Original Title: Robinhood’s Vlad Tenev Claps Back: From Memecoins to Real Assets

Original Source: Unchained

Original Translation: Sleepy.txt

Editor's Note: In this episode of the podcast, Robinhood co-founder Vlad Tenev systematically elaborated on Robinhood's future strategy from multiple perspectives, including products, markets, and regulation.

When discussing Robinhood Chain, he emphasized that this is a cross-industry attempt, aiming not only to serve crypto users but also to bring traditional financial assets onto the chain. In his vision, stablecoins serve as the best reference. Just as stablecoins have exploded in markets outside the U.S., the implementation of tokenized stocks and private equity may also first succeed overseas before benefiting the domestic market.

In the latter half of the program, he shifted the topic to AI. The Aristotle model of the Harmonic project has already achieved gold medal-level performance on international math Olympiad problems and can output machine-verifiable formal proofs using Lean. In Vlad's view, this is not just a mathematical breakthrough but relates to fundamental issues in finance and the crypto industry. When the scale of AI-generated code and contracts surpasses human review capabilities, verification will become a new infrastructure. The value of formal verification will rapidly amplify, from smart contracts to financial derivatives, and even in healthcare and aerospace.

This conversation provides three key insights worth noting: how Robinhood Chain positions and implements itself; how private equity tokenization seeks mechanisms for breakthroughs within compliance boundaries; and how formal verification can establish a secure foundation for AI and the crypto industry. Together, they point to a longer-term issue: in the next decade, how retail investment, on-chain finance, and machine intelligence may intersect and reshape market order.

The following is the full content of the conversation:

Haseeb: Hello everyone, welcome to The Chopping Block. Today we have a special guest—Vlad, a futurist in fintech and the founder of Robinhood.

Vlad: Hello everyone.

Haseeb: Vlad, you've been very busy lately. Over the past year and a half, we've discussed Robinhood's rise many times on the show. But this time, I'm glad to have you here to directly address the community's most pressing question: What exactly is Robinhood Chain? Why are you building Robinhood Chain as a Layer 2 instead of directly creating a Layer 1?

Vlad: I think this topic has already been discussed on Twitter. Someone described it well, saying it's really a choice between being a "landlord" or a "tenant." We believe that having our own chain will ultimately allow us to control all the details of the entire platform from start to finish.

Of course, this may come with some interoperability costs at the beginning. But I believe that in the long run, all chains will have tokenized products, including stocks and other derivatives. Developers will build cross-chain capabilities, and in the end, the question of "which main chain to choose" may not be that important.

But for us, having a chain from the start, where we can control the entire experience from smart contracts to the Sequencer mechanism, is very important.

Haseeb: But the issue is, we are now hearing that Stripe is also building their own chain—although Stripe hasn't officially confirmed it, there have been media reports. Additionally, Circle has clearly stated they will launch their own chain. Both of these chains will be Layer 1, like Ethereum or Solana, while Robinhood Chain, like Base, has chosen Layer 2.

Do you think this difference is crucial for Robinhood's goals? Or is it irrelevant as long as it works and users can use it?

Vlad: I think the advantage of Layer 2 is that it is easier to achieve interoperability. Because there is already a lot of related infrastructure, some of which is even built by our friends at Offchain Labs (the Arbitrum development team), which means we don't have to "reinvent the wheel."

Of course, there is also a rationale for choosing Layer 1, as that allows you to completely rebuild everything from scratch. I don't think that's an unreasonable choice. But when we evaluate, we ask ourselves what we want to do now and what we want to do two or three years from now. Ultimately, we believe Layer 2 can meet the current needs.

Haseeb: Understood. So the story of Robinhood Layer 2 that we see now is roughly that this chain will issue tokenized stocks, which can be traded internally on Robinhood as well as directly on the chain. Users can even transfer tokenized stocks from Robinhood to the chain. Please correct me if I'm wrong—but this is probably the highest goal for the V0 and V1 stages of this chain, right?

Now, assuming we push the timeline forward five or ten years, if Robinhood Chain achieves great success, what would that look like? Would it become the next Ethereum or the next Solana? Would it have more users than these chains? Can you help me paint that picture?

Vlad: Let me say this first: when we decided to create this chain, we looked at the landscape at that time. Many companies were trying to build the so-called best chain for degens. But that space is already very crowded, and we didn't feel we had a unique competitive advantage in that area. You could argue that we have some advantages, but I believe Robinhood's core advantage lies in being the intersection of traditional finance and crypto.

