Lennix Lai of OKX on tokenization and AI agents
2 Sept 2026 · 36:35
Lennix Lai, global chief commercial officer at OKX, tells Milk Road host John Gillen that tokenization and AI agents could bring crypto's next wave of capital after ICOs, stablecoins and Bitcoin ETFs. He explains the partnership with Intercontinental Exchange, owner of the New York Stock Exchange, tokenized equities and commodities, and why putting assets onchain enables new products. He also covers agent-driven trading and safeguards that keep humans in the loop.
Speakers
Topics
Transcript
So AI will become the intellectual layer and crypto will become the economic layer. What it means is >> everyone is talking about tokenization and AI, but what does that mean for crypto? Is there going to be a distinction between trady and crypto anymore? And what does it mean to have an exchange for AI agents? Hello and welcome to the Milkroad Show, the podcast that's starting to think we're going to have to ask AI to figure out what to do about itself and politely ask it to make no mistakes. I'm your host, John Gill, and today is Wednesday, September 2nd, and today we are joined by Linux Live. Linux is the global chief commercial officer at OKX, where he has led international strategy, commercial operations, and the exchanges expansion into DeFi and onchain infrastructure.
Linux has had a front row seat to OKX's New York Stock Exchange ambitions, onchain activities, AI, a whole lot of other things. He's full of alpha. He's going to tell us a whole bunch of information in this episode. If that sounds good to you, make sure you like and subscribe. Share this episode with somebody who's going to enjoy it. And a reminder, our podcast today is free and that would not be possible without our wonderful partners at Saber. Money, the stable coin payments platform built for Asia. So keep an ear out for more information on them later in the episode, but for now, welcome to the Milk Road Show, Linux. How are you, sir? >> Hello, John. Thank you for having me. >> I'm really excited to have you on the show. There's a lot >> Yeah, there's a lot of exciting things going on at OKX these days, and I want
to hear about all of it. Um, I thought a good place to start this conversation would actually be not in the east where you are, but over here in the west, the parent company of the New York Stock Exchange invested in OKX at a $25 billion valuation and they took a board seat. I'm curious to hear your thoughts on what this strategic partnership means, what it unlocks for OKX, and just what you're seeing happening here. >> I think it's it's also it's a very important strategic partnerships for between OKX and ICE. I think uh we have a share visions in relation to our tokenizations of the financial asset classes. I think the eyes partnership give give us I would say a practical
route for us to connect to a regulated trading venue. So that uh going forward I think with that partnerships we have a very strong fundamental and uh with a with a very strong partner to potentially uh substantially increase the distributions and the rich on tokenization equities commodities derivatives products. on the contrary that uh the crypto product probably in one day uh being able to tra be become a tradable instrument in a regulated uh liquidative venue like ice for for instance. So I think it serve a very important bridges between these two uh
legacy isolated uh system. >> Could you talk to me a little bit more about how these two sides of of of this partnership are going to change each other? In other words, crypto is coming to Tradfi. Tradfi is coming to crypto. How are they changing each other as they start to collide here? And what are you seeing happen as those those two sides those two worlds run into each other? >> The two world is I think yeah what we are directly or indirectly influence each others not express not not particularly OKX versus ICE, right? I think I'm I'm I'm talking in in general on crypto ecosystem versus coll with the traditional uh ecosystem. You can see traditional
exchanges starting to add trading hours. There's an obvious example that being influenced by crypto exchange for instance for example in OKX because in crypto we are trading 24 hours 7 day 300 we never we were the never the markets never stop the market never stop is a very important technological advantage because it means the market the matching engine the risk engine or everything has to run non-stop so so so it's also symbolic that the archite the architectural fail within OKX or within crypto exchange has to be really robust to handle all kind of abnormalities, right? So the the traditional exchange need to catch up because we are building
