Robinhood’s Tenev Targets Private Markets to Upend Retail Investing

Robinhood特内夫发力私募市场，欲颠覆散户投资 Robinhood计划通过将OpenAI和SpaceX等私募股权代币化来绕过公开披露障碍，并认为在人工智能时代，阻碍散户接触顶尖初创企业的监管框架已经过时。 在本次对话中，Robinhood首席执行官Vlad Tenev分享了20世纪90年代保加利亚恶性通货膨胀的童年经历如何塑造了他对金融、稳定币以及跨国财富保值的看法。他针对批评声音为订单流支付（PFOF）进行了辩护，回应了围绕GameStop交易狂潮的公众舆论，并概述了Robinhood向传统银行业务和私募资产代币化拓展的战略规划。此外，Tenev还详细阐述了他通过内部精英团队实现产品快速落地推进的方法，并探讨了他的数学人工智能创业项目Harmonic。 Vlad Tenev在20世纪90年代保加利亚度过的童年经历了一场严重的恶性通胀危机，该危机在1997年达到约2,000%的峰值，导致他最初价值2,000美元的童年储蓄账户缩水至仅剩10美元。 (1:20) 由于20世纪90年代的保加利亚缺乏散户投资平台，Tenev的祖父在公寓壁橱中囤积铜制炊具作为实物价值储藏手段，以保全个人财富。 (1:20) 在面临恶性通胀的现代新兴市场中，由于美元作为记账单位具有强大的全球品牌效应、安全性与熟悉度，海外公民更倾向于选择稳定币而非比特币。 (4:08) 尽管多元化的美股投资组合可以作为国际公民摆脱母国偏好、实现财富储藏的极佳途径，但落后的传统金融基础设施限制了其在全球范围内的普及。 (4:08) Robinhood通过免除7.95美元至9.95美元不等的佣金颠覆了传统折扣券商模式，消除了以往将起步资金较少的新手投资者拒之门外的高昂投资组合拖累。 (7:00)

Robinhood co-founder and CEO Vlad Tenev argues that modern financial infrastructure and automation are necessary to dismantle legacy barriers, allowing retail investors global access to wealth-building assets from equities to private markets. Tenev contends that outdated regulatory frameworks, such as accredited investor restrictions, penalize ordinary consumers while modern technological tools make traditional investor-protection rationales obsolete. Monetary Instability and Global Asset Demand (1:20) Tenev links his perspective on financial access to his childhood during Bulgaria's 1997 hyperinflation crisis, when inflation surpassed 2,000% and his family's $2,000 equivalent savings account diminished to $10. Without access to retail investment platforms, his grandfather hoarded copper cookware as a physical store of value. This background informs Tenev's view of international cryptocurrency adoption: while Bitcoin was envisioned as an alternative store of value, users in developing markets gravitate primarily toward stablecoins pegged to the US dollar because of the currency's global recognition and stability. Tenev argues that a diversified portfolio of US equities could serve the same wealth-preservation purpose for international citizens facing local currency depreciation. However, cross-border participation has historically been limited by poor legacy infrastructure and home-country bias, which Robinhood intends to address by connecting users directly to major global exchanges like the London Stock Exchange and Hong Kong. Automation, Brokerage Disruption, and Revenue Diversification (7:00) Robinhood entered the brokerage industry by eliminating $7.95 to $9.95 commissions and account minimums, targeting retail beginners and active traders who faced substantial cost drag. Drawing on their background running a high-frequency trading firm, Tenev and his co-founder realized that automated software built from scratch could replace the expensive operations departments and mainframe computers used by incumbent firms. This technical automation enabled instantaneous, low-cost onboarding and clearing. Addressing the controversy surrounding Payment for Order Flow (PFOF), Tenev notes that market makers pay for retail orders because retail flow lacks the toxic adverse selection risk associated with institutional trades. While PFOF became a primary revenue driver through high transaction volumes, Robinhood has since diversified into interest income, options, crypto, and Robinhood Gold. Operating on a single-tier subscription model analogous to Costco, Gold bundles multiple financial services to establish Robinhood as a default primary financial account ahead of its banking and AI-assisted migration rollouts. Retail Influence, Private Tokenization, and Regulatory Friction (26:41) Tenev maintains that direct retail participation provides a stabilizing counterweight to modern finance, where institutional ETF rebalancing and macroeconomic announcements often trigger collateral sell-offs in fundamentally sound equities. To expand retail access, Robinhood is testing the tokenization of late-stage private companies such as SpaceX and OpenAI, backing promotional positions directly with underlying shares on its balance sheet while aggregating retail demand. This strategy faces regulatory and structural disagreements. While Stripe co-founder John Collison notes that the SEC views selling private equity to the public unfavorably and points out that everyday citizens already hold indirect stakes via pension funds and university endowments, Tenev argues that accredited investor rules are outdated. Tenev highlights the contradiction where retail investors can freely speculate on volatile memecoins but are legally barred from investing in productive private companies, asserting that modern AI tools now provide retail investors with institutional-grade risk analysis. Operational Realignment and Organizational Restructuring (55:42) Following the volatility of the 2021 pandemic trading surge, Robinhood conducted an internal exercise imagining what an external turnaround CEO would alter, allowing leadership to reverse past decisions without personal attachment. Consequently, the company reversed its 2022 remote-first policy in favor of a tiered in-office requirement—mandating five days for executives, four for managers, and three for individual contributors—and discontinued products misaligned with its core user base, such as its debit card for paycheck-to-paycheck consumers. To eliminate operational bottlenecks, Tenev restructured Robinhood from a functional hierarchy into a General Manager model with decentralized P&L accountability. This shift allowed the firm to reprioritize high-volume active traders who drove disproportionate revenue but were previously churning due to beginner-oriented infrastructure. Internally, leadership instituted a 150-person 'Founder's Community' based strictly on impact. Outside of Robinhood, Tenev is pursuing mathematical super-intelligence with his startup Harmonic, translating natural-language mathematics into Lean code to formally verify complex proofs. The discussion leaves unresolved whether private equity tokenization can overcome regulatory scrutiny, as established companies like Stripe prefer capital-light, employee-aligned ownership models while platforms like Robinhood lobby the SEC for modernized tokenization frameworks and accredited investor reforms.

Robinhood 联合创始人兼首席执行官 Vlad Tenev 认为，通过全自动化基础设施消除交易佣金并拓展订阅与衍生品业务，不仅重塑了零售经纪体系，也揭示出全球散户追求美元资产与直接投资权力的底层诉求。 技术重构与零佣金的底层经济学 (7:00) 在 Robinhood 推出前，传统券商如 E-Trade 等普遍向每笔交易收取 7.95 至 9.95 美元的佣金，并设立约 2,000 美元的账户门槛，这使散户面临极高的进出成本，同时也让高频交易者月度手续费高达数万美元。Tenev 指出，传统券商依赖落后数代的专有大型机架构，需维持庞大的运营与审批部门；而他与合伙人早年运营高频交易机构的经历证明，仅凭三名工程师即可编写管理数十亿美元日成交额的软件。Robinhood 据此从零开发底层系统，全面自动化开户、审核与资金链接，将边际成本压缩至极低水平，从而支撑了零佣金模式。 针对订单流支付（PFOF）的争议，Tenev 说明其创业初期商业计划聚焦于 API 访问、融资融券与增值订阅，并未预期 PFOF 会成为主要支柱，早期甚至依托 Apex 与 Instinet 清算。做市商愿意为散户订单付费，核心在于散户订单不具备机构投资者的不对称信息风险，即“逆向选择”极低。Tenev 认为，迈克尔·刘易斯（Michael Lewis）2013年出版的《闪电侠》（Flash Boys）将技术落后的华尔街大行塑造成受害者，并误导公众认为 PFOF 损害散户利益。但事实是，当行业跟进免佣后，TD Ameritrade 股价单日暴跌 35%，最终促成了嘉信理财（Charles Schwab）收购 TD Ameritrade 以及摩根士丹利（Morgan Stanley）收购 E-Trade 的行业并购潮。 通胀记忆、美元硬需求与全球资产配置 (1:20) Tenev 的金融理念深受童年经历影响。1990 年代保加利亚经历恶性通货膨胀，1997 年通胀率高达 2,000% 或以上，他在出生时存有相当于 2,000 美元的储蓄账户到 1997 年贬值至仅剩 10 美元；由于缺乏正规金融投资渠道，其祖父甚至囤积整壁橱的铜制锅具充当实物价值储备。这段背景让他意识到民众逃避本币贬值的刚需，但也观察到海外用户并未如比特币早期叙事那样将 BTC 视作避险首选，而是蜂拥买入与美元挂钩的稳定币，凸显出美元品牌的全球信任度。 除了加密资产，Tenev 认为多元化美股组合同样具备财富储备功能，但受限于落后的跨境金融基础设施，海外散户参与度较低。此外，全球零售投资者普遍存在“本土偏好”（如荷兰投资者只买壳牌、爱尔兰投资者只买本土银行股），而这些本地标的长期表现往往大幅落后于全球指数。为打破这一瓶颈，Robinhood 正规划接入伦敦证券交易所与香港等主要海外市场，满足国际投资者对特斯拉、英伟达等头部资产的直投需求。 商业模式演进与单一订阅生态 (16:12) 随着业务规模扩张，Robinhood 逐步将单一的股票交易与佣金依赖，分散至期权交易、加密货币、净利息收入以及名为 Robinhood Gold 的订阅服务。针对利息收入的周期敏感性，Tenev 强调金融服务本质上依赖交易费、利息差或订阅费三类来源，而引入订阅模式大幅平滑了市场交易量波动。与 Revolut 等机构的多层级会员不同，Robinhood 坚持推行单一层级的 Gold 体系，将八项权益打包提供。 这种机制借鉴了“Costco 效应”或“亚马逊 Prime 模式”的心理机制：一旦用户开始为订阅付费，就会主动寻找机会最大化会费价值。许多用户最初仅因买入股票或加密币进入平台，但在成为 Gold 订阅会员后，其后续的借记消费、高息存款或 3% 返现信用卡需求会优先沉淀在 Robinhood。Robinhood 即将推出涵盖高年化收益率（APY）及大额现金配送的银行服务，并计划运用定制迁移工具与人工智能技术，实现一键转移工资代发与账单支付，以此确立其作为用户主银行账户的定位。 散户资本角色、非公开资产代币化与预测市场 (26:41) Tenev 强调，Robinhood 平台沉淀的资金主要来自于散户削减零售娱乐开支后的“新增资本”，而非分流自传统公募基金，因而批评对手将该平台的产品优势污名化为“恶意的游戏化设计”。对于养老金与机构层层转投 ETF 导致的金融抽象化，散户直接持有股票提供了关键的纠偏力量，例如在宏观恐慌导致优质标的被误杀时充当“抄底买方”。关于股东代理投票，受制于内容晦涩且散户享有极低的退出成本（不满意即可直接卖出），其参与度始终低于机构。 在非公开市场股权方面，Tenev 主张利用完全资产质押（100% 由公司资产负债表持有）的模式代币化非上市资产（如 SpaceX 与 OpenAI），并采用聚合需求的“IPO Access”机制解决流动性限制。他指出，目前散户可以无门槛押注高度投机的 Meme 币，却被合格投资者法规禁止投资顶级非上市创新公司，这种监管保护逻辑在 AI 工具普及的当下已然滞后。在与 Stripe 联合创始人 John Collison 的探讨中，Collison 指出 Stripe 这类轻资产公司并不缺钱且更倾向内部员工持股，且风险资本的有限合伙人（LP）本就包含大学捐赠基金与公共养老金；但 Tenev 仍认为应当允许普通民众直接参与二级要约回购。此外，Tenev 将预测市场定义为超越娱乐的“真实机器”，因其能够剔除媒体炒作，以概率形式实时反映如网球比赛、拳击赛事或大选结果等客观趋势。 组织反思、全面重组与数学超智能探索 (55:42) 面对疫情后期的扩张失误，Tenev 与管理层曾进行“假设由 Frank Slootman 这样铁腕 CEO 接管公司会怎么做”的反思练习。这项演练促使 Robinhood 摆脱沉没成本谬误，全面推翻了 2022 年确立的“远程优先”政策，改为按职级强制到岗（高管每周五天、经理四天、普通员工三天），并建立严格按影响力评选的“Top 150”创始人社群以取代僵化的官僚层级。业务端，公司果断砍掉了针对低收入月光群体的借记卡（Cash Card）等非核心项目，将研发资源重新集中到贡献绝大部分收入但曾被系统忽视的高频交易用户，并将组织架构彻底从职能型划分为由总经理（GM）独立承担盈亏的业务线。 在经纪业务之外，Tenev 介绍了其联合创立的商业数学人工智能实验室 Harmonic。该项目旨在通过将自然语言表述的数学命题转化为程序语言 Lean，实现复杂数学证明的自动化绝对检验，以解决数论领域极难人工核查百页证明的瓶颈。Tenev 表示 Harmonic 模型的数学水准已超越其个人的纯数博士研究阶段，长远目标是实现对黎曼猜想等顶尖难题的机器化形式验证与破解。 Vlad Tenev 描绘了一个由自动化系统、高透明度预测工具与代币化资产驱动的零售金融图景。然而，非公开市场代币化的推进仍面临监管机构对公开披露规则的坚持，且平台如何在全球合规扩张中持续平衡机构利益与散户直投需求，仍有待市场检验。

[4:08] Vlad Tenev: Everyone just wants US dollars [5:41] Vlad Tenev: If you talk to someone in Europe, they also would like to invest in Tesla, Nvidia [7:00] Job candidates from legacy brokerages: How do you do it? We've been trying to do this for decades. [22:04] Elon Musk: You're being an asshole to the people. What you've gotta do is full transparency, 100% put it all out there, play-by-play. [22:04] Vlad Tenev: like these people's father in a sense [22:04] Vlad Tenev: it's kind of time for them to just come home... I think it's enough rebelling. [268:26] Vlad Tenev: derivatives in the ultimate sense [268:26] Vlad Tenev: truth machines [54:39] Vlad Tenev: If there was a disaster or some kind of a apocalyptic scenario, and we had to rebuild Robinhood with 150 people, you would be in that group. [55:42] Vlad Tenev: If they brought in a CEO who really cleaned this place up, what would they do? [57:36] Vlad Tenev: Where are you making your money? How much are we investing? Why are we investing so much in this thing that doesn't make any money and very little in the actual core of the business?

