Sean Neville of Catena Labs discusses Catena Labs' mission to build a banking platform purpose-built for artificial intelligence, AI actors. Neville outlines how the platform establishes agentic identity, implements Know Your Agent and Know Your Business, KYA and KYB onboarding and supports programmable money and standards for agent verification. The conversation occurs at the AI Agent Conference 2026 at Agentic Studio hosted by theCUBE and NYSE Wired. Gemma Allen of theCUBE Research and NYSE Wired moderates the discussion and explores practical integration paths for automated agents into payment, treasury and foreign exchange, FX workflows.
Key takeaways emphasize the need for standards-based agent identity and auditable controls so good bots participate safely. Neville states that KYA and KYB processes, interoperable protocols and programmable dollars such as stablecoins enable new commerce while reducing risk from bad actors; they call for early industry cooperation in governance and standards to unlock broader financial inclusion and innovation.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
theCUBE + NYSE Wired: AI Agent Conference. If you don’t think you received an email check your
spam folder.
Sign in to theCUBE + NYSE Wired: AI Agent Conference.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open this link to automatically sign into the site.
Register For theCUBE + NYSE Wired: AI Agent Conference
Please fill out the information below. You will recieve an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for theCUBE + NYSE Wired: AI Agent Conference.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Select your Interests!
add
Upload your photo
Uploading..
OR
Connect via Twitter
Connect via Linkedin
EDIT PASSWORD
Share
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
theCUBE + NYSE Wired: AI Agent Conference. If you don’t think you received an email check your
spam folder.
Sign in to theCUBE + NYSE Wired: AI Agent Conference.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open this link to automatically sign into the site.
Sign in to gain access to theCUBE + NYSE Wired: AI Agent Conference
Please sign in with LinkedIn to continue to theCUBE + NYSE Wired: AI Agent Conference. Signing in with LinkedIn ensures a professional environment.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
Qingyun Wu, AG2ai
Sean Neville of Catena Labs discusses Catena Labs' mission to build a banking platform purpose-built for artificial intelligence, AI actors. Neville outlines how the platform establishes agentic identity, implements Know Your Agent and Know Your Business, KYA and KYB onboarding and supports programmable money and standards for agent verification. The conversation occurs at the AI Agent Conference 2026 at Agentic Studio hosted by theCUBE and NYSE Wired. Gemma Allen of theCUBE Research and NYSE Wired moderates the discussion and explores practical integration paths for automated agents into payment, treasury and foreign exchange, FX workflows.
Key takeaways emphasize the need for standards-based agent identity and auditable controls so good bots participate safely. Neville states that KYA and KYB processes, interoperable protocols and programmable dollars such as stablecoins enable new commerce while reducing risk from bad actors; they call for early industry cooperation in governance and standards to unlock broader financial inclusion and innovation.
In this interview from the AI Agent Conference in New York, Qingyun Wu, founder and chief executive officer of AG2ai Inc., joins theCUBE + NYSE Wired's Gemma Allen to discuss how an open source multi-agent framework is accelerating the shift from AI experimentation to production-grade agentic systems. Wu explains that AG2ai's SDK now reaches 1.4 million monthly downloads, with adoption evolving through three stages — from early developer exploration to enterprise deployment to a striking new phase where AI agents themselves are using the framework. Production...Read more
exploreKeep Exploring
Who is downloading your SDK, and what are the primary use cases for it?add
What enterprise-level penetration or opportunity are you seeing—are organizations building enterprise-grade deployments, and can you share commercial examples?add
How will developer loyalty in 2026 compare with that in 2006—particularly in the agentic AI ecosystem with competing frameworks like LangChain and Crew AI—and what trends are currently being observed and expected?add
>> I'm Gemma Allen with theCUBE and NYSE Wired, coming to you from the Agentic Studio here at the AI Conference in New York, where we are talking all things, builders, bots, buyers, and breakers with this next wave of tech. Joining me now is Qingyun Wu, CEO and co-founder at AG2ai. Welcome, Qingyun.
Qingyun Wu
>> Thank you for having me, Gemma. It's my great pleasure to be here.
Gemma Allen
>> So you are all in on the open source protocol. That has been, I guess, your baby, your swan song for quite a while now. Maybe just fill us in a little bit on first what exactly AG2ai does, and also a little about your own background that brought you to, I guess, co-found this company.
Qingyun Wu
>> Yeah. So at AG2ai, we are enabling the next generation of AI agent systems, especially multi-agent systems. And we are starting that from the open source. So we lay the open source foundation. And right now, we have an open source framework that is supporting millions of developers to build multi-agent systems every day.
