Benchmark普通合伙人Everett Randle：为何巨型基金不会带来良好回报
AI应用经济学打破SaaS熟悉的80%毛利率评分卡。 Randle认为，投资人应该基于终局5–7年的毛利率、毛利润倍数，以及单客户绝对毛利润进行承销：一款毛利率50%的AI产品，如果每客户贡献50万美元毛利润，可能优于毛利率75%、但每客户仅贡献20万美元毛利润的SaaS产品。他直言：“我们今天不应该把那么多注意力放在毛利率上”，尤其是高推理成本可能恰恰反映了真实的AI使用量。
Coding已经是“黄金品类”，即使Cursor的市场份额仍在下滑。 Randle估算，代码生成在约2.5年内从几乎为零增长到60–70亿美元ARR，今年还可能再增加40–50亿美元；Cursor的份额可能已从约80%降至25–30%，但它仍在服务一个大得多的市场。使用量最高的产品——包括Cursor、Claude Code和Codex——也会通过部署最快迭代，并可能“把所有人甩在身后”。
按最新报价，Randle会以5000亿美元选择OpenAI，而不是3500亿美元的Anthropic。 Anthropic在Coding和B2B商业化上仍略占优势，OpenAI则凭借Codex追回了差距；决定性资产是ChatGPT，Randle认为其增长轨迹几乎不可能被打断。他曾因非营利结构和稀释担忧，以320亿美元的估值错过OpenAI，如今预计OpenAI明年就可能成为万亿美元公司：“我只见树木，不见森林。”
Benchmark的小基金目标是最大化倍数，而不是赢下每一轮巨型融资。 Randle表示，其上一支基金最好的5笔投资按上一轮价格计，约为1笔60倍、2笔30倍和2笔20倍——ChatGPT之后的OpenAI融资没有一轮能匹敌。Benchmark因此不必买入每一家实验室，其两大北极星是成为创始人最亲近、ROI最高的合作伙伴，并为LP的创投组合带来最高的资金倍数回报。
巨型基金可以赚到巨额绝对利润，却仍无法通过创投回报测试。 Randle的论点是结构性的：“基金多大，你就得投多大”，因此70–100亿美元的基金必须开出巨额支票，而这些支票最终必然成为主要产品和组织重点。他怀疑巨型基金管理人能否可信地承诺，在相关基金组合上实现5倍净回报；Harry的反驳是，前所未有的结果或许仍能拯救这一模式，Randle接受其美元层面的可能性，但不认为倍数也必然成立。
AI的护城河仍然是技术，而不只是分发。 分发只能让公司获得构建产品的机会，但Randle认为，优秀AI产品需要稀缺人才、精细的模型管线和工作流设计，而不是在文本框旁边“接入OpenAI API”。实验室把每月20美元或200美元的体验设为基准，因此应用公司若要收取更高价格，就必须创造足够深度、能够穿越下一次模型迭代的差异化工作流价值。
绝对毛利润可能比毛利率百分比更重要。 Randle偏好的比较是：一家类似ServiceTitan的企业，每客户贡献20万美元毛利润、毛利率75%；一家AI公司每客户贡献50万美元毛利润、毛利率50%。他会选择后者，因为它覆盖了更广泛的客户关系，也获取了更多经济价值。AWS是他的类比——更低的毛利率可以与大得多的客户支出并存。
只有当应用越过实验室设定的基准，增长才是真增长。 AI公司可以在不到1年内从0增长到100，但Randle始终牢记“来得快，去得也快”。Jasper曾快速增长，随后因GPT-4让其输出看起来与ChatGPT的20美元套餐过于相似而萎缩；它后来依靠更具差异化的工作流软件实现复苏。应用公司必须超越用户直接从实验室获得的产品能力。
商品化AI基础设施可能压倒质量层面的顾虑。 Randle改变了对AI云的看法：他起初认为CoreWeave只是转售商品化算力的中间商，后来天文数字般的推理需求压倒了这一质疑。他称CoreWeave的市值约为600亿美元，Nebius约为300亿美元，公开市场相关公司合计超过1000亿美元。他仍预计CoreWeave及类似公司最终可能下跌70%，但认为需求足以支撑顺势投资。
Benchmark按基金规模进行投资，而不是追逐每一家实验室。 Randle将Conway定律应用于创投：机构会“交付”自己的基金规模、团队和结构。一支拥有50名投资人的70亿美元基金必须参与巨型融资，因为十亿美元级支票是少数能有效部署其资本产品的方式。Benchmark规模更小，可以追求集中投资和更高的潜在现金回报倍数。
持股比例是投入，合作关系和回报才是产出。 Benchmark的两大北极星，是成为每位创始人最亲近、ROI最高的合作伙伴，并在LP的创投持仓中创造最高的资金倍数回报。惯常的20%持股目标本身并不属于其中。Randle表示，即使Benchmark在Mercor的持股比例更低，只要仍是公司最具影响力的创投伙伴，也能实现卓越回报。
价格只有放在公司自身的上行空间中才有意义。 Randle的排序是人、产品、市场：人是上游引擎，产品是检验其质量的最强证据，而市场最具可替代性，因为公司可以转向。SpaceX在1500亿美元时、Rippling在2.5亿美元A轮时，以及Figma在400万美元ARR对应4亿美元估值时，都教会他去掉那些吓人的零，基于TAM、竞争地位和上行空间进行承销，而不是服从市场惯例。
模型可以检验信念，但过度详细的预测会制造虚假精确。 Randle先列出基准情景——也许就是其他投资人承销的3–5倍路径——再问自己的定性判断是否认为公司会“彻底碾压这些预测”。超过这一尺度后，模型就不再特别有用。他在2021年的《Playing Different Games》文章中预测了创投行业的分化，如今又看到“6、7个Tigers”在追求投资速度。
Tiger或许最终会被证明是正确的，但下一次崩盘仍要求生存。 Randle同意，考虑到Tiger持有Databricks和OpenAI，以及部分失败投资中的优先股保护，其最终表现可能远好于市场声誉。他预计许多AI公司会归零或下跌90%，而少数赢家将复利增长数十年。Benchmark的答案是约束基金规模、谨慎使用资本，以熬过不可避免的崩盘。
Benchmark最大的风险是停滞，而不是错过某一个周期。 Randle称，未来20年最大的威胁是停滞：Benchmark必须随着资产类别演进，同时守住两大北极星，并继续触达最优秀的创始人。他的长期乐观来自AI提高人均GDP，而人口增长正在放缓——以“继续把蛋糕做大”为基础，建立一个运转良好、零和色彩更弱的社会。
1. 伟大投资人把流程变成信念
Mary Meeker的量化声誉掩盖了Randle认为她真正的天赋：她是他合作过的“最定性的投资人”。她读公司历史和预测数据时“就像在读《黑客帝国》”，看到的是一条8–10年的叙事——对DoorDash而言，不是抽象的增长率，而可能是20%的家庭每月下单。
这段经历教会他，应该用数字推动投资故事，而不是被模型困住。Meeker的序列数据让她能够想象公司最终会变成什么，把量化纪律与对采用率、用户行为和市场地位的定性判断结合起来。
在Randle看来，Peter Thiel的天才同样体现在机构设计，而不仅是选股。Founders Fund允许员工与基金并肩个人投资，形成了一项隐蔽的信念测试：如果投资人宁愿把钱放在标普500，“我们凭什么让LP拿到这个配置？”一些年轻员工甚至会使用无担保授信参与投资。
Founders Fund以高强度著称的投委会之所以有效，是因为关系足够稳固，能够容纳“毫无保留、追求完整真相”。Randle可以对Keith Rabois猛烈抨击，而不必担心层级介入；分歧更像是和兄弟姐妹争吵，而不是在政治官僚体系中周旋。
2. Mamoon Hamid通过让卓越触手可及来传授品味
Mamoon Hamid最核心的教训是，年轻投资人必须近距离见过卓越。如果早期没有接触过非凡的创始人、管理团队和董事会，他们就无法在真实环境中稳定识别这一标准，也无法要求较弱的被投团队达到同样水平。
Randle认为，Figma、Glean和Rippling有一条共同主线：面向企业的B2B软件，却拥有消费级产品体验、异常强的用户喜爱和参与度，以及能够达到相应高产品标准的团队。
Hamid的优势在于对人、产品和市场都拥有“无可挑剔的品味”，并在一个刻意收窄的优势领域内不断打磨。Hamid的指导鼓励Randle形成同样具体的品味，而不是照搬一套泛化的创投打法。
Harry也给出了夸张但精准的背书：Hamid邀请他参与一笔交易时，他告诉团队尽调已经结束——“是B2B，有点像PLG，是Mamoon”——这说明投资人多年积累的品味能够携带多大的信息价值。
3. 320亿美元错过OpenAI，成为改变Randle直觉的那一笔
Randle一开始就喜欢ChatGPT，但因为非营利结构、员工份额和可能出现的稀释看起来“真的很棘手”，最终错过了OpenAI那轮320亿美元融资。这些风险并非虚构——公司结构一度接近失稳，招募人才也带来了严重稀释——但与其前所未有的增长和效用相比，它们并不重要。
他的复盘毫不留情：“我被吓住了，只见树木，不见森林。”私募股权训练赋予了他分析纪律，却也让他在产品本身已经提供非凡证据时，过度放大了结构复杂性。
Josh Kushner对Spotify和Instagram的本能反应，如今成了Randle的反向模型：当一个产品让人感觉势不可挡时，要足够相信这种直觉，避免让次要缺陷主导决策。OpenAI仍是他最大的错失，“直到今天都很痛”。
Randle预计，OpenAI明年可能成为万亿美元公司，并且在这一估值水平融资“完全没问题”。Harry从Kushner那里学到的相关规则是：如果投资人为了迁就别人，愿意接受更少的配额，那么自身意愿被削弱本身就在说：“别做这笔交易。”
4. ChatGPT让OpenAI占据优势，Coding仍在争夺
当被要求在5000亿美元的OpenAI和3500亿美元的Anthropic之间二选一时，Randle选择了OpenAI，不过他认为两者都可能是好投资。他的下行情景分析从ChatGPT开始：他看不到什么会打断其增长轨迹，或阻止它成为未来5年最重要的消费目的地和应用。
Anthropic在B2B商业化上可能仍略占优势，因为它投入了更多时间和资源做企业销售。在Coding方面，它凭借Claude Code、Sonnet和更完整的模型组合，也仍然略领先。
但OpenAI凭借Codex“追回了大把进展”，让Coding变成短兵相接，而不再是Anthropic一边倒的优势。Randle偏好的估值逻辑，取决于ChatGPT的消费端地位，而不是OpenAI赢下每一种工作负载的判断。
对于Cursor的命运，Randle给出了诚实的非答案：“我不知道。”他反对的是把份额下滑推导为机会收缩：即使Cursor被Claude Code、Codex和Cognition抢走相对份额，它仍可能成长为一个规模大得多的绝对市场。
5. 代码生成扩张速度快于份额碎片化
随着竞争者进入，Cursor可能已从最初市场约80%的份额降至25–30%。但Randle估算，代码生成市场在约2.5年内从几乎为零扩张到60–70亿美元ARR。
他过去用“黄金品类”标准识别市场：一年新增10亿美元净ARR，规模大到多阶段基金实际上必须持有一个仓位。仅今年，Coding就可能在不同产品和服务中增加40–50亿美元。
Harry追问，AI是否会让每个品类都变成黄金品类，以及门槛是否应从10亿美元提高到100亿美元。Randle承认，AI会扩大许多市场，尤其是软件替代人工的领域，但并非全部如此：例如面向兽医的AI细分市场，可能仍缺乏足够的客户和预算。
使用量本身也具有战略意义。Randle预计，Claude Code、Codex和Cursor等产品会进步最快，因为AI产品会通过使用不断变好；那些资金充足、却还没有把产品交到开发者手中的公司，可能会迎来“一次不太愉快的觉醒”。
6. 劳动力预算要求AI公司采用新的分类体系
Randle举的具体例子，是一家由Kleiner Perkins投资、提供全天候接待员的家政服务语音AI公司。一名客户在7款ServiceTitan产品上每年支出约25万美元，而在这款刚上线的AI产品上也花费了同样金额。
这套经济模型能够成立，是因为客户可以把接待员从3人减至2人，同时实现全天候接听电话和预约，而不是只能在9点到16点工作。AI产品既降低了劳动力成本，也捕获了原本会流失的收入。
如果硬把这家公司塞进SaaS模板，就无法理解这种客户关系。Randle曾经工作的第一家公司Vista Equity Partners的CEO Robert F. Smith过去常说，SaaS“尝起来都像鸡肉”：这些企业足够相似，可以套用一套可复制的运营打法。