we already successfully build continuous markets for years. So on the other hand as crypto exchange for instant like OKXs uh we're adding more instruments aside from crypto native product like Bitcoin, Ethereum or other else we're starting to adding tokenization instruments agree or everything into the exact account structure the risk structure account structure governance structure on boarding infrastructure within the same system. So these two world are collide and communal within each other. That's the trend that we are seeing and we are seeing that we have both influenced to each other. The market has become uh as
communal as a huge giant market. >> Hi everyone, this is John. Listen, our Milkroad Pro analysts nailed this AI bull market. They we had dozens of calls pay off in a big way for our community. If you want to see how our entire team of analysts are positioning for the crypto bull market that's going to come after this bare market finally ends, you can join us now at Milkroad Pro. It's just a dollar. Link is in the description. You'll get access to all of our analyst trades, their watch list, and see how they're navigating this in their portfolios and get access to ask us questions in Discord anytime, day or night. It's just a dollar. Link is in the description. Milk Road Pro. Join now. I'll see you there. So talk to me about what you see happening in like let's say for example the DeFi space. As
we start to get more tokenized assets, more real world assets from Trady on blockchain rails. Do you see these moving into the DeFi space and and how does that evolve? What what changes there as that starts to happen? >> Yeah, the that's very interesting and a little bit in personally a little bit surprised development on in my perspective. uh we we I personally was was worried about hey what's the reason that a user trade Apple share listed on NASDAQ versus are a tokenized Apple share traded on OK for OKX for instant right so so there will be distinction between between uh liquor and market tab and all
that but I can see right now tokenized instruments starting to become an important collateral or fundamental fundamental asset class within the define space. So define historical is just borrowing lending protocol a lot of their borrow lending protocol with AMM and uh liquidity professions but right now we seeing a lot of uh liquidity pool or AMM protocol starting to add tokenized treasuries as the basic yield generation instruments and we're seeing that people particularly integrating MIMS coin versus tokenization of
securities as part of the uh building blocks in device spaces. So I'm not surprising there will be an other new hunting opportunities on device but with the basis of tokenizes instruments. >> So I want to talk about some other things that are happening at OKX that are part of the evolution of of crypto and tradi here. Um I I wanted to bring up AI agents. A few months ago, you gave a presentation that I I saw at the uh web 3 festival in Hong Kong. And this was the title of the talk was onchain OS, the operating system for the agentic economy. I would love for you to tell us a little bit more about this. What is this operating system? How does this bring AI agents on chain? What are you
guys building here? What what's what's going on? >> In short, it's uh it's open to kits. So OKX is is a massive ecosystem is we call it cryptochain but it's massive ecosystem where we we have wallet function we have central exchange function we have market data we have payment rail we have dexis routers we have all kinds of API function call uh so what we really want to achieve here is whether or not people a simple users or general general users can interact with this massive features and product within plain language plain natural language with the capability of
of what LMM AI model that's building right now. So for for give me a give me a very simple instance. It's like could an user could could a simple users basically tell tell us hey I want to buy bitcoin as the dip in plain language then we we can translate this kind of uh natural language process it and we distill and become a executionable uh trade route with callable API according to all our function that is the basic concept so as a as a tooth kit cell developer can use uh AI skill for natural language workflow. Uh you the developer can use the MCP native agent
integration API with full programmatic control. So the key point is orchestration. The tooth kit allow developers and users with plain language and we on the back on behind we have orchestrate all kind of OKX product and features with security layer so that so that um the the agents that the user build with the capability to uh seemingly using all kind of features and product that's allowed in OKX ecosystem >> is that live today in other words if I go on OKX It's live right now. So I could go on OKX. I could hook up my agent and in plain language say buy Bitcoin and it'll go in OKX and be able to do that.