Gemma Allen
>> Wow. And you have an impressive track record. I mean, I think I read you have a million monthly downloads on your SDK?
Qingyun Wu
>> Yes, exactly, 1.4 million downloads.
Gemma Allen
>> Wow. So maybe-
Qingyun Wu
>> Monthly downloads....
Gemma Allen
>> help me understand. 1.4 million downloads, who are these folks downloading? Are they enterprise developers? Are they weekend explorers, vibe cutters? Break it down.
Qingyun Wu
>> So basically there are around three stages of our development. At the initial stage are those power users, developers. They are very early adopters of AI. They are exploring those developer tools, for example, through . And we have exploded growth around a year ago. And then we attracted a lot of enterprise interest. And in the most recent three months, we see a very interested trend translating to using agents. Agents are actually using our SDK.
Gemma Allen
>> Wow. Oh my God, agents. That is fascinating.
Qingyun Wu
>> Yes. Yes.
Gemma Allen
>> Give me some use cases. Explain to me how this SDK is being used, what sorts of benefits it's generating. Bring it to life for us.
Qingyun Wu
>> Yeah. So initially are just people building all kinds of toy projects, research-related projects. And then on the enterprise side, we see a lot of use cases around the customers, customer support, customer success, customer voice of customers, and also research, because those are the systems that may benefit from multi-agent design because you can paralyze a lot of tasks. At the same time, folks, you can hear a customer voice in one hour, which may take in the past one week or one month to do a customer survey.
Gemma Allen
>> Wow. And what sorts of penetration or opportunity are you seeing at an enterprise level? Are you seeing many folks building applications for enterprise grade deployments? What sorts of, I suppose, commercial examples are you seeing?
Qingyun Wu
>> Commercial examples. So we have a flagship adoption from Walmart.com. So basically, they are using our SDK to build their recommendation and the content writing system for Wordbound.com. So that's one flagship use case. And we have another use case adopted by a pharmaceutical company. They are using it in their production system about research, autonomous research, and also a voice of customer-related use cases.
Gemma Allen
>> Wow. That's interesting. Walmart, I mean, what a fascinating company, especially in terms of what Agentic AI means for recommender systems, right?
Qingyun Wu
>> Yup. Yeah.
Gemma Allen
>> That is a very broad and opportunistic space.
Qingyun Wu
>> Yes, exactly.
Gemma Allen
>> So Qingyun, your own background, you are a researcher.
Qingyun Wu
>> Yes.
Gemma Allen
>> You were at the Microsoft research arm. You also, I know, do some work in the field and academia. You, I guess, really are somebody who preaches a lot about the kind of longer term value and the collective opportunity of open source.
Qingyun Wu
>> Yeah.
Gemma Allen
>> Talk to me a little bit about how things have evolved for you, especially as you think about reaching new audiences, engaging on a collective thesis at a time when things feel very uncertain, and there's such a race in play.
Qingyun Wu
>> Yes, exactly. That's a very good point. So actually, I think at heart, I'm a researcher. So I got my PhD training from University of Virginia, and I've been a researcher for almost 10 years. So one main characteristic of researchers is they are used to uncertainty. The nature of the work is to explore uncertainty. So for me, myself, I started with open source. I do research around open source projects because I think that's a good way to make real world impact because as a researcher, it's easy to got trapped to your own kind of intellectual game or intellectual things that may not turn into something valuable to the world. So I wanted to make sure I bring the real value to the community and to the industry. So that's why I put all my research on the open source so that I can immediately see the adoption and bring value through that adoption.
Gemma Allen
>> Wow. And I mean, the world of research has always been quite thoughtful, right? It's intentional. And sometimes, it's almost slow by design, right?
Qingyun Wu
>> Yes.
Gemma Allen
>> People want to be very, very thorough and calculated in what they reveal and what they publish.
Qingyun Wu
>> Yes.
Gemma Allen
>> But in this world, especially as we think about what's happening in the world of tech and the race for AI and for agentic AI, it seems as though there is no time for pause, right? There is no time for any sort of collective moment of pause. What are your thoughts? Do you think that's a challenge, especially when you think about things like ethics and risks and all of those other narratives that are keeping a lot of folks up at night?
Qingyun Wu
>> Yeah. I think the way to do research has to almost fundamentally change. I think it could be in a very good way because these days, we can do certain things very, very, much, much faster than before. And, of course, we want to be critical on the other part as well. Folks, in research, there are a lot of more, lower level tedious work about folks collecting data, doing experimentation and verifying some hypothesis. And those kind of work actually can already be automated. So there are actually, in the most recent months, a few projects about auto research and co-scientist things are already being developed. So those are the signals that those kind of work can be readily automated. But there are other parts that agents cannot help us, for example, proposing some hypothesis and on the high level approach and high level methodology. So we, human researchers, still need to put ourselves into it and be critical about it and also be critical about the results returned by the AI agents. So I think certain part of it can be speed up 10 times, 100 times. But the other part, we need to spend even more time. And I think that's fun because you have more time because the certain part of work you don't need to do, and then we put all of our work to the more critical thinking part.
