Tesla FSD says 25% of the world has access. Only a third of that is real.

What the data actually says

Regional & Velocity Metrics Highlighted in This Project

To understand FSD adoption beyond raw numbers, I analyzed both regional differences and velocity metrics—two critical dimensions that Tesla strategy teams use internally to evaluate autonomy readiness.

The analysis in charts

Global Population with FSD Access (Cumulative)

Each country's approval adds to the cumulative population share. Note the 2026 European wave: five countries in ~60 days, showing that regulatory approvals behave like a network domino effect once a reference decision (Dutch RDW) is established.

Headline vs. Reality: Population by Rollout Status

Live (green): Active rollout (639M people) · Limited (gold): China's restricted access (1.41B people) · Pending (red): Approved but not yet rolled out (13M people). The 25% headline overstates usable access by ~3x because two-thirds of the approved population is locked in China's data-compliance stall.

Approved FSD Population by Region & Status

A regional breakdown shows how FSD availability is shaped by local laws. North America is almost fully live, Asia-Pacific is dominated by China's limited status, and Europe has approvals but remains mostly pending rollout.

Regional Regulatory Readiness Index (0-5 Scale)

Quantifying readiness across regions on a 0-5 scale. China leads in policy support (4.0/5), the US leads in testing permissions (3.5/5), while the European Union remains the most restrictive (2.0/5) due to strict L2+ testing caps and GDPR data residency burdens.

US NHTSA FSD Investigation Scope Over Time

Investigation scope grew 33% from Oct 2024 (2.4M vehicles) to Mar 2026 (3.2M vehicles) — now in the Engineering Analysis phase, the formal step before a recall. Camera-only degradation under glare and debris remains the primary regulatory hazard in the US.

Safety Reality vs. Public Narrative: Crash Miles vs. Media Scrutiny

Comparing actual safety data (miles between crashes under FSD vs. US national average) against relative negative media narrative volume. Empirical safety shows FSD is 10.5x safer than the US average, but media scrutiny of FSD crashes remains extremely high (95 pts index). This reveals that the biggest barrier to adoption isn't capability — it's trust.

FSD Subscription Index Over Time: Key Release Milestones

Tracking the subscription index shows how software releases (such as v12.3 introducing end-to-end neural nets, accompanied by a 1-month free trial and a price cut to $99/mo) and regulatory approvals (such as the EU wave) drive massive adoption velocity.

FSD Expansion Pipeline: Top Untapped Markets

Combined, the untapped pipeline represents 2.2B people. India (1.45B) and Brazil (216M) represent the largest untapped populations but currently have no announced regulatory framework, highlighting that shaping local standards is the biggest scaling barrier.

How I built it

Curated three datasets from primary public reporting (regulatory approval tracker, NHTSA investigation timeline, expansion pipeline), loaded them into SQLite, and wrote 8 SQL analyses — window functions for cumulative access and approval velocity, status splits for the headline-vs-reality gap. A reproducible Python script generates the processed tables and charts; Tableau Public hosts the interactive dashboard.

What I'd tell the strategy team

• United StatesTreat the Engineering Analysis as the critical path: ship proactive camera-degradation alerts before a recall mandates them, and automate the regulatory data pipeline — slow responses compound distrust.
• EuropeRide the RDW template: sequence Nordics and Benelux first, then push a single EU type-approval instead of 27 national campaigns.
• ChinaShip the rebranded v14 with full in-country data residency — treat the rebrand as a compliance product, not marketing.
• New marketsReuse the UAE playbook: 12–18 months of localized road testing shared with the regulator before launch; in India and Brazil, help write the supervised-system standard rather than wait for one.