Strava tightens data access to block AI scrapers before IPO

Strava’s new posture is another data point in a clear trend: as AI labs vacuum up the open web, high‑value platforms are shrinking their public surface area and moving toward paid, controlled access. By putting most profiles and club data behind login and levying a flat developer fee, Strava is trying to curb unsanctioned scraping while signalling "data discipline" to prospective IPO investors. The company explicitly links the change to AI companies that ignore robots.txt and route crawlers through aggregators; Perplexity is called out as an offender by name.([techcrunch.com](https://techcrunch.com/2026/06/01/strava-declares-war-on-scrapers-ahead-of-ipo/))

For the AGI race, this clamps down on a class of behavioral and geospatial data that would be extremely valuable for training embodied agents and health‑related models. As more sites follow Reddit and now Strava in charging for structured access, frontier labs will lean harder on licensing deals, synthetic data and closed‑loop user interactions inside their own products. That likely tilts the advantage further toward well‑capitalized players who can afford large content deals and compute. At the same time, it may improve privacy outcomes by pushing the ecosystem from "scrape first, ask forgiveness later" toward explicit contracts – but at the cost of reducing the diversity of real‑world data available to smaller labs and open‑source efforts.