投资人习惯了约80%的毛利率、高80%区间的毛利留存率、120%以上的净留存率，以及很少的资本开支。AI应用把推理成本直接纳入COGS。因此，更低的毛利率可能意味着真实使用量；异常高的AI应用毛利率，反而可能意味着客户几乎没在使用AI功能。Randle希望投资人从终局经济学出发，而不是奖励一个好看但可能空洞的百分比。
7. 绝对毛利润可能比毛利率百分比更重要
Randle偏好的比较是：如果ServiceTitan每客户贡献20万美元毛利润、毛利率75%，而一家AI公司每客户贡献50万美元毛利润、毛利率50%，那么“我不在乎”后者的百分比更低。它覆盖了更广泛的客户关系，也获取了更多经济价值。
分析单位应该转向毛利润倍数和单客户绝对毛利润，同时从第一性原理出发，评估5–7年后的毛利率。训练成本和自研模型，也会进一步拉大AI公司与传统SaaS之间的差异。
AWS是他的类比。他估计AWS毛利率可能为50–60%，经营利润率约30%，但它通常是软件公司的最大支出项目，远高于Salesforce、Workday或Adobe，因为基础设施支出规模极其庞大。
2010年代初，一家营收1.5亿美元的软件公司可能会出现3000万美元、令人咋舌的AWS COGS。相比AWS的百分比毛利率，更重要的是客户在它身上支出的倍数；如果从Amazon中拆分出来，Randle认为AWS可能是一家万亿美元的独立公司。
8. 只有当应用越过实验室基准，增长才是真增长
AI公司可以在不到1年内从0增长到100，但Randle始终牢记“来得快，去得也快”。Jasper曾快速增长，随后因为早期收入缺乏足够的支撑和持久的客户价值而萎缩。
GPT-4暴露了其中的弱点：客户认为Jasper的输出与ChatGPT每月20美元提供的内容相似。Randle认为，Jasper后来通过把LLM嵌入差异化的营销工作流，而不是简单包装模型访问能力，实现了复苏。
如今，实验室产品设定了客户体验的最低基准。一款应用如果收费明显高于ChatGPT每月20美元或200美元的价格，就必须提供足够强的超额价值，才能支撑企业分发、客户留存和可持续的商业方程式。
Randle反对护城河已经整体从技术转向分发的说法。分发提供的是“构建差异化技术的权利”，但优秀AI产品仍需要稀缺人才、精心设计的模型管线和有品味的工作流整合，而不是把API塞进一个文本框。
9. AI人才和产品工艺仍是稀缺技术
经典的“七大能力”并未消失；应用扩张更快、实验室更快改善并分发自身产品，使竞争 stakes 更高。可持续增长仍取决于能够穿越一轮轮模型升级的差异化。
Randle将技术护城河重新定义为部分由人才构成的护城河。很少有人能够决定LLM应该嵌入工作流的何处、如何改善输出，以及如何让完整产品的表现足够优秀，从而胜过实验室应用。
这份稀缺性解释了为什么研究人员能拿到十亿美元级合同和“LeBron级别的薪酬”。没有这些构建者，分发或许能带来触达，但无法造出足以守住客户关系的卓越产品。
因此，Harry关于分发和数据的论点遭到直接反驳：Randle表示，护城河从根本上仍然是技术。产品质量同样是判断哪些人有能力构建它的证据。
10. 商品化AI基础设施可能压倒质量层面的顾虑
Randle对AI云的看法转变最为剧烈。他起初认为CoreWeave只是转售商品化算力的中间商，是一个结构性低毛利的经纪商，但天文数字般的推理需求压倒了这一关于业务质量的质疑。
他回忆，CoreWeave当时私募融资约30亿美元；在节目所述时点，它已成为一家市值约600亿美元的上市公司，Nebius约300亿美元，整个板块的公开市场市值超过1000亿美元，还不包括迅速成长的私营公司。
Randle仍认为，CoreWeave及类似公司某个时候“可能会下跌70%”。但早期投资人已经在从他拒绝投资的价位上涨20倍后获得流动性，这使他原本关于商品化业务的批评在经济层面失去了意义。
这个教训有意令人不适：当需求看起来像第一轮超大规模云厂商浪潮，而AI推理可能呈现更陡峭的群体增长曲线时，“有时候你就是得闭上嘴，顺着动量投资”。
11. Benchmark按基金规模进行投资，而不是追逐每一家实验室
Randle将Conway定律应用于创投：机构会“交付”自己的基金规模、团队和结构。一支拥有50名投资人的70亿美元基金必须参与巨型融资，因为十亿美元级支票是少数能有效部署其资本的方式。
Benchmark规模更小，可以玩另一种游戏。其上一支基金最领先的5笔投资按上一轮价格计，约为1笔60倍、2笔30倍和2笔20倍；Randle表示，ChatGPT发布以来，没有任何一轮OpenAI融资达到这些倍数。
Harry计算，按320亿美元估值投资OpenAI，账面涨幅可能约为12–15倍，但考虑稀释后更接近6–8倍。他将其与Benchmark在Lovable、LangChain、Sierra、Mercor和Fireworks上的持仓作比较：对于小基金而言，现金回报倍数才是产品，声望不是。
Randle承认，跳过实验室可能威胁到机构的相关性和资源获取。眼下他的反证是Benchmark与Bret Taylor等人的关系——他称其为AI应用领域的教父——以及他提到的Mercor的Brendan，后者代表AI基础设施方向。
12. 持股比例是投入，合作关系和回报才是产出
Benchmark的两大北极星，是成为每位创始人最亲近、ROI最高的合作伙伴，并在LP的创投持仓中创造最高的资金倍数回报。惯常的20%持股目标本身并不属于其中。
Harry援引一篇文章称，Benchmark在Mercor的持股约为10%，低于历史惯例。Randle回应称，更大的潜在结果意味着更多实现卓越回报的路径，同时仍能成为公司最具影响力的创投伙伴：批评者“把投入和产出混为一谈”。
Harry以Delian Asparouhov关于Benchmark解雇创始人的批评，挑战其“最佳合作伙伴”说法。Randle指出，替换创始人曾经是常规操作——Google的投资人当年立即寻找职业经理人CEO——但他也表示，如今治理方式以及董事会与创始人的关系已经显著改善。
创始人忠诚不能凌驾于法律、伦理或受托责任之上。Harry描述了一些董事会为保护创始人NPS而牺牲股权结构的情况；Randle表示同意，并补充称，伟大的创始人不需要唯唯诺诺的人，也不需要“董事会里的GPT-4o”，而需要愿意交锋、推动公司进步的成年人。
13. 价格只有放在公司自身的上行空间中才有意义
Benchmark的4位普通合伙人各代表合伙企业25%的权益，并保有各自鲜明的风格：Eric Vishria偏好从公司创立之初介入，而Randle预计自己最初会更多投资Series A、B及更后期阶段。共同约束不是阶段，而是卓越的创始人合作关系和回报潜力。
Randle承认，自己作为一名成长型投资人加入Benchmark时曾感到不安。Vishria回应称，Bill Gurley在Benchmark之前也曾是公开市场分析师；Randle还以Pat Grady为例，说明最优秀的投资人正越来越超越阶段边界，而不是被困在某个组织分类中。
他的排序是人、产品、市场。人是上游引擎，产品是判断其质量的最强证据，市场排在第三，因为它最具可替代性。团队可以转向，就像Slack一样；但一个并不卓越的创始人或产品组织则更难改变。
SpaceX在1500亿美元时教会Randle去掉那些吓人的零，比较TAM、竞争地位和上行空间。Rippling的2.5亿美元A轮、Figma在400万美元ARR对应的4亿美元估值，相对于市场惯例都看似荒谬；但如果只看相对价格，就会把这些卓越公司筛掉。
14. 模型检验信念，但巨型基金结构会重塑行为
Randle用模型列出基准情景——也许是其他成长型投资人承销的3–5倍路径——然后再问自己的定性判断是否认为公司会“彻底碾压这些预测”。超过这把尺子后，详细预测就会变成虚假精确。
Figma说明了这一陷阱：任何准确预测其最终存续时间、增长和盈利能力的模型，看起来都会像是为了通过投委会审批而人为设计。简单的市场规模测算同样失效，因为只统计设计师，忽略了产品人员和其他角色的采用。
Randle在2021年的文章《Playing Different Games》中，预测了Tiger高速度、对创始人友好、低接触的模式，与Benchmark集中、高接触、重工艺模式之间的分化。缺乏特色的中间地带成了他的“J.C. Penney基金”；Tiger受挫后，“我们又多了6、7个Tigers”。
Harry反对把Tiger的北极星套用到Thrive、Lightspeed或General Catalyst身上。Randle弱化了对Thrive的描述，但提出了一个组织测试：去问负责人、初级合伙人和助理，投资部署是否会影响晋升。当十亿美元级支票可能贡献95%的利润时，他认为这些支票必然成为主要产品。
15. 巨型基金可以赢得美元，却输掉创投回报测试
Randle接受Harry的核心反驳：OpenAI、Anthropic和Cursor可能比任何人曾经想象的都大得多，让投资人赚到巨额绝对收益。分歧在于，这是否能转化为LP期待的创投倍数。
他怀疑最大型机构的负责人能否“面不改色地”告诉LP，其按同等条件参与的基金组合能够实现5倍净回报。他认为，在80–100亿美元基金上实现4倍净回报，已经接近“违背物理定律”的规模。
LP对私营科技资产的需求，可能因为机构目前愿意用较低回报换取配置机会而推迟清算。但Randle指出，LP已经有私募股权来获得较低回报的敞口，而且流动性通常更好；创投的独特作用，是创造异常高的资金倍数回报。
在一个拥有50名投资人的平台里，第23位合伙人可能继承30个尚可的公司关系，需要做成几笔交易才能晋升，再寄希望于其中一笔带来任期保障。Randle认为，这更像投行或大型并购机构，而不是“结识真正有趣的创始人，只做最好的投资”。
16. Tiger或许会被证明是正确的，但下一次崩盘仍要求生存
Randle同意，Tiger在2021年的投资组合最终表现可能远好于其市场声誉。Databricks和OpenAI的大额持仓，加上许多失败项目中的优先股清算保护，可能让“#JusticeForJohnCurtius”不再只是玩笑。
他的条件式情景非常明确：如果Databricks成为4000–5000亿美元公司，OpenAI成为数万亿美元公司，那么Tiger的基金可能“还不错”——不是LP曾拥有过的最佳组合，但也远非一台烧钱机器。
那些过度仍然令人记忆犹新。2021年12月或2022年1月的一场迈阿密派对上，可能还有Vanilla Ice在场，Randle看着公开市场科技股下跌30–40%，想起《The Dark Knight Rises》：“Gotham正在燃烧”，而行业正在参加最后一场奢靡庆典。
今天可能与互联网泡沫周期相似：许多AI公司会归零或下跌90%，而少数真正持久的赢家将复利增长20–30年。Benchmark的答案是约束基金规模、谨慎使用资本，保留穿越崩盘的能力，而不是迫使LP逃离。
17. Benchmark最大的风险是停滞，而不是错过某一个周期
一位合伙人曾建议Randle，第一次Benchmark投资失败或许会让人获得解放：一旦项目失败、 “LP仍然喜欢我们”，也没有人被解雇，第二次决策的心理负担就会减轻。相反，早期的表面成功可能不断加剧压力。
当被问到，在曾经效力过的机构中会把钱投向哪里以获得最高现金回报时，Randle选择了Founders Fund，因为孵化项目能在购买股权变得过于竞争激烈之前获得股权。Anduril展示了可能的基金层面结果；每隔几支基金，或者每5–10年，这家公司就能孵化出一家令人难以置信的企业。
Benchmark自身必须保持动态，同时守住两大北极星。Randle称，“停滞”是未来20年最大的威胁：与最优秀创始人的关系是这个资产类别的货币，传统不能让“尾巴摇狗”。
他对未来10年的乐观判断来自宏观经济。社交媒体让资本主义把人类注意力优化成盯着屏幕，而AI可以在人口出生率下降之际提高人均GDP。沿用Peter Thiel的框架，Randle认为经济增长——“继续把蛋糕做大”——是繁荣与和谐、非零和社会的必要条件。
Ev Randle
I think we should not be placing that much emphasis on margins today. We need a new taxonomy for AI companies.