So in the past, when we saw B2B opportunities, we launched Robinhood Connect, which is an on/off-ramp tool. Many leading non-custodial wallets and DApp providers have integrated Robinhood Connect because it is the most convenient and cost-effective way to convert fiat to crypto and vice versa. If you have crypto assets and want to convert them to fiat, we can provide a very smooth mechanism.

This actually comes from the capabilities we have already accumulated: as a large consumer retail platform, we have a lot of experience in seamless account opening and are very good at fraud prevention. The platform offers a diverse range of assets and products for users to choose from. Over time, my vision is that—regardless of which financial market or asset, whether you want to custody or trade, you can achieve it in the easiest way on Robinhood.

Therefore, when we talk about Robinhood Chain, our core argument is whether we can make it the most suitable chain for real-world assets. I believe the answer is yes. We can start with existing assets, such as U.S. stocks.

You can imagine that any asset offered on Robinhood, as regulation matures, could potentially be issued in on-chain form, primarily targeting markets outside the U.S., but there is also potential in the domestic market.

In addition, there are some things that are currently not easy or even impossible to do, such as the private market. We can also open them up through on-chain methods. This is exactly what we showcased at our own event, Robinhood Con: on one hand, the expected tokenization of public stocks; on the other hand, the unexpected application of the same technology to private companies, which were previously almost completely inaccessible to retail investors.

Haseeb: So one of the points that caused a stir previously was the tokenization of OpenAI and SpaceX. When the news broke, the internet exploded, and everyone was excited, thinking Robinhood was going to bring stocks onto the chain and even include private companies.

Then OpenAI quickly posted a tweet saying, "Guys, this is not us. We don't know who holds these shares. Robinhood, this has nothing to do with us."

I'm quite curious, did Robinhood provide an official response at that time? Because this actually reveals your structural model: essentially, any financial asset, regardless of what it is or whether the related company is aware, can be turned into a tradable, real-time holdable token on-chain.

Can you help us understand that incident? How should we interpret it? Does this mean private companies will explicitly resist this kind of tokenization? Or do you think there will be more open private companies than OpenAI?

Vlad: I think there are a few points worth noting. First, there have been attempts in the past to do similar things, such as other companies trying to provide "exposure" or using the Reg A+ mechanism in the U.S. to do limited forms of crowdfunding. The problem is that these models often encounter "adverse selection." In other words, companies willing to open shares to retail investors are often those that do not have many other financing options.

Of course, there are exceptions. Some companies do care about retail investor participation, especially in later stages, and they want retail investors to have the opportunity to participate. This has become increasingly common in recent years. We can also discuss our observations in the IPO Access business.

IPO Access allows companies to open subscriptions directly to retail investors during their IPO. I think this is good for retail investors and good for the companies themselves. When we launched this service in 2021, we did a lot of work to persuade companies to understand its value. Now, almost the best companies actively come to us, willing to give retail investors a higher allocation ratio.

So we see that companies are gradually realizing that having a diverse and large retail shareholder base is a strength in itself. This is an evolutionary process.

But when it comes to private companies, the "adverse selection" problem still exists. Companies like OpenAI and SpaceX can freely raise hundreds of billions of dollars from private market institutional investors. Many times, they are unwilling to deal with complex new mechanisms, especially when these mechanisms lack mature cases and precedents.

Therefore, private company tokenization is a completely new thing, and retail investors participating in private company equity is also new.

And when these companies themselves have a lot of financing channels, their core focus remains on serving customers and building products; they do not want to become innovators in the capital markets. But we are willing to, as we believe this is a huge issue. Because these private companies that are changing the world and leading industries may only allow retail investors to participate when they go public, and their valuations have reached hundreds of billions or even trillions of dollars.

If this issue is not resolved, I think the risks are very high. Ordinary people may completely miss out on the value distribution of these technologies and instead develop a sense of exclusion towards them. This could even lead to social-level conflicts and hinder overall progress.

For example, if AI really changes the world and even impacts the labor market, replacing your job, but you have no ownership in these companies, you might tend to oppose it. However, if you can hold relevant shares, you would, to some extent, hope for its success. We have already seen this situation in the public market, and I believe it will be the same in the private market. This is why we care so much about providing opportunities for retail investors.

Tom: This actually relates to a recent case where Anthropic, in their latest round of financing, required all new investors to participate without using SPVs (Special Purpose Vehicles). In other words, not only retail investors but even qualified investors, if they are not funds, are excluded.

