TanStack rolled out the beta of Table v9, a data-grid library that lets developers opt-in only the features they need. Bundles shrink to roughly 5 KB and interaction-to-next-paint (INP) delays that made sorting or filtering feel sticky disappear.

Why the change matters

In v8 the library ships every piece of grid logic—sorting, filtering, pagination, row selection, grouping—whether a project uses any of them. That extra code runs on the main thread, inflates bundle size and adds latency to user actions. For dashboards that need fast, responsive tables, a few milliseconds of lag can push the INP score into a poor range.

What v9 does differently

  • Opt-in feature modules – Import only the pieces you actually use. Skip sorting and the sorting code never makes it into the bundle.
  • TanStack Store integration – Manage state with a fine-grained store, so updating one row doesn’t trigger a full re-render of the filter bar or other unrelated UI.
  • Reduced memory footprint – Fewer objects and arrays ease pressure on the JavaScript heap during long sessions.

These changes translate into a smaller download (the library can sit near 5 KB for a simple list) and smoother interaction when the grid is busy.

Who stands to gain

  • Frontend teams building internal tools, admin panels, or SaaS dashboards where tables are the primary UI element.
  • Performance-focused sites that monitor Core Web Vitals; a lower INP directly improves the metric.

What v9 does not fix

The improvements target the code you control. They won’t magically speed up a page that pulls a massive JSON payload, nor will they offset the cost of a heavyweight third-party script. Large data sets still need proper pagination, and network latency remains a separate concern.

A practical migration path

  1. Identify the heaviest tables – Look for order lists, inventory grids, or CRM views that already show noticeable lag.
  2. Capture baseline metrics – Record INP and long-task durations on those pages before any changes.
  3. Migrate one grid at a time – Replace the v8 import with the v9 module set, enabling only the features the screen actually uses.
  4. Avoid the catch-all stockFeatures bundle – Pulling the default feature set defeats the size-saving purpose.
  5. Retest interactions – Measure sorting, filtering, and selection again to confirm the performance gain.

Counter-point: not a silver bullet

Some developers may expect v9 to solve every sluggish UI problem. In reality, the library’s gains are bounded by the amount of custom code you can prune. If a table’s bottleneck is the sheer volume of rows or an inefficient server API, the bundle-size reduction will have limited impact.

What to watch next

Takeaway: TanStack Table v9 gives you a concrete way to stop paying for unused grid functionality. By importing only the needed modules and using a fine-grained store, you can shave several kilobytes off your bundles and deliver noticeably snappier table interactions—provided you pair the upgrade with sensible data-handling strategies.