I'm thrilled to welcome Benchmark's newest partner, Ev Randle. Benchmark is one of the best firms in venture.
Ev Randle
Tiger died, and we got 6 or 7 more Tigers. I don't think Roelof or Hemant or even Ben and Marc, at this point, can go to LPs and say, "Hey, we're going to get you 5x net on that." When you're writing billion-dollar checks, that is your main product. Go talk to the principals, the junior partners, and the associates at those firms, and you tell me that capital velocity is not the north star of those firms.
I think Tiger's going to end up much better than anyone thought they were going to end up.
What do you think the biggest threat is to Benchmark being successful in the next 5 years?
Ev, I am so excited for this. I cannot believe we have not done this before. I think I personally timed it pretty well, if I'm honest. I'm rather chuffed with myself. Thank you so much for joining me today.
Ev Randle
Thank you, Harry. I have actually been listening to 20VC since 2017, which, ironically, I think is the year that you had Peter on for the first time. It's just been so fun to watch the show and the platform that you've built grow this way. It's almost like watching a startup become an IPO-worthy company or something. So, congrats to you, Harry.
Do you know what? I've had a man crush on Peter Fenton since that first show. I remember he told me that price is a litmus test for your conviction, and I think about that at least on a weekly basis. I've repeated it to my team many, many times.
Before we dive into Benchmark, you've worked with some of the best: Peter Thiel, obviously, at Founders Fund; Mary Meeker at BOND; and Mamoon Hamid, one of my big bros at Kleiner Perkins. If I were to ask you for your biggest takeaway from each, what would you say your biggest investing takeaway is from each of them?
Ev Randle
One of the things I really love about the asset class that we practice our craft in is that there are so many different ways that you can be successful at it, and there are so many different strategies and frameworks that you can employ and still generate amazing returns. Each of the people that you just mentioned has a very different style and a very different way of practicing their craft.
If I was to lay out what I learned specifically from Mary, Peter, and Mamoon, I think with Mary, she does such an incredible job. Everyone thinks of her as this quantitative investor. She had her time as an equity researcher at Morgan Stanley during the dot-com bubble, and then she came to Kleiner Perkins. Everyone talks about the DCF models she creates and all the numbers that she does, but she's really the most qualitative investor that I've ever worked with. It's probably a surprise to hear that, but what she does is almost like she's reading the Matrix.
She lays out all the sequential numbers historically for a company and then all the numbers going forward. It's almost like she's reading the Matrix code as it comes down, and she's seeing what the company will become on an 8- to 10-year time horizon when she sees what the numbers are.
She'll look at a DoorDash model, and that was an investment that we had led at KP out of the growth fund at the time. She won't see 7 years out, 80% growth or something like that. She'll see that 20% of households are going to be ordering from DoorDash on a monthly basis, and she can visualize that.
From her, I just learned that when you use numbers in venture and growth, and when you want to be quantitatively driven, don't get stuck in a quantitative lens with it. Actually use that to drive the narrative and drive the story of an investment. That's been an incredible mental framing that I've used with someone like Peter Thiel.
Peter, I think so much of his cleverness and so much of his genius is actually in the way that he builds his firms, rather than even his investments. The way that he's designed Founders Fund is that he creates all these incentive structures and mechanisms to constantly be testing your conviction.
There's a program, for example, at Founders Fund where anyone that works on an investment—or if you're leading an investment—you can personally invest alongside the firm in that investment, almost as if you're angel investing. At first glance, it just looks like this amazing perk that you can have by being an investor at Founders Fund, but deeper down, it's a conviction test.
If you're sponsoring some pro rata of a company that's doing okay but not great, and the founder really wants you to do the pro rata so as not to blow up the round, but you're not doing some of your portion on the individual side of that investment—your angel investment—Peter can go to you and say, "Do you not think this is better than having your money in the S&P? Why would we give our LPs this allocation in this round if you don't even want to put your own money in this round?"
There are 100 different things like that that exist in Founders Fund that aren't explicit, like, "Do you have high conviction?" They test your conviction in deeper ways.
I absolutely love that as a conviction test. Do you ever, just reflecting on that, have the fear that if you had that with a younger person—say, when you were at Founders Fund—if you don't have that much liquid cash, it is a lot when you have rent and bills?
I'm thinking through this as an active partner with you now because I'd love to implement that in 20VC, but I would hate for people to be scared and then say no to something because they didn't have the cash. That could be great. What do you think?
Ev Randle
It's super valid. I think, again, if you're at Founders Fund, you are all in. Most of us that were young at Founders Fund at the time all had debt lines—unsecured debt lines—that we were using to make these side, personal investments.
By the way, it's turned out to be an unbelievable portfolio for myself personally, and it's all worked out. So I'm very glad that I had it. But I think that's part of how, throughout his entire career, he has really designed his organization so people are all in.
He had a bonus system for PayPal employees. If they lived within a couple miles of the office, he'd give them more money. He just designs organizations this way. There's less pressure for the young folks that don't have much net worth yet, for sure, but he still expects you to be scrappy and find a way to do it.
What do you think no one knows about the inner workings of Founders Fund that they should know from the outside in?
Ev Randle
From the outside in, obviously, Founders Fund is a bit of a black box. Everyone's like, "Wow, the returns are amazing. There's a bunch of weird personalities within that place. How does it all happen?"
When I was actually doing backchannel references on Founders Fund before joining, something that everyone said to me that they thought was a negative but ended up being a huge positive was, "Oh, you really got to watch out for the culture because I've heard that they yell at each other during ICs, or investment committee meetings, and they get super intense."
A few months into the actual job at Founders Fund, I realized that, yes, sometimes people did yell at each other at ICs, but it was because you were yelling at your brother, yelling at your sister, or yelling at your best friend. Everyone was so secure in themselves and the relationships that they had with each other—and they all have extremely deep relationships with each other—that you could actually just be extremely truth-seeking.
You weren't afraid to step on toes. You weren't afraid to do anything that might be seen as, "Oh, you shouldn't say that to a GP." It was just no holds barred, complete truth-seeking, everyone trying to get to the best answer.
I remember a few months into the job, I was on an email thread and just teed off on Keith Rabois, our good friend. At any other firm, that might be a fireable offense, or you might get a tongue-lashing for doing that. But at Founders Fund, it was a pat on the back. It was like, "Yes, that is how we do things here." It's flat. We're just trying to get to the truth. We're not trying to uphold some political bureaucracy or something like that.
And then Keith fired you. (Laughter.)
Dude, that, to be fair, is a bold take for a younger person in your first years. Well done. That's conviction, going up against Keith in that way.
If we go to Mamoon, what are the takeaways from Mamoon? I think Mamoon is just one of the greats. He's done so well with KP. What are the takeaways from Mamoon?
Ev Randle
Yeah, Mamoon Hamid—I have learned so much from Mamoon. He's a wonderful mentor. We were talking before the show, Harry, about the kindness that he showed you when you were young, and he did the same thing for me.
I think the biggest thing that Mamoon has taught me—and this is a reflection of what he did and has done in his career—is that you need to, early in your career, see excellence up close. In terms of a company, a management team, or a founder, you need to see how the absolute best operate and do the job of company-building.
If you don't see that relatively early in your career, it's much, much harder to spot it in the wild, and you also don't know the bar to hold your other founders and your other management teams to. He does a very good job of getting younger folks that work at Kleiner Perkins, or even back at Social Capital, involved in the very best companies and in those boardrooms, seeing how they operate.
He thinks that once you've seen it, and once you know the "it" of what makes an A++ team tick, you can, one, much more easily see that in the wild, and two, you can really hold the rest of your management teams and founders that you work with to that really high standard.
I think Mamoon, more than anyone, has developed this impeccable taste around a mix of product, market, and people.
If you think about his huge, huge winners, whether it's Figma, Glean, or Rippling, they all have a common throughline. It's B2B software, but it's almost like consumer-like software that demands really high user love and engagement. He's just developed this really, really tight understanding of where he shines and where he has a really deep understanding of companies.
He's really sharpened his taste in doing that. So I think he's definitely encouraged me and encourages people that he works with to really develop a specific form of taste around the people, the products, and the companies that you think are going to be the big ones.
He very kindly messaged me the other day and said, “Hey, I'd love to bring you into one of my deals, Ev. The founder is amazing, and you'd be great for it.” I messaged my team, just being like, “Hey, we're doing a deal. It's amazing. Mamoon's bringing us in. We're done. Diligence over.”
They're like, “Harry, no, you can't be serious.” I'm like, “It's B2B. It's kind of PLG. It's Mamoon. Would you like your job tomorrow?” So I totally agree and get you there.
Can I just ask, before we move to Benchmark—you mentioned Mary Meeker and the mental plasticity that she had around numbers and what the future could be. Where were you not mentally plastic where you should have been, and what did you learn from that?
An example for me would be: I met Alex at Deel when it was 2 on 10, and I looked at Paychex and ADP and I was like, “Nah, shit market, incumbents, distribution advantage, crap investment.” What a mistake. I wasn't mentally plastic and I should have been. What would yours be?
Ev Randle
An instance where I haven't, and where I haven't exuded neuroplasticity enough, I actually have a very recent example of this. It was the OpenAI round at $32 billion.
I started my career in private equity, which gave me a lot of strengths, but it also definitely gave me some blind spots in venture that I've needed to unlearn a little bit. When I was at Founders Fund, I was actually extremely positive on OpenAI. I had left Founders Fund right after ChatGPT came out, and ChatGPT, when it came out, was one of those moments where you're like, “This product is it. This is so unbelievably cool.” You could just tell that it was going to be a massive, massive product.
Then the $32 billion round of OpenAI came around when I was at Kleiner Perkins, and all of a sudden I was like, “Oh, man, this structure seems really gnarly. They're going to have to convert this somehow. It's a nonprofit. They're selling these employee units, and I think they're going to dilute the hell out of the investor base.”
I got spooked and missed the forest for the trees, both in terms of the structure of the company at the time and the potential future dilution. By the way, both of those things were very valid risks. The structure at certain points has almost taken the company down, and they've diluted a ton, given that they've had to attract all these AI researchers and all this incredible talent.
But it didn't end up mattering. None of that ended up mattering. What ended up mattering is that it's had the strongest and highest growth trajectory of any technology company in history. It's probably the best and most useful product that anyone who uses it has in their pocket.
