When we set out to create Starburst, our ambition was to turn the data industry upside down, allowing companies to perform data warehousing analytics, without the data warehouse. That spirit of disruption has positioned us well to deliver the data foundation for Enterprise AI.
Today, as we announce crossing the $100 million ARR milestone, what stands out to me is not just the number, but also how we got here. This major milestone validates our long-standing belief that enterprises need governed, federated access to their data wherever it lives. Since our founding, this demand has been driven by BI dashboards, reporting, and ad-hoc SQL use cases. However, more recently, we have seen a noticeable shift towards AI-driven applications where agents are the new data consumer. Starting from zero around this time last year, AI revenue now accounts for a $20M run rate for Starburst and we are excited to see where it goes from here.
From analytics to AI, the revolutionary nature of our mission has only grown and deepened. Last year, I set out a vision for Starburst as the data foundation for AI, placing us at the very heart of the AI transition. Today, our technology powers AI workloads that drive some of the most critical insights and agentic workflows on the planet, and we’re just getting started.
Why AI workloads are driving this moment
First, let’s look at the numbers. Starburst’s success is broad-based, cutting across diverse customers, geographies, and data workloads. Over the past year, we grew overall ARR by nearly 40%, and our financial services business grew by over 80%. Meanwhile, our business outside the US more than doubled.
Those numbers are not an accident. They are the direct result of enterprise customers who need a federated context platform that supports AI.
There’s another shift underway as well. Increasingly, organizations are moving beyond traditional BI workflows toward agentic AI systems that can answer questions, retrieve context, and support decision-making directly. That change is less about replacing dashboards for the sake of it and more about how people want to interact with data.
Why AI needs access and context
For that to work, AI needs context. Specifically, it needs governed, reliable access to data across systems. Many BI environments evolved as layered, disconnected tools over time. Moving to an AI-driven model requires something more coherent. In other words requires a strong data foundation with universal access across data sources.
That transition is not cosmetic. It represents a structural shift from a static approach to data with BI, to a conversational, interactive one with AI.
Financial services drives shift towards agentic workloads
This is exactly what we’re seeing, and we see it most clearly in regulated industries, but the effect is mirrored in every organization and every business vertical.
At some of our largest financial customers, they are aggressively deploying AI for fraud detection, customer engagement, and risk management. But they cannot afford the compliance risk or latency of large-scale data migration. Into this space, data federation provides the ideal solution–the ability to access data anywhere and use it to drive AI workloads towards a new model of conversational, interactive reporting.
Starburst leads the way
As it takes hold, it is creating new demands on enterprise data architecture. In this sense, there is a convergence between our technology’s capabilities and the unique needs of the AI revolution taking hold around the world. As AI systems move into production, the true bottleneck is no longer the models themselves. Instead, the limiting factor is the contextual data needed to make the models operate effectively, and especially access to governed, contextual data drawn from heterogeneous environments. And we’re going further, with AIDA, a conversational interface built directly into the Starburst product that truly pushes the potential of AI beyond BI.
We see this playing out in several ways, and it’s worth reflecting on the ways that Starburst uniquely meets the moment in terms of AI and Agentic AI workloads.
1) Starburst overturns old data paradigms around data access
Starburst’s mission was always a revolutionary one, and it has always centered around data access. When we started, old narratives around default data centralization persisted, holding back the potential for workloads of all types. Data was becoming more valuable to businesses every year, but also exponentially harder to access and harder to use. It lived in more places, across more systems, under tighter governance constraints.
In answer to this, the industry’s default answer was still centralization. Move the data, copy it, rebuild pipelines, then try to govern what you created.
The importance of data foundations for AI
That model was built for a different time, and it always had its problems. With the shift to AI, it broke apart entirely. Data federation, already an answer for analytics, is perfectly suited to the needs of AI.
AI requires access to contextual data to operate correctly, and it’s that data access that is proving the major bottleneck. In other words, it’s not ever-more performant AI models that are constraining AI growth. It’s the foundational access to the data to make the models operate.
The data foundation is the place where AI success is being made or broken, and it’s an area uniquely suited to Starburst, our technology, and our mission. In this sense, we’ve always been focused on foundational transformation of the data architecture that powers our world. In doing so, we created technology that not only solves historic bottlenecks around data centralization using data federation, but also creates the perfect foundation for universal data access needed for AI.
2) Starburst’s technology makes AI projects achievable
To support enterprise data foundations, you have to start with strong core technology.
