The rise of AI agents dramatically increases query loads on data warehouses, exposing closed data stacks to steep cost and performance challenges. Fivetran’s Chief Product Officer calls for open data infrastructure and smarter workload routing to sustain analytics reliability and cost-effectiveness in this new era.
- AI agents increase query volumes drastically, stressing closed single-engine data stacks
- Open infrastructure allows routing queries to cheaper or specialized engines for cost control
- Rejecting lockdown instinct in favor of openness and semantic discipline unlocks productivity
Infrastructure signal
The arrival of AI agents that run exponentially more queries than human analysts is revealing critical shortcomings in legacy closed data stacks. These environments often funnel all analytical queries through a single, often expensive compute engine that is ill-suited for sustained high-volume query execution. This leads to disproportionately high costs and inefficiencies, analogous to using a high-performance sports car for routine yard work.
Fivetran’s approach emphasizes enabling an open data infrastructure where multiple compute engines coexist and query routing dynamically directs workloads to the most cost-effective option. By implementing this strategy, organizations can mitigate the escalating compute expenses while improving infrastructure resilience and flexibility, essential for the AI-driven analytics era.
Developer impact
Developers and data teams face a paradigm shift in how analytics workflows operate under agentic loads. The spike in query volume demands new thinking around query efficiency, orchestration, and platform openness. Instead of restricting access and throttling queries to control costs, teams are encouraged to innovate through more sophisticated semantic modeling and smarter data orchestration that leverages cheaper compute tiers when possible.
Working within an open ecosystem also broadens the choice of integration points and tooling options, reducing friction and enabling developers to build more scalable, maintainable data pipelines. The emphasis on semantic discipline—clearly defined business logic and metadata—further enhances the accuracy and cost-effectiveness of AI-driven queries, directly benefiting developers tasked with supporting complex, agentic analytics workloads.
What teams should watch
Data and infrastructure teams should closely monitor the economic and operational impact of rapidly growing agentic query volumes, particularly on cloud compute expenditure. They must evaluate their current data stack’s flexibility to route diverse analytical workloads across multiple compute backends and consider investing in open architectures that facilitate this capability.
Additionally, teams should avoid premature cost-control measures that could stifle innovation and instead accelerate the adoption of semantic frameworks and benchmarking tools that provide transparency into query routing efficiency and cost attribution. Watching vendor initiatives—such as Fivetran's Open Data Infrastructure Data Access Benchmark—will also help teams track industry best practices for managing AI-driven analytics cost pressures.