I think Josh Kushner actually does probably the best job of this. He talks about his intuitions, and he saw Spotify and just kind of knew that, no matter what, he needed to invest in the company. The same with Instagram. I still need to learn to trust my intuitions more, because sometimes I let silly things like that cloud my thinking.
Josh taught me one of the most valuable lessons, actually. He taught me that if you're ever willing to do less in a deal, don't do the deal. “I'm happy to take 10% if it means giving my buddy 3%.” Don't do that deal. That's a bad signal.
I remember Vinod came on the show when he did that deal, and he said, “Harry, listen, if it's a trillion-dollar company, we'll all make money.” We laughed at the time, and now it's like, “Oh, it might be a trillion-dollar company.”
Ev Randle
Three trillion. Who knows?
Yeah. Do you think it'll be a trillion-dollar company next year?
Ev Randle
I think it'll be a trillion-dollar company next year, yes. I think they could probably raise at the end of the year—Q2, I think OpenAI could raise at a trillion dollars, no problem.
Would you rather be in OpenAI at $500 billion or Anthropic at $350 billion? Obviously, at Kleiner Perkins we invested in Anthropic, and we had this debate a lot internally. Everyone kind of thinks this is a fun debate: OpenAI or Anthropic at the last-round price.
Ev Randle
I think they represent relatively different things. In terms of downside risk, it's hard to imagine anything that could knock ChatGPT off of its growth trajectory. I don't know what could stop ChatGPT from growing at the rate that it's growing. That asset alone is unbelievably valuable and completely locked in.
There's no way that it's not going to be the most important kind of consumer destination and consumer app over the next 5 years. I think where everything else is still hand-to-hand combat is obviously in coding.
I think OpenAI has actually done an incredible job with Codex and made up a bunch of progress against Anthropic that they didn't have before. On everything on the B2B side, Anthropic probably has a bit of an edge right now. They've spent a lot more time and resources toward really mastering the commercialization effort there.
In coding, Anthropic, with Claude Code, Sonnet, and all the models that they have, is probably still a little bit ahead of OpenAI. But given ChatGPT, I think I would probably rather do OpenAI at $500 billion than Anthropic at $350 billion. I think both are relatively good investments even today.
I would be thrilled with both, just in case Dario or Sam are listening. Very happy to take some shares.
Ev Randle
Very happy.
If you want to help me out here. Sam, I'll buy Brad Gerstner's if you want that one, because he doesn't want to. I won't ask any questions. I'll just wear a Sam T-shirt.
What happens to Cursor? You see Codex crushing it, as you said there, and Claude Code has done so well. What happens to Cursor? I don't know. I'm purely lost on that one.
Ev Randle
I think the thing that everyone has underestimated thus far is just how immense a potential market code can be. When you think about Cursor, a lot of people are like, “Well, their relative market share has gone down a lot,” because at first it was really just them, then Claude Code came out, then Codex came out, and now Cognition is scaling.
Instead of, I don't know, 80% of the market or something, maybe they have 25% to 30% of the overall ARR in the market today. Again, what people are missing is that the market for code generation, over the last 2.5 years, has gone from essentially $0 to probably $6 billion or $7 billion of ARR.
Something we used to do at KP and Founders Fund was try to identify the golden categories. A golden category is a category where the entire market for a single product adds $1 billion of net new ARR in a single year. If you find a golden category, especially if you're a multistage fund, you have to have a bet in that category, because it means that it's going to produce really big outcomes.
Instead of adding $1 billion of net new this year, I think code generation is going to add $4 billion or $5 billion of net new across every single product and service that's available for people to buy, both on the B2B and B2C side.
Can I ask: does AI not make every category a golden category? I don't mean that stupidly, but customer service is, of course, tens of billions of dollars. Even if you think about much more verticalized software plays, could you not apply “golden category” to everything then? Should we not move $1 billion to $10 billion?
Ev Randle
It does for a lot of categories. It remains to be seen, right? Let's say you're doing AI for vets—veterinarians. Maybe there just aren't enough vets that have enough money to actually create $1 billion of net new in a given year.
But I do think that, for so many categories that seemed like they were middling in size, a lot of what AI has been able to do—especially if it can touch something that a labor force within a category was doing before—is creating much, much bigger markets.
As one example of this impact, at KP we were invested in a home services AI business that was essentially a 24/7 receptionist. Its first product is a 24/7 receptionist for HVAC people, home services, and anyone that would be a ServiceTitan customer.
We were calling customers and asking, “Okay, how much do you spend on ServiceTitan?” They're like, “You know, $250,000.”
It's like, okay, well, how much are you spending on this company? And they're like, "$250K." It's like, okay, you have 7 products from ServiceTitan, from SaaS 2.0, and you have 1 product that's just out of beta from this new startup in voice AI, and you're spending as much on that as you are on ServiceTitan, the system of record for everything that you're doing.
And they're like, "Yeah, well, we no longer have to staff 3 receptionists; we can staff 2, and we're now able to actually accept calls and book appointments 24/7 rather than the 9-to-4 schedule that our receptionists were sitting there." And so it's driving more revenue and more impact than even ServiceTitan was doing, given that the capabilities are just so much broader and real than the impacts that SaaS can have on companies.
Ev, are we gonna be frenemies?
Ev Randle
Was that ProBook?
No, no, this one, Lerer Hippeau led a round in Aloha, is the company's name.
Ev Randle
Oh, thank God. I lost this deal and I didn't know who I lost it to. And it's exactly the same way. You're like, "How much do you spend on ServiceTitan?" And then it's the same, if not more. And you're like, "Oh my God, that is 1 incredibly valuable segment that we're covering."
Yeah, and I think it gets to something that I desperately want us to do in the venture industry, which is we need a new taxonomy for AI companies. What I mean by that is AI app companies are meaningfully different from SaaS companies in a dozen different ways, yet we keep trying to shove all the metrics from these AI app companies into the frameworks that we created for SaaS.
What was your pause on that? What metrics do we try and shove in that we shouldn't?
Ev Randle
If you just think about the P&L of a SaaS company, Robert F. Smith, the CEO of the first firm that I ever worked at, Vista Equity Partners, always used to say—probably still says—"SaaS is great because it tastes like chicken." All the businesses are the same. And the whole thesis behind Vista was that SaaS companies are so similar that you can do the same exact things to each of them in the whole Vista playbook style and make them way more profitable and run a lot more efficiently.
And so we're used to, "Gross margins need to be 80%, gross retention should be in the high 80s%, net retention should be over 120%, there should be very little CapEx," and that's what makes a good company. And I think what you're seeing with AI app companies is a very different situation where, if they're good companies with a lot of usage, you have a lot of AI inference in your COGS that you don't for normal SaaS companies.
And so people are like, "Oh, these are worse companies because they have worse gross margins." But if your average gross profit per customer can be 4 or 5x that of a normal SaaS company, then you actually have much more absolute dollars of gross profit per customer and potentially a much, much larger market than you do for SaaS companies as well.
So instead of talking about gross margins and revenue multiples, I hope that someday we talk about gross profit multiples and absolute gross profit dollars per customer. If your relationship with a customer can be much, much broader because you're taking part of their labor budget or you're giving them more economic value than you would if you were a SaaS company, it's just not appropriate to be grading them on a metric like, "Do they have 80% gross margins?"
It's like, well, if ServiceTitan has $200,000 of gross profit per customer and this other company has $500,000 of gross profit per customer, I don't care that that second company has 50% gross margins and ServiceTitan has 75% gross margins. It doesn't matter.
So I think that's the biggest example: the contract sizes can be much larger even if the gross margins are lower. But I think there are several others. Some train their own models, and so there are training costs and other various changes as well.
So, help me out: should we not place such emphasis on margins?
Ev Randle
I think we should not be placing that much emphasis on margins today. I think the work that we should be doing is trying to understand what the terminal gross-margin structure looks like for these businesses and also what the absolute gross profit dollars are in each of these categories that these companies can represent.
Again, I think the folks over at Andreessen Horowitz have done a lot of good work in terms of evangelizing this idea that if you have high gross margins as an AI app company right now, it probably means that you have very little AI inference expense in your COGS, which means no one's actually using your AI features.
It's not the easiest thing to understand what these AI app gross-margin profiles are going to look like in 5 to 7 years. But I think that at least trying to go from first principles and reason about what the gross profit dollar per customer and the gross margins of these companies in 5 to 7 years look like is so much more worth doing, and it's such a better intellectual exercise than trying to compare it to SaaS, which is just a very, very different business and has a very different pricing and business model that isn't going to be as relevant, I think, over the next 10 years.
It's so interesting. Rory O'Driscoll from Scale, who's basically like my adopted father—he doesn't know that, so you've just gained a son—he's listening to this show like, "Wow, this is a productive show." But he always tells me that, fundamentally, whether we make money from AI or not will be predicated on whether we see the movement from human labor budgets to AI software spend.
I think exactly to your point there, for everyone who's trying to understand absolute dollars in terms of profit, your margin can be lower, but because the spend is 5x, your absolute profit is significantly higher on a per-customer basis. Correct?
Ev Randle
Exactly. So let's think about AWS, for example. AWS—I actually don't know their exact gross margins—but they're not as high. They're not 80%; let's say they're 50% or 60%. I know that their operating margins, I think, are at about 30%.
The thing about AWS is that it is the largest line item for essentially any large software business versus anything else that they pay for. You're paying more for AWS than you're paying for Salesforce, Workday, or any other SaaS company, by a wide, wide margin.
In the early 2010s, you had companies doing $150 million of revenue, and people started to be like, "What is this $30 million COGS line to Amazon Web Services? What in the hell is this?" And I think that's an amazing example of, yeah, does AWS have lower gross margins than Adobe? Of course it does. But everyone that uses AWS and is a core customer of AWS spends multiples more on AWS than they do on Adobe, which is why it's such an unbelievably large business, probably a trillion-dollar business if it was spun out of Amazon.
So that is the idea that I think we need to all get in our heads: it's not going to be every company, and it's not going to be every market. But for the right AI companies in the right markets, the size of their revenue per customer is going to be so much larger than SaaS that, even if they have lower gross margins, they're going to be much, much more valuable companies.
Going to that as well, what is AWS? It's a commodity. And that's what I find so interesting. You were like, "Oh, your models won't make money because they're just commodity businesses." And then you look at Google Cloud, you look at Azure, and you look at AWS, and you're going, "Wow, maybe the best business in the world is a commodities business."
To your point, one of the things I've changed my mind on over the last 2 years is these AI inference cloud businesses. So who's going to be the AWS, GCP, or Azure of the AI era? When CoreWeave was first raising in private markets, I was like, "Oh my God, they're reselling a commodity. They're a middleman. They're a broker of compute. It's going to be a low-margin, yada yada yada."
How wrong was I? I mean, maybe the market's down a little bit, but last time I checked, it was a $60 billion public company. Nebius is a $30 billion public company. There is over $100 billion in public market cap, and there are several private players that are growing astronomically as well in this AI inference cloud.
So I think sometimes we can twist our minds in knots over, "Oh, is the business quality okay?" When you have demand like this, like you had for the initial hyperscaler clouds—and I think we're seeing an even greater cohorted demand curve for AI inference—sometimes you just got to shut your mind up and invest with the momentum.
Ev Randle
I totally agree in terms of "shut your mind up" and investing with the momentum, but it brings me to the other element, which is different than ever before. You mentioned there the change from margins to a focus on absolute gross dollars per customer. The thing that's different is growth rates.
And I think the thing that I'm struggling with is sustainable versus unsustainable, but also being a sucker for momentum and high, high numbers. How do you think about the importance of growth rate—optimizing for it versus sustainability? And do we need a new taxonomy around growth rate as well?