[0:00] No calorie? [0:01] No, I wouldn't say no calorie. [0:01] It's Guinness. It's a meal in and of itself. [0:04] Nobody gives a shit about on-chain issuance. [0:07] They just want a button to deliver money, [0:09] so that they can raise [0:10] for their venture as easy as possible. [0:12] Were there any moments during that saga where you're like, [0:15] "I can't believe I'm having this call right now?" [0:17] Elon was like, "You're being an asshole to the people. [0:20] What you've gotta do is full transparency, 100%." [0:24] You know that I'd love to tokenize [0:26] or otherwise make your stock available to retail. [0:29] And you know that I'm skeptical. [0:30] I wanna see this handstand pushup now. [0:33] Yeah, it's fine. [0:35] 60 calories, so pretty light. [0:38] We'll get Vlad mic'd. [0:40] Vlad Tenev is the cofounder [0:41] and CEO of Robinhood, the company [0:42] that turned zero-commission trading [0:44] into an industry-wide phenom. [0:46] He went from a math PhD to running one [0:48] of the fastest-growing consumer finance apps to date. [0:51] Cheers. [0:51] Cheers. [0:54] Thanks for coming. Good to see you. [0:55] Thanks for having me, John. [0:57] Mm. Okay. [0:58] So, you grew up in Bulgaria, [1:00] moved to the US. You're a math nerd. [1:02] You studied math undergrad, you dropped out of a PhD? [1:05] I dropped out, yeah. [1:06] And then you started an HFT firm that [1:08] became-ish Robinhood over time. [1:11] Yeah, sort of. [1:12] I mean, that's putting it— [1:14] It was a relatively fast failure, the HFT firm. [1:17] But we learned a lot. [1:19] But the experience was useful. [1:20] We'll get to Robinhood, but first, I wanna know, [1:22] tell me about the hoarding of copper pots [1:24] in Bulgaria in the '90s. [1:26] When I was a child, [1:30] Bulgaria had the unfortunate distinction [1:33] of having the highest inflation rate in the world. [1:37] I think in 1997, we went through hyper-inflation. [1:40] What was the inflation rate? [1:42] I think it was something like 2,000% [1:44] Wow. if I remember correctly, [1:46] and maybe even higher than that. [1:48] There were a few echoes of high inflation [1:52] before the big moment. [1:55] Ultimately, Bulgaria ended up having a lot [1:57] of debts in foreign-denominated currency, [2:00] which creates a problem. [2:01] Then if your currency devalues even a little bit, [2:05] it creates this death spiral. [2:08] The Berlin Wall fell, [2:09] communism kind of collapsed in Eastern Europe. [2:12] And in 1991, [2:14] when I was still living there, I was four years old, [2:17] the inflation rate was like 150%. [2:19] Yes. [2:20] Then my grandfather was actually like, [2:23] "This country's not going in a good place." [2:25] He told my parents, "You should try to figure out [2:27] how to get out of here." [2:29] The good thing was, [2:31] after the Iron Curtain was lifted, my dad had an opportunity [2:35] to go to the University of Delaware for a master's. [2:38] We were able to eventually all move over there. [2:42] It was first my dad, then my mom, [2:43] and then I lived [2:44] with my grandparents back in Bulgaria for a bit. [2:47] But then they had enough resources to send for me. [2:50] About five years in, [2:53] we had this hyper-inflation bout. [2:56] So my grandparents' pensions, they had recently retired, [2:59] went to essentially zero. [3:03] I had a savings account in Bulgaria [3:05] that they had opened for me. [3:06] It was funny, I was in Washington a couple weeks ago [3:09] for the Invest America Child Investment Accounts Initiative. [3:14] I opened up an old briefcase that my dad had, [3:16] and it had the slip for my savings account, [3:19] which basically showed [3:20] when I was born they put [3:22] $2,000 equivalent in there. [3:25] And then they added money over time. [3:28] And then in 1997 it was down to like $10. [3:31] Yep. And then they [3:32] just stopped adding money. So pensions went to zero. [3:36] My grandfather at the time, he was a doctor [3:39] for the Bulgarian Maritime forces, kind of like the Navy. [3:45] He knew some people on port [3:49] that had copper cookware. [3:51] So he figured out— [3:53] he didn't have anything like Robinhood. [3:54] He didn't have any way to invest in stocks. [3:57] We had this closet in his apartment [4:00] in Varna, Bulgaria where you just opened the door [4:02] and copper cookware would fall out. [4:04] So, that was the store of wealth. [4:06] That was the store of value, yep. [4:08] This presumably must inform your views of crypto, [4:11] where there's a general trend in a bunch of emerging markets [4:15] with currencies that have historically seen hyperinflation, [4:18] where people want to keep their money not in the currency [4:21] because of just the experience you had [4:22] with the savings account. [4:24] It feels like that is a big part [4:25] of the early product-market fit around the world. [4:27] The surprising thing about what you said, to me, [4:31] is that Bitcoin was originally thought of [4:35] as the store of value [4:37] and this way to escape local hyperinflation. [4:41] But people don't seem to be using it [4:44] for that, or quite as much. [4:46] If you look at stablecoins. [4:48] Stablecoins have gotten a lot of traction overseas [4:51] because you have this [4:53] need to have US dollars. [4:56] Everyone just wants US dollars. [4:57] Yes. It has a very strong brand [5:00] as a safe, stable currency. Yeah. [5:02] Part of that might be an education thing, [5:04] but it's hard to counteract. [5:06] I think partly people in these countries [5:08] have some familiarity with US dollars already. [5:11] So you've less of a unit [5:12] of account confusion and things like that. [5:14] I think that's right. [5:16] I think crypto has become a really good mechanism [5:18] to store your wealth, but it's not necessarily the only one. [5:23] I think US stocks [5:25] and a diversified portfolio [5:27] of companies could also achieve the same goal. [5:30] The problem is it's lagged [5:34] because the infrastructure has been so bad [5:35] that it's actually quite difficult for you [5:37] to get access to US stocks if you live outside the US. [5:41] I was gonna ask about that. [5:43] There's a strong home country bias in investing. [5:47] Pieter Levels was here, [5:49] and he talked about his parents buying [5:51] a Royal Dutch Shell stock in the Netherlands. [5:54] My parents would've owned Irish bank stocks [5:57] in Ireland, and so, traditionally that's what people do. [5:59] A lot of markets have underperformed, [6:03] and maybe there's more desire for people to own [6:05] either the S&P 500, or the MSCI, Global Index. [6:09] Is part of Robinhood's strategy [6:12] letting people escape the home country bias in investing? [6:15] The plan is to give you access to every asset. [6:18] Eventually, over time, we're gonna integrate [6:21] with the London Stock Exchange, Hong Kong, [6:23] all the major markets. [6:24] But doesn't that mean you are escaping [6:25] the home country bias, because again, [6:27] like I would posit, the people invest too much [6:29] in their home country today. [6:30] Yeah, I think that, you know, [6:32] there's a conflicting interests here to some degree [6:35] because if you talk to the governments, they always want [6:37] to encourage investment in the home country. [6:39] Yes. [6:41] I think that's reasonable because, at the end of the day, [6:44] the citizens should benefit from that, is the idea. [6:48] But from the customer standpoint, [6:50] the best customer experience is just to get the best stuff. [6:53] If you talk to someone in Europe, [6:56] they also would like to invest in Tesla, Nvidia— [6:59] Totally. Yeah. [7:00] As I think about how Robinhood broke out [7:04] and became really popular. Clearly, [7:07] a mobile-first experience is a big part of it. [7:12] Again, Stripe was not the first payments company. [7:14] Robinhood was not the first online brokerage. [7:17] Even discount brokerage guys like E-Trade, [7:18] and stuff like that. Yeah. [7:19] And so, when you come into a market [7:21] where you're not the first, you need [7:25] a significantly better value proposition. [7:27] Having a really good mobile [7:28] experience was a big part of that. [7:30] What else allowed you to take so much market share versus [7:34] E-Trade and all those other guys? [7:36] The zero-commissions was a big thing. [7:38] I think that we did two things that were difficult [7:43] and differentiated simultaneously. [7:45] One was lowering the commissions to zero. [7:49] When we started, it cost $7.95 to $9.95 [7:53] to place a single trade, which I think was a problem [7:58] for two types of customers. [8:00] One is someone getting started with $100 bucks. [8:03] If $9.95 is your commission. [8:06] Once you get in and out of trade [8:09] 20% drag on your portfolio. Yes. [8:11] That's a problem. [8:12] Basically made it so that you couldn't get started [8:14] with under $2,000. [8:17] And by the way, most of them had account minimums [8:18] of $2,000 anyway. [8:20] So we got rid of that when we eliminated the commissions. [8:24] Then the second group of people [8:27] that it's a problem for is the active trader. [8:29] If you're trading somewhat actively, [8:32] you could be paying tens of thousands of dollars [8:34] or more a month in commission fees. [8:36] Yes. When we launched, [8:39] those types of customers [8:40] who are really poorly served by the existing offerings [8:44] just kind of flocked to it. [8:46] That was one thing, innovating on the pricing [8:48] and the structure, which led to a bunch [8:51] of other downstream things. [8:52] From the very beginning we were like, in order [8:54] to make this work, we have to make our costs as low [8:57] as possible. Which means we have to automate everything, [9:00] and we have to streamline all processes [9:04] from onboarding customers to customer service elements, [9:09] to processing payments, linking bank accounts. [9:13] It was actually quite innovative for many years [9:17] to do automated account approvals and brokerage. [9:19] We would interview people from other brokers. [9:24] When we launched, they all [9:25] wanted to come join us. [9:26] They were like, "How do you do it? [9:28] We've been trying to do this for decades." [9:30] So the tech [9:31] and computerization revolution