Europe wants an AI champion. Can Mistral deliver?

20:28Arjun Kharpal:Owning the chips may come. I think it should come at some point. But for now, we are relying on NVIDIA, which is a great partner to us. And we're testing a few things here and there.

20:38Arthur Mensch:So in terms of self-designed chips, that's more right now just research phase, seeing how they work with your models rather than obviously being deployed in your data centers.

20:46Arjun Kharpal:Yeah, this is not something we know how to do. We know to spec the things that we want. We know the kind of memory bandwidth we need. We know the kind of network requirements we need. We know the kind of memory chips that we need. But then we think other companies actually do a much better job than we do. Maybe one day. So I would say maybe one day,

21:09Arthur Mensch:but really optimizing for cost is going to matter in the entire industry. You have to start thinking of cool names. All these chips have cool names. So you have to start brainstorming some of those.

21:17Arjun Kharpal:We found mistrial already, which I think is pretty cool. But finding others, we're leaving to others to do it.

21:26Arthur Mensch:There is this idea of agents. As AI gets more and more sophisticated, it's going to start doing more things autonomously on our behalf so you can ask it to do longer tasks, etc. But there's a key part of that, and that is orchestration. Think about a conductor in an orchestra. There's loads of different parts that need to be joined together seamlessly. Agents connecting to company systems, their data, the humans in the loop as well. This is all about orchestration, getting this all working within an organization with the ultimate goal of this agent effectively kind of being a digital helper.

22:05Arthur Mensch:So let's talk about the agentic experience. And Arthur, one of the other new products you've launched is called Vibe. So here you're combining two existing products, as I understand it, into what you call this agentic enterprise product. Just run us through kind of what this new product is for the market and how it's going to make an impact in the business.

22:23Arjun Kharpal:So it's an agentic platform that is based on our open source models. So I think that's a really pretty key aspect compared to others, I would say. What is an agentic platform? First of all, it's a place where you can, well, use the models with chat interaction the same way you've been doing it for the last three years. but truly where you start to get value out of AI systems like the one we're doing is when you start delegating some tasks to it so if you're a software engineer it means downloading one of your connecting to one of your pull requests and making it better it means taking a PRD and turning it into a pull request it means orchestrating multitude of agents that are actually doing your jobs on your behalf in a way where the models are well connected to the entire context of software engineering so that means the product surface where the product gets designed.

23:12Arjun Kharpal:That means the documentation of the company. That means the customer requirements, etc. So we are betting on the fact that really the job of delegating tasks as a developer is going to be a very similar job of delegating tasks as someone who is actually not technical. So that's the reason why we brought together two products. One of them, which is our coding agent platform with a common line interface, which is the thing that developers like to use. and then we took Le Chat, which we named in a French way, but if you pronounce it in English, you get various results. And so we realized that we could actually turn it into something quite unified.

23:53Arjun Kharpal:So it's an agentic platform. What is there in an agentic platform? You have the models, you have the business context that is constantly updated over time. You connect them to your various system of records and then you launch agents that are building intermediary representation of what's actually happening in your company. And then you have an execution layer. So you can say to an agent to actually run tasks all the time with triggers on a regular basis, etc. So that execution layer matters. What is kind of specific to what we do, I would say the first thing is it's built on open source model, which means that you benefit from the cost optimization that we brought.

24:30Arjun Kharpal:The second thing is that you get way more control than with other providers. All of the state, all of the data, all of the customization to the user and to the organization can actually be hosted on our customer tenant. We can connect what's deployed on our customer tenant to our token generator and we can connect them to deployment on their GPUs. It's sometimes hard for our customers to deploy on their GPUs. So we've combined a stateful hosted component on our customer tenant and a stateless GPU hosted by Mistral to have the best of both worlds. Control over your data, but then efficiency and cost efficiency and latency improvement.