Ev Randle
I think we do. I think the things that we need to hold in our heads when we're thinking about this are that we have companies going from 0 to 100 in less than a year. We've never seen that. At the same time, is it easy come, easy go? We had early examples of this.
I'm comfortable saying this because now the company has rebounded and, to my knowledge, is doing really well. I remember when people were talking about Jasper. The 2 AI investments that started the wave were Stability AI and Jasper AI. Stability is a different story, but Jasper went from 0 to 100 very, very quickly and then actually started shrinking. It was sort of easy come, easy go with the revenue, and they hadn't built enough scaffolding or enough actual, true value to really sustain the customer relationships they had and sustain their growth rates.
The way I've been thinking about this, especially as it relates to app-layer companies, is that the other aspect is: what is the risk for a lot of these app-layer companies, and who are they in danger from? It's the labs. The labs are creating apps, creating more value via the models, and giving that value directly to users. Oftentimes, as an app company, you need to be doing better than what $20 a month can get you from ChatGPT.
For a lot of these categories, the labs set the baseline in terms of customer experience. They're your competition at the base layer. Whatever you can get from ChatGPT, or whatever you can get directly from the labs' apps themselves, you need to be sufficiently differentiated from that because they're happy to charge $20 or $200 per month per user. A lot of these AI companies want to charge a lot more than that in order to have a sustainable business equation and actually be able to do B2B distribution.
When you think about Jasper at first, the issue they ran into was that when GPT-4 came out, people started saying, "Wow, the outputs I'm getting from Jasper are kind of the same as what I'm getting for $20 a month from OpenAI. I'm not going to pay however much more for Jasper. I'm just going to use ChatGPT." What they've done now is build sufficiently differentiated workflow software and tie in LLMs through the lifecycle of how their users work and operate, in a way that is sufficiently differentiated and gives them more of a moat.
I don't think the sources of moats have changed from SaaS to AI, necessarily. 7 Powers are still 7 Powers. All of the same ways to build differentiation are still there. The stakes are just much higher because the growth rates are much higher, and the labs are getting so much better so quickly, especially at delivering applications.
How do you feel about people who say the moats have changed? The moat that was technology is now fundamentally distribution—in terms of access to customers and data, and access to data—and it shifted from technology to those 2 things. Do you disagree with that, or do you agree with that?
Ev Randle
I definitely disagree with that. I think the moat is still fundamentally in technology, not in distribution. Distribution obviously gives you the right to build differentiated technology, but one of the huge learnings we've had as an industry is how damn hard it is to build good AI products.
A good AI product is so much different to build than a good SaaS product. You need different people, and there are so many different parts of a good pipeline. Where do you bring in LLMs? How do you improve them? How does it fit within a general workflow? It's not just bringing in the OpenAI API and using it within the text box or something. It's extremely nuanced and complex to build an exceptional AI product, especially one that's going to outshine the labs' applications themselves.
I still think it's technology. It might just be different in that maybe it's not a tech moat in terms of having a unique database that no one's ever built before that's more efficient for X, Y, and Z use cases. It's really a talent-scarcity and talent-tech moat, where there just aren't that many people who know how to build these products and build off of these models in a super-intelligent, tasteful way. That's why you're also seeing people go for billion-dollar contracts and make LeBron money as an AI researcher.
How do you at Benchmark think about that? I struggled with this one, too. My fund is $400 million. Benchmark, I believe, is $500 million to $600 million. You guys never really announce funds in the way that most people do because it's probably mostly just your money at this stage.
My question to you is: when you see a Mira Murati or Periodic Labs—great and very talented people—but these are $300 million rounds and $2 billion rounds, do you just accept that that is not a world that you play in?
Ev Randle
This gets to a question that I think some people have. I don't think you've had it, Harry. You've been very kind to us. But I think some people have asked the question: did Benchmark miss AI? Did Benchmark not get in on the AI wave because they're not in one of the labs, or they weren't in Mira's, they weren't in Thinking Machines, or any of these investments?
I'm a big believer in Conway's law. Conway's law is a programming concept that, when it's super-dumbed down for people like us, Harry, says you ship your org chart, or the product you ship looks like your organizational structure. I'm a huge believer in that for venture capital firms as well. I think you ship your fund size, or you invest your fund size, and your team structure.
If you have a $7 billion fund and you have 50 people, you definitively need to get in on these mega-rounds. It is the only way that you can put $1 billion of capital to work productively in a single shot. If you don't and it ends up being successful, you are left in the dust, while all of your megafund brethren got those returns. Now you're benchmarked poorly against them because you missed one of those things.
For a firm like Benchmark, it might not make any sense at all to invest in a $5 billion financing in a lab, even though those are good investments, because of our fund structure and our lean size. But our lean size and our smaller fund size also allow us to do other things that we think could generate even better returns.
Out of our last fund, our 5 best investments today, held at LRP—last-round price—are about a 60x. We have 2 30xs, and we have 2 20xs. Since ChatGPT was released, there isn't an OpenAI round that touches that return multiple and that money-on-money multiple. We have 5 of them, and each of them, I think, has a fair amount of upside even from here today, or maybe a lot of upside even from here today.
So I think you have to choose the game that you're going to play, and it's based on how big your fund is and how many people you have on your investment team. There are so many different ways that we can play the game and generate maybe even better money-on-money returns than folks who are investing in the labs. I do think the labs have obviously been amazing investments.
Your fund size dictates the problem that you're solving for. When you said something about missing OpenAI at $30 billion, transparently, all I thought was, that's like a 15x on a blunt multiple to where it is today. But with dilution—
Ev Randle
Okay, let's say 14x. Let's say 12x.
But with actual dilution, you're looking at more like a 6x to 8x, which, don't get me wrong, is fantastic. But when you do a comparison to your Lovable, your LangChain, your Sierra, your Mercor, and your Fireworks, I mean, we're focused on cash-on-cash.
Ev Randle
100%. Again, I think the lab investments are amazing as well, but our job is to—if we're going to stay small—the only way we're going to impress LPs is by having incredible cash-on-cash returns.
Do you worry you need them to stay relevant? I agree with you on LPs, and I agree on cash-on-cash, but just in terms of relevance with founders and the community, do you worry that you need them to stay relevant?
Ev Randle
I think it's a question that we need to constantly be asking ourselves. If we ever find that our network access, the close relationships we have, and the people we have access to are slipping, or we're not getting access to the right people or the right network nodes, it's something that we always need to be sharp on and revisit.
But if you think about the cultural-touchstone founders of today's AI era, is there anyone more than Bret Taylor who represents this wave of AI applications? He's the godfather of AI apps right now. When you think about these really cracked young teams in AI, who do people look up to more than Brendan at Mercor and what they've done on the AI infrastructure side?
At least thus far, even with our strategy and the trade-offs that mean we can't invest in every single good round, we've still been able to attract and partner with—and I think build really great relationships with—a lot of the founders that people look up to in this AI wave.
I think our network thus far has been exceptional. But I do think it's an ongoing question, because if all there is left is these billion-dollar raises in order to build relationships with these people, then that's an ongoing question.
Well, if that happens, then all of us will either work for the North Korean army or for Andreessen Horowitz. One or the other. [laughter]
Ev Randle
They're one and the same to me, Harry. They're one and the same to me. Marc and Ben. He said it. He said it. [laughter]
You said the word “slip,” and then you said about Mercor. I'm in Mercor a little bit after you guys, sadly, but I did see the article that said the ownership Benchmark had in Mercor was obviously much less than traditional. I think it was about 10%, give or take.
I'm not asking specifics about the company. I'm just intrigued: How do you think about discipline around ownership in a new AI world where everyone's ownership is trending down?
Ev Randle
When I think about Benchmark's north stars—what we really care about in our investment strategy—and this relates to the ownership that we get in our investments versus everyone else gets in their investments, we have 2 north stars that we think about.
We want to be the highest-ROI and closest partner to the founders that we partner with. We want to be their most meaningful VC and partner from the moment that we partner with them until the company no longer exists. It can go public; we'll still be on the board, but until the company is no longer a going concern. And we want to generate the highest money-on-money returns that any of our LPs have in their venture portfolio.
But that's basically it. As the asset class evolves, there are ways to really serve those 2 north stars without necessarily having to get 20% ownership, or around there, every single time.
I think we're all believers—I think you're a believer, and we're certainly believers—that the outcomes are much, much, much larger in today's technology landscape than they were 10 or 15 years ago. There are more bites at potential $100 billion and trillion-dollar companies. There's just so many bites at the apple in terms of how you can both be a really meaningful partner to the founders and, 2, generate really exceptional returns.
So, in the Mercor case, yes, we didn't get high teens or 20% ownership, but I think if you talk to the Mercor folks about who their most impactful VC partner has been, I think they would say Benchmark. And I think that's going to generate unbelievable returns for our LPs. It's one of those cases where you can do the math on what the money-on-money return has been thus far from our ownership stake.
I think sometimes people confuse the inputs for the outputs at Benchmark. It's like, “Oh, they have to have 20% ownership, and they only want to invest at 100 post,” and all these things. I think that couldn't be further from the truth. We really have those 2 north stars, and whatever the asset class allows for in terms of the relationships that we build and how we can deliver the best for LPs and our founders, that is what we're serving toward and that's what we're optimizing for—not some vanilla percentage-ownership number.
Ironically, I think the last thing I'd say on that is, if you pulled any of our founders and said, “Do you regret the percentage ownership that you gave to Benchmark?” I don't think you'd get a single one of them to say, “No, we gave Benchmark too much.”
I think that's one of the really special things about the history of the partnership. It's my third week, so obviously I've contributed nothing to that. I'm just speaking to the amazing work that all of our current and historical GPs have done for the platform. But I also do think that even when we get—and we still often do get—really high ownership stakes, I don't think a single founder regrets that partnership.
Ever. It seems that despite many years in venture, you still need a lesson from me, which is: regardless of what you did, it was all credit to you for the brilliance that happened before. [laughter]
Okay, it was me. Yeah, yeah, yeah. Cool. I remember doing eBay back in the day. Pierre and I were hanging out. We basically co-founded the business together.
Ev Randle
So good. Yeah, I do need to work on that.
You said about—well, we can take this out if you want to—I pry, and you can take out one of your old partners. Delian is quite vocal about Benchmark. I mean, it's kind of the popcorn GIF, you know.
Ev Randle
It is.
And you say about being the best partner. I think if you can answer it, it's helpful, because Delian's commentary doesn't help when he says, “Well, you just fire founders continuously.” Is that not slightly incongruous, being the firm that fires founders and also your best partner?
Ev Randle
I'll start with Delian's media strategy. Delian—and, you know, I love Delian. He's a close buddy of mine, so I hope he doesn't—I don't think he'll mind me saying this, because he certainly busts my balls more than enough.
Delian has always found an amazing media and Twitter strategy, which is: go find someone with a stalwart brand, or go find the biggest person on the playground and punch them in the face. People love it, it gets a lot of likes and clicks, it helps raise your profile, and it almost elevates you to their positioning. He's done it to Sequoia an immense amount. He's done it to Andre over the years, and we have not been spared the clickbait Delian tweets either.
No, I think every story has an immense amount of nuance in what happens between a board, a founder, and a management team, and there's an immense amount that goes into every single one of those decisions.
Again, I think times also completely change. In the '90s and early 2000s, if you think about or read about the Google investment, it's like Kleiner and Sequoia do the Google investment and immediately start searching for a professional CEO. It used to be the absolute norm. It wasn't even like, “Oh, we're going to push the founders out.” It was like, “No, you invest, and then you all together go look to recruit a CEO.”
2025 is immensely different than 2000. It's immensely different than 2010. It's even immensely different than 2015. The relationships between boards and founders have changed, and the relationships between venture firms and management teams and companies have changed a lot. I think it's for the better.