had not fully hit brokerages [9:35] where they hadn't the level of automation [9:38] and technology, which would allow for a lower cost structure [9:42] and more scale that they should have had [9:44] if they were taking this transit this way. [9:45] Not at all. [9:45] These were all mainframe. [9:47] I think a lot of them are still running on mainframe. [9:50] They're like three technology revolutions behind. [9:54] Actually, that was our observation [9:56] from the first business, the high-frequency trading firm [9:59] that ultimately wasn't successful. [10:01] Our observation was, well you've got our software [10:07] and then a team of three engineers that can manage a book [10:11] that's trading billions of dollars a day. [10:13] So why do these brokerages have these gigantic [10:18] operations departments, [10:20] and why are they charging customers $9.95? [10:24] Part of it is certainly [10:27] the technology just wasn't there, [10:29] so it cost them much more to process each trade. [10:32] Part of it is also [10:33] that there wasn't any new entrant in the space [10:35] that was actually providing pricing pressure. [10:38] I think we were able to do that. [10:40] It's so obvious when you say it in hindsight. [10:42] You created sophisticated tech platform that allowed [10:47] for lower marginal cost [10:49] and passed that onto the consumer in the form [10:50] of lower fees. [10:52] Then you had a good mobile app and– [10:55] Yeah. That's probably 80% of it. [10:57] Mobile, I thought was big too. [11:01] Each of these things could have probably led [11:04] to a good decent-sized business [11:06] if we just had the best mobile app [11:08] and we charged $4.95 a trade [11:11] or if we just had zero commissions. [11:14] But it worked really well, [11:15] and the mobile app maybe was shit. I think [11:18] that probably also would've worked. [11:19] But two of them together led [11:22] to an outlier level of success. [11:24] The thing that allows you, [11:26] or at least allowed you in the beginning, [11:28] to do the zero-commission trades is the payment [11:33] for order flow. [11:33] I find the whole narrative around payment [11:35] for order flow crazy, [11:38] where you'll have politicians out there [11:41] on the pulpit railing against this front running [11:45] that's happening and how this is kind [11:47] of unfair to consumers. [11:49] Whereas, actually, [11:50] Matt Levine has written extensively about it, [11:53] it seems pretty clear that the reason [11:56] that payment for order flow exists is [11:57] the adverse selection problem. [11:59] Where when you are trading you care about [12:01] who your counterparty is, [12:03] and if your counterparty is Ken Griffin you're like, [12:05] "Ha, what does he know that I don't?" [12:07] Whereas for the kind [12:09] of consumer retail order flow you have, [12:12] people are very happy [12:14] to trade with those parties. [12:16] As a result, you can just get— [12:17] Well you're required [12:20] to get the best price. [12:21] But you can get a better price by selling the order flow [12:24] because there isn't that adverse selection problem versus [12:27] what is out there on the market, where you don't know, [12:29] is it someone who's super sophisticated [12:31] who knows something you don't. [12:32] I would say one thing. [12:33] I think the narrative got a little bit away from us. [12:37] The truth is, at the beginning, we didn't care about payment [12:40] for order flow, and we actually did not think [12:42] that it was gonna be a big source of revenue [12:45] because it was so small. [12:47] Actually, for the first several years, [12:50] we weren't self-clearing, and we were relying [12:52] on a third party to even route our orders. [12:55] So we cleared through Apex. [12:57] Apex had a vendor relationship with Instinet, [13:01] which routed our orders. [13:02] If you look at the early pitch decks for Robinhood, [13:06] we thought we would make money from API access [13:09] and a premium subscription offering [13:12] that was yet to be undefined and margin lending. [13:16] Then we saw [13:18] that we had all these active traders [13:20] and then payment for order flow became a larger [13:23] and larger revenue stream. [13:25] But that was really just because we got so much bigger, [13:28] so much faster on the trading side [13:30] that even this tiny revenue stream became [13:34] quite meaningful over time. [13:36] All of the competitors offered it [13:40] or took advantage of payment for order flow. [13:41] If you look at TD Ameritrade [13:43] before they moved to zero commissions, [13:45] their average revenue per trade was something like $12. [13:49] And it was like, or maybe $11, $10 of it was commissioned, [13:53] $1 was payment for order flow. [13:55] Yes. [13:56] Of course we deleted the $10 bit, [14:00] and when they had to match our pricing, [14:02] their stock went down 35% on a day, which was crazy. [14:07] They couldn't survive as an independent company. [14:10] They had to be merged into Schwab, [14:12] E-Trade got gobbled up by Morgan Stanley. [14:15] Then the narrative somehow became, "Oh, [14:19] they're making even more money from this other [14:21] payment for order flow thing than they would've made [14:24] if they were charging commissions." [14:26] That's my question on payment for order flow. [14:27] How did the debate get so wacky [14:29] and so disconnected from how the markets actually work? [14:31] I think the way it happened, [14:33] and this was a little bit [14:34] of a unfortunate timing on our part. [14:36] When we announced Robinhood, our initial waitlist in 2013, [14:41] there was this book that came out, "Flash Boys." [14:44] And it was basically coincident with the announcement [14:47] of Robinhood—the Michael Lewis book. [14:49] Sure, I've read it. [14:51] Which it was interesting [14:51] because as a high-frequency trader, [14:54] a former person in the space. Yes. [14:57] I read that book and I was like, [14:58] "Man, this is a sensationalized account." [15:02] But basically, if you think about— [15:04] he was setting this up as a good versus evil battle. [15:08] But the good people in this story were the big banks, [15:14] who were unable to keep up with this technology [15:16] that were getting picked off by the MIT kids [15:19] with all the new software. [15:21] The evil MIT kids. [15:22] I was like, "Man, this is a amazing piece of jujitsu, [15:26] how you can make the big legacy banks [15:28] that are making tens of billions of dollars [15:30] of revenue feel like the victim here." [15:32] Yes, yes. [15:33] He threw this payment [15:34] for order flow thing in there [15:36] that I think, somehow, made you think [15:39] that the victim was a retail consumer. [15:42] Because you have to tie it to make the story. [15:43] That kind of started the meme, okay, that's impress— [15:45] I think that poisoned the well a little bit [15:47] around this whole thing. [15:49] It made it something that we had to contend with, [15:52] even from the time we launched. [15:54] From the time we launched [15:56] you look at Reddit threads or Hacker News threads, [15:58] and it's like, oh, there's payment for order flow. [16:01] I just read "Flash Boys" by Michael Lewis. [16:04] You gotta get this book if you [16:06] really want to know what's going on. [16:12] So now the equities part of the business [16:16] and payment for order flow is one [16:18] of the smaller business lines [16:19] where options is bigger, interest income is bigger, [16:22] maybe crypto is bigger. I don't know. [16:23] And so you've created many other successful revenue lines. [16:26] Interest income is a significant part of the revenue stream. [16:29] Doesn't that make the business very interest-rate sensitive? [16:33] Maybe that's just like, yes, we're like a bank [16:35] in that way, and that's just a fact of life [16:37] for the business, but how do you think about the fact [16:38] that you have kind of now more volatility in the business? [16:42] I mean, I guess you had volatility previously [16:43] with trading volumes, but— [16:44] It's more volatility before. [16:46] If you zoom out, [16:48] in financial services, there's not [16:52] so many ways that you can make money. [16:54] You can make money on the transaction. [16:56] You can make money [16:57] on interest on the assets, whether it's cash, [17:01] and you're earning a small spread on that [17:03] or securities lending for securities that we hold. [17:08] Then for the industry, [17:11] that's pretty much been it. [17:13] All these things that we've discussed kind [17:15] of fall into one of those two buckets. [17:16] We also have subscription through Robinhood Gold, [17:20] which is not common in the brokerage industry. [17:23] I think since we've kind of rolled that out, [17:25] people have tried to replicate aspects of it. [17:28] In the beginning when we started [17:30] with subscription, everyone's like, "No, [17:32] in the brokerage industry it doesn't work. [17:33] Nobody does it." [17:35] I said, "Well, let's just see. Let's see if it works." [17:40] It's interesting to me how it feels like many [17:43] of the neo-financial services are moving [17:47] towards subscription models as a way [17:50] to segment the high-quality customers [17:51] essentially. Like Revolut has a bunch [17:53] of different subscription products, you guys do, [17:56] and it seems like maybe a good idea [17:57] because previously banks would not do that. [18:02] As a result, I think they didn't have a way [18:04] of like actually segmenting the product offerings [18:06] in this good way. [18:07] I don't know if it's segmentation. [18:09] We have one subscription tier. [18:13] And a lot of people— [18:15] I think that subscription is more of a- [18:19] there's a loyalty element to it, [18:22] which is a little bit more psychological. [18:24] People feel like they've invested [18:25] in this product, and they want to use it. [18:26] Exactly. So our general flow is someone comes [18:31] to Robinhood and usually they come [18:33] because they wanna buy stock [18:35] or they wanna buy some crypto. [18:37] There might be some other reasons, [18:38] but those are the simple dominant ones. [18:41] Then we tell them about the Robinhood Gold subscription, [18:43] we make it