25:18Arjun Kharpal:I'd say the third thing is really around the customization. You get some benefit by deploying personal agents that are doing things on your behalf. You get much more benefit if you're building a procurement system with the right interface and with the right orchestration that is pinging multiple people at the same time. And so we've built Vibe in a way where you can host your custom application. So your procurement agent, your customer service agent, all hosted in the same place with the same observability, with the right governance. So connected to the same data and you can control what kind of employees get access to what kind of data.

25:54Arjun Kharpal:And we've built it so that our forward deployed engineers can actually quickly build applications that are business applications and that are easily accessible by the customer we work with.

26:04Arthur Mensch:When we have sort of this term, agentic, a lot of the way it's being described is sort of these highly autonomous kind of systems. With Vibe, how are you approaching this idea of kind of autonomy and execution of tasks from the agent versus kind of, how much is human in the loop?

26:24Arjun Kharpal:So autonomy is the one thing that matters because you get more and more leverage when you actually can delegate a long task to an agent. When you're talking about autonomy, one critical aspect in terms of infrastructure is how do you connect the GPUs that are producing the tokens to the execution layer. And you want to do it in a way where you have supervisors on top that are making sure that your agent is only doing the things that it is allowed to do. So for this, we're using the sandboxes that the Koyeb team has built for us, and we're deploying them. It will soon be available in studio as well.

26:57Arjun Kharpal:these sandboxes make sure that you basically have a serverless interface to spin up and down a lot of agents at the same time so autonomy requires the right infrastructure and the infrastructure is both the GPUs, the CPUs and you want serverless deployment there and also the state management when you're deploying an agent it's going to create some state it's going to learn things it's going to write memories it's going to use file systems to organize the thing it is learning whenever it has a query And so the management of state and personalization, the persistence in between two sessions of delegation matters.

27:33Arjun Kharpal:And so Vibe is actually doing it for you. Now, the second thing that matters when you're deploying end-to-end process automation in the company, that in general, it's not only about autonomy, it's also about validation by humans. And so you need a combination of dynamic deployment of agents that are dealing with dynamic inputs. But you also need deterministic gait. You need a human validator in certain places. You need your procurement and your bill to be validated by a human so that someone actually takes the responsibility of validating an agent process. And so what that means is that you need durable process execution that can interrupt themselves and go and ask for human permission, go and send a message to a human so that they can pursue the process with the right indication.

28:19Arjun Kharpal:And you need to do that in a way where the system, the agent can actually interact with multiple humans. so that's why we are using what we call workflows which is the combination of deterministic behaviors and dynamic behaviors in a way that makes the CIO happy in that they know what's happening and they know that the process is being followed but it also makes the business users happy because it works and we are all making that through in a way where when you're deploying an agent in Vibe the the ground source of how it's working is code so you can actually take that code and modify it and a developer can actually understand what's going on, which means you can maintain it.

28:58Arjun Kharpal:Because if you build agent platforms in a way where agents are only defined in the proprietary language of a provider, you're not going to be able to maintain the thing over time. So we've made that vibe is allowing to go from a deployment for a non-tech user to deployment for a tech user so that you can maintain the thing over time. All right.

29:21Arthur Mensch:And as sort of more and more agents get into the enterprise, where do you think some of the biggest changes in terms of enterprise organizational structures or workflows are going to be?

29:32Arjun Kharpal:I think two things. First of all, it's disrupting completely the SaaS business in that overall, once you have the execution layer, once agents can actually operate a lot of things for you, the one thing you should care about as a buyer in an enterprise is to make sure that all of the SaaS providers that actually owns some of your knowledge are making that knowledge accessible to the agents. So there is effectively a growing realization that the connection of models to the system of records of the company is the one thing that matters. So once this is done, you can actually start doing a lot of automation.

30:07Arjun Kharpal:Now, a big problem in enterprises when they are deploying systems like this is that pretty quickly, they cease to be bottlenecked by the intelligence of models and they become bottlenecked by their own organization. So back to the point I was making, you need to orchestrate agents with humans in a way that is observable and in a way where the systems evolve over time and get autocorrect over time. It means that you need to readapt the organization you've had centered around agents that are orchestrating the processes. So the way to do it is basically to look at your core processes and to think about how you would make them faster, how you can automate them more, and where should be the human gates where you maintain the quality and you maintain the innovation.

