seed round at an unusually high valuation, above $30 million and perhaps around $40 million post-money. He believed his experience in payments and view of the competitive market justified the price. His prediction that Ramp would outperform competitors is an investor’s opinion about a company in which he has a stake, not an established outcome.

Capital needs and future dilution also affect his willingness to pay. A company that must reach a costly, discrete technical milestone may fail if a high entry price and insufficient resources leave it unable to raise the next round. He says the shape of the business matters as much as the headline valuation.

On reserves, Rabois describes Khosla’s more planned allocation process and Founders Fund’s case-by-case approach. He personally leans toward making follow-on decisions as opportunities arise, assessing progress, founder ability, the incoming investor and price. He presents this as difficult judgment rather than a proven universal rule.

Stebbings worries founders might prefer a fund that promises reserves in advance. Rabois says he did not see the absence of a formal reserve policy become a practical problem at Founders Fund. Both speakers are venture investors, and their preferences reflect their own fund strategies.

They agree that ordinary seed deals do not necessarily compensate for their risks, though exceptional outliers can. Rabois views a $5 million investment at a $25 million pre-money valuation as expensive for the information available and says he would require extraordinary reasons, even though he has done deals at similar terms.

One reason to accept a larger seed round is that the amount is genuinely needed to reach the next meaningful milestone, rather than reverse-engineered from a desired valuation. He says a Miami company he backed needed $5 million and was later supported again by Founders Fund; this is his account of an investment with a financial interest.

A useful investor, he says, helps a founder identify the achievements that will unlock future financing and estimate the time and capital required. Too little money can prevent a valid test; too much can burden the company. He recalls Opendoor raising $10 million at seed because testing home purchases and resales required substantial capital.

Series A can offer a strong risk-reward trade-off because investors have more information than at seed. Rabois nonetheless prefers leading seed: he thinks assessing a team and a presentation is a comparative advantage, competition is often less intense, and early involvement lets him influence decisions before processes harden.

Stebbings suggests some multistage investors have become more active at seed while pulling back at Series A. Rabois says strong competitors remain active and he often encounters the same one or two investors. Even if Series A has attractive pricing relative to the information gained, competition can make winning the deal less certain.

Stebbings asks whether former operators risk backing companies simply because they want to be needed. Rabois says helping is useful only if it does not crowd out the company’s own capabilities. He offers frameworks, lessons from other companies and introductions rather than assuming he should make every decision.

He recalls founders such as Patrick Collison arriving at meetings with their hardest unresolved questions. Even adding value on one or two was meaningful to him. He describes founder–investor collaboration as helping a founder work through trade-offs, rather than supplying an answer by default.

Rabois criticizes many growth investors for pursuing momentum without understanding price or company building, and says growth is not his main comparative advantage. This is a broad industry opinion, not evidence that every growth fund failed. He has occasionally invested later when he believed he had unusually strong knowledge.

At Stripe’s Series C, he says experience at PayPal and Square helped him assess the business. The price was unusually high for Khosla at the time, but partners encouraged him to make his own call. He similarly cites long familiarity with Faire and Ultima Genomics as reasons for later-stage investments; all are examples involving his investment interests.

Asked about approximately $3 billion raised by Khosla, Rabois breaks the total into roughly $415 million for seed, $1.5 billion for venture and $800 million to $900 million for growth. These are his approximate fund-size figures, not a statement that the entire sum belongs to one fund or strategy.

He says appropriate fund size depends on investment stage, opportunities, ownership goals and the number of partners capable of leading deals. A seed portfolio of roughly 30 to 50 companies is, in his view, within common guidance. The arithmetic depends on fees, reserves, check sizes and allocation decisions; none of it guarantees returns.

He compares the weighting with Founders Fund, where he estimates a smaller venture pool and a much larger growth pool. Strategy, team capacity and fund size must fit together rather than being judged from a headline total. His numbers are approximate and specific to the funds discussed at that time.

Rabois says he does not yet know enough about Khosla’s younger team to judge whether it should be younger. One lesson he hopes to bring from Founders Fund is greater analytical rigor for later-stage and pro-rata decisions. Even an early-stage investor must decide whether to follow a successful portfolio company at each new price.

