Why Attend TiDB SCaiLE 2026: Tracks, Format, and Focus

Key Takeaways

TiDB SCaiLE 2026 runs October 15 at the Computer History Museum in Mountain View, CA, as a single-day program.

The agenda splits into two tracks, AI-Native: Infinite Agents and Enterprise: Infinite Data, that sit on the same data foundation so sessions connect rather than compete.

Sessions are practitioner-led and selected for real numbers and real tradeoffs, including talks from teams that evaluated TiDB and chose something else.

The call for presentations closes September 11, 2026. Registration is now open.

A single user action in an agentic application no longer maps to a single database query. It spawns agent instances that branch context in milliseconds, hold memory across sessions, and provision their own tenants without waiting for a human. At Manus, agents create more than 90% of new database clusters. The stacks underneath them were designed for a user who clicks, waits, and clicks again.

TiDB SCaiLE 2026 is a one-day event built around that gap. If you are asking why attend TiDB SCaiLE 2026, the answer lies in how the program is organized: two tracks, both grounded in production systems, covering the enterprise teams that accumulated complexity over a decade and the AI-native teams generating it inside a single workflow. This blog covers the tracks, the format, who the day is for, and how to get on the stage.

Agents Changed the Clock Speed, Not the Complexity

Enterprise scale and agent scale look like two different problems. But they are the same problem running at different clock speeds. An enterprise accumulates tenants, tables, regions, and compliance boundaries over ten years. An agent platform generates the same shape of complexity inside a single workflow.

The numbers make the comparison concrete. Atlassian consolidated more than 750 PostgreSQL clusters into 16 TiDB clusters to run its Forge platform, hosting over 3 million tables and validating 500K concurrent connections per cluster. That complexity took years of growth to create. Manus migrated to TiDB Cloud in roughly two weeks and reached close to one million database tenants within three months, with agents rather than humans creating more than 90% of them.

Same architectural pressure, different timescale. TiDB Co-Founder and CTO Ed Huang’s argument that AI agents are becoming the primary database user is the primary thesis the 2026 program is built on.

Two Tracks, One Foundation

TiDB SCaiLE 2026 runs two tracks across one day. Both sit on the same data foundation, so the sessions connect instead of splitting the audience into two unrelated events. Follow one path or move between them.

AI-Native: Infinite Agents

This track covers workloads where software, not a person, is the primary user. Expect sessions on agent-driven provisioning, per-agent and per-tenant isolation, vector and semantic retrieval, memory layers, sub-second database creation, and cost control when instance counts run into the millions. If you have read about why agent memory has to write back and want to see how teams implement that loop in production, this is the track.

Enterprise: Infinite Data

This track covers transactional and analytical workloads at scale across global, compliance-sensitive, or high-throughput environments. Expect consolidation projects, zero-downtime migrations, multi-region resilience, observability, and the operational practices that held under load. The talks here read closer to postmortems than to roadmaps.

Why Attend TiDB SCaiLE 2026

Four things separate this event from a webinar series with a venue attached.

Sessions open with the problem. Original, practitioner-led talks are featured with real numbers, tradeoffs, and production scars. We explicitly invite talks on why a team evaluated TiDB and chose something else. That constraint shapes what makes the agenda.
Product news lands on this stage first. TiDB used the 2025 keynote to announce TiDB X, the architecture that puts object storage at the backbone of TiDB. Roadmap conversations happen in the room, not in a follow-up email.
The people in the hallway are solving your problem. Attendees are engineering, platform, data, and AI infrastructure leaders running these systems in production. At a one-day event, the unstructured time is worth as much as the sessions.
One day, one venue. The full program runs October 15 at the Computer History Museum in Mountain View. It costs you a day, not a whole travel week.

Who Should Attend TiDB SCaiLE 2026

The program assumes you operate a data layer rather than evaluate one from a slide. It is built for:

Platform and infrastructure engineers running multi-tenant systems at high tenant counts.
Data architects deciding whether to consolidate a fragmented stack of transactional, analytical, and vector stores.
AI infrastructure teams building memory, retrieval, and state layers behind production agents.
SREs and DBAs responsible for migrations, multi-region resilience, and uptime under unpredictable load.
Engineering leaders setting the data layer strategy for the next three years.

If you are earlier in your evaluation journey, our Database for AI Agents overview and What is TiDB explainer page cover the architectural background the sessions assume.

How to Speak at TiDB SCaiLE 2026

The call for presentations is open until September 11, 2026. Talks run 30 minutes including Q&A, and panels and roundtables on emerging data infrastructure topics are welcome. The program committee is looking for submissions on:

Where TiDB fits in your stack, or why you evaluated it and chose something else
Distributed SQL in production: what worked, what broke, what you would do differently
Data architectures behind agentic and AI-native applications
Multi-tenancy at high tenant counts, and scaling storage and compute independently
Lessons from outages, performance tuning, or migration projects

You can make a session submission through our online proposal form. Check out last year’s lineup for TiDB SCaiLE 2025 as a reasonable calibration point for scope and depth.

Registration is Now Open

Agent workloads and enterprise consolidation are converging on the same data layer problem, and the teams furthest along are the ones worth listening to for a day.

Don’t wait! Save your seat for this can’t-miss event via the event registration page.