30:50Arjun Kharpal:So that means thinking about the organizations in a way that is quite top-down, in that you want to take every function, every core business, and think about how to reorchestrate all of the people that are involved in that process around an AI system where the agent is the orchestrator and the coordinator. So that's, I think, a very important aspect. And maybe the second aspect, or third, is that it's changing profoundly the way information is actually being shared in companies. and so you no longer need to ask your colleagues about what's happening provided you have the right context we call it the context engine that connects the models to the different system of records and to the documentation in companies and so that allows you to gain a lot of time because you have a much faster turnaround what it requires though is something that is not technology you need to ask as a business leader, you need to ask your reports and their reports to actually document what they know.

31:50Arjun Kharpal:Because if you don't document what the process should look like, if you don't say what is in your head and what is easy to transmit to your co-workers that are nearby, your agents are going to be a little lost. And that's actually extremely important. It means that as humans, as employees, we do need to give as much context as possible to the agents that will become orchestrators of core processes.

32:21Arthur Mensch:You're about to now, for this final part of the episode, hear a term that maybe you've heard before. Artificial general intelligence or AGI. Maybe a term you heard before. The problem is there are so many definitions of it, varying definitions from depending on who you speak to. More broadly, it's taken to mean AI that is as smart or smarter than humans and not just limited to carrying out one narrow function. Arthur Mench's view is interesting. He says it's not some sort of magical moment that just happens. Rather, it's the direction of progress, not the final destination. His view is that AGI used to be this kind of vague future vision that leaders used to talk about.

33:01Arthur Mensch:But actually, as the technology is developed, the real challenge is much more practical and messy.

33:11Arthur Mensch:When you think about more and more agents, I was speaking to the creator of Claude Code the other day. And I asked him, well, what do you think is going to be the biggest theme this year in terms of AI? He goes, cybersecurity. And I get that because Anthropic has mythos, and that's been a big topic of discussion as well. But cybersecurity, of course, is a top of mind for enterprise. And as more and more agents get into the enterprise, get on sensitive data, perhaps act more autonomously, how, as Mistral, are you thinking about cyber? Is there a product you have or are thinking about that can complement your agentic product, some sort of mythos-style product that can help enterprises deal with some of the cyber challenges that could arise from more and more agentic use?

33:55Arjun Kharpal:So we do because our customers are more and more asking us for multiple things on the cyber security side. Cyber security in AI has multiple pillars. So the first is that you can use AI to detect vulnerabilities. You can use AI to do faster updates of your code bases. And you can use AI to guard against penetration attacks. It's not only about the models. Models matter, of course. And we've seen huge improvements in every category. And I should say our models are a few months behind, but they're really catching up very quickly. Those models are able to detect vulnerabilities, to propose exploits.

34:33Arjun Kharpal:And those models are also able to be used for heuristics to guard against cyber attacks that are penetration attacks against networks. So you need the models, but you also need the harnesses that use the right tools, the right network attack tool. You need the red team harness to test your systems with pen testing. You need the blue team harness to defend your systems. And so those things are products that we're working on with customers that are asking us to provide them urgently with a solution. And we are building that on open source models. And it works very well. and we think that as anything cyber security at the end of the day you want the systems to be open source it's the case for encryption it will be the case for open source it will be the case for ai as well so we're betting on open source technology we're betting on making our models available to everyone so that everyone can understand the capabilities they may have even for attackers and we think that's going to lead us to to a safer system that's the first That's the first pillar.