So I have a question: what do you think about the lack of information disclosure from private companies? Could this create valuation discrepancies between private financing and the public market? Because there is a strange situation now: the disclosures to retail shareholders are often less than those to private investors (who hold preferred shares). There is always an issue of information asymmetry. How do you think this situation will ultimately be resolved? In my view, this is one of the biggest structural challenges in the "private equity derivatives market."

Vlad: Yes, my view is that in the private market, if a company is oversubscribed by 30 to 40 times in a financing round, then for SPVs, some institutional investors, private wealth management clients, or the private banking departments of large banks, they won't do much due diligence when they get their shares.

Imagine a company is oversubscribed by 40 times, and everyone is scrambling for allocations; the company won't spend time allowing every SPV investor to conduct in-depth due diligence. For many investors, this is fine. They see top-tier funds like SoftBank or Altimeter investing and are happy to follow, treating their due diligence as a stamp of approval.

So, overall, I support information disclosure. But I also think many people will say: I understand the risks, and I am willing to bear them. I have used ChatGPT, I have some understanding of the aerospace industry, I can see the public information, and I fully understand that this investment could go to zero, but I still want to invest. As a believer in free markets, I think people should have the right to make that decision for themselves.

In the U.S., we currently have standards for "qualified investors." But we have been pushing for reforms to this set of rules. I believe the foundation of these standards is outdated—they were established before the internet, when information about companies was very limited. Now, with the development of AI and social media, public information is easily accessible, and we have tools to digest and understand that information.

So I think regulation should be re-examined and evolve towards a model based on disclosure and self-certification by investors. Otherwise, if you force companies to make various additional, non-standard disclosures for retail shareholders, they may simply refuse to allow retail participation and only want to deal with a small group of large fund managers. In that case, we would ultimately return to the original problem we wanted to break: market opportunities being highly concentrated in the hands of a few institutions.

Tom: Exactly. Now, let's bring the topic back to the current regulatory rules. Because you mentioned before that your products will primarily target markets outside the U.S., rather than the domestic market. Is this because the existing regulatory restrictions in the U.S. are too strict? When you actually launch, how do you plan to design the structure of overseas products? For example, will you restrict U.S. users through KYC and geographic fencing? Or will you set the blockchain itself to be permissionless, with assets on the chain also being completely permissionless?

In other words, based on the current regulatory environment, what is the actual structure you envision? I know you hope the rules will change, but given the current situation, what would you do?

Vlad: It actually depends on the regulatory maturity of each market. Overall, the model we anticipate is that Robinhood will first acquire or hold some shares or positions in private companies, and then we will package them and offer them to retail investors in different ways. It could be direct offerings or through some diversified investment tools, depending on the compliance requirements of different markets.

Currently, I think a good analogy is stablecoins. Stablecoins are actually the earliest tokenized assets. In the U.S., the capital markets and financial infrastructure are relatively sound, and people can easily transfer money through ACH, debit cards, and credit cards. Therefore, stablecoins are more often used as tools for cross-border transmission of U.S. dollars, especially in markets outside the U.S.

I think a similar situation will arise with the tokenized version of U.S. stocks: in the domestic market, because the capital markets are already well-developed, the acceptance speed may be relatively slow; but outside the U.S., in markets with higher entry barriers and more limited channels, the tokenized version may become the main channel for investing in U.S. assets.

Of course, the specific direction for the U.S. is hard to predict. Currently, tokenization is not widely allowed as an investment tool for retail investors. But this situation may change. And once it changes, I believe the most direct and significant improvement will be retail investors' exposure to private companies. Because tokenization can solve a core issue here: liquidity.

In the past, in the private market, even if you could obtain shares in private companies, you would face many restrictions, such as transfer limitations and lack of pricing transparency. Although some secondary markets and trading platforms have emerged, they must build their own liquidity markets, which often struggle to attract enough buyers and sellers.

But once connected to the global crypto network, the liquidity issue will be significantly alleviated. So I believe that as regulation matures and allows tokenization, this will be the biggest change.

Robert: I understand what you mean by "tokenization essentially solves the accessibility issue," and stablecoins are indeed a good example.

But I also notice that other companies in the crypto space are leveraging on-chain technology to accelerate product market entry and enhance the accessibility of financial services. For example, Coinbase directly offers Bitcoin lending services in their wallet through Morpho; and Phantom integrates Hyperliquid perpetual contracts directly into their product.

They interact with on-chain DeFi applications through this self-custody model, thereby accelerating their go-to-market strategy. Essentially, they are just a wallet or service provider. I want to ask if Robinhood has considered a similar path? What do you think of this model?