clear that that's basically a no-brainer. [18:47] You get eight things. [18:49] And they might not be interested in each [18:50] of those eight things, [18:52] but then they become a Robinhood Gold member. [18:54] Then I think the natural tendency is, [18:57] "I'm already a Robinhood Gold member, [18:59] as long as I'm peripherally aware [19:01] that Robinhood offers a product in the category that I need, [19:07] I'm gonna check that first [19:08] because I feel like I'm already invested." [19:10] People wanna get use from the bundle [19:12] that they have purchased. I wanna get use [19:14] from the bundle. [19:15] I wanna get a good deal. [19:17] If we can communicate that, [19:19] if you're a Robinhood Gold member, [19:21] you don't really have to think about anything. [19:24] You should just assume [19:25] that you're getting the best deal in the market. [19:29] Then I think you get Amazon Prime, Costco effect [19:31] where that just becomes your go-to place [19:33] for financial services. [19:35] You have investing. You have now the card [19:38] with Robinhood Gold and the 3% cash back. [19:41] Do you at some point [19:43] essentially become people's primary bank account [19:47] or primary financial account, [19:48] however you wanna call it, [19:49] where they get their paycheck deposited there [19:52] and then they have a card that they can spend that money on [19:54] or they can go invest it? [19:56] It seems like you have a lot of the ingredients already [19:59] to become people's—consumers' primary financial home, [20:02] which again, I would characterize as the place [20:03] that they're getting their paycheck deposited. [20:05] Yeah, we've been thinking about this for a while, [20:08] and we're launching Robinhood Banking, [20:11] in the coming months. [20:12] Oh, that's cool. We announced [20:13] that at our Gold Event a couple of months ago [20:16] with a few innovative features including cash delivery, [20:19] which I'm very excited about. [20:20] The armored truck rolls up to your house. [20:22] Yeah. Or the small Tesla model. [20:26] I want an armored truck. [20:27] Yeah, that's better aesthetics. [20:29] We would have to do the armored truck if you want a delivery [20:32] of like $250,000. Exactly. [20:34] Yeah, it's very expensive for the $100 [20:36] to $200 average ticket charge for a ATM transaction. [20:40] Okay, so you are doing consumer banking, [20:43] and the features that you think are important [20:45] for that are—cash delivery's one, what else? [20:48] I think having a good, high APY. [20:51] Also there's an opportunity to make it really easy [20:55] to switch from one bank to the other. [20:58] Part of the reason it's hard to make it [21:02] as a bank that doesn't cater to the low end is you have [21:06] to move all of your stuff from your existing bank [21:09] to the new one. And you've got a lot of bills. [21:12] You have direct deposit from your paycheck. [21:15] If we can make that process frictionless, [21:19] if you can just push a button, [21:20] and we take care of everything— [21:22] I think that would be really powerful. [21:23] Sorry, how do you press that button [21:25] and take care of everything? [21:27] Have you built migration software from other banks? [21:30] What does the button do? [21:31] It's not out yet, [21:35] but what it will do initially is, it's a lot [21:38] of custom migration software [21:40] that links into your payroll providers [21:42] and things like that. [21:44] But over time, as these AI models get better [21:46] and better, you should think of it [21:48] as like your family office CFO. [21:52] You don't have to do all [21:55] of the migrations for you. [21:57] You just call this person and then they [21:57] deal with it. [22:00] I think there's no reason we can't deliver [22:02] that experience within the next one to two years. [22:04] Do you have any funny stories [22:06] from the GameStop saga? [22:08] You must have had some just surreal moments in there [22:11] that make for a good memory now— [22:14] type 2 fun. [22:15] I guess to come up [22:17] with something maybe a little bit more interesting [22:21] that hasn't been reported on too widely before. [22:24] Most people [22:25] that follow know about the capital requirements [22:28] and the regulatory stuff. [22:30] I think one of the interesting things was [22:32] back in 2020 when COVID started [22:36] and there was a huge influx of retail investing- [22:41] one of the things that differentiated Robinhood was [22:44] that we were giving free stocks to people, [22:46] and we still do this. [22:47] But it was a great way to get started. [22:49] You get your free stock. [22:50] If you refer someone, it was bidirectional [22:54] and you get one share. Now, one of the stocks [22:57] that was actually very popular at that time was GameStop. [23:01] So if you joined in 2020, [23:05] there was a good chance that [23:08] GameStop was the stock that welcomed you to Robinhood. [23:12] We had millions of customers joining in 2020. [23:17] In hindsight, this whole movement— [23:19] and a lot of people are still angry at me, as I could tell, [23:23] every time I talk to another journalist [23:25] or anchor from a channel [23:28] that I haven't been on a bunch of times, [23:31] the first reaction I have is, [23:32] "Whoa, you've got a lot of fans [23:34] that are all about GameStop. [23:37] Maybe I should spend more time with you." [23:39] To some extent that, [23:40] that whole thing wouldn't have happened without us. [23:42] We made possible, this whole thing. [23:45] As in you feel a lot of long pressure on the GameStop stock. [23:50] I wouldn't even put it on the technical market aspects, [23:53] but I kind of feel like these people's father in a sense. [23:55] Just psychologically, [23:59] I feel like they grew up with our product— [24:01] You gave a lot of people their first GameStop share. [24:03] I think it's time for—some of them are angry, [24:05] they've left home, they've gone out in the real world. [24:08] We've grown too, so I think it's kind of time [24:10] for them to just come home. [24:11] I think it's enough rebelling. [24:15] Were there any moments during [24:16] that saga where you're like, [24:18] "I can't believe I'm having this call right now. [24:20] I can't believe I'm in this position." [24:23] The whole thing was also very weird [24:25] because it was all remote. [24:26] And at that time it was sort of like one year into COVID. [24:31] I, at least, was itching [24:32] for some real human contact. [24:35] I think it was maybe two days [24:36] after the GameStop—the trading restrictions. [24:41] It was a Saturday or Sunday. [24:44] One of the really cool things was, [24:46] my investors, my friends, they were calling me. [24:49] They were offering help. [24:50] A bunch of people that probably wouldn't have given me [24:53] the time of day were offering to get on calls [24:56] and help me navigate this whole situation. [24:59] Marc Benioff was very nice. [25:03] Mark Zuckerberg. I never talked to Mark Zuckerberg, [25:05] but he's like, "Hey, maybe I can help you with this thing. [25:08] I've had my share of PR issues." [25:09] We workshopped some stuff. [25:13] So these were people reaching out to be supportive? [25:14] Yeah. Elon Musk connected with me too. [25:18] That's really nice. [25:18] I was always a big Elon Musk fan, early Tesla adopter. [25:23] I think I got connected to Antonio Gracias [25:27] through one of my investors. [25:29] He was like, "Well, I can help you with this. [25:31] You're probably gonna get asked to testify [25:34] before Congress. [25:36] You should try to see if you can get out of that." [25:38] Ultimately, I didn't end up getting out of it. [25:40] That was a great conversation. He was very helpful. [25:43] Then he called me back 10 minutes later, [25:45] and he's like, "I've got Elon here. [25:48] He wants to talk to you." [25:51] And Elon was very intense. [25:52] He's like, "You're being an asshole to the people. [25:57] What you've gotta do is full transparency, [26:01] 100% put it all out there, play-by-play, [26:04] full transparency." [26:06] Then he was doing this Clubhouse, [26:08] I don't know if you remember it. [26:09] It was his— [26:10] Yeah, The Vlad the Impaler Clubhouse. [26:12] Yes. So that was a scheduled thing [26:15] with Sriram and Aarthi on that "Good Time Show." [26:19] They were blowing up— [26:20] Back when we had scheduled Clubhouses, [26:22] back in that period of the pandemic. [26:24] Everyone was very, very excited about this. [26:27] And so I text him after my calls, like, [26:28] "I've got this crazy idea. [26:30] How about I just go on the Clubhouse [26:31] and we lay it all on the line?" [26:34] That was a spur-of-the-moment decision to hop on [26:37] that 1:00 AM clubhouse. [26:39] But I think it ended up being quite good. [26:41] As we talk about this kinda GameStop saga, [26:45] professional investors say that the rise of retail [26:50] and their increased relevance has really changed [26:54] the investing business. [26:55] What's your view on that, or do you have any kind [26:57] of perspective from your seat? [26:58] It's just interesting to me as you hear people [26:59] who run hedge funds or, you got people [27:01] who are professional investors, that [27:03] it's different versus 5 or 10 years ago. [27:05] One of the values [27:07] that we had on the Robinhood website [27:10] from the very beginning was the importance [27:14] of individual direct participation in the markets. [27:18] We had this feeling [27:19] that everything was getting a little bit abstracted. [27:21] You're not investing in a company, [27:25] you're investing in some retirement fund or pension fund, [27:28] and then that's being handled by [27:30] an asset manager. [27:32] Maybe they're putting it into an ETF. [27:34] So you're probably four steps removed [27:37] from actually investing in the real company. [27:40] I thought that was kind [27:41] of a degenerate case when you think about it, [27:43] because you have all sorts [27:45] of weird things happening. [27:47] Like something happens with tariffs, [27:51] and now suddenly, a stock like Palantir [27:55] or something that's completely unrelated [27:56] to tariff action is taking a huge hit. [28:00] And you can understand it when