Obviously, I spent a fair amount of time in my career at Founders Fund. I love the idea of never firing founders and having them lead their companies from the moment you partner until the IPO and beyond. But at the end of the day, I'm also a believer in basic governance. If you do end up investing in someone who breaks the law or someone who's crossed ethical lines, it is your responsibility as a board member to potentially take remediation and action on behalf of all of the shareholders, all of the employees, and the company.
If you take a board seat like we do and you do have governance, you ultimately do have at least basic ethical and moral responsibilities. I think it's actually a convenience for Delian and some of the Founders Fund folks to absolve themselves of that weight and responsibility just by being like, “Oh, it's not part of our thing.” But I think it's almost out of laziness sometimes, more than it is some duty that they feel to founders.
100%. I completely agree. I also think we're actually too kind on the flip side, where we now do not adhere to our fiduciary responsibility because we do not want to lose NPS so much. I'm on a company now that I'm invested in where the board is deliberately obfuscating their fiduciary responsibility just to preserve founder NPS in case they say something bad.
They are not looking after the cap table just because they do not want to piss the CEO off. That is a deliberate obfuscation of your responsibilities to protect shareholders and do what's best for them.
Ev Randle
100%. I completely agree. I also think the best founders don't want sycophants in the boardroom. They don't want GPT-4o in the boardroom telling them that everything they do is perfect, that they walk on water, and that they do nothing wrong. They actually want other adults in the room who are going to push them, spar with them, and make the company better.
Can I ask you—you said about multiple bites at the apple. I wrote down “apple bites.” When we look at Benchmark over the years, Fenton did Airtable's Series C. I think Gecko was Series C. LangChain was a seed.
To what extent will you push the partnership now to expand the boundaries of what we call an A and what Benchmark does, to take more broad bites at the cherry or apple?
Ev Randle
Yeah, I think historically, again going back to those north stars that we talked about, when you think about the Benchmark investment strategy, it's helpful to marry the north stars that we talked about: every investment needs to be potentially absolutely astronomical money-on-money returns for LPs, and we want to be the most meaningful partner to our founders.
There are a lot of different ways to do that. So I think you marry those north stars with the personal investing style of each of the GPs. Each of us is 25% of Benchmark, and we work really well together; we’re a super tight-knit team. But each of us has our own styles. So even if Eric Vishria tends to love getting in right at inception and being the first check in, being really in the primordial-soup phase of a startup, it doesn’t mean that Chetan, Peter, or I are always going to operate exactly at that stage. We all have our particular preferences.
I think Peter does an amazing job of just following his founder conviction. He doesn’t think about stages. When he finds a Howie, when he finds a Brett, when he finds any of these founders, that is what guides him. He’s like, “I’m going to find a way to become the most meaningful partner to this founder, and I’m going to find a way for the investment to make them a lot of money for our LPs.” And so I think that is the mindset that we all have.
Historically, I’ve done more growth. I’ve focused more on Series B and beyond than on early stage. So will I do more Series A, B, or whatever we call it these days, than inception- or seed-stage investing, especially at first? Probably. But again, we’re guided by those north stars and finding founders that we really resonate with. Typically, we’re able to find ways to make it work on the back end and for our investments to generate exceptional returns.
So I think, unlike a huge mega-fund that’s like, “We have our Series A partners, we have our Series B partners, we have our Series C partners. They do fintech, they do healthcare, they do blah blah blah,” we don’t think about those things at all. We really just think about our north stars. I think we’re realizing more and more that there are so many different ways that you can have 10x, 20x, or 30x returns.
I spoke to one of your former colleagues, and they said, “Ev is a phenomenal growth investor, but he’s a growth investor.” When I think about what matters at different stages, for me, in the early stages it’s people, and in the later stages it’s the market—actually, market sizing, depth, and just how big something can be.
We recently did Airwallex late at $4 billion. Incredible. But why? Because, dude, B2B payments—it’s a big market. We’ve got a lot more room to run. How do you think about that shift earlier? Are you nervous about making it? And what changes in what matters in your mind?
Ev Randle
I’ll be vulnerable with you, Harry, and say that this was a dinner I was having with Eric Vishria on our team. I was having a moment of insecurity when I was talking to him about the role of being a GP at Benchmark and saying, “Hey, I’ve mostly done growth.”
He was like, “Dude, Bill Gurley was a public-markets analyst before he came to Benchmark. You certainly are not going to be the most off-the-wall hire that Benchmark has made. That’s actually more par for the course for Benchmark.”
I think also, when you look at who you think are the amazing investors today, they transcend stage. If you look at Pat Grady, does Pat Grady think of himself as a growth investor, or does Pat Grady think of himself as just an amazing investor? Maybe he’s too humble. He’s a pretty humble guy, so maybe he doesn’t think about himself as an amazing investor at all. But I look at what Pat does, and I’m like, he finds incredible founders in investments that he thinks have an immense amount of upside, and he goes and partners with those founders.
I wouldn’t say that I’m an amazing investor yet. I don’t have the track record yet to say that I am, but that is my north star and that is my goal, and I’m going to work my ass off to do that. I think I’ve been able to tune my intuition, and even though I’ve used it to execute on growth-stage investing, if you look at a lot of the people that we think are growth-stage investors, they’re doing earlier-stage companies now and doing a lot of different stages at the same time.
I thought Pat just did the deals his wife did. [laughter]
Ev Randle
I’m just kidding.
He said it. He said it, not me, Pat. He said it. [laughter]
Ev Randle
Dude, I’ve known him for 10 years. I’ve said this shit. You can get away with that.
I’ve said it for years. And this is why I think he’s just like, “I’ve never met Harry. Don’t know who this guy is. No idea.” A thing that does change, obviously, is price, and it does matter at different stages. How do you think about your own relationship to price?
Ev Randle
I think starting my career as a growth investor actually really helps me. One of the first investments that I did at Kleiner Perkins when I came back in 2022 was SpaceX at $150 billion. At the time, it was like, “Oh my God, $150 billion entry price.” The absolute numbers were, “Can we really make a good return on this investment?”
Having to go through the process of saying, “Hey, let’s not focus on just some large absolute figure. Let’s look at the TAM. Let’s look at their competitive position in their market. Let’s look at what happens if this goes right, and let’s look at the probability of it going right and who could potentially knock them off their perch to make it not go right.”
When you actually zoomed back and said, “Let’s take a few zeros off every single number—the TAM, the valuation, the revenue, everything,” if you were to look at it as a vanilla WidgetCo and reduce every number by 2 orders of magnitude, you’d be like, “This is an absolute no-brainer investment with a 10x upside case.”
I think doing later-stage investing can really help you think about price even at the earlier stage, because it makes you think, “I’m going to ignore what feels like a large entry price relative to the market.” If you’re in a market where everyone’s like, “Oh, the Series A market’s $100 million post-money, and if you do something at $200 million post-money, you’re an idiot because that’s 2x more expensive,” then you miss Rippling at $250 million.
The famous Series A that Mamoon did, where everyone’s like, “This guy’s out of his mind. He just paid $250 million for a Series A company that barely has any revenue.” Obviously, you miss Parker’s excellence, you miss the TAM that he’s going after, you miss the product sequencing and the differentiation that he’s going to build, and you miss the exceptional team that he had built.
I think if you can always try to isolate, “Hey, I’m not going to care about what’s going on in the market. I’m going to care about what matters for an investment and how much upside I think there is in a vacuum,” that matters a lot more. It’s something that you can actually get if you start your career in growth.
Do you remember when Andrew Reed did Figma, and they were at $4 million in ARR and he did it at $400 million, and everyone was like, “This guy is 100x—what? Nuts.”
Ev Randle
One of the first 100x deals, I think, in SaaS. People were like, “100x ARR? What the hell?” And obviously it ended up being, I don’t know, 30x, 40x—an unbelievable investment.
Do you outcome-scenario-plan, though? Because you said there about market analysis and trying to do top-down versus bottom-up. How do you think about that, and do you not worry that it can mislead you in the wrong direction?
Ev Randle
This is a lesson I think I learned from Mary mostly, and it’s one of my most important frameworks. You should understand what the base case, or the base-rate future, of the company looks like.
If you were an equity analyst and this was your 100th company that you were doing a little forward model for, and you weren’t paying that much attention, you’d be like, “Okay, just triple, triple, so it’s going to double, double, double,” or whatever. If you just did, “Hey, this is what the market thinks is sort of the baseline of what this company should do,” it’s actually extremely helpful to lay that all out and visualize that.
I don’t say, “This is the bull case, this is the base case, and this is the bear case.” I lay out what people are underwriting to. Let’s say at the growth stage people are underwriting to a 3x–5x. What does that look like on paper? And then how does that jibe with my mental framing of how important this company is going to be for its customers, for its market, and for the US economy, in some cases?
When you have a really strong intuition about a company in the middle of an inflection that’s about to absolutely explode, you look at the numbers that people are underwriting to in order to get to their 3x–5x, and you say, “This company’s going to absolutely smoke these projections.” It happens very rarely, but it’s really nice because it gives you the amount of conviction when you look and say, “Oh, everyone’s going to underestimate this thing.”
I think the other reason why models are not useful beyond that simple framing is that with every successful investment, you just feel stupid if you were to model Figma’s growth. Everyone would make fun of you. They’d be like, “Dude, come on. You’re just trying to get this deal done. Why would you model it growing this fast, for this long, this profitably? It’s never happened in SaaS. You’re crazy, or you’re just doing the IC a disservice.”
And so I think beyond being a yardstick to test your conviction, models aren’t that useful.
But for that, they're really, really good. I always remember Ernie from Carvana coming on and saying, “The amount of investors that would be like, ‘Hmm, the biggest car showroom is like a $300 million market cap, so this is a bad business.’” I always think market comps are such a dangerous thing to rotate your mind around when investing.
100%. Or Figma with designers. I think David George—maybe it was on this show or another—talked about how he underestimated and understated the TAM of Figma because he went back to the team and said, “Well, look at how many designers there are in the world. If you just do the P times Q—the price times the quantity of designers—you don't get that big of a business.” Obviously, Figma then ended up penetrating a lot more roles within a company beyond designers.
People, product, and market: rank 1 through 3 in order of priority for you?
Ev Randle
The way you said it, honestly: people, product, and market. I think the people define everything else. They are the upstream engine that makes everything go. They're the most important piece.
Second, I think the product that the people build tells you so much about the people. It's the greatest evidence of the quality of the people: the product that they build.
Then market, third. Obviously, I am a believer that the market you're in ends up defining the size, and then the founder determines what percent of that size you can get in your exit. But I just think it's the most fungible. I don't think you can turn a non-exceptional person into an exceptional person. I don't think you can take a team that can't build a good product and make them a team that can build a good product. But you can change markets, especially early on in a company's life.
Most of the amazing companies and exit stories had some pivot along the road, whether you're talking about Slack or any of these others. I truly think that because it's the most fungible, market is the least important of those 3 things.
You said that you can change markets. It was on this show where Doug Leone said, “Venture capital has transitioned from a high-margin boutique community to a low-margin, commoditized industry.” Tears ran down my face with my $400 million fund, which seems quite paltry. Do you agree with him in that statement?
Ev Randle
I think Doug might have gotten that idea from me. I'm half kidding. I wrote this piece back in 2021, and I think it's the reason why we first DM'd. It was called “Playing Different Games.”
Ostensibly, the piece was about the rise of Tiger. But what the piece was really about was the rise of a firm-level strategy that centered itself around increasing investment velocity as the core strategy. The idea was that you could make more money as a firm and as a GP if you invested a lot more money per year, even if you thought the forward returns were going to be lower on average per investment.
