2026-06-17 — views
Physical AI Master Scorecard — Who Is Winning the Robot Ramp in Mid-2026
Tesla, Waymo, and China mapped across 10 competitive dimensions — one unified scorecard for the physical AI race at mid-2026.
The physical AI race in one table
Eight benchmark articles. Three major players. Ten competitive dimensions. This synthesis delivers the unified verdict: who leads today, who leads structurally, and what the rest of 2026 will decide.
Master Scorecard
| Dimension | Tesla | Waymo | China (Baidu) | Winner |
|---|---|---|---|---|
| Commercial rides today | Austin launch (~50 vehicles) | 150K+ /week, 4 cities | 100K+ /week est., 10 cities | Waymo (US) / Baidu (global) |
| Fleet scale | 6M+ FSD-capable (supervised) | ~1,500 driverless | ~1,000+ driverless | Tesla (scale) |
| Sensor cost per vehicle | ~$650 (est.) | ~$4,000–6,000 (est.) | ~$2,000–4,000 (est.) | Tesla |
| Regulatory approvals | Texas (Austin only) | CA, AZ, TX, GA (pending) | 10 Chinese cities + Abu Dhabi | Waymo (US) / Baidu (global) |
| Training data flywheel | 6M cars x 24/7 | ~1,500 vehicles | ~1,000 vehicles + road cams | Tesla |
| Humanoid ramp | Optimus 5K–10K units (est.) | None | 20+ startups, state-backed | Tesla (US humanoid) |
| Compute independence | Full (HW4 + Dojo) | Partial (Google TPUs) | Growing (Horizon Robotics) | Tesla |
| Unit economics path | ~$0.60–1.00/mile (est.) | ~$1.15–1.80/mile (est.) | Unknown | Tesla |
| Investment backing | Self-funded (~$1T market cap) | $45–50B standalone est. | State + VC backed | Tesla (scale) |
| Regulatory risk | Medium (FMVSS waiver needed) | Low (permits in hand) | Low in China | Waymo (near-term) |
Sensor cost, unit economics, and ride count estimates are based on public reporting and analyst estimates as of mid-2026. Baidu weekly rides labeled estimate.
Section 1 — The three ramp theses
Thesis A: Waymo wins the near-term commercial ramp
Waymo has the permits, the driverless fleet, and the operational experience. No competitor in the US is closer to scaling paid driverless rides today. With 150,000+ weekly rides across four cities and active expansion into Atlanta and Texas, Waymo’s moat is operational depth — the accumulated permits, incident response protocols, and dispatch infrastructure built over a decade in