you go [28:02] through this indirect thing. [28:04] Someone's making a macro decision, [28:06] they're rebalancing their portfolio [28:07] and these individual companies that maybe have nothing to do [28:11] with it are just collateral damage. [28:14] I think as we were going to this more abstracted world, [28:19] the signal of, "I'm buying this stock [28:21] because I believe in this company [28:22] and I think their future prospects are very good," [28:25] was getting more and more diluted. [28:27] I think retail provides a counterbalance to that. [28:29] That's why recently with the whole tariff saga, [28:33] retail ended up doing quite well because— [28:36] Sometimes this is criticism, sometimes it's kind [28:40] of praised, but retail's known as dip buyers. [28:43] Because they're like, "Oh, well, [28:44] this stock is at a discount. [28:45] I believe in it, I'm gonna buy it." [28:48] I think they end up looking quite smart, right? [28:50] Because they take advantage of this like macro headwind [28:53] that ends up disproportionately impacting stocks [28:56] that really don't get harmed from it. [28:59] Then I think when like that stuff washes out, [29:04] they've ended up benefiting disproportionately. [29:06] So I'm a fan of it. [29:08] I think more direct ownership is [29:10] generally healthier and less stacks [29:13] and stacks of misdirection. [29:15] Do Robinhood users vote their [29:15] stock ownership in the proxies? [29:20] They do. I mean, [29:22] they have the ability to. [29:24] Because that seems like a good trend [29:26] where it's been much decried, the fact [29:29] that everything's in index right now, and then you have ISS, [29:31] and like the shareholder services companies [29:35] where you have like a very small number [29:37] of random firms making recommendations [29:39] on how mutual fund managers should vote [29:41] versus people directly making— [29:43] The voting thing. [29:44] I think there's a lot of people, including us, that are keen [29:48] to figure out how to make that a bigger thing. [29:50] I think there's two problems with it. [29:53] One is that most proxy votes are just not super interesting. [29:58] If you look at it, it's sort [30:00] of incomprehensible things [30:02] that don't really get retail very excited. [30:06] If there's some kind of like proxy war [30:09] you're voting the CEO out— [30:10] But it's usually not that. [30:11] It's usually something totally— [30:12] It's usually not that. [30:14] Maybe you'll get like one per year. [30:16] Or one every two years. [30:17] The other thing is the sort of decision [30:21] to vote on something versus just selling the stock. [30:25] As retail, you have the option [30:27] where you can just sell the stock. [30:29] In some ways, that can be an easier outlet [30:31] for the feeling that you're not happy with the company. [30:33] Whereas if you're [30:35] a fund manager, [30:37] you're in that stock because you don't have a choice. [30:39] You're subject to some algorithm that says you have [30:42] to buy it, and then you end up accumulating [30:44] so many of the shares, [30:46] but you can't just sell easily. [30:47] That's why those guys are particularly motivated to vote [30:51] because that's actually their only outlet [30:53] to expressing a point of view about the company. [30:56] In order for voting to become bigger, we have [30:58] to solve both of those problems. [31:00] It might be possible if you make it more delightful [31:04] and if you incentivize it to solve the first one. [31:07] But you're competing with users [31:08] being able to sell the stock, [31:12] which is a powerful alternative. [31:14] Should I go first? [31:16] Yeah. [31:19] So, 15, 27. [31:21] 12, 27, Okay. [31:23] If we think about who owns companies in America, [31:26] there's insider ownership, there's ETFs, [31:29] there's direct retail holdings, [31:30] there's mutual funds, and hedge funds— [31:32] maybe you merge those, [31:34] lump those last two together, [31:35] but those are kind of the way in which stocks are held. [31:38] Where is Robinhood taking share from? [31:40] Are people taking dollars [31:42] that they would've invested in a mutual fund [31:44] and investing is instead directly in companies. [31:47] Is it new dollars coming into the stock market? [31:49] I'm just curious how I should think about that mix. [31:52] It's been a while [31:54] since we've done this analysis. [31:55] I think the last time we did it, we found [31:57] that it was mostly new dollars coming in. [31:59] That money that customers are investing would've [32:04] very likely been spent or consumed in some way. [32:07] And then maybe at some point [32:09] someone's investing it, [32:10] but not that retail investor, which is why [32:15] the whole gamification narrative kind [32:17] of rubbed me the wrong way from the beginning [32:19] because it's like, "Oh, are you making it too easy?" [32:22] For a while it's like, [32:23] that money would've just been spent. [32:25] That would've been [32:26] spent on entertainment. [32:28] It would've just been- [32:29] people were buying trinkets on Amazon. [32:32] We're taking money from the spending bucket [32:36] and moving it into investing. [32:38] Now, of course, over time, as we do more and more, [32:41] we're helping you with your spending. [32:43] We're helping you with your retirement. [32:44] So it's becoming, like you said, [32:46] more a comprehensive financial services company. [32:48] But at least at the beginning [32:50] that was always like my irk with the gamification thing. [32:54] Presumably part of what triggers [32:58] those narratives is, it feels like every few years [33:00] there's like a Robinhood headline [33:02] around zero-day options [33:03] or Super Bowl contracts or whatever, [33:06] or the stuff that's on kind of the more speculative end [33:08] of the investment market, [33:10] and that's what ends up causing these narratives? [33:13] I think that's true, [33:16] you're probably right. [33:17] Although I think the narrative originated [33:19] well before we had any of these things. [33:22] I think it actually originated [33:25] right when our zero-commission model became universal. [33:31] It was end of 2019. [33:34] Everyone dropped commissions to zero. [33:37] And then I think before then, the answer [33:39] for why Robinhood was successful was pretty simple. [33:42] It's like, "Oh, they're not charging commissions. [33:45] Everyone else is." [33:46] Then the industry matched us, [33:48] In 2020, we grew our market share [33:52] at an accelerated pace, [33:53] and the business itself grew 3-5x. [33:56] Then the story for why Robinhood is successful became [34:00] a little bit more complicated. [34:01] It's no longer the commissions. [34:03] Maybe it's a better product, [34:05] but saying it's a better product is not as salacious [34:08] as saying, "Oh, this gamification, they gamify [34:11] and they," it makes it much more nefarious. [34:14] So I think the media kind of ran with that narrative [34:16] because it was more effective. [34:19] So you wanna tokenize private companies. [34:21] I have a lot of questions. [34:22] Yeah, let's talk about that. [34:26] First off, how do you actually do it? [34:30] Because aren't you kind of nakedly short the position [34:34] where you owe the customer the appreciation [34:38] of the underlying company, [34:39] but you can't guarantee that you're gonna be able [34:41] to get your hands on the stock [34:42] because it's a private company. [34:44] How do you actually mechanically manage that? [34:46] First, before we tokenize it, we have [34:49] to make sure we have the underlying stock or exposure [34:53] if there is no stock, whatever the equivalent is [34:56] that we're giving the customer. [34:57] In all cases thus far, which is [35:02] the tokenized public stock where it's easy, [35:05] and also the SpaceX and OpenAI giveaway, [35:09] that's backed by holdings that Robinhood- [35:13] has on the balance sheet. [35:14] But doesn't that mean [35:15] that you're then limited in the rate [35:17] in which people can come into these positions? [35:19] Because if you get a billion dollars [35:21] of interest floating in for OpenAI, you have [35:24] to go acquire a billion dollars of OpenAI stock, [35:27] which you might not be able to do at the drop of a hat. [35:30] There's a similar product [35:32] that we have called IPO Access. [35:35] The way IPO access works is [35:37] kind of a similar situation. [35:39] Retail expresses interest. [35:42] We collect it. [35:44] Typically it's massively oversubscribed. [35:48] And then the company hands us out. [35:52] Most of the time they don't give us very much. [35:55] Other times, if they like us, if they like the idea [35:57] of retail, they give us a little bit more. [36:00] We've had some IPOs [36:02] that customers have wanted over the years. [36:06] I think it's a good offering for us. [36:08] But generally speaking, [36:10] we don't fill the retail demand. [36:11] It's massively oversubscribed. [36:13] I see. So there are cutbacks. [36:14] And so you have the concept of, "There's only [36:16] so much allocation that we can grant here." [36:18] But we are thinking about this. [36:20] And there's a way to... [36:20] It's kind of a chicken and an egg problem. [36:24] Do we collect the retail demand first? [36:26] Do we get the supply? [36:30] And I think one of our strengths [36:31] is we have a large balance sheet, [36:32] so we have some flexibility here. [36:34] You can act as an underwriter. [36:36] Or a venture fund in the private sense. [36:39] So we're thinking a lot about how to do this. [36:42] You know that I'd love to tokenize [36:44] or otherwise make your stock available to retail. [36:48] And you know that I'm skeptical. [36:49] And I guess, I've a number of skepticisms, [36:51] but one of them is we have decided at a policy level, [36:57] as society, [36:59] that there's a certain set of disclosures [37:02] and regulations that companies selling stock [37:05] to the public should be subject to. [37:07] They need to release their [37:08] quarterly financials in this way, [37:10] and they need to make these adaptations. [37:11] They need to run in this certain way, and all these things. [37:13] Don't you end up just end running around