The idea was that Tiger was really the first one to take this idea and really, really run with it. They raised $15 billion or whatever they did in 2021. John Curtius basically deployed it all over that 18-month period.
At the very bottom—this is the ironic part of that piece—I said venture capital was going to bifurcate. On one end, you're going to have the Tiger model, which is high capital velocity, a lot of money out the door every single year, low touch, good prices—giving founders really good prices. On the other end, who did I have? I had Benchmark, ironically. That was going to be the craft: high-touch, the best signal that you can get if you're a founder, and very, very involved.
In the middle, I said we have the J.C. Penney funds, which are the dead zone. I think the crazy thing to me is what's happened over the last 4 years. Harry, how many firms have moved towards the Tiger side of the spectrum? Tiger died, and we got 6 or 7 more Tigers out of it in the last 4 years.
Obviously, a lot of these firms are running different strategies. Thrive and Founders Fund are doing extremely concentrated investments in really high-quality companies. You have the mega-funds, like Lightspeed and GC, doing their thing. It's a lot of different flavors of capital velocity—investment velocity—as their north star. But there are 6 to 8 firms now doing capital velocity, investment velocity, as their north star.
That was one of the reasons why I was really confident in high conviction and joining Benchmark, because if you look at how many of those tier-one brands have moved more toward the Benchmark side of the scale, there's basically none.
Can I just push you? Do you think they are doing capital velocity as their north star? I think Josh would ardently push back on that from Thrive, and I don't even think you could apply it to Lightspeed and GC. I think they're solving for large checks, which is why they have to be in these mega-companies, because they need to deploy $500 million in some cases. I don't feel like they're solving for velocity in the same way that Tiger was.
Ev Randle
I do. I would push back. There are 2 things. Obviously, there are different subsegments of this now. So let's take the actual mega-funds, the people that I think are most following this strategy. If you were to say, “Well, for GC or Lightspeed or some of these mega-funds, is investment velocity the north star of the strategy?” to answer that question, I would have you go talk to the principals, the junior partners, and the associates at those firms.
You interview 10 of those people, and you tell me that capital velocity is not the north star of those firms, and I will cede victory to you, Harry. When you actually look at what's going on at the ground floor, it doesn't matter what Ravi Mhatre is saying. When you actually look at what's going on with the people who are actually doing these investments, they feel it. They feel that they need to get money out the door, and that's the only way that they're getting promoted up those organizations.
On the Thrive side, I agree with you. I think Josh would resent that characterization, and it was probably too blunt of a characterization. But again, I think subconsciously, as a firm, it's very, very hard to care about something that's not your main product.
When you're writing billion-dollar checks, that is your main product. That's what's going to make you all the money. If you put $3 billion in OpenAI and it's going to turn into $12 billion, it is unbelievably hard, whether it's conscious or not, to then go and say, “We're also going to be the best Series A firm, and we care just as much about Series A,” because why would you?
Ninety-five percent of the profit that you're going to make and the money in your pocket is going to come from the billion dollars you put in Databricks, the $3 billion you put in OpenAI, or any of those things that have ended up being your main product.
So I think you just can't focus on everything and give it your all. Even though it's less conscious for firms like Founders Fund and Thrive, it has become their main product and their main focus subconsciously.
If we accept that, people then often move to the idea that they're going to do worse. They accept a lower rate of return because the Norwegian sovereign wealth fund wants 4% a year, and so that's what they're going for.
Then you actually look at outcome scenarios and outcome sizes: OpenAI, which will be a $1 trillion company next year; Anthropic, which will definitely be a $600–700 billion company; and Cursor, which hits $1 billion in ARR insanely fast. Outcomes are so much larger than we ever anticipated. I think they will make a huge amount of money because the outcome sizes have continuously expanded. Do you agree?
Ev Randle
Oh, yeah. They're all going to make an immense amount of money. But again, let's change the framework from absolute dollars to what you're giving each stakeholder of the 3 legs of the venture stool. Venture has 3 stakeholders: you have your LPs, you have your founders, and you have each other as GPs within a firm.
I don't think, as Ravi, Hemant, or even Ben and Marc at this point, that they can go to LPs—one of those legs of the stool—and say, “Hey, this basket of funds that we're making you invest pari passu across, we're going to get you 5x net on that.” I don't think they can say that, or they at least can't say that with a straight face. If you look at the recent return data, I think it suggests that.
I think they'll be able to make an immense amount of money on an absolute basis, but I think a lot of these LPs are in venture to make high money-on-money returns. They have PE for the low-return stuff, and they probably get better liquidity from PE. They're here for the high money-on-money returns.
This is one of the reasons why I'm extremely excited about Benchmark's competitive position in today's market, because we can go to LPs and say, “Hey, we're shooting for higher than 5x. We have the historical track record to back it up, and we have the fund sizes to back it up as well.”
I mean, you had Miles from Carnegie Mellon come on here and do the awesome math, the very clear math, of, “Hey, do you know how hard it is to return 4x net on $8 billion, $10 billion?” It is immensely hard, and it defies the laws of physics.
So, I think there's a difference between whether they're going to make a ton of money and whether they're going to produce the returns that LPs really want this asset class to produce. Those are 2 very, very different things. But for now, the rubber won't meet the road because, as you mentioned, there's just so much global demand from LPs for exposure to private technology, and they're happy to take lower returns. So I don't think there's any end in sight, but I think on a relative basis, between all of these different constituents and all these different GPs, there's a huge delta and huge differentiation between who can actually produce venture-like returns.
Tiger. Mhm. I think Tiger will do much better than anyone anticipated when you look at their positions in Scale AI, OpenAI, and the protection they're going to get from a lot of liquidation preferences that they actually have, meaning a lot of them will get 1x plus a little bit, maybe. Do you think I'm wrong and being too optimistic, or do you think the whole ecosystem shat on them a little bit too early?
Ev Randle
I completely agree. I think Tiger's going to end up much better than anyone thought they were going to end up. I jokingly texted some of my friends, and I was like, “Hashtag justice for John Curtius.” I actually think everyone put him as kind of this pariah, the personification of the excesses of 2021, but, again, it might have proven—his strategy might have proven prudent and the correct strategy all along because they got really big stakes in Databricks. They invested in OpenAI very, very early. I think they have a large position in OpenAI. They actually have large positions in a lot of these amazing companies that could continue to compound 5x more.
Again, they'll probably benefit from the liquidation preferences and the beauty of having preferred stock for a lot of the things that don't work. In the fullness of time, I'm sure it's not going to be the best portfolio that any LP has ever gotten, but I definitely don't think it's going to be a money-incinerating fund by any means. I actually think it might end up being pretty okay once Databricks is a $400–$500 billion company and OpenAI is a multitrillion-dollar company.
It is hilarious. I do think people gave them too hard a time, probably, and I do think they might end up being okay.
I love that hashtag. I'm sure John will listen to this and be like, “Yes, thanks, guys.” [laughter] When we were chatting about multistage funds before and going back and forth over email, you said how it sucks to be in a mega-fund. Why does it suck to be in a mega-fund, Ev? From the outside, building a firm, having mega-fees, mega-offices, Fiji Water in unlimited supply, [laughter] and more EAs than you have investors seems pretty good. Can you help me out here, dude?
Ev Randle
Okay, so maybe I should caveat by saying all on a relative basis, these people definitely aren't going to the coal mines and laboring all day under the hot sun or something. But I know I have so many friends at these funds, and some of them are probably going to kill me for this part of the conversation. I've mentored a lot of people who are either coming out of private equity or thinking about moving firms in venture growth, and one of the first things I say to them is, think about the day-to-day that I know exists in a lot of these mega-funds.
If there are 50 investors, if you come in and you're the 23rd partner at ICONIQ or one of these places, what companies do you get to cover? That's the first problem: you end up getting a very small sliver of the overall market because so many people have already laid claim and are the point person on the very best companies with the very best founders. So you end up being focused on this local maximum where you're like, “Okay, I have 30 pretty good companies where I am the point person on the relationship.”
I also really need to do investments because that's how people get promoted here. I really need to get a couple of these in the portfolio. To me, sometimes it just feels like a different job than the craft of venture capital, where you're almost playing the lottery. You're like, “Okay, I have these 30 names that I own. I'm going to try to do 2 of them, and then if 1 of them hits and is a huge success, then I'm going to get tenure and I'll get to be a GP, and then I'll get more coverage, and then everything will be okay.”
But it feels a little bit more like investment banking or a large private equity firm than it does what people think of when they think of being at a venture capital firm, which is meeting really interesting founders, building genuine relationships with them, and only doing the very best investments and partnerships. I just think it's really gotten away from that, and I think it's inevitable. Again, it's to that Conway's Law point of venture capital firms. It's just based on the fund sizes and team structures of these places.
I agree. And if they get fed up with the private chefs and the Aesop soap in the bathrooms, then they can always go and build their own funds and toil away and do the painful, hard yards, in which case I wish them well. [laughter] I do agree with you there. I have to ask you—you mentioned doing those 2 deals out of the 30. The first deal is really hard, dude. How do you think about your first deal at Benchmark? You can fall on 2 sides. Just get it done; it may not be your best, but it's kind of like the first check. [laughter]
Just get it done. You promised me spicy, Harry, and you delivered. You delivered. [laughter]
Or it's like, “You know what? I'm going to wait until I find the perfect company, and only when I find the perfect company am I going to commit to it.” Which side do you sit on?
Ev Randle
Yeah, my first investment. I won't lie to you, Harry. The weight of being a Benchmark GP exists. I definitely feel it. You're like, “Wow.” The people who have walked these halls—Bill Gurley, Mitch Lasky, Matt Cohler—incredible people with incredible track records, and you feel a lot of pressure to live up to the history of this place and the history of the brand.
I think I got really good advice. I won't name the partner, just so no one can trace back to what company they're talking about. One of the partners coming in was like, “Look, there's going to be nothing better for you than if your first investment sucks, because once you do 1 and it fails and you realize it's not the end of the world, life goes on, LPs still love us, and you're not fired, then you feel really comfortable and you start getting into a really good rhythm.”
Whereas if your first one's pretty good or looks really good, you can then feel even more pressure on the second one. So they were like, “There is something beautiful in having your first be a failure. It doesn't mean that I'm going to be looking for a failure out of my first investment, but it was really relieving, and it was amazing advice to get: it's all okay, and if anything, failing can help you feel more relaxed as you go up to bat the next time.”
Just do 1 of them. [laughter]
Ev Randle
Exactly. Yeah. I'll do a Swedish satellite antenna company or something that Delian did, and then I'll be guaranteed a zero. [laughter] Having said that, obviously, Delian—I'm joking. Delian's track record is pretty good. You look at Sword Health, where you're like, “Really? That's a non-obvious pick,” and, “Wow, what a business.”
The thing I always give Delian crap for is that he does have an incredible track record, but a lot of it is software. He loves to shit on software companies and people that invest in software and all these things. And I'm like, “Dude, you sourced the seed of Ramp. Khosla owns an ungodly amount of Sword Health, which is an amazing company. And, yes, you obviously incubated Varda, which is a great company as well, and all these things, but a lot of your track record is in software.”
But no, honestly, I think a lot of people at Founders Fund have very underrated track records. I think Mathias Vantiani—I think he's the most underrated venture capitalist that exists today. He's so quiet. He's never online boasting about himself or anything, but DolarApp, Trade Republic—
Why? [snorts] I met him when he was in London, and we had dinner, and he was doing the Trade Republic deal then.
Ev Randle
Yeah.
Why do you say he's the most underrated?
Ev Randle
I just don't think a lot of people know about him. He's not online. I think he might—
What's he got to his name, though? Trade Republic. So, Trade Republic, DolarApp, which I'm sure you know, Inter down in Latin America, and then he's instrumental—the growth team over there is quite small.