35:35Arjun Kharpal:Now, cyber security is also about the way you're deploying agents. Because it's very easy to vibe code. When you have an idea, models are great at dealing with a lot of different heterogeneous data sources. So, if you use even open source tools as just a single user, you can create pretty powerful deployment and agents that are doing things on your behalf, or your personal assistants, etc. That's easy if you're a single person. and if you're comfortable giving your data to a close source provider. It becomes much harder if you are an enterprise because you have inherently a tension in between making your employees and your managers able to build applications that are very adapted to the business processes you want to automate, but you want to make sure that they are not doing things that are weakening your security posture.

36:27Arjun Kharpal:So you want isolation, you want monitoring, you want guardrails, You want to make sure that models are accessing only the data that you're giving them access to. And even if you're respecting the access control that the IT is providing, you want to also make sure that you don't have need-to-know problems. When you're connecting your models to all of your system of records and documentation, you're bound to find places where certain employees actually have access to things they shouldn't know about. and it was already the case but because you're reducing friction and because agents can just look for everything you actually need to worry about those things and you need to have dynamic access control systems in place so you have a variety of primitives that needs to be set for an enterprise to be comfortable doing a strong delegation to AI systems and so that requires a combination of systems and model capabilities and so we are at work building them with our customers because oftentimes you need high level of customization, you need the models to deeply understand the legacy systems overall just the pure IT architecture that is oftentimes quite messy in enterprises and so customization, deployment of our security engineers is actually something that enable our customers to go faster so we'll have a few more announcements in the upcoming months about that thing

37:55Arthur Mensch:Arthur as we wrap up I just want to get your take on some of the sort of bigger picture things happening around AI in markets. There feels to me, I don't know if you feel this too, some kind of euphoria right now. I think when you look at perhaps some of the public markets, we've seen some of these memory stocks run up 800 % over the last year. People on X are talking about sort of where to put their money to ride this AI wave. I was speaking to a chip CEO who said even a cab driver in Korea was asking him about when the memory crunch is going to end. Does this kind of euphoria concern you right now?

38:33Arthur Mensch:Because if there is any kind of collapse or you believe there's a bubble that pops, we've often seen in the past ripple waves across things like investment in infrastructure building and other kind of areas. What are you feeling right now?

38:47Arjun Kharpal:Well, we like to think that we're resilient to this kind of euphoria. I mean, there are reasons to be very happy in that the models are really getting stronger. We see in certain domains in enterprises that it's starting to pick up. There's still a lot of work. There's still a lot of viscosity in adoption in enterprises, which means that there's still, let's say, a lot of value creation to be had. Software engineers, they can use AI systems. Industrial engineers, they cannot use AI systems. And so if you actually want to go and tap the 30 trillion market of manufacturing with artificial intelligence, generative AI, you actually need to solve a lot of new problems that we haven't solved yet you need the models to understand tools that are very complex you need the models to understand physics that's recent announcements that we've made investing in a acquiring a company that is building models that understand physics so you know when we talk about agi super intelligence it's all going to be simple i think it's too simple an idea you need strong intelligence in language but you also need a very strong understanding of the physical space.

39:51Arjun Kharpal:We're not there yet. So viscosity is high. Enterprises will make strides in the coming years, but they need to change their organization. There are a multitude of domains where AI is not having an impact yet for lack of capabilities. So we need to build those capabilities. We can. The recipe is there. You need more data. You need more compute. You need expertise. And you can do it. But there's a lot of work to be done. Now, on the question of whether there might be things on the market, etc., and where to invest, etc., we think that as long as we're focused on the creation value for our customers, and as long as we take the right bets in terms of providing them with the compute they need, in terms of making them happy and making them understand that this is not just a technology, it's actually a true industrial revolution, we think we're good.

40:42Arjun Kharpal:We think we're doing the right bets. We think we're resilient to anything that may happen in the market. And of course, we tell our customers that they should care about the supply chain because it's effectively getting more expensive because there's a frenzy of buying. And so that also means we need to accelerate on our compute facilities and that we need to make sure they get good costs and that we get good costs so that they get good prices. So still a lot of work to do. Again, it needs to be full stack. It needs to be on the model side, on the infrastructure side, on the product side. Still a lot of work, still a lot of work for our customers as well.