>> You of course you need an account in OKX. they will start we so the beauty part is it's not basically hey and an agent add an agents and buy bitcoin as div and and we we add the the core part here is not integrating with agent or the the the the crux of this product I would say is adding security layer so that we keep adding checkpoint on every single step we need human control and critical element right before the agent starting to deduct your from within the account. We need user authorization before they making a critical movement of any kind of transfer access or movement of withdrawal. We need uh
mandatory authorization from users. I think that is the c the core part of the agentic trading ecosystem right now. >> Okay. So making sure humans stay in the loop and that it's done safely and securely. How much adoption are you seeing for this? right now. Is this something that you guys have added but that people are hesitant to try? Are you seeing a lot of users move to to adopt this? Like what's what's the the traction been so far on on this product? >> We see is very interesting example. Uh it's a it's a relatively new product. Uh I think I would say uh people are starting to adopt it still a little bit hesitate because it's a relatively new thing. It it it got technological barrier and background. But the
interesting figure to me is um most of the early user within this product uh I would say is retail with no coding experience. I think that is something that we want to see right now. What I think we got around monthly trading volume generated by this kind of agent agent agent product static maybe $15 million US$ 17 million US uh around that. But the the what really exciting at is the retail starting to add automation capability because because what we can see the problem of retail trading are mostly on they are mostly affected by emotion right you can
imagine right so the emotion affect trading so much and affect eventually affect trading performance so if the the user or general user is starting to add hey screen the markets do your back testing execute on an average basis or t b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b b bas on a single trade or hash my portfolio using option give me a give give me a parameters with with with with agent in bill so I think it's add a lot of capability for the user what exciting us is a majority of this user this product a general retails with no experience it's not exactly institutional professional whales that's I think that's I think that might be
trending the trading behavior going forward if it's pick Do you think that this is helping these users with no technical experience who are now leveraging this? Is this helping them get more out of OKX? In other words, are their agents able to leverage things that you guys offer, products and services that a human user might, you know, miss that the value of that like because I I feel like that's a big unlock that AI agents can bring because they are able to see the whole landscape of products and services and understand what what's available there. They can find things and do things that a human user might miss. Have you seen any of that happening or or how do you think about that?
>> I think uh I see people might be underrated their own capability pivously. They they can't do things. They can't do things. Uh for example, um a lot of user story uh classic user stories look like this. Hey, I have a years of experience on reading certain pattern of the markets. However, I don't know how to codify between an automated trading execution agent and very importantly I don't have experience or sufficient knowledge in terms of risk management and positional management
using tools instruments like option because I have zero understanding zero knowledge on option. So I need this kind of I they they would translate the basic trade ideas talk to the agent talk to their model and the agent starting to step by step guide them to become to build a professional investment strategies with risk management tools monitoring and starting on even a best execution exercise. This is something that cannot imagine without a professional institutional to step in.
Uh so right now it's a capability that everyone can use. We we think we everyone will have a huge potential by basically learn the market translated into uh this capability into into their plain language to agent empower the agents. OKX serve at the security and infrastructure layer and help them to to to to to trade. >> So let's take this a step further. You said that right now a big feature a big focus of the platform is to keep humans in the loop. Make sure that humans are approving everything and they have you know secure and and uh scalable trustable agentic operations. What about unsupervised agentic activity? Is that
something that you envision being on the horizon as something you want to work towards? In other words, I can just give my agent a wallet and the keys and like kind of my investment objectives and let that agent manage my capital for me without my supervision or do you not think that that's a direction that you want to go or or just how do you think about that process? I think it's an automation with boundary with boundary means uh you the the agents are and a platform itself as a whole they have to give or identify high risk movement of those agent and make sure humans in the loop and also the platform
for example like OKX we have to enable. We have at least have a features enable for the for for users. For example, these agents can not train. The only capability for this particular agent is sticking for instant for instance and these agents have no capability to withdraw. This agent can only have a read only or trade only authorizations. So this kind of permission control as a whole we have to bail it and let you the agents to call it to use that and let human in the loop. Of course, maybe maybe maybe one day in the future the
