In a crowded AI startup space, Mozilla is placing a bet on enterprise appetite with Mozilla.ai, which is building open source tools for prototyping, evaluating and managing governance.
The company launched Mozilla.ai in 2024 with $30 million in funding as part of its effort to lock into the AI boom spawned by OpenAI's ChatGPT success. The goal, Mozilla said, was to create a trustworthy and independent open source AI ecosystem, billing Mozilla.ai as a startup and a community.
Since then, the company has launched Blueprints (templates for AI prototyping), the Any-Agent interface to evaluate AI agents, and Lumigator, a developer tool used for AI model evaluation. Big tech firms like Google, Microsoft and Amazon have heavily invested in proprietary AI ecosystems. Mozilla.ai's open source focus targets enterprises that favor developing in-house solutions.
In an interview with Informa TechTarget, Mozilla.ai CEO John Dickerson discusses the state of enterprise AI, his open source philosophy, and the future of ROI as companies pour billions into AI efforts.
Obviously, there's a lot going on with AI agents and enterprise adoption. Can you give us an overview of what Mozilla.ai is doing to capture interest?
John Dickerso
John Dickerson: It's an exciting time to be in the space. You hear a lot of hype around agents right now. But allowing these agent systems to take actions with some human oversight is delivering value in some verticals. But I wouldn't say that's true for all verticals right now. I think it will be an interesting year, where we will figure out exactly where this tech is working and where it's not working moving forward.
We are trying to build a sustainable business around agentic AI, but we're also mission-aligned with Mozilla. That means supporting open source. That means supporting ownership of data. That means supporting privacy preservation. We think we can thread that needle through a combination of our products and a product that we'll be launching that is aimed firmly at democratizing access to AI in a way that allows folks who are not experts to get some of the advantages of AI technology that's currently being deployed. We're also building a firm open source layer that is delivered to traditional machine learning engineers, software engineers, and coders which allows for folks to have access to any open source model or closed source model, any agentic framework, any style of guardrail they might need to use in terms of evaluation.
Can you talk a little bit about Mozilla.ai's role in reshaping enterprise IT infrastructure as these workloads take shape?
One of the things that we're going to see changing is how data is consumed and how knowledge is consumed.
John DickersonCEO, Mozilla.ai
JD: One of the things that we're going to see changing is how data is consumed and how knowledge is consumed. You may not go to a traditional internal or external website in the same way that you used to. You might have an agent do that. But that agent interacts via open protocols with internal or external data at your company. That will change the way the economics of the internet works. That will increase the speed at which you can aggregate information -- and that can be on external or internal systems. On the infrastructure side, we're starting to see a lot of enterprises getting much more comfortable with using cloud to do everything, but not everybody is doing that. I think we're going to see more internal infrastructure buildouts for on-premises as well within some enterprises, especially the highly regulated ones like banks and healthcare.
And that's a space Mozilla would want to play in because we believe in local models and we believe in data ownership and model ownership. That's where open models and open software can go a long way -- because enterprises that have those concerns also want to be able to have control and understand exactly what they're deploying.
Let's drill down a little on ROI because that's a huge pain point for a lot of companies that have spent a lot of money in the last couple of years trying to implement different AI tools. What is your experience with the ROI part of the discussion?
JD: I realize that I'm leading an AI company that is selling an AI product. So, I should say that AI can solve all problems. But if anyone ever tells you that, it's complete B.S.
What you ended up with was a bunch of enterprises setting budgets for 2023 -- in October, November of 2022 -- basically shifting or lowering their IT spend. And then ChatGPT launched. I'm convinced that what happened was the CEOs, CFOs and the non-technical C-suite just got really turned on by ChatGPT. They went home over the holidays and showed their families, "Hey, look, this is awesome." But they still had frozen IT budgets, but discretionary spending opened up more specifically for AI. And it was a giant hammer used for every single nail out there. When they started deploying these sorts of projects in 2024, that's when the ROI discussion started coming up. And I think you're going to start seeing the same ROI discussion come up in 2026 for the agentic AI tools.
I think you're seeing a lag in measurement. The ROI isn't flowing fully downstream yet. But are we going to see that in 1 or 2 years? I think so.
Shane Snider, a veteran journalist with more than 20 years of experience, covers IT infrastructure at Informa TechTarget.
John Dickerso