Vlad: We have Robinhood Wallet, which has grown rapidly in the past, and we are continuing to invest in it. Robinhood Chain is clearly another important component.

We believe that relying on tokenization, especially in overseas markets—of course, we also hope to do this in the U.S. in the future—Robinhood can provide a variety of truly valuable financial products, starting with stocks and the private market, and eventually expanding to all asset types on our platform. We hope to bring these assets on-chain in a permissionless manner.

I believe we have a unique advantage because we may be the only financial services platform with large-scale operations in both crypto assets and traditional assets. The essence of tokenization is the intersection of the two.

On-chain tokenization still requires a traditional financial company to custody all TradFi assets to complete the minting and burning processes and put them on the chain.

So there are two vertical integration paths here. The first is in the blockchain dimension; if you control the chain, wallet, and intermediary, you can provide better pricing to users. The second is in the custody dimension; if you can control the custody layer of traditional assets, you will have more room in pricing and economic models, offering users lower costs and a better experience.

Tarun: I have a question: in the future, if 10%, 25%, or even 50% of stocks are circulating on-chain, what will happen to institutions like transfer agents and DTCC (Depository Trust & Clearing Corporation)?

This is interesting because once this structure reaches a larger scale on existing underlying assets, will it completely change the existing brokerage business logic? For example, will handling transfers and connecting the entire ecosystem (those websites that don't even have dual verification) be completely disrupted?

Vlad: My guess is that the final form may resemble the model of ADRs (American Depositary Receipts) or ETF custodians. In other words, traditional assets will still be held by a TradFi custodian, and when better prices can be obtained on traditional exchanges, trading will still occur there—this is the system design we have already showcased at Robinhood Con.

At the same time, tokenized assets will form a secondary market. The initial selling point is that tokens can be traded during traditional market off-hours, such as weekends and holidays. Over time, this market will gradually expand.

Haseeb: So Vlad, expanding from daily trading hours of 9:30–16:00 to being able to trade 24/7—stocks being able to be bought and sold on weekends and during off-hours—makes complete sense to me. I think this is almost inevitable, especially considering Robinhood's long-standing product direction of making trading easier for retail investors.

However, I find it even more interesting that you mentioned the private market. From your previous answers, I gather that you believe the private market line, in the long run, is a bigger story than simply making the public market more efficient.

As someone who has also invested in private companies, I feel strongly about this: there has always been a tug-of-war between the private and public markets. Companies are staying private for longer periods. You mentioned earlier that maybe someone will be unemployed because of OpenAI, but at least they can buy some OpenAI stock.

But the reality is that while private companies want looser listing rules and lower costs, they are also very clear about what they dislike, and they are much smarter than in the Facebook era. For example, they are increasingly disallowing secondary market trading and the transfer of shares between SPVs.

So there is a confrontational game here: as long as there are enough SPVs in the market, Robinhood can find one of them to tokenize the shares it holds, allowing retail investors access and liquidity. What do you think the endpoint of this game will be? After all, companies can simply say, "We don't like what you're doing, go away," and even remove you from the cap table.

Vlad: Yes, I think there are two points. First, the mechanism of tokenization must be able to operate smoothly without relying on the company to "choose to participate" actively.

Haseeb: Wait, how is that possible?

Vlad: We have actually demonstrated this. As you just said, if you obtain indirect exposure through an SPV or some institution, you can put these traditional assets into a pool and then tokenize them, or use other tools to allow retail investors to access them.

For example, in the public market, if every company had to approve every shareholder change or every transaction, the entire market simply couldn't operate. Similarly, if in the private market, tokenization required the company to approve each instance, it wouldn't work either.

So I believe the mechanism must be independent of the company's direct participation. Not only because the company may be unwilling, but also because they likely do not want to take on the extra workload and processes.

Haseeb: I think many times, companies really just don't want this. They truly do not wish to have a fluctuating price marked every day. You should understand this reason well: Robinhood experienced the pressure of the public market deeply before and after going public—stock prices fluctuate daily based on your tweets, your words, and your actions.

In the private market, they actually prefer the state of insufficient liquidity and valuations "being pushed up." They like this opacity; no one can say, "Oh, it seems the company is worth less today than it was yesterday."

Vlad: Many people will bring this up, saying, "When you were still a private company, you probably didn't want this kind of liquidity either."

I think it varies by company; each private company has different points of focus. Personally, I can fully accept it—clearly, I am not the best counterexample.

But there are also many companies willing to do this, and not just those without other financing channels; some companies are very aware of the value of retail investors.