For us, that began with Trino, a distributed query engine built to access data across systems at scale.
From there, we built a platform around that core, with a goal towards achieving a universal data foundation capable of:
- Data governance from the ground up
- Performance at scale
- Support for a variety of data sources
- Support for data lakehouse technology, particularly Iceberg
- Reliable data ingestion and data maintenance
- Interoperability across different data platforms
At the same time, workloads were evolving. AI workloads sit at the far end of the use-case curve. They are continuous and context-driven. They require broad access, fresh data, historical depth, and strict policy enforcement. When the foundation is weak, they stall quickly.
Throughout, our objective has remained straightforward. Build data infrastructure capable of supporting the most demanding enterprise workloads. Today, those workloads are AI.
3) Starburst is built for the future of agentic workloads
This brings us to the next point, the rise of agentic AI. As AI evolves, so do the demands placed on data.
Agentic workloads are not passive. They reason, retrieve, and act. They require continuous, governed access to enterprise data across systems, clouds, and environments. They cannot operate inside silos, and they cannot depend on old approaches to data centralization.
This is exactly why Starburst was built, and the rise of agentic AI has merely brought the core problems we have always solved into the foreground.
To meet this, we have embraced AI at every level, both in terms of the workloads we support, and with the product. For example, we have extended our platform with an agentic intelligence layer embedded directly into the data foundation. AI-powered retrieval, SQL generation, and agent APIs now operate within the same federated, policy-aware architecture that enterprises rely on every day.
At the same time, data products are becoming the de facto, stable interface for intelligent systems. AI performs best when it operates against governed, well-defined assets aligned to business domains. AIDA is part of this approach, bringing an interactive, conversational interface that directly replaces static BI dashboards with something more human.
The future of AI will not be built on isolated models. It will be built on strong, accessible data foundations that allow for business context and interactive conversation.
4) Starburst is versatile and built for our customers
Technology only matters if it works in production.
In this area, we take our insights from our customers, and they are very clear. AI initiatives succeed when they are grounded in governed, federated access to real enterprise data backed up by the requisite business context.
This is exactly what we see with all of our customers, but let’s look at two major examples from complex enterprises that use Starburst in production.
Citi
“As a strategic investor in Starburst, we are confident in their vision of enabling federated data access for enterprise AI,” said Siris Singh, Managing Director and Global Head of Markets Strategic Investments at Citi. “In an industry where data access and governance are paramount, Starburst’s approach eliminates the need for costly, risky data migrations while accelerating AI initiatives at scale. This partnership allows us to innovate at the intersection of technology and financial services, and we’re excited to support Starburst as they continue to grow and set the standard for AI-ready data infrastructure.”Our collaboration with Citi reflects this alignment. As Siris Singh, Managing Director and Global Head of Markets Strategic Investments at Citi, notes, “As a strategic investor in Starburst, we are confident in their vision of enabling federated data access for enterprise AI”.
He goes on to say, “In an industry where data access and governance are paramount, Starburst’s approach eliminates the need for costly, risky data migrations while accelerating AI initiatives at scale. This partnership allows us to innovate at the intersection of technology and financial services, and we’re excited to support Starburst as they continue to grow and set the standard for AI-ready data infrastructure.”
With Citi, the alignment around architecture and governance is what allows AI to move from isolated use cases into the core of how large institutions operate.
Switch
At Switch, Starburst underpins intelligent infrastructure. As Zia Syed, Chief Technology Officer for Switch, states, “Starburst is foundational for how we’re rolling out AI within our own operations. It gives us the trusted data foundation we need to automate with confidence.”
Switch has built Starburst into its data foundation from the ground up, providing a solid layer of data access and governance that helps AI succeed in its production environment.
Where is Starburst going next?
Crossing $100 million in ARR reinforces something we have believed for a long time. Federated, governed access to distributed data is not a niche architectural choice. It is becoming a baseline requirement.
Nothing meaningful happens in analytics or AI without a strong data foundation. As workloads become more intelligent and more operational, and as AI workloads supplant other approaches like BI dashboards, that foundation matters even more.
We are also seeing something broader. Organizations are beginning to rethink not just where data lives, but how insight is surfaced and how intelligence is embedded into everyday workflows. That evolution is still early, but it is accelerating.
We will continue building toward that future.
For now, I want to thank our customers, partners, investors, and the Starburst team. This milestone reflects the work we’ve done together and the trust placed in the architecture we’re building.