He treats founder–investor selection as matchmaking. A founder with a strong cultural philosophy needs a partner who can understand and challenge decisions within that framework. A design-led founder such as Jack Dorsey, he argues, may not work well with an investor who dismisses design as insufficiently measurable.

He sees meaningful overlap between Founders Fund and Khosla portfolios, naming several shared investments. In his view both look closely at founders, while Khosla may sometimes start more from technology. Portfolio overlap is his observation and does not establish that the firms make identical decisions.

Rabois considered starting his own fund at various points but says the administrative and organizational burden was unattractive. He prefers discovering founders, supplying capital and advice, and working with them over time. Even if later operating costs become manageable, he sees the initial setup as a significant fixed cost.

Asked what motivates him beyond money, he says he wants to have a meaningful impact on people’s lives. In venture that means helping founders build companies they might not otherwise have built. This is his personal account of motivation, not a measure of an investor’s actual contribution.

His advice to investors is to identify a genuine comparative advantage: why should this founder take your money and why can you judge this opportunity? He believes broad venture returns are poor outside a small top tier, though he does not provide supporting data for the percentages he cites.

His examples include knowing Faire’s founders from Square and drawing on payments experience at Stripe. If he is not the best partner for a founder, he says, he looks for a colleague who may be a better match; at Khosla, founders could sometimes meet several partners and choose. These are his descriptions of firms in which he has worked.

Stebbings asks whether a board member’s words can carry too much weight. Rabois agrees and credits Roelof Botha with teaching him to ask questions rather than prescribe answers. He also tries to state how confident he is, explain the logic behind a view and let the founder evaluate it.

He says he rarely tells a founder what must be done. One exception was advising a frugal founder to pay for office air conditioning after hours rather than make employees work in the heat. The anecdote illustrates his preference for restraint with occasional direct advice.

On selling positions, Rabois admits he has not mastered the timing. He recalls Vinod Khosla arguing that the firm should hold Square shares and says the decision produced better results. Rabois had personal Square shares and says he recused himself from the firm’s sale decision; the account does not establish a general exit rule.

He admires Khosla’s ability to trace the implications of emerging technology and his willingness to spend long hours with founders and researchers. He cites Khosla’s early interest in AI and medicine as an example of foresight. These are Rabois’s assessments of a colleague and the firm he is rejoining.

Having worked at two firms, he hopes to distinguish inherent features of venture from choices about culture, hiring and decision-making. Founders Fund’s growth-investing discipline, he says, helped him assess valuation and follow-on decisions. He wants to carry those lessons into his next role.

As a parent, he believes young children absorb more from their environment than adults may realize. He worries about raising children without entitlement when they have substantial advantages, and seeks advice from other parents. These are his personal observations, not conclusions established by the interview.

Stebbings worries that intense professional ambition may conflict with being present for children. Rabois believes exceptional achievement usually requires deliberate trade-offs and a great deal of work, while acknowledging that different people allocate time differently. This is his view, not a universal prescription for family life.

Asked about Bitcoin, Rabois says he does not know what will happen. His theory is that adoption and value may rise where confidence in the rule of law weakens, while greater stability might reduce that demand. He expects 2024 to be turbulent; neither the causal theory nor the price outcome is established by the interview.

He predicts Trump will not be the Republican presidential nominee and says he lacks enough information to judge the legal cases. That was a January 2024 forecast, not a current account of events. He also offers his own antitrust views on the proposed Adobe–Figma and Visa–Plaid transactions, which should not be treated as legal findings.

Rabois expects IPO opportunities in 2024 and argues that companies can benefit from going public earlier, with the required financial criteria changing over time. This is his then-current prediction and advice, not a guarantee that any named company would list or achieve a particular outcome.

He identifies choosing which first meetings to accept as a persistent weakness: seed investors receive more opportunities than they can personally evaluate. He recalls declining meetings with companies that later did well and is still looking for a better screening method. Delegation can help but introduces another person’s judgment.

In quick fire, he mentions the Thiel Fellowship as an educational path he admires and offers his opinion about Europe’s economy. He praises Stebbings’s combination of media and investing as differentiated. These brief comments are personal views rather than an analysis of every alternative.

Looking ten years ahead, Rabois says he does not expect to retire or start his own fund, though he leaves open the possibility of work outside technology. He says Khosla’s limited partners committed before knowing he would rejoin. Both his future plans and his assessment of the move are statements made at the time of the interview.