that [37:18] and having regulatory arbitrage where private companies get [37:22] to act as public company [37:24] or get to sell stock to the public, [37:25] which is not something that we want to have happen? [37:29] I disagree across a few axes. [37:34] I think these rules, [37:36] these public company disclosure rules, which were created— [37:39] Everyone agrees that we should improve [37:42] But I think that's very, very hard. [37:46] Once a process gets to a certain level [37:49] of sophistication, you basically have to [37:54] sort of rip it up and make it efficient again. [37:57] And I think there's two things working [38:01] against companies going public more. [38:05] One of them is the process [38:07] and the costs of going public have started outweigh [38:11] the benefits in a lot of cases. [38:13] I'm sure you guys probably experienced that directly. [38:15] And I'm one of the cases [38:16] where I think the benefits outweigh the cost, actually. [38:19] Because my entire business— [38:20] Exactly. You are a public stock trading company. [38:23] But for you guys—for many companies, [38:25] it's probably a little bit more [38:26] of a nuisance, at least now. [38:29] The other thing is, it's just easier [38:31] to raise private capital. [38:32] You can get infinite private capital, [38:35] and the only thing that would maybe compel you [38:38] to go public is restrictions on a number of shareholders. [38:41] Things like option grants, [38:43] kind of these minor technical things. [38:45] You look at retail. [38:46] So you gave like a retail-protection argument, [38:49] which is that you don't want retail investors getting [38:54] into things where they don't have [38:55] the full disclosures and information. [38:58] What I'd tell you is that retail investors [39:02] have the unrestricted ability to YOLO into memecoins. [39:06] It's kind of a silly juxtaposition [39:09] to say, "Any meme coin is okay, [39:12] open season, there's hundreds of thousands of them, [39:14] put however much you want into this. [39:18] But SpaceX and open AI and Stripe are too risky." [39:24] There are problems with disclosure. [39:29] And I think the origin of these accredited investor rules, [39:32] the thinking behind it is, "Okay, [39:34] if you have a million dollars, you can hire someone... [39:37] You can hire a trained professional [39:39] to think about this for you," [39:41] which confers an element of protection. [39:44] But now you have AI, you have all these tools [39:46] that give intelligence and data basically for free to you, [39:50] so they do need to rethink. [39:53] And I think we should figure out [39:54] how to safely give retail investors access to these names, [39:59] which are some of the most innovative companies. [40:02] I think we can handle the [40:04] disclosure elements fairly neatly actually. [40:07] How far do you want to go with this? [40:09] Should Chick-fil-A be tokenized? [40:13] Should a family farm in California be tokenized? [40:16] What is this universe [40:18] of companies that should be tokenized? [40:19] There's two things that are interesting. [40:23] One is late-stage privates, [40:26] and I think the customer base is a little bit different [40:28] because late-stage privates is more like employees [40:31] that are looking for secondary. [40:33] You, as the founder, are probably not the person [40:37] that's championing for that [40:39] unless you really like retail access. [40:41] Some people believe in that very, very strongly [40:45] and they're excited by it. [40:47] That's kind of the late-stage private opportunity. [40:51] Then there's early stage, [40:52] which is almost like capital as a service. [40:56] There's some talk about on-chain issuance, [40:58] and, "Oh, this will never work without on-chain issuance." [41:01] I think that's all bullshit. [41:02] Nobody gives a shit about on-chain issuance. [41:05] They just want a button to deliver money [41:08] so that they can raise for their venture [41:10] as easy as possible. [41:12] If blockchain technology is the easiest [41:15] and most direct way to do that, then it'll win. [41:19] But nobody's buying on-chain issuance [41:23] at the earlier stages. [41:24] They're just worried about growing their business. [41:26] So you would like to see this as a tool [41:29] for capital formation, where startups are able [41:32] to raise money in this way. [41:34] Yeah, I think that's the end state. [41:35] We have seen countries play with frameworks for this, [41:37] like equity crowdfunding in the UK, [41:39] some frameworks around that. But there are— [41:43] It's been very effective in some cases. [41:44] It's still a little clunky [41:46] and it's limited to that market. [41:49] But Revolut, for example, did [41:52] multiple equity crowdfunding rounds [41:54] in the earliest stages, [41:56] which ended up working quite well for those. [42:00] Let's talk about Stripe. [42:01] Have you thought about whether there would be some mechanism [42:06] where we could either tokenize your stocks [42:09] or make it available to retail? [42:12] I can present a couple of options to you. [42:14] A menu of options. [42:17] I have a menu of options. [42:18] This is my podcast. [42:19] One is, obviously we can tokenize it directly. [42:23] The other is, you can put it in a diversified fund, [42:27] a little bit less exciting. [42:28] So kind of less price action essentially? [42:31] Yeah, but price action would be obscured [42:34] because there would be five companies, let's say. [42:37] Yes. [42:40] I think— [42:43] You seem excited. [42:44] Exactly, you can tell I'm incredibly excited. [42:47] There's multiple things. [42:48] One, again, we do not sell equity [42:53] to the general public today. [42:55] Our view, just as a practical legal matter, [42:59] is that the SEC takes a dim view of that. [43:05] And so, that's one set of obstacles. [43:09] But then, the deeper one is just that Stripe is [43:14] not a capital-consumptive business. [43:18] We're not out there raising many billions of dollars [43:22] to build out, presumably more data centers. [43:25] That million-GPU data center, right? [43:28] Exactly. And so, there are these [43:29] highly capital-intensive endeavors [43:33] that, as a result, it probably does make sense to go out [43:35] and raise from kind of the deepest pools of capital. [43:37] Stripe is not that. It is a pretty capital-light business. [43:42] And so the current set of owners, if you leave aside [43:45] early investors, are employees. [43:48] And we kind of like that model. [43:49] You know, we like the idea that it's owned by the set [43:52] of people who are building it and creating the value. [43:54] And then we've solved the employee liquidity question [43:58] through tenders. And so, I think it would be [44:01] a solution looking for a problem. [44:03] But the problem is [44:05] normal people can't get access to it, right? [44:08] So what if the tenders, instead of being purchased [44:11] by institutional venture capitalists, could be [44:15] a little bit more democratized? [44:17] I'm never that sympathetic to that argument. [44:21] The main reason is just, if you look [44:23] at the underlying ownership, [44:25] it's normal people through a different lens. [44:27] Who are the LPs in a company like Stripe? [44:32] It's generally college endowments. [44:34] It's CalPERS, [44:36] it's pretty broad ownership underneath. [44:38] It's not, when you're saying being owned [44:41] in a diversified vehicle, [44:43] that's essentially what Stripe is today, [44:45] where the underlying owners that are not people who kind [44:48] of work at the company or have worked at the company- [44:50] it's generally funds. [44:52] And those funds are generally people's money managed [44:55] in some way. We also have a few kind [44:57] of mutual fund owners or something like that. [44:59] It actually strikes me [45:01] as a pretty sensible ownership structure. [45:02] The other thing that we're kind [45:03] of getting into in these discussions, [45:06] and kind of when you announced the OpenAI tokenization, [45:09] it kind of struck me as a prediction market on the future [45:13] of a specific company. [45:14] I'm curious, you're getting now more [45:16] into prediction markets. [45:18] I'm curious what the world looks like [45:21] where prediction markets become much more relevant [45:24] because that's one of the biggest changes [45:25] of the last five years. [45:26] Yeah, I think prediction markets are awesome. [45:30] It is fundamentally a little bit different than this. [45:32] We'll get into the prediction markets thing. [45:35] But one of the criticisms that [45:38] maybe the tokenized stocks [45:40] or the stock tokens offering has had is, [45:41] while these are just counterparty [45:42] swaps with Robinhood, [45:48] but it's not exactly that [45:50] because they are backed by the underlying asset. [45:53] In a sense, if you think about stablecoin, [45:58] it's the same mechanism. [46:00] There's different paperwork surrounding that, [46:03] which we can debate— [46:04] You have ownership [46:05] of an underlying pot of assets. [46:06] But, you've got a bucket [46:08] that you put some stuff in, [46:10] and then you put tokens against that. [46:12] And that becomes a tradable instrument on blockchains. [46:15] I actually don't think there's any criticism [46:18] that you can apply to stock tokens [46:20] that doesn't apply to stablecoins. [46:22] Stablecoins have sort of crossed [46:25] the Overton window, where nobody's really talking [46:27] about these things. [46:28] I think effectively, [46:31] from a practical process standpoint, they're identical. [46:35] Prediction markets, to your point, [46:36] are like derivatives in the ultimate sense. [46:42] And I think what's exciting about prediction markets is [46:46] not just the fact that customers can trade them, [46:50] although that is great for us [46:51] and it's been a growing business. [46:53] But also that there's a use case [46:56] that's mass market outside of trading. [46:58] You can just look at it, [46:59] and it becomes an alternate source of news. [47:04] That's why I resonate with this idea [47:06] that prediction markets are truth machines. [47:09] Where do you use prediction markets outside [47:11] of