The ideal situation is that we can collaborate with these companies, and they recognize the value of what we are doing and open up to retail investors at increasingly earlier stages. Ultimately, retail investors could even become part of the company's founding.

In other words, at the seed round stage, companies could include retail investors in the cap table, allowing them to participate from the very beginning.

Haseeb: Sounds very much like the crypto world. It resembles ICOs.

Tom: But in this model, the company is actively choosing to "join," right? They are excited about this and believe that having a large number of retail investors on the cap table is a good thing, so they open up voluntarily.

This is completely different from being passively pulled onto the chain by Robinhood without consent. Right? I understand it as opt-in, rather than you "forcing it."

Haseeb: However, it sounds like some are indeed opt-in, while others are more like, "We are going to do this. For example, OpenAI, such an important company, retail investors should have access."

Vlad: There is actually a subtle difference here, depending on who the "client" issuing the stock is.

In the early stages, the so-called "capital as a service" clients are entrepreneurs. Entrepreneurs raise funds from the primary market, which is a typical opt-in—they actively decide whether to open up. But as companies enter later stages, the clients are no longer entrepreneurs but senior management, employees, and early investors, who seek liquidity in the secondary market.

At this level, whether employees and early investors can freely sell their stocks is itself a controversial topic. Different companies have different attitudes: some are very strict, while others are relatively lenient. In those lenient cases, investors can sell shares without the company's consent—meaning the company does not need to actively "opt-in." So, each case varies.

You can also understand this subtlety. But if you ask me whether I think "all employees and early shareholders must have company approval to transfer shares," I wouldn't say so. That feels more like a legal decision at the company level.

Haseeb: I actually don't care where Robinhood's tokenized shares come from—whether from employees or some early investors. I believe there are always people on the edge of the cap table whom the founders don't care much about, and whether they can cash out is irrelevant.

Tarun: But there will also be some shares that the founders care very much about.

Haseeb: Right, there are indeed shares that the company cares about. I think Robinhood is unlikely to really get that portion.

What I'm more interested in is the point you just mentioned—Robinhood's story has always carried a certain "Promethean" quality: bringing down the once lofty spark, allowing retail investors to access any financial asset, any stock, any financial tools they were originally not allowed to touch. Now they can.

I understand this narrative, and it aligns well with the spirit of the crypto world. But at the same time, it blurs the lines between "private companies" and "public companies." After all, establishing such a complete system is essentially meant to delineate the two.

If a company is private, then regulation can be relatively relaxed, allowing them to operate more freely; but if a company is public, everything must be open and transparent, and all disclosure requirements for retail investors must be met.

So, when retail investors can access both public and private stocks, even if only partially attractive private companies, does this, from a societal perspective, weaken the significance of the "public vs. private" system?

In the world you envision, how do you think this system should be transformed?

Vlad: I do not believe this will weaken the significance of the system. I don't think anyone has ever seriously designed a system that actively prevents 80% of ordinary Americans from participating in these value appreciation opportunities. This seems more like an accidental result of historical evolution. Initially, it was indeed based on investor protection—allowing companies to operate and start more easily without bearing excessive disclosure requirements, which is reasonable.

But over the past few decades, the situation has changed. Going public has become increasingly difficult, with processes and requirements becoming more cumbersome; meanwhile, the private market has become easier for companies to enter, with more VC and PE providing funding.

The combination of these two factors has led many companies to be unwilling to go public or to wait until they have extremely high valuations when they do. The ones who suffer are retail investors, who are kept out of these appreciation opportunities. So I believe it is entirely possible to address some regulatory concerns while also allowing retail investors to participate in the early opportunities of these excellent companies.

Tarun: In fact, what you mentioned reminds me of an interesting phenomenon in the crypto world. To me, it feels like an inevitable tug-of-war between the private and public markets—who can get the earlier shares first.

This situation occurs repeatedly in the crypto space. For example, during bull markets, many private rounds are partially opened to retail investors, leading to the emergence of platforms like Echo or Legion, where companies complete private financing first and then split part of the allocation into AngelList SPVs to sell to retail investors.

But this model often expands faster than the actual growth of the company, resulting in a quick reversal once the hype dies down, completely privatizing and stopping access to retail investors. The uniqueness of crypto lies in its rapid cycle speed, equivalent to seeing twenty cycles of the public market in one hour.

So there is a key question: if this model crosses a critical point, will there be a reversal? Or in the path you envision, is this actually a one-way street that, once opened, will not reverse?