user feel is so comfortable. Uh I can delegate a part of the features and part of the author authority to another agents. But I believe the concept is still the same right human need to be in the loop in every single step. But if if you're comfortable enough, you can delegate some of the features, some of the authorization, certain agent at a check and balance. Maybe it will evolve sometime. Gotcha. Okay, that's helpful framing on that. Um, the last question I had on this is, you know, we talked about human supervised training and then if a human gives the permission to the agent, it can do some things without human supervision, but that's up to the human's discretion. But what about agentto agent transactions and and uh
economic activity that's happening? How do you think about approaching that and what's the outlook on that? very important because I think uh eventually the financial eos financial system it going to be evolved and upgrade. So AI will become the intellectual layer and agents or agents will become or or or say crypto crypto will become the economic layer. Uh what it what it means is the agents with because it's empowered by AI right we will will have intellectual elements well within the agent. So the agents starting to need to make decisions and going forward agents need to talk to the other agents or even buy a subscribed service with
another agent. Very simple example I I was a quant trader before and I need to normalize the market data from different kind of market data feed. So there's a agent built for that. It's a it's a very streamlined op trading operation. I think everyone would need it as a quant trader. So I starting to hey this is a agents that I built for years. It works so far for for OKX market data plus IC market data for instance. I want to sell this normalization agent service to another uh quan trader where will we share the same problem. So the the the other quant trader will think hey you you works we this features work just fine this agent web can I just borrow
subscribe this kind of service and but this subscription service already facing a lot of difficulties today with the current payment system it need fiat we need KYC it needs a payment rail banking rail acceptance so uh crypto is I think fundamentally built for agents versus agent trading because it has a identity as an escrow fund. We have limit authorization control bell within we can pay per call rail even have a reputational system within blockchain right so so I think uh going forward with AI as the intellectual layer agents
can control the identity and start paying money we will literally see more agents to interact with each other than helmet eventually >> you know one thing we've talked about a lot on this show is that crypto is quickly becoming a huge part of the global payments infrastructure. And nowhere is that more obvious than in Asia. But if you're actually running a remittance company or a payment business, you know that the hard part isn't moving the stable coins. It's dealing with local banking partners, compliance, liquidity, and all of the operational headaches that come with sending money into places like India and Southeast Asia. That's why today's partner is Saber. They give payment companies stable coinpowered infrastructure to collect and make payouts across Asia without having to build all of that complexity themselves.
They've already processed more than $3 billion in transactions across 40 different countries. So, this isn't just a concept. These guys are actually doing this for real. If you're building payment infrastructure or expanding into Asia, make sure you check out saber.money. Yeah. And there's a lot to unpack in all of what you just said there, but I think that's a really interesting perspective on this. Um, okay. I have a lot of other questions, but I want to kind of just like drill down a bit more on this agentic um agents as the intelligence layer. What does the product look like in that world? Because I think you know right now when we think of an exchange it's built for human beings and all these user interfaces and all these products and services are designed specifically for human beings. But in a world that is primarily driven by a Gentic operators
and AI operating in in trady and crypto and making these transactions. Does the exchange just become a wallet? Does it just become like an an a API layer that an agent will call when when a human speaks to it and tells it to? Like how do you see OKX invol evolving in that world and and what does an exchange mean in a world where Agentic operators are the ones conducting most of the transactions? >> Probably OKX and also the other company not just OKX but also another SAS company for example like Salesforce for instance they need to starting to become more agent agent friendly. So this look so the the fundamental infrastructure
doesn't change. So meaning the matching engine needs to be robust need to be resilient. KYC modules AML essential model need to be uh still need to be robust and ful fulfill all the regulatory uh requirement. The product need to have sufficient and comprehensive API features that allow uh people to call and automated uh to run automated agents or within this those kind of API features and call. So uh we will likely to say there will be a disappear in terms of trading interface. Uh so trading in we we I think OKX including the other peers we spend a really a lot
of time in terms of user trading interface charting buttons cutler tone on this kind of thing but uh going forward including OKX we have to think about in an agent perspective which mean whether or not we have sufficient features for the agents to deploy whether or not we can support sufficient model or whether or not uh we can maintain the balance of optimations uh private key security and also whether or not we can f we can open up the compliance layer to