elections to get news? [47:13] I was at Wimbledon a couple weeks ago, [47:15] and we had a great Wimbledon prediction market. [47:20] There's a Bulgarian tennis player [47:23] who I happen to know a little bit, Grigor Dimitrov. [47:24] I don't know if you saw this match. [47:28] He was playing Jannik Sinner who ended up winning. [47:32] Now, Grigor Dimitrov was up two sets to zero, [47:38] right, and he was actually winning the third set. [47:40] So he was going to clean the match up. [47:45] And then, you look at the prediction markets, [47:46] I think the time, it was like 55:45. [47:50] So it was very, very close. [47:53] Anyone who was a casual observer— [47:55] Just watching the score. [47:56] Would be like, "it's gotta be 95:5. [47:59] Nobody ever comes back from this." [48:02] But it was pretty much even. [48:05] And I think that's interesting. [48:07] I think that the score itself doesn't tell you [48:10] the full information. [48:11] There's all sorts of information like injury history, [48:14] prior experience, that if you [48:17] tie it into one number becomes a useful prediction. [48:21] I remember watching the Jake Paul-Mike Tyson fight [48:24] as well. If you're watching the fight— [48:28] I was watching it on Netflix at the time. [48:31] It was the third round, [48:33] and they were like, "Oh, Mike Tyson could win this [48:36] with one punch. [48:38] It's very, very close." [48:39] They were trying to keep you entertained, [48:41] so you keep watching it. [48:42] You look at the prediction market, [48:44] I think I was looking at Polymarket at the time, [48:46] it was like 92% Jake Paul. [48:49] So I think it's broadly applicable. [48:52] I think you can also look at it for these one-time events. [48:55] The Pope one was very popular. [48:59] There's a lot of AI-prediction markets, [49:01] which I'm personally very keen on. [49:03] Like when is AI gonna achieve different milestones? [49:07] Sometimes, at least for me, [49:09] I just wanna know what the answer is. [49:11] What the most likely answer [49:13] to a very specific question. [49:15] You just can't get that from the news, [49:17] because the news has become a little bit corrupted [49:20] because the model is to keep you engaged [49:22] and entertained, which is contrary [49:24] to giving you the immediate answer [49:27] to the question that you're looking for. [49:29] What would you change if you were running the SEC? [49:32] I think Paul Atkins is great. [49:35] I have a lot of confidence that he's gonna do [49:39] all of the things that I think are important. [49:42] In general, I've been pretty impressed [49:44] by the efforts coming out of the administration. [49:48] I know it's difficult to say something like that, [49:51] but when you compare that [49:53] to the previous administration, [49:56] I think they're focused on the right things. [49:57] Are you thinking about the Invest America Initiative? [50:00] I think the Invest America Initiative, [50:02] that originated in private sector with Brad Gerstner. [50:06] The administration kind of took that on, [50:09] and I'm hugely bullish. [50:11] But I don't think it's just that. [50:12] I think the AI plan that came out today was very good. [50:17] I know you and I are in the same chat groups [50:20] that discussed that, but like to think [50:22] that the US is like pushing open source... [50:26] It's just a very, very good thing. [50:28] On the SEC side, I think private markets [50:32] and tokenization are the biggest things that I care about. [50:37] I think they're keen...I mean, the past couple [50:41] of weeks there's been a lot of talk. [50:44] The SEC is sort of like very receptive [50:47] to trying to make that work. [50:50] So I'm feeling good about it. [50:51] But if we end the year, [50:54] or even if it happens next year, [50:55] with comprehensive crypto asset security [50:58] tokenization framework [51:00] and some solution that's workable, [51:01] better than what we have today, [51:02] about private company access. [51:08] I think a full redo of accreditation, frankly, [51:11] is what's needed. [51:12] It seems like everyone agrees [51:14] that the current accredited investor regime is not [51:17] that sensible. [51:18] How does Robinhood ship so fast? [51:21] I think that it is just a matter [51:24] of where we're focused. [51:27] Let me think what the useful thing [51:29] to say is without spilling all of my secrets [51:32] to my competitors. [51:33] No, please spill the secrets. [51:35] If you let me tokenize your shares, I will. [51:39] I think you have to want to ship fast. [51:41] A lot of people say that they want things, [51:47] but then, when they actually break it down [51:49] into what that entails, [51:51] they're unwilling to do those things. [51:53] You have to be confronted with the harsh reality [51:56] that, sometimes the things that are needed [52:00] to ship fast are not comfortable things. [52:03] What are the uncomfortable things [52:05] people are unwilling to do? [52:07] You have to create culture [52:11] with the right talent. [52:13] You have to hold yourself to a really high performance bar. [52:18] You have to have the right environment for people, [52:20] which I think, in our case, [52:22] we emphasize the in-person experience. [52:26] We found that that makes a big difference. [52:28] We've made mistakes across all of these vectors [52:32] that we've then had to revert, which is the benefit [52:36] of doing this for a while. [52:38] You get to kind of try a bunch of things [52:39] and quickly change. [52:41] We announced we were a remote-first company in 2022, [52:44] and then I regretted that pretty much immediately. [52:46] I was like, "Oh, gosh. [52:48] That was the wrong decision." [52:50] Everyone said it was a one-way door, [52:52] but it turns out it's a two-way door. [52:53] You can reverse pretty much anything. [52:55] Are you guys fully in-office now? [52:58] We emphasize in-office. [53:00] It's obviously not 100%. [53:01] But it's the main way that you work. [53:04] Yeah. We do this thing where if you're a senior leader, [53:08] an executive, you're five days; if you're a manager, [53:11] four days; if you're in IC, three days. [53:12] I like that. [53:13] And I think It's good because if [53:14] you're an individual contributor [53:18] and you're doing work, it's very nice to know [53:20] that your manager is going through more pain than you. [53:22] Yes. [53:23] That's also a general principle [53:26] for shipping fast, or for motivating. [53:28] It has to start at the top. [53:30] You have to be willing to do yourself, [53:31] to an even higher degree, [53:37] what you ask of everyone else. [53:39] What we do is, we very much concentrate [53:40] rewards with our top performers. [53:46] If you're one of the best people in the company, [53:49] you get rewarded very well. [53:51] Promotions are zeroed in on those people. [53:56] I have a founder's community, too, [53:58] which isn't just the people that report to me. [54:00] I think that's the sort [54:02] of degenerate case where you end up creating a group [54:05] of like the most important people of the company, [54:07] and it ends up being the people at the top [54:09] of the— [54:10] Org chart. [54:11] The org chart. [54:12] We have to have systems that fight against that. [54:15] How does this founder's community work? [54:16] Basically, if you're in the top 150 people [54:20] in the company by impact across all levels, [54:23] it's a real community, so it started out as compensations. [54:25] Is this modeled after the Apple top 100? [54:27] They do this event that's the top 100 people do [54:30] some offsite, I think, once a year, [54:32] and again, by impact and across all levels. [54:35] And I think gets a lot of noses out of joint. [54:38] I have heard about that. [54:39] But the idea is [54:40] that it started out as [54:41] being compensation-centered. [54:46] So it was like, "Okay, if you're the best people, [54:48] we have discretionary budget that we allocate towards that. [54:52] But over time, it's grown into an actual community [54:55] where, if I go through the updated strategy [55:00] or vision, we get those folks together, [55:02] we get their feedback. [55:03] We try to incorporate. We do events. [55:05] We have events in each city, [55:07] and we get them together for dinners. [55:09] Oh, that's cool. [55:11] Is it a way to make Robinhood, [55:13] as a larger company still feel like a smaller company? [55:15] When I induct people into this community, [55:20] I usually give a spiel, [55:22] which is basically. "If there was a disaster or some kind [55:27] of a apocalyptic scenario, [55:29] and we had to rebuild Robinhood with 150 people, [55:33] you would be in that group." [55:35] That's how I think about it. [55:37] I think that you always have to be aware [55:40] of who your best people are. [55:42] You described in 2021 growing too fast [55:47] and then doing the visualization exercise [55:49] of, "what would Frank Slootman do [55:50] if he took over my company?" [55:52] Oh, yes. [55:53] What did you do keeping Frank Slootman in mind? [55:56] That was a funny exercise. [55:58] We'd just had, it was either a bad earnings [56:01] or a bad board meeting. [56:02] One of the ones where you just feel like got punched [56:06] from all directions. [56:07] So I sat with my team and I was like, [56:09] "Well if they brought in..." [56:10] it was kind of a morbid question, [56:12] but "If they brought in a CEO [56:15] who really cleaned this place up, what would they do?" [56:18] And there were a few observations. [56:20] I think the first thing that came to mind is, [56:22] Frank Slootman would be like, [56:23] "What is this remote work nonsense? [56:25] Let's get people in the office and working." [56:27] It was a good exercise to figure out. [56:30] If you made a decision, [56:34] it's hard to reverse that decision. [56:36] I think there's a natural sort of reluctance [56:39] to not feel like a flip-flopper. [56:41] Totally, you don't need to feel a prisoner [56:43] to your previous decisions. [56:45] That's one of the things [56:46] that's actually a benefit to swapping people. [56:50] Yes, yes, totally. [56:51] If I take myself out of it, [56:53] and I'm like, "All right, I'm a independent board [56:56] of