Vlad: Yes, it is difficult to predict where it will ultimately lead. My view is that the biggest issue in the crypto space is that we have never been allowed to truly connect crypto assets to companies, equity, or what I call "things with basic utility."

As a result, due to the lack of this connection, market activity has largely flowed toward meme coins. They are popular precisely because they are disconnected from real utility, and this is something that is allowed.

At the same time, there is constant emphasis on "investor protection," saying we need to prevent investors from making poor decisions and losing money. But the reality is: you can buy meme coins at will, while OpenAI and SpaceX are deemed too risky. This seems absurd to me and may also be an unintended consequence.

Tarun: Since we are talking about limited accessibility, I want to ask what you think about digital asset treasury companies?

They are somewhat like the hybrid you just mentioned: these vehicles themselves are not traditional companies, but they package some assets and allow you to buy them in the form of stocks. Additionally, private equity has also seen similar tools, such as closed-end funds, claiming to hold a bunch of Stripe shares. If you buy their ETF, you indirectly hold shares of Stripe. This kind of "wrapper" product has grown rapidly in the past year. What do you think of these strange financial instruments?

Vlad: I don't want to comment too much on specific investment targets, but I can say: I believe in the free market.

If someone legally and compliantly issues a stock, a certain type of financial instrument, or even cryptocurrency, I believe it should be open to retail investors in principle. The beauty of the free market is that it allows the market to decide for itself. If a product is truly attractive to consumers, over time, the fundamentals will reveal themselves, and the market will naturally provide the answer. So I won't analyze each one as "bullish or bearish," whether it's treasury companies, closed-end funds, or open-end funds.

For me, our goal is to allow users to access all types of financial assets and transactions. We believe Robinhood can become a platform that is both safe and compliant, offers a good user experience, and maintains low costs. So regardless of the asset, we have the opportunity to bring them in.

Haseeb: It seems we still can't get you to directly discuss the topic of "digital asset treasury," but that's okay.

However, there is an interesting phenomenon: Robinhood has basically been a platform for trading U.S. stocks in the past. And over the past few years, an increasing portion of your revenue and profits has actually come from crypto business. Meanwhile, one of your biggest competitors, Coinbase, which used to be purely crypto, is now starting to talk about increasing stock trading—almost a form of "cross-evolution."

Does this indicate that as Millennials, Gen Z, and even Generation Alpha grow up, their future trading venues will increasingly resemble a fusion of crypto and traditional finance?

What do you think about Robinhood's positioning in this transition? I guess you didn't expect crypto to grow so quickly in your business, right? As this part of the business increases, has your understanding and positioning of Robinhood changed?

Vlad: I believe the fundamental argument of Robinhood has not changed. Our goal has always been to become a financial super app, helping users meet all their financial needs: to custody any asset and conduct any financial transaction.

Of course, as you mentioned, the proportion of crypto in our business has become very significant. Last quarter, our revenue was close to $1 billion, with crypto accounting for about 20%–30%. As crypto technology increasingly resembles infrastructure and Robinhood's business gradually globalizes, I think the boundaries between traditional finance and crypto will become increasingly blurred.

For example, if you complete a tokenized stock transaction on-chain, or when we custody stock tokens or private tokens, should you count that as part of the crypto business or the TradFi business?

This classification actually encounters many boundary issues. Just like the concept of the "tech industry" is becoming increasingly irrelevant—almost no company can claim not to be a tech company because all companies rely on technology. Similarly, over time, the distinction between "crypto vs. TradFi" will also become less important.

As a Robinhood user, you may increasingly use crypto technology, but it won't be the "explicit part" of the experience.

Of course, if you are buying Bitcoin, it's clear that you are purchasing a crypto asset; but if you are buying tokenized assets, crypto here is more of an infrastructure and transmission mechanism, and ultimately what you receive is the financial tool you truly want.

Tarun: This is actually related to the previous question: Robinhood started with stocks and then added crypto, while Coinbase started with crypto and is looking to expand into stocks. Another difference between the two companies lies in the institutional/infrastructure level.

For example, in terms of custody, staking, and supporting institutional clients, Coinbase tends to prefer "full-stack control." However, in the public market, it's hard to have the entire full stack. But in the crypto space, strangely, a single entity can actually provide a complete full stack for end users.

What do you think about this? As Robinhood aims to serve both crypto and non-crypto "full markets," will you gradually take control of more infrastructure yourself? Because I think this is a significant shift. For instance, if the trading volume of tokenized stocks far exceeds your existing traditional stock trading, will Robinhood also need to become "full-stack"?