the agents so that they have the authority that they are allowed to access those kind those kind of products and market from the the OKX doesn't really change but uh it's become a really big wrapper layer to serve the future thousand or even millions of agent. So we need to think about throttle as well and and and and the capacity that how how many agents that we can interact. >> Gotcha. Okay. So, it's no longer just being built with humans in mind and human users, but you also now have to take into account the AI users and making sure you're building things that are friendly for them and that make it OKX attractive as an exchange to AI
users. Um, really interesting perspective here. I want to kind of zoom this out a little bit. I really enjoyed this conversation by the way. So, this is fascinating to me. Um but there are a lot of other exchanges as you said that are trying to move in this direction of pursuing tokenization bringing more real world assets on chain and then also trying to come up with their own solutions for for AI agents and you know this crypto AI integration that we're seeing. Could you differentiate for me a little bit just in your perspective what makes OKX's strategy on these things stand out or different from maybe a Coinbase or Robin Hood or some other incumbents from Wall Street? Um just what what makes OKX's uh strategy here distinct >> think about that what is the ultimate purpose of tokenization because
financial market doesn't change still we got stock we got IPO we got everything with commodities the exchange we're still operating the ultimate purpose of tokenization in my opinion I think is distributions or interconnected with other ecosystem so OKX X I think is play a role and within a distributor for the tokenization product in a global scale for we distribute the strategies very simple we have a really strong breath of the range of product from exchange from wallets
from X layer chain and not just products and but also different kind of onshore regulated entity in global scale. So think about tokenization might be happen in US and we OKX representing we will we will become the bridges to distribute this product to or integrated this tokenizing product within all of the OKX ecosystem from central exchange from apps from X layer ecosystem and tailor made it to the regulated market that allowed it for to those
users. So eventually you will see maybe maybe Coinbase or some of our peers maybe really focus on some market for instant in US our capability as we distribute we distribute through products plus robo regulators like framework. >> Gotcha. Okay. So the the OKX secret sauce is huge global distribution but also a huge amount of regulatory compliance and you know getting access to more jurisdictions and more distribution by being compliant that that makes a lot of sense there. I think that strategy um >> and product and product you need to have a a really strong diversity of product and we put those tokenization instrument
into all this product and make sure everyone can use it for example payment product a card product right a card product ok and OKX pay product uh one of the example is we're putting stable coin as a usable balance within with in some jurisdiction in in a card product this this a perfect example that's how we and distribute a tokenized US trade review within a dayto-day payment ecosystem daily life right using okay against product plus our regulatory framework it's can never can never happen in pivously right can think about >> gota yes product is definitely important too okay this is a crypto show I do want
to bring this conversation back to crypto and digital assets because >> tokenization is a big part of of all of this But I think a lot of people in our audience are seeing all these things happening and these new products, things being launched, but they're wondering, okay, how does this actually involve things like Bitcoin or Ethereum or Salana? And I would love for you to tie that tie all of this conversation back to some of these blockchain ecosystems and the digital assets and just what your outlook is on on that. >> The outlook on that would be people the the the inter the two world startings to influence within each other. Take it for example for our I think I I I call it a stock plus MIMS
ecosystem of building blocks within device space. That's very to be very very surprised to me but uh uh but but you can we can see the mechanics look like this. People want create certain names that tie to certain equities right and uh they put it in a defy building blocks. So people using USDC or other stable coin they want to buy those memes. But however in order to buy buy the memes you have you you or subscribe those mean or receive the air drops in names of those memes you have to buy or you have to own those kind of tokenized act. So uh
wherever have the authority to ask though to to to to to owns and distribute those kind of tokenized equity is starting to buy more shares in regulated markets. Right. So so it look like mems create an other and other rungs of demand to a regulated stock a stock or equity list on the regulated exchange. It's really fun. So what what what we imagine here can we can we can we we can we can look we can look beyond that what what happened is not mess but certain certain um corporate action events arbitrage safe strategies yield
enhancement tokens. Um so in order to receive those kind of benefit in define space you have to put liquidity into a liquidity pool with a collateral or the tokenized equity that subsequently influence the other equity in in in in regulated exchanges. So these two worlds starting to I think reinforce each others what underlying technologies that we're using Ethereum Salaners arbitrums. So these kind of area in this kind of fundamental blockchain layer and directly benefits from this from this uh new area or new flow of capital I would