directors here making a decision about the company," [56:58] one of the benefits of swapping someone out is they can [57:00] just reverse all of this guy's bad decisions, [57:04] which they don't want to redo. [57:05] But if you're willing to do that yourself, [57:06] I Think that could be very powerful. [57:08] Patrick's and my version of this joke was, [57:10] when private equity takes over a company [57:12] and they're like, "Wow, [57:13] look at what these previous idiot managers did." [57:15] And there's no reason we can't say that. [57:16] It's like, "Get a load [57:17] of what these previous management did. [57:19] Isn't that us? Shhh..." [57:22] People are unwilling to do that. [57:24] You can just decide to be inconsistent. [57:28] Oh, good, I've got a shot. [57:31] Okay, 26. What did I get? [57:33] 27. [57:34] Oh. [57:36] I think the other thing which was also true is, [57:38] there were some obvious things [57:40] that we knew were the case [57:42] that we were just ignoring. [57:44] I think the big one for us is the active trader market. [57:47] We noticed fairly early on, maybe as early as 2019, [57:51] that we have a big active trader business [57:54] and sort of a disproportionate amount [57:56] of revenue was generated [57:59] by relatively small group of customers. [58:02] We had kind of attracted them incidentally. [58:05] So we built a product targeting first-time investors. [58:08] We wanted to make it as easy as possible [58:09] to get started with $100. [58:12] Incidentally, that value prop also appealed to someone [58:16] who is paying the $10,000 a month in commissions, [58:19] and was trading quite actively. [58:21] And so we attracted all these people. [58:24] But then, of course, the market moved. [58:25] Everyone matched our commission offering, and [58:27] we had these competitors [58:29] that were building great, tailored, [58:31] active trader experiences. [58:32] So your pro-offering wasn't good enough, essentially? [58:35] It wasn't good enough. [58:36] And at first, we didn't have a pro-offering. [58:38] It was just pro customers in our regular offering. [58:41] So your Pro offering was nonexistent. [58:43] Yeah, we didn't have a Pro offering. [58:45] Actually, there were all sorts of scenarios [58:48] where if you got too active on our platform [58:52] and hence generated a lot of revenue for us, [58:54] we would force you to churn. [58:55] That seems bad, that [58:56] you were rejecting your best customers. [58:58] Yeah, we didn't have good handling [58:59] and our entire business was geared towards [59:03] the first-time investor. [59:04] So these customers didn't even [59:06] really show up in the data. [59:08] Every time we do our standard surveying process, [59:10] we'd be like, "Oh, a very small percentage [59:12] of our customers cares about this issue." [59:15] So we did some very simple things first. [59:17] We were like, "We care about active traders. [59:19] This is what's supporting the entire business. [59:22] We have to make sure they're happy." [59:25] It was the reverse. [59:25] The more active you were, [59:26] the less happy you were with the platform. [59:28] When we grokked that, [59:30] it was a nightmare scenario. [59:32] So I think that's another Frank Slootman thing. [59:34] You'd be like, "Where are you making your money? [59:39] How much are we investing? [59:41] Why are we investing so much in this thing [59:43] that doesn't make any money [59:46] and very little in the actual core of the business?" [59:49] What were the non-core things you cut at that time? [59:51] We were working on a lot of things [59:54] that I think were good, reasonably well-motivated, [59:59] but it just wasn't the right time for. [60:00] Our cash card offering, probably the best example. [60:05] We had a debit card tied [60:08] with a money transmitter account. [60:11] Sort of similar in nature to like a Cash App. [60:14] Chime. [60:16] It was a product for the paycheck-to-paycheck customer. [60:19] It didn't work very well. [60:21] I mean, it worked a little bit, [60:22] but the Robinhood customers not [60:24] a paycheck-to-paycheck customer, generally. [60:26] You have to have a little bit more than that in order [60:29] to become interested in investing. [60:32] Our best customers weren't benefiting from the offering. [60:35] They tend to be credit primary. [60:37] Once you go to credit primary, it's very difficult. [60:40] You're basically not going back [60:42] to moving your spend onto a debit card, generally speaking. [60:45] I think that didn't work. [60:47] There were also a bunch of little things. [60:49] If we were peanut butter [60:51] and spread across all these initiatives, [60:53] we were generally inefficient. [60:55] We were functionally organized. [60:57] So, at that time, I moved us to a GM structure. [61:00] My feeling there was, I was feeling a lot of stress [61:04] because everything kind of rolled up into me, [61:06] all these other businesses. [61:07] And I had to decide which one was more important. [61:10] The product people cared about [61:11] the product being good. [61:13] The finance folks, they care about costs [61:16] for the particular product area. [61:20] The engineers just cared about it getting built. [61:22] I wanted more people to feel— [61:23] to get the sleepless nights of having [61:26] the product not working. [61:29] I think with the GMs, we got that. [61:30] When we moved to GM structure, [61:32] and we gave full accountability to the business lines [61:36] and the revenue to the GMs, it immediately improved. [61:40] Because I could look around, I could see Johann, [61:42] our Crypto GM, he's sweating. [61:44] He's like, "If crypto's not going-" [61:46] So you have individual accountability. [61:47] Yeah. And that's worked very well for us as well. [61:51] The events have worked well. [61:53] That's more of a recent thing. [61:55] The marketing events? Yeah. [61:56] And the product launch events. [61:59] You ask what's worked well [62:02] to get shipping cadence. [62:04] I think it's always good to have a, [62:06] on top of all the other things that you need to do, [62:09] to make it possible. [62:10] It's always good to have an event— [62:12] Stake in the ground. [62:12] Stake in the ground. [62:14] You're either gonna make it for this event [62:16] or you're not. [62:18] I think that can be pretty motivating to people. [62:20] Totally. [62:21] Because everyone wants to present well. [62:22] They don't want to be the folks [62:22] letting down the other people in the event. [62:27] We find that with Sessions, [62:28] our headline customer event as well. [62:28] It's very clarifying for getting products out the door. [62:33] For sure. [62:33] We started with one last year, [62:34] and now we're doing 4+ a year. [62:35] That's cool. [62:40] Maybe last question, how's it going with Harmonic? [62:43] You started a mathematical super-intelligence lab. [62:45] It's going very well. [62:47] Those guys have been— [62:49] You've cracked super-intelligence. [62:50] Not yet. [62:52] We're getting there, though. [62:53] I think our intelligence- [62:54] We've cracked my intelligence [62:57] at solving math problems. [63:00] And you know, I was a Math PhD student so I went to UCLA [63:04] to do a pure math PhD to study with Terry Tao. [63:08] But when I was in high school, I got to like AIME level. [63:11] If you're familiar with that, it's two rungs below [63:15] the International Math Olympiad. [63:18] In the past year, we've exceeded [63:18] my intelligence by a considerable margin. [63:22] So I think we're on the way. [63:22] Probably another 10X and we'll be at [63:26] something that you can consider super-intelligence. [63:29] Is it a for-profit endeavor? [63:30] Is it a non-profit? And what's the revenue model? [63:33] It's a for-profit endeavor. [63:35] Okay. [63:36] We want to get really, [63:37] really good at solving math problems. [63:39] The unique spin on it is [63:42] that we're using formal mathematics. [63:45] We use a language called Lean, [63:47] we translate natural-language math into Lean. [63:50] And what that gives you is verification. [63:54] And I think verification has been [63:55] a big problem outside of the AI space. [63:58] Obviously, if you are relying on AI [64:00] for something mission-critical, [64:01] verification's important. [64:03] But even in math, [64:05] some of these proofs can be hundreds of pages. [64:07] And mathematicians have [64:09] to debate in a colloquium for years. [64:12] Like Fermat's Last Theorem, [64:13] probably a good example. [64:14] It took two years for that proof to be certified. [64:18] The verification problem is a real problem, [64:19] and it's only gonna get worse as the output [64:23] of artificial intelligence systems becomes [64:25] more and more voluminous. [64:27] Why is the proof of Fermat's Last Theorem [64:29] so long when the theorem itself is so short? [64:32] Doesn't that suggest to us [64:36] that it's not the final proof, [64:37] that it's just, a shorter one should be possible? [64:39] Not necessarily. [64:40] If you think about what's widely considered [64:45] the deepest theorem in mathematics, [64:48] the Riemann hypothesis. [64:50] You can write the Riemann hypothesis. [64:52] It's like one sentence on a chalkboard. [64:55] But you ask any mathematician [64:58] and they're like, "We're not even close." [65:00] We don't even have a strategy. [65:02] New math has to be invented to solve this thing." [65:05] So that's the holy grail. [65:07] That's actually what motivated Tudor [65:10] and I to start Harmonic. [65:11] The vision would be— [65:12] The Riemann hypothesis specifically? [65:13] To have a smartphone app [65:16] with our AI model to take a picture of a statement [65:20] of the Riemann hypothesis on a chalkboard, [65:22] and then it just cranks it out [65:24] and gives you a formally verified proof. [65:26] That'd be a pretty handy app. [65:29] That would be a nice viral moment [65:31] when all these conjectures end up going on social media. [65:34] Exactly. For your a '27 product launch event, [65:37] we're launching the Riemann hypothesis is now [65:39] the Riemann proof. [65:40] Hopefully it gets some retweets. [65:42] Yeah, alrighty. [65:44] Vlad, thank you. [65:45] Thank you so much. [65:45] Good stuff.