Vlad: I see the development of Robinhood in three phases: short-term, mid-term, and long-term.

Short-term: The goal is to become the number one platform for all active traders, covering all asset classes we offer—cryptocurrencies, stocks, options, futures, etc.

Mid-term: This is the stage we are building towards, becoming the No. 1 financial super app for customers. Your entire wealth is held in Robinhood, and all transactions are completed through Robinhood. The first two phases are primarily retail-focused; we believe this part of the business does not need to be fully vertically integrated at the institutional level.

Long-term: As we continue to build these capabilities (such as 24/7 trading, extremely low margin rates, asset tokenization), these features themselves will be attractive to enterprise and institutional clients. After all, institutional investors also want to trade 24/7, want to access both crypto and traditional assets in one place, and want the lowest margin rates.

These features will naturally give rise to opportunities in B2B and institutional business. Looking ahead ten years, I believe Robinhood's international business could surpass its domestic business, and the scale of institutional business could also exceed retail business. This excites me greatly because we are just getting started. I can see at least several paths that could allow Robinhood's business to grow tenfold, and we can pursue multiple paths simultaneously.

Haseeb: So now Robinhood's role is quite special:

On one hand, you are becoming builders of the crypto ecosystem through products like Robinhood Chain; on the other hand, you are also "external users," feeling the established elements of the entire industry through collaborations with entities like Arbitrum.

I imagine you must have some feelings about this, such as: "People are doing poorly in this area and need improvement." Many entrepreneurs, developers, and builders are among our audience. I would love to hear Vlad's thoughts: where do you think the crypto industry is lacking? What areas should be improved? What do you hope to see developers invest more in?

Vlad: I actually don't see myself as a "customer of the crypto industry." I believe Robinhood is a very active participant. In terms of market share, trading volume, and revenue, especially in the U.S., we are already one of the largest companies in the crypto space.

I think what has been missing in the past is that crypto and the traditional financial system are almost two disconnected worlds, lacking a connection in between. Stablecoins are one of the few exceptions; they have started to serve as a bridge. But my attitude is not to say "you fix it," but rather—we are actively working to improve it ourselves, leveraging Robinhood's capabilities and advantages to make that connection work well.

Robert: Of course, there is also a time factor here. For example, you are now doing Robinhood Chain as an L2, but perhaps you wanted to do this years ago.

I believe there are certainly things you think, "I really wish I could do this now," but at that time, the technology wasn't mature enough, the product experience was too poor, or there was insufficient liquidity. This even includes the private market assets you are now exploring or some attempts at prediction markets. It feels like the timing has just matured now. What do you think are the things you want to do but haven't had the right timing for?

Vlad: I think there is a lot of discussion now around AI Agents: will these AIs use cryptocurrencies to pay each other and complete transactions in the future?

But beyond that, there are real-world AIs, such as robots, medical devices, etc. Will these real-world AI Agents also use cryptocurrencies to pay each other in the future? What would that look like? I believe we are only a step away from both scenarios, and taking another step forward would mean that crypto truly becomes a tool for Agentic Commerce.

Haseeb: Since you mentioned AI, let me ask: you are also working on a company called Harmonic. If I understand correctly, you are building a foundational model that uses Lean and theorem proving to verify mathematics, thereby validating the answers generated by the model. As the CEO of Robinhood, how do you still have time to do this? What level of involvement do you have in this company? Why did you choose to pursue this direction?

Vlad: I am the chairman of Harmonic. It is a completely independent company from Robinhood; I do not have a daily operational role, but I am the founder and care deeply about it.

A few weeks ago, Harmonic announced that our model Aristotle achieved gold medal level at the International Mathematical Olympiad (IMO). As you may know, this is an extremely difficult math exam. Off-the-shelf large models like GPT-5 and Grok typically cannot even solve a complete problem in this context and perform very poorly.

There are two aspects to the problem. One is that these problems require a "flash of inspiration" creativity; if you can't find that entry point, you can't solve it at all. The second is that they usually require more than ten steps of logical reasoning; if any one step is wrong, the entire proof is incorrect, and the result is wrong.

So, it requires both creativity and logical rigor. Moreover, in this scenario, the AI's "hallucination" problem can compound. Therefore, this result is a very good validation of our technology.

In fact, we completed it in a formalized manner. You just mentioned the Lean theorem prover; we indeed used it to generate formal proofs. This means the results do not need human manual verification; as long as you input it into the Lean kernel, the system will gradually check whether each step of the proof is correct.

