Analyst(s): Brad Shimmin
Publication Date: April 17, 2026
Starburst announced AIDA (AI Data Assistant), an AI assistant designed to reason across distributed enterprise data sources without requiring data movement or centralization[1]. The product targets a real bottleneck: most enterprise agentic AI initiatives stall not because models are weak, but because they can’t reach the right data. With accuracy and hallucination risk still the top GenAI reservation among data leaders, AIDA’s bet on federated reasoning over data consolidation is a direct challenge to the data lakehouse orthodoxy pushed by data platform players such as Databricks and Snowflake.
What is Covered in This Article:
- Starburst’s AIDA and the federated data reasoning approach versus centralized data platforms
- Why enterprise AI accuracy depends more on data architecture than model selection
- Competitive implications for Databricks, Snowflake, and the semantic layer ecosystem
- The gap between agentic AI ambition and the write-back infrastructure reality
The News: Starburst launched AIDA, which the company describes as the first AI assistant that reasons across all enterprise data for fast, accurate decisions. Rather than requiring organizations to copy or move data into a single repository, AIDA uses Starburst’s federated query engine to let AI models access data wherever it lives, across data lakes, warehouses, SaaS applications, and on-premises systems. The product is positioned as a business-user-facing tool, not a developer platform, aiming to let non-technical users ask complex analytical questions and receive answers grounded in actual enterprise data.
The timing is notable. According to Futurum Group’s 1H 2026 Data Intelligence, Analytics, and Infrastructure Decision Maker Survey (n=818), accuracy/hallucination risk remains the top GenAI reservation at 24.9%, with explainability (20.2%) close behind. Meanwhile, 56.7% of organizations are actively optimizing inference costs, suggesting that the economics of querying data in place rather than replicating it into centralized stores could resonate with budget-conscious buyers.
Can Starburst’s AIDA Crack the Enterprise AI Data Access Problem?
Analyst Take: Starburst is making a structural argument, not just a product announcement. AIDA’s thesis is that the dominant pattern of centralizing data before applying AI is architecturally wrong for most enterprises. If federated reasoning can deliver comparable accuracy without the cost and latency of data consolidation, it undermines the gravitational pull that Databricks and Snowflake depend on to grow storage and compute revenue.
Why Agentic AI Data Architecture Matters More Than Model Choice Right Now
The enterprise AI conversation has been disproportionately focused on which foundation model to use. OpenAI leads adoption at 64%, followed by Azure OpenAI at 62% and Google Gemini at 54%, according to Futurum Group’s 1H 2026 AI Platforms Decision Maker Survey (n=820). But model selection is becoming commoditized. The harder problem is getting models access to the right data with enough context to produce trustworthy answers. Futurum found that LLMs reasoning over an OSI-governed semantic layer achieves up to 3x higher accuracy compared to those parsing raw data tables (‘The Semantic Layer is Finally Code, Not Just a Concept,’ March 2026). AIDA’s federated approach sidesteps the data movement tax, but Starburst will need to demonstrate that query-in-place can match the accuracy gains that semantic layers and curated datasets deliver. That’s the real benchmark, not speed or cost alone.
The Write-Back Wall That Agentic AI and Federated Reasoning Won’t Solve
AIDA is a read-oriented tool. It reasons across data to answer questions. But the direction of enterprise AI is moving toward agents that don’t just analyze but act. According to Futurum Group’s 1H 2026 Data Intelligence, Analytics, and Infrastructure Decision Maker Survey (n=818), the top infrastructure bottleneck for agentic AI is that agents cannot write back to systems of record (24.6%). Starburst’s federated query engine was never designed for transactional writes, and AIDA inherits that limitation. For knowledge management and analytical queries, this is fine. For the 72% of organizations researching, piloting, or deploying agentic AI per Futurum Group’s 1H 2026 AI Platforms Decision Maker Survey (n=820), a read-only agentic AI assistant will eventually feel like half a solution. Starburst needs a credible roadmap for how AIDA evolves from answering questions to triggering actions.