41:21Arjun Kharpal:But overall, the future looks bright because you can actually go to, we are going to enter a society that can grow faster, that can solve problems that we have been unable to solve at global warming, that can solve a lot of health issues that we're running into because of longevity, that can actually build better planes. We've announced the partnership with Airbus today that can actually better lithography machines, that can build better cars, safer cars. We've announced the BMW partnerships today as well. And the future is bright, but we all need to work together and we need to make sure that everything is built in a way that is fair.

41:58Arjun Kharpal:And to build fairness, building an open source is probably the right bet.

42:02Arthur Mensch:Does the OpenAI IPO upcoming, whenever that happens, mean something for the industry?

42:09Arjun Kharpal:does it mean something i think uh i mean of course ipor always looked at uh we as a company are a private company and can stay a private company for longer uh we uh but of course that's

42:22Arthur Mensch:something that will look with interest for sure and just finally um as we wrap up arthur you mentioned agi very briefly and we haven't got time to go into it in depth but this artificial general intelligence idea was something that was spoken about a lot in the last, say, two, three years by some of the top AI labs out there as well. It feels like it's gone out of fashion a little bit at this point. You mentioned something interesting, like to get there, we need to understand the physical world. There's a lot of talk of things like world models emerging as well in order to get to AGI. What is the current state of, I guess, debate and progress in terms of the industry achieving what it believes to be AGI, given I know there's multiple definitions of the phrase.

43:07Arjun Kharpal:I think AGI was a PowerPoint concept. And so the reason why we've heard about AGI for like 15 years in between 2010 and 2025 is because the thing was not really working yet. But now it's actually working and we can turn it into something that has value for customers. and so suddenly it becomes it's not no longer a PowerPoint term but suddenly the thing becomes way more complex because if you're if you're talking about something that does not exist yet you're talking about something abstract you would rather have like a single term instead of saying it's all going to be very complex it's going to have to understand many different things etc nowhere at the stage where actually you just need to take the models and to connect them to business data and to understand what the people wants to do with it to actually build value but it means you're running into a lot of things that were not anticipated in the PowerPoints of 2010 so you need them to understand physics you need them to understand the human behavior you need them to be connected to all sorts of the 50 years legacy of software that we've been building so it's all very complex it's all very messy it's holding to organization and people which is even more messy than what you can build with technology and so at the end of the day I don't think intelligence matters I think what matters is empowerment.

44:21Arjun Kharpal:It's a technology that can do many things, but there's a lot of plumbing to be done to make it work. So suddenly you move from a very abstract concept and what I think is a little bit messianic as well into something that is very concrete and where you need the combination of people that do not understand technology but understand their business with people that understand their technology but do not understand the business. So the business of their customers. So they need to share our knowledge. and that's the way we can build AI that is actually useful. That's the way we can accelerate technological progress.

44:53Arjun Kharpal:So if you take AGI as the definition of accelerating technological progress, which is rather a direction than a point of arrival, then really we are building AGI. But it's really a direction and it's a very messy direction. There's a lot of things to be built and the frontier is enormous.

45:09Arthur Mensch:Arthur, I've got a whole list of things we need to talk about next time. We catch up robotics, physical AI and a ton of other things, but we'll save that for the next conversation. Happy to have it. Thank you so much for joining me. We are going full steam ahead with the Tech Download podcast. We have some big, big interviews. I mean big, coming up over the coming weeks as well. So make sure you subscribe, follow, do whatever you need to do to keep up to date. It'll be on YouTube, it'll be on cnbc.com and anywhere you get your podcasts. If you want to talk to me about some of the things that we're discussing in this episode or have suggestions on what we should talk about next, You can get in touch with me directly.

45:47Arthur Mensch:Just search Arjun Karpal on LinkedIn, on TikTok, Instagram or X. Thanks for listening and watching and we'll catch you next time.

46:16Thank you.