Last week, StarTree Cloud officially reached 1B Average Daily Queries. Those sub-second insights power social media engagement metrics, real-time game leaderboards, LLM token observability, fintech fraud alerts, digital advertising funnel analysis, real-time logistics tracking, personalization engines, network telemetry analysis, live customer-experience monitoring, and more. For teams building real-time, customer-facing analytics at scale, it’s a validation of architectural choices, product maturity, and sustained customer expansion across production workloads.
Here’s what’s behind the number:
Where We Came From: Growth and Momentum
One year ago, StarTree Cloud was processing approximately 1.5 billion queries per week. Today, that number has grown to just over 7 billion queries per week, representing a 367% year-over-year increase in weekly query volume.
This growth has not been driven by a single breakout customer or short-lived traffic spike. Instead, it has come from new customers who have moved from batch to real-time insights, and existing customers steadily expanding their footprint on the platform—adding new use cases and onboarding additional teams. As those workloads scaled, query traffic scaled organically alongside them.
How We Calculate “Average Daily Queries”
The 1B Average Daily Query figure reflects sustained production load over time. We calculate ADQ by measuring total query volume across a rolling seven-day window and dividing by seven to arrive at a true daily average. Using this approach, StarTree Cloud now consistently averages approximately one billion queries per day. This method smooths out natural day-to-day volatility. While the platform has processed well over one billion queries on certain individual days, ADQ represents durable, repeatable scale rather than short-lived surges.
Ingest Volume and Storage Footprint
Query volume tells only part of the story—sustaining more than one billion daily queries requires an equally strong ingestion and storage backbone. In 2025, StarTree Cloud is handling peak ingestion rates of approximately 40 million messages per second, with an average sustained rate of 30 million messages per second. This level of continuous ingest supports high-velocity event streams across observability, customer analytics, fraud detection, and real-time machine-learning feature delivery. On the storage side, a single customer footprint has grown to over a petabyte, reflecting both higher ingest volumes and longer data retention across production workloads.
The 1B ADQ milestone reflects a mix of StarTree-managed SaaS and BYOC deployments. However the majority of today’s query volume is driven by BYOC environments. This mirrors what we see across large enterprises: organizations want the operational simplicity of StarTree Cloud while retaining direct control over networking, security, and data residency in their own cloud accounts. As those environments scale into core production systems, query volumes scale rapidly with them.
Why This Matters and What’s Next
From the beginning, Apache Pinot and StarTree were built around a core belief: data creates value when it produces insights, and insights emerge through rapid, repeated interaction with data. The true measure isn’t how much data you store, but how many questions you can ask, how quickly you get answers, and how many people can ask at once. Queries—not terabytes—are the real proxy for value.
Reaching one billion average daily queries is powerful evidence that this founding vision is being realized. Founded in 2012, Snowflake reached one billion queries per day nine years later by riding the cloud wave. StarTree achieved the milestone in just six years by riding the real-time wave—delivering fast and fresh insights with the concurrency required for customer-facing applications and AI agents.
This milestone demonstrates that:
- Customers trust StarTree Cloud to power customer-facing, revenue-impacting applications, not just internal analytics
- The platform sustains high concurrency, consistently low latency, and continuous ingest, enabling fast, iterative exploration where one question naturally leads to the next
- Growth is driven by both new customer adoption and expanding use cases, as teams unlock richer, more interactive data experiences
Historically, this kind of insight-driven interaction wasn’t feasible at scale. High cost per query forced organizations to limit access, pre-aggregate aggressively, or reserve analytics for internal users. StarTree changed that dynamic—making it economically viable to support massive query volumes and, critically, to give customers their data back in real time through highly concurrent, interactive applications.
One billion queries isn’t an endpoint—it’s a new foundation. With recent advancements like Model Context Protocol (MCP) support and native Iceberg integration, early adopters are already extending StarTree Cloud into real-time AI systems, agent-driven workflows, and interactive customer-facing data products. As these use cases accelerate in 2026, we expect query volumes to grow far beyond today’s scale and further validating the original vision: when insights are instant and access is unconstrained, value compounds.