Databricks and Snowflake Should Pay Attention, Not Panic
Starburst won’t displace Databricks or Snowflake overnight. Both have massive installed bases, deep ecosystem integrations, and increasingly sophisticated AI capabilities of their own. But AIDA exposes a vulnerability in the centralization thesis: enterprises with data spread across dozens of systems don’t want to hear that the answer is to move everything onto one more platform. The global Data Intelligence, Analytics, & Infrastructure (DIAI) market is projected to grow at a 16.9% CAGR to surpass US$1.2 trillion by 2031, according to Futurum’s 1H 2026 Data Intelligence, Analytics, & Infrastructure Market Sizing & Five-Year Forecast. There’s room for a federated approach to carve out a meaningful share, particularly among organizations with strict data residency requirements or complex multi-cloud estates. The competitive question isn’t whether federated or centralized wins; it’s whether Starburst can build enough AI-native features on top of its query engine to keep pace with platforms that control the full stack.
The Agentic AI Future: Beyond Chatbots to Autonomous Intelligence
Perhaps the most forward-looking aspect of this news is the introduction of the agentic layer and the support for the Model Context Protocol (MCP). The industry is rapidly moving away from simple chatbots toward agentic AI workflows where AI can take actions. Starburst’s decision to expose its engine via an MCP server is a brilliant strategic move. It ensures that AIDA can function as part of a larger agentic AI ecosystem, connecting to third-party agents and proprietary applications.
This bring your own model (BYOM) philosophy is essential. During demonstrations, Starburst used Claude 4.6 Sonnet to analyze complex Formula 1 and financial datasets. However, the platform remains model-agnostic, allowing businesses to optimize for cost, performance, or accuracy based on their specific needs.
Furthermore, Starburst is introducing skills and actions within AIDA. This allows the agentic AI assistant to move beyond merely answering questions to executing tasks. For instance, a user can define a workflow where AIDA identifies a budget anomaly and triggers a downstream action in a custom system. This transition from conversational intelligence to intelligent action, facilitated by an Agent Gateway, is the next frontier of enterprise productivity.
What to Watch:
- Accuracy Benchmarks: Will Starburst publish independent accuracy comparisons between AIDA’s federated reasoning and centralized alternatives such as Databricks and Snowflake within six months, or will this remain a marketing claim?
- Semantic Layer Integration: Does Starburst partner with semantic layer vendors such as AtScale or dbt, or try to build its own? The 19% projected growth in the semantic layer market makes this a strategic fork in the road.
- Agentic Roadmap: Can Starburst extend AIDA beyond read-only analytics into write-back and action-triggering capabilities before agentic AI expectations leave it behind?
- Enterprise Proof Points: Will AIDA gain traction with data-sovereign industries such as financial services and healthcare, where federated access has natural advantages, or will it struggle against incumbents with deeper vertical solutions?
- Model Context Protocol Adoption: The success of the agentic layer will depend heavily on the MCP’s long-term success. As Starburst positions itself as the gateway to enterprise data, its value grows in proportion to the number of third-party agents that can seamlessly plug into its servers.
Read the full press release for more information.
Disclosure: Futurum is a research and advisory firm that engages or has engaged in research, analysis, and advisory services with many technology companies, including those mentioned in this article. The author does not hold any equity positions with any company mentioned in this article.
Analysis and opinions expressed herein are specific to the analyst individually and data and other information that might have been provided for validation, not those of Futurum as a whole.
Other Insights from Futurum:
Can AgentExchange Cement Salesforce’s Lead in the Agentic AI Platform Race?
Does Horiemon Ai’s Simplicity Signal a New Minimalist Trend in Web AI?
Does CodeRabbit’s Codex Plugin Signal the End of Context-Switching in Code Review?
Author Information
Brad Shimmin is Vice President and Practice Lead, Data Intelligence, Analytics, & Infrastructure at Futurum. He provides strategic direction and market analysis to help organizations maximize their investments in data and analytics. Currently, Brad is focused on helping companies establish an AI-first data strategy.
With over 30 years of experience in enterprise IT and emerging technologies, Brad is a distinguished thought leader specializing in data, analytics, artificial intelligence, and enterprise software development. Consulting with Fortune 100 vendors, Brad specializes in industry thought leadership, worldwide market analysis, client development, and strategic advisory services.
Brad earned his Bachelor of Arts from Utah State University, where he graduated Magna Cum Laude. Brad lives in Longmeadow, MA, with his beautiful wife and far too many LEGO sets.