Next, you might ask: what are the application scenarios? We call it "mathematical superintelligence."

Its significance lies in the fact that AI can not only generate answers but also verify the correctness of those answers to a very high standard. Think about how AI is already generating a lot of code. The work of senior software engineers is shifting from "writing code" to "verifying the code written by AI." If it's front-end UI, it's relatively easy to verify; you can visually confirm whether it meets design specifications. But if it's a back-end system or a smart contract, especially in scenarios where failures could lead to hundreds of millions of dollars in losses, humans must verify it rigorously.

This creates a market for formal verification. So we are very excited about this. I believe its applications are not limited to mathematics but will extend to all software and even hardware.

Haseeb: So you believe this is a better path that can build "software engineering agents" faster than routes like Anthropic. You think that using theorem proving and mathematics as a foundation is a necessary prerequisite for achieving true AGI.

Vlad: I believe this is not only a better path but also an inevitable one. Because in a world where AI generates a lot of content, these outputs have become so numerous that humans cannot possibly review them all; we need a new way to ensure they are correct.

There are two points here: we must ensure the results are correct; they don't even need to be expressed in a language that humans can directly read, because humans no longer review line by line. So, the entire underlying logic will change. The question becomes: how can we quickly verify whether AI's results meet expectations without manually checking line by line?

Tarun: I have two related questions. The first is a kind of aesthetic criticism of formal verification.

You dropped out of college years ago but became deeply involved in mathematical research, perhaps because you were attracted by the beauty of certain proofs. I feel the same way—some proofs of theorems are concise and elegant, making them feel extremely beautiful. Mathematics has always had a tension: it's hard for formal computation and aesthetic intuition to coexist.

The proofs generated by Lean might be extremely lengthy and unappealing, like the computer proof of the "four color theorem," whereas human indirect proofs are more elegant. How do you view this situation? How can we preserve the beauty of mathematics?

Vlad: Have you seen the proofs generated by Aristotle in Lean?

Tarun: I've seen one or two, but I haven't seen a complete collection.

Vlad: I actually think they are beautiful. Moreover, converting Lean proofs into natural language is not very difficult; it can almost be mechanized.

Because Lean's functions and definitions are descriptive, it is easy to convert highly formalized details into more descriptive English narratives. I think this is also why Lean is relatively easier to use compared to previous formal languages, and it is one of the reasons it has become popular not only in the AI field but also among mathematicians.

On the other hand, the reverse is much harder. Converting non-formal proofs into formal proofs is a very labor-intensive task. Currently, at Imperial College London, Professor Kevin Buzzard is leading a large project to fully formalize Fermat's Last Theorem. This is a massive undertaking that requires hundreds of mathematicians and several years to complete manually. By comparing these two, you can see the difference in difficulty.

Tarun: My last question is: besides the Millennium Prize Problems, what problems do you hope Aristotle or its successor models will solve?

The Millennium Prize Problems are too obvious, such as the Riemann Hypothesis and P=NP; everyone is discussing them. But do you have any "personal obsession problems"? If they were solved, you would feel, "Okay, this is my greatest achievement."

Vlad: Well, let's not talk about the Riemann Hypothesis.

Tarun: Right, that's too obvious.

Vlad: But I actually quite like the Riemann Hypothesis (laughs). If I were to give another example, I think creating a "benevolent version of HAL 9000" would be very cool. I would love to build a logical core and control center for a spacecraft, provided it doesn't go rogue and take off on its own like in the movies.

Haseeb: Uh… are you sure this is the example you want to share with everyone?

Vlad: (laughs) Yes, I mean a spacecraft control system that can be formally verified. A truly "benevolent AI." We do need a reliable control core. I think the initial goals were indeed separate, but now overlaps are starting to appear. For example, smart contract verification.

Smart contracts are essentially a relatively independent piece of code running on languages like Solidity and Rust. We must ensure some basic properties, such as contracts cannot stall, and there cannot be issues like "double spending." Because once an error occurs, it could lead to hundreds of millions of dollars in losses, and there have been painful cases. Currently, smart contract companies can only rely on manual audits—spending hundreds of thousands or even millions of dollars to hire external companies to check the code line by line.

So in the crypto space, I believe once formal verification becomes feasible, it will be very powerful. There are many other scenarios, such as those requiring complex mathematical calculations—margin calculations, options pricing, etc. These are all software where the cost of errors is extremely high. Therefore, I believe that the financial services industry, as well as sectors like healthcare, automotive, aerospace, and robotics, are all very suitable for formal verification.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。