Ev Randle
You know, you got Napoleon, you got Matis, and you got a few other people like Amin. He's just played a pretty big role in a lot of the really good growth investments as well. So I just think, for someone who is ostensibly a growth investor, he's done a ton of really good early-stage things.
Yeah. And he's humble and nice. He's like one of those perfect kids at school. You're like, “Oh, God.”
Ev Randle
I don't know about nice. Some people don't think he's very nice, but he's a sweetie at heart.
Ouch. [gasps] Do you know what? I spoke to Henry from Stord before, and he said, “You got to ask, what's the most ridiculous story you remember from the 2021 times?”
Ev Randle
Oh my gosh. There were so many absolutely absurd ones. Again, I was at Founders Fund, so we were spending a fair amount of time in Miami. I remember distinctly, I think it might have been the 3rd Miami Tech Week or something. I think it was December of '21 or maybe January of '22, when it was pretty clear that the bubble was bursting from COVID and equity valuations were starting to get slashed 30% to 40% in public markets.
We were at some very decadent party where I think Vanilla Ice was performing or something. We were in Miami, and there were all these crypto people. I was just like, “Oh my God.” I was sitting around, and I was like, “This reminds me exactly of the scene in The Dark Knight Rises where Anne Hathaway is dancing with Bruce Wayne, and they're at this fancy party.” She's like, “I don't know how you could think that you guys could do this glamorous, decadent stuff while Gotham is burning.”
I was like, “Wow, we are at that party today. Gotham is burning. It's about to come to us, but for this time, this is the last decadent thing that we're going to be doing.” I just feel like 2021 was all like that. There were so many ridiculous things where I look back and it's like, why did we ever think, 1, this investment was a good idea, or 2, why were we doing these very decadent things in Miami? It just seems ridiculous in hindsight.
I am so here for a Dark Knight reference, by the way. I love that scene. Well done, dude. Love it.
The final one before we do a quick-fire: you can think that Gotham is burning today, actually, in a lot of ways when you look at the state of the world. You can also look at it and go, “Christ, we're so early in the adoption and inflection of AI that this is just the start.” I hold these 2 opposing thoughts in my mind, and I'm kind of stuck on which one to adopt. How do you feel?
Ev Randle
I feel the same. I think today, relative to the dot-com boom and bust, if you really think deep down about what happened in that era—let's say you did the Amazon Series A, for example—there was a point in time 4 years later where it had IPO'd and then was down 80% from its IPO. But if you had held to today—and I forget if the Amazon Series A was $40 post or whatever it was—you went from $40 post to multiple trillions of dollars in value.
I think today is very similar in that there's going to be a ton of companies that are pump fakes that end up going to 0 or go down 90%. But I think it's really important to position yourself so that you can survive the inevitable crash on the other side. If you end up in these really incredible companies that still endure and define the next 20 to 30 years of technology, you're going to be paid in so many multiples of what you would get in a normal cycle.
We think and stay up all night thinking about, well, what is going to be the Amazon, the Google, and the Microsoft of this era? I think it also goes to our strategy. We're like, let's constrain our fund sizes. Let's be careful about what we do so that we don't get too far over our skis, where we can easily weather a crash and LPs aren't going to go fleeing once there's a crash because we've been very careful with our capital and we haven't incinerated billions of dollars or something like that.
I think Benchmark could invest in Elizabeth Holmes doing Theranos Take 2, and you'd still get LPs queuing out the door, thinking that Benchmark has seen something no one else has seen.
Ev Randle
Maybe we should. Maybe we should. I think she's still in jail, but whenever she comes out, maybe we should.
It's about being contrarian and right. The terrifying thing is she has quote-tweeted me agreeing with me more times than I like, and I'm always like, “This is a bad sign.”
Ev Randle
You're like, “Wait a minute. Apparently that's not her.” At least, I saw something on Twitter where someone's impersonating her. Apparently, she doesn't actually have access. So maybe it's one of your superfans.
That's reassuring if so, because I'm always like, “[Expletive], delete tweet, delete tweet.”
Dude, I could talk to you all day. I want to do a quick-fire round. I say a short statement, and you give me your immediate thoughts. What have you changed your mind on most in the last 12 months?
Ev Randle
I think, honestly, the quality of the AI cloud business model. Again, I was very negative on it when CoreWeave was first coming up. I was like, “Oh, this is reselling a commodity.” I actually think there's a lot of interesting things that people are doing, and the demand for AI inference is just so astronomical that, at least for now and for the next few years, I think it's going to overcome all business-quality and business-economics concerns.
I think at some point CoreWeave and all these things will probably go down 70%. But obviously, I thought that back when it was raising at $3 billion, and now it's a $60 billion public company where the investors have been able to get liquidity. So I was definitely wrong.
Which pumped company today will have the steepest fall, do you think?
Ev Randle
We're investors in Cursor. I am not a believer in some of the companies that have raised a ton of money and have not released a product or haven't had products that a lot of developers are using. I'm a huge believer that you have to get developers' hands on the product.
I think there are a couple of companies that have raised billions of dollars and are like, “We're going to build the best thing ever,” but they don't actually have products that a lot of developers are using. I think that's going to be a rude awakening, because AI products get better via usage. Oftentimes, if you have the right environments and Claude Code, Codex, and Cursor, the companies and apps with the highest amounts of usage are going to improve the fastest and leave everybody in the dust.
Tell me, you've got BOND, you've got Founders Fund, you've got KP—all fantastic firms—but if you had to put your money in 1 firm for the highest cash-on-cash, which one do you go with?
Ev Randle
Maybe Founders Fund, just because they have a very unique ability to incubate companies. When you think about Anduril, the fund that Anduril is in has got to be such an ungodly return on that capital.
I think it's become a really, really competitive market. The only way that you can fend off how hard it is to buy equity is to sell equity or produce equity. The way you produce equity is by incubating companies. Every few funds, or every 5 to 10 years, they've incubated an unbelievable company. Obviously, Scott Nolan over there is the most recent to do it. I think that's just a way to get differentiated returns that are hard to produce from anyone else.
Totally agree with you. Phenomenal. When you look back, no one has so reliably had such good-performing funds at scale—them and Index. Yeah, unbelievable there.
Okay, totally get you. Can you take me to the moment where you said in your head, “Yeah, I'm going to do Benchmark”? Was it a dinner? Was it a coffee? When did you go, “Yeah, I'll do it”?
Ev Randle
I think I knew I wanted to join Benchmark when I was 22, entering the industry. Honestly, you enter the industry and you read eBoys. I'm reading all of Bill Gurley's blog posts. It's a mythical place.
You enter the industry as a young investor and think, if I really work my ass off and get pretty lucky, maybe one day I'll be able to compete for a seat there. So when it comes true and they give you the envelope with the offer in it, it's almost like a childhood fantasy of joining the Yankees or something. It's super surreal.
Obviously, the people matter most, and that was really important during the recruiting process. I felt unbelievable alignment and really a level of closeness with Peter, Eric, and Chetan. But beyond that, I think the firm itself is just one of these mythical seats that you dream about from the day that you enter the asset class.
It's like Real Madrid and going to the Bernabéu. I'm a—
Ev Randle
You tell me. I played FIFA in high school, but I don't know. I think it's like playing for Real. You grow up watching Ronaldo and you're like, “Oh my God, I could go play for Real one day.”
We all chat shit that I'm so happy at Chelsea. I love being at Man U.
And then Ronaldo gets the Real Madrid offer and you're like, “Yeah, Old Trafford's not so great.” [laughter] Not so great, is it?
Ev Randle
Not as sunny as...
No, I totally love that. Tell me, what's the biggest miss for you, dude? And how did that change your mindset?
Ev Randle
Biggest miss, we've talked about it a little bit, but the biggest miss has to be OpenAI at $32B. I think, obviously, it was kind of a hard-to-fill round. Obviously, they did it, but it was very non-obvious at the time, and it was just one of those ones that is so unbelievably painful because you missed the forest for the trees.
You let the structure thing and the dilution thing trick you out of investing in what is maybe going to be the largest tech company of all time. And also, just being in that ecosystem, it's such an unbelievable group of people that even if it was just an okay return, you'd still want to be involved with Brad and Sam and all the people over there that have defined a lot of what the AI industry is today. That one hurts to this day.
What do you think the biggest threat is to Benchmark being successful in the next 5 years?
Ev Randle
I think the thing that is most dangerous, and the biggest risk to Benchmark not being successful over the next 2 decades, is stasis. I think we need to be dynamic. We need to always be evolving with the asset class while staying true to our North Stars.
I am very much a believer that we don't have to leave our North Stars or bastardize our true north in order to continue to be involved with the very, very best companies. But at the end of the day, being involved with the very, very best companies is the currency by which we all live in this asset class. That has to be the most important thing.
If there's ever a situation where we're letting our North Stars, or we're letting something else—the tail—wag the dog, with the dog being getting involved with the very, very best founders building the best companies, that's when we need to reevaluate our strategy and what we're doing.
Penultimate one. Are you ready to lose all of your friends in your new partnership? Who's the best picker in Benchmark?
Ev Randle
I'll go with a data-driven one. Honestly, I think Eric is the most underrated picker. When you look at some of the things he's done, he's just done some really low-key things that have ended up being unbelievable.
He'll have Cerebras, which will IPO at some point in the future. That's going to be an unbelievable company. He has a lot of these sneaky absolute bangers. But if you go back and look at the percentage of Series A investments that ended up being generational companies, I think it's got to be Peter.
Peter also has the advantage of being in the game for 20-plus years, but it's got to be Peter.
I also think Peter is possibly the greatest salesman I've ever met. His ability to manipulate language to sell his position is really beautiful.
Ev Randle
I went into my first pitch with him in my first week, and my jaw was on the floor more than the founder. I was like, “I have so much to learn from this guy.” I've never seen someone who practices the craft of infinite EQ and hospitality and just wordcraft like he does. It's unbelievable to witness.
You weren't like, “Wait a minute, this is just like being in a room with Delian.” [laughter]
Delian's more of a blunt instrument, I would say.
Oh, dude, he's going to [expletive] either hate me or love me for this show. I have no idea.
Ev Randle
Probably a mix of both.
Dude. Final one for you. I like optimism. What are you most excited about for the next 10 years?
Ev Randle
I think over the next 10 years, the only thing that's ever made me less of a capitalist than I am is realizing that capitalism is really, really good at optimizing things and making them more efficient. When the laser beam of capitalism moved from cars, TVs, and electric goods, making them cheaper, to the minds of people, I think it actually had a lot of negative consequences.
When social media had its rise, the thing it was optimizing for was, “How do we get people to glue their faces to the screen for as long as possible?” I think that's actually been a significant negative and probably the only negative that technology's had on society thus far.
When we look at what's happening in AI, Peter Thiel always talks about how the most important thing for keeping our society harmonious and functional is growth. As soon as this pie stops growing, things get a lot worse and people treat each other a lot worse because it's zero-sum. I think, unfortunately, you're seeing a fair amount of that where you live, Harry.
When you think about the determinants of GDP growth being basically population and then GDP per capita, and how much the birth rate is slowing, I think AI is going to be unbelievably good at continuing GDP growth. I think continuing GDP growth—the growth of the economy, continuing to grow the pie, having the middle class grow, and having everyone feel more and more prosperous over time—is the single most important variable in continuing a harmonious, functional society. I think it's going to do that in spades over the next 10 years.
You know, Ev, honestly, I love doing this show. But I've done it for 10, 11 years, and not every show is as brilliant as this, by any means. It's shows like this that make me go, “This is why I still love doing what I do.”
So thank you for being so brilliant. Seriously, this was so much fun, and I couldn't be more thrilled with this show.
Ev Randle
Thank you, Harry. It's been so awesome just to hang with you, and next time we'll do a pint in London when I'm out there.