The Announcement
Starburst used its annual AI & Datanova 2026 conference in Miami to recognize ten customers and partners through its 2026 AI & Data Visionary Awards. The honorees span financial services, healthcare, retail, and digital banking, and include household names such as GEICO, Citizens Financial Group, and Highmark Health alongside partners Dell Technologies and NVIDIA. The awards are not ceremonial gestures. They are a deliberate showcase of production-grade deployments built on Starburst’s federated query platform, and they offer a useful window into how enterprises are actually operationalizing AI at scale rather than nursing proofs of concept.
Our Analysis
Data Mesh and Federation Are Moving from Aspiration to Architecture
The most telling theme across the customer winners is not the AI itself. It is the data foundation underneath it. Vizient built an internal data marketplace on a data mesh architecture, enabling clinical, supply chain, and quality teams to discover and publish governed data products independently. Citizens Financial Group deployed Starburst not as a query tool but as the federated access layer for an entire enterprise data platform, with governance and an agentic conversational interface layered on top. Inter&Co replaced a constrained query environment with federated access to more than 34 petabytes of data, giving thousands of data consumers self-service capability without routing every request through a central data team.
These are not isolated experiments. They are a consistent answer to the same underlying problem: enterprises cannot build AI systems on data that is siloed, ungoverned, or too expensive to query at scale. ECI Research’s 2024 analysis found that the average enterprise now uses more than two public cloud platforms, with Kubernetes, Snowflake, and GenAI often coexisting across a patchwork of teams, workloads, and tools. Federated query architecture is the practical response to that fragmentation. You cannot consolidate all that data into a single warehouse economically or quickly, so you query it in place with governance applied at the access layer.
What This Means for ITDMs
For IT decision-makers, the Highmark Health case is the one worth studying. Nik Acheson’s approach, fund AI innovation through efficiency gains, rationalize the platform, and commit to open architecture, is a financially disciplined model that most enterprise AI programs have not yet internalized. It avoids the common trap of seeking a separate AI budget, instead treating AI as something that earns its own runway by freeing up resources elsewhere. That framing matters when boards are asking AI programs to justify their spend.
The Spreetail story adds a different dimension. Overhauling a cost-heavy, siloed data architecture and migrating to Starburst Galaxy as the unified query engine for AI, ML, and internal applications in six months is an unusually fast execution. It suggests that the migration complexity argument, which vendors and consultants have used to extend timelines for years, may be overstated for organizations that have made clean architectural decisions upfront.
For ITDMs evaluating their own positions, the GEICO recognition signals something worth watching. GEICO is orchestrating AI agents across internal systems and customer products using multi-agent design and open standards like Model Context Protocols. Most enterprises are still connecting isolated AI use cases. GEICO is operating at a coordination layer above that. The gap between where GEICO is and where most enterprises are is not a technology gap. It is a governance and architecture gap.
What This Means for Developers
The partner awards reveal where the technical investment is concentrating. The NVIDIA partnership is the clearest signal: agentic AI workloads create new performance requirements for data infrastructure, and GPU-accelerated query execution is being positioned as a production necessity rather than an optimization. Developers building agentic systems need to account for the data retrieval layer, not just the inference layer, as a performance constraint.
The Dell OEM partnership addresses a different pain point: the assumption that enterprise AI requires cloud-first infrastructure. The ability to run Starburst analytics directly on Dell’s AI Data Platform, without migrating data to a cloud data warehouse, matters to the significant share of enterprise workloads that cannot or will not move to public cloud for cost, latency, or regulatory reasons. According to ECI Research, 61.8% of enterprises run hybrid deployments, while an additional 10% operate in multi-cloud environments. Building for a single deployment model means building for the minority.
For developers specifically, the Artefact partner recognition points to a skills and methodology gap. Artefact is described as covering the full journey from data foundations to autonomous agents that reason and execute end-to-end workflows. The existence of a dedicated AI-native services partner category at this conference is an acknowledgment that most engineering teams do not have the internal expertise to go from data architecture to agentic production systems without external help. That is consistent with broader market data: according to ECI Research, 82% of AI/ML teams report skill gaps in AI/ML operations, with 31.3% describing these gaps as extremely prevalent. The managed services and implementation partner ecosystem around platforms like Starburst is going to grow significantly as a result.
Competitive Positioning
Starburst is making a clear architectural bet: open standards win. The platform is built on Trino and Apache Iceberg, and the GEICO recognition explicitly calls out the use of open standards like MCPs. This is a deliberate counter-positioning against proprietary lock-in from hyperscaler data platforms. The bet is that enterprises with complex, distributed data estates will favor portability over integration convenience. Given the hybrid and multi-cloud deployment reality described above, that bet could be well-calibrated.
What’s Next
The Agentic AI Inflection Point Is a Data Infrastructure Problem
The GEICO deployment is the most forward-looking indicator in this award set. Multi-agent orchestration at enterprise scale requires not just capable models but reliable, low-latency, governed data access at query time. That is an infrastructure problem before it is a model problem. ECI Research’s 2025 AI Builder Summit survey found that two-thirds of enterprise AI leaders have already implemented multi-agent collaboration in live or pilot workflows. If that adoption rate is accurate, the demand for data infrastructure that can serve agents reliably, with appropriate governance and access controls, is arriving faster than most data platform vendors have prepared for.
The Federation Model Will Face Its First Serious Governance Test
As federated data architectures mature in production, the governance question becomes acute. Querying data in place across multiple clouds, on-premises systems, and partner environments is operationally appealing, but it distributes the governance surface area significantly. The organizations that will extend their leads in 2026 and 2027 are those that treat governance as a first-class architectural requirement rather than a compliance overlay. Vizient’s data mesh model and Citizens Financial Group’s agentic-ready governance layer both suggest that the most mature Starburst deployments are already treating this seriously. The broader market has not caught up. That gap is where competitive differentiation will be won or lost over the next 18 months.
Stay Ahead of Application Development Trends
Get weekly analyst insights, research notes, event coverage, and AppDevANGLE updates delivered directly to your inbox.
Subscribe for Weekly Insights
Join technology leaders, practitioners, and GTM teams following the trends shaping modern software delivery.
Looking for deeper research access?
Explore ECI Research reports, survey insights, and market analysis through the ECI Research Portal.