Gemma Allen
>> So let's talk open source for a second and what that means in 2026. We know the traditional model for Linux and other open source protocols, it made a lot of sense, right?
Qingyun Wu
>> Yes.
Gemma Allen
>> It was a collective mission, but also had stickiness.
Qingyun Wu
>> Yes.
Gemma Allen
>> Devs tended to have a loyalty to a particular company, brand, technology, and they built on that for decades, right?
Qingyun Wu
>> Yes.
Gemma Allen
>> In this world, especially in agentic AI, there's like competitors like LangChain and Crew AI. Apologies. What are you seeing? What are you expecting? What does dev loyalty look like in 2026 versus 2006?
Qingyun Wu
>> Yeah. I think the loyalty may still exist, but it will return to its essence. I think the essence is about the value add from the open source software or open source infrastructure. This day is much, much easier for people to switch to another framework, another software, or even people create their own wines. But I think I still believe in collective intelligence and collective wisdom and folks, some of these protocols, for example, this A2A, MCP, those kind of protocols are universally needed almost for anyone. And it does not make sense for me, myself, to recreate another one if there is already a good design. If I have opinions, I can express my opinions to that community and try to integrate my designs. So I think that still works but moves in a much, much faster speed.
Gemma Allen
>> So when we think about open source and we think about the globalization of that, especially the geopolitical conversations that are also happening now between open weight models in China, DeepSeek and others versus what's happening here in the US, there's a lot of value or potential opportunity to see more technology from outside of the US come in, right? It has certainly apparently got some cost advantages, especially as the world of tokenization and the tokenomics is just becoming somewhat crazed, right?
Qingyun Wu
>> Yes.
Gemma Allen
>> But there's a challenge because everything you would build would have to be stored locally. You're not going to send data to China to send it back again. And then you have all sorts of other sorts of infrastructure challenges in allowing that to actually really scale. What do you think? How do you think things will play out in the world of China versus the US and the future of what it really means to be truly open source and have real dev and customer benefits?
Qingyun Wu
>> Yeah. I think China open source has been starting to lead almost, and that kind of, I think, create healthy pressure to all the rest of the world because you may want to make sure you play an important role in the open source world. And I think the good thing about open source is the geopolitical part is a factor. But, ultimately, we look at the value and we look at, if it is convenient for us to use and it is allowed to use. There are open source lessons that we of course need to respect for. And as long as it is within that permission, I think people are adopting models, infrastructures from everywhere, and I think that's healthy enough.
Gemma Allen
>> When we think about what it means to be a builder in this next decade ahead, obviously, cost is huge, right?
Qingyun Wu
>> Yep.
Gemma Allen
>> And again, is cost going to favor the already wealthy, those who've already been building for a long time? What are your thoughts in terms of how people who are really using this technology and building on open source frameworks from the perspective of just true innovation can continue to play and continue to, I guess, also have impact when tokenomics become so hard to kind of navigate, right?
Qingyun Wu
>> Mm-hmm.
Gemma Allen
>> So I guess cost beneficial, it's really not there right now, right?
Qingyun Wu
>> Yeah. Yeah. I think the cost part will come later. That's my opinion. So at this stage, a lot of the companies I know, a lot of builders, I know they... of course, under their own situation, constraints from their own situation, cost is not a constraining factor at the beginning because we are all trying to unlock the fullest capability from the most capable models, and we are just push the frontier capabilities. Of course, once we know, okay, what it can help us do, then we folks must switch to the open source models and do some evaluation and see if they can reach the same level of capability. That's kind of the typical or the effective path I see. And in terms of like, this is natural cost for everyone, for builders, but if you think about the value they add, for example, you should compare that cost to the cost of folks having one additional employee, because they are effectively doing the kind of work that would need a human to do, and that makes it either your time or your employees time.
Gemma Allen
>> Wow. Well, that's certainly a very interesting way of looking at it. I mean, I can't but agree. Qingyun, thank you so much for coming on theCUBE theCUBE and chatting to us. Wish you and the team all the success for the year ahead.
Qingyun Wu
>> Thank you very much, Gemma. Thank you.
Gemma Allen
>> I'm Gemma Allen here at the Agentic Studio at the AI Agent Conference in New York talking all things the next frontier of tech. Stay tuned.
Gemma Allen