say. >> Okay. And the other question that naturally follows from this I think a lot of people are wondering and you know this is just your opinion. You're not speaking for OKX or anybody. I just want to know what you think about this. Where do you think the value from that starts to acrue? Does it, you know, collect in some of these deep defy protocols where maybe some of this mimemetic trading and some of this uh these tokenized assets are going? Does it go to some of these L1s like Salana, like Ethereum? Where do you think that value starts to finally acrue as all of this this new economy and new tokenized activity moves into crypto and onto crypto rails? I think is give us an example that hey it's an other level of use case because people used to think it's just an other equity
right so people are wonder why hey to the people starting previously we compare the liquidity of tokenized equity versus traditional equity but right now people starting to starting to think about hey this is merely a collateral and with that collateral I can receive different kind of air drops. So that will be a new wave of innovation and creation and product innovations using that mimmetic trading model. Mims is a start but how about uh recycle borrowing and lending using that as a collateral. So tokenized equity treasurer will become the fundamental
building block as a start where from yield perspective and and the counterparty underlying perspective with that we will see a lot more innovation product by reference of mmetic trading. >> Okay. So you're saying that mimemetic trading is going to just basically bring a whole bunch of new things that we can't anticipate right now and we still have to kind of see where that goes. Is that kind of what you're saying there? we we start to think there will be another wave of innovation that sub that subsequently benefits all kind of fundamental layer that we we have just mentioned whether Ethereum solen L1 or so we we receiving new per record high of TVL and record high transaction volume and also with uh a
record high of interaction and probably we'll see a new wave of different other D apps compared with the D5 1.0 zero, right? So, because everything changes today and we'll see a lot more isolated uh uh element within within Dexis for example that only accept regulated entry and we will see that this kind of evolvement eventually. >> Okay. So, you're just basically saying there will be a huge wave of innovation, rising tide lifts all boats and there's going to be a lot of benefit to go around here. I think that perspective makes a lot of sense. Um, it sounds to me, Lennox, like you're you're kind of implying here that you think we've entered a new crypto bull market and that there's a a huge wave of new
products and new capital coming. Do you think that crypto has bottomed and do you think we are starting a new bull market? Is that is that what you would you think has happened here? >> I'm getting a a bit more optimistic compared with the last bearish cycle. So, think about think about the wave of capital. The phase one capital I would say driven by ICO with right. Stage two capital I would say probably driven by stable fa phase three capital probably driven by bitcoin ETF and instment adoption that's we talked about two two years ago. So right now probably a new wave of new capital that's starting to interest people is like a tokenizations plus composibly define. So every all the all the define polices
layer one chain layer two chain starting to update and upgrade their product and features to cope with this new way of changes and capital and all of course um regulatory charity starting to accept uh tokenization traditional chain starting to build tokenization platform rapper or different kind of current execution starting give a lot more regulatory carity for institutional access for this kind of product. Uh it will become a new wave of capital flowing to crypto space.
Um so that's why I think is uh a lot more optimistic than before in terms of if you're asking about crypto awareness or cryptocato. >> All right. I have a very specific ask for you guys. Somebody listening to this knows the person running marketing at a crypto or AI company that should be advertising with Milk Road. We reach more than half a million investors across everything we do and Q4 is filling up. So don't be the brand that waited too long. And the cool part about sponsoring the show is this episode is still going to be getting plays months from now. So send this to them. Look like a genius and tell them to go to milkroad.com/sponsor.
>> Lennux, I have really enjoyed this conversation. It's really a pleasure to get you on the show and to share all of your insights about all these major things that are happening. And like I said, you have a front row seat to what's going on. So your your input is really appreciated. Thanks for being on the Milkro show. Where can we send people to find more of you and your work online? >> Follow me on Twitter. Thank you so much. >> Yeah, thank you for being here. Linux Live, global chief commercial officer at OKX. Thanks for being on the Milk Road Show. I really enjoyed it. >> Thank you so much. >> And thank you all for joining us. I hope you all learned something today. So until next time, stay safe, stay educated, stay bullish, and we will see you all in the next episode of the Milk
Road Show. Thanks for being here, everyone. Bye. >> Thanks for listening to Milk Road. If you enjoyed the show, make sure you like and subscribe. And if you're struggling to find winners in the market, that's exactly what Milkroad Pro is built for. Our analysts have called some of the biggest winners early, and Pro lets you see what they're buying next, every trade they make, and the research behind every position. Check out Milkroad Pro at the link below. Everything you hear on Milkroad is forformational purposes only. These are our personal opinions, not financial advice, and we may own some of the investments we talk about. Always do your own research and make the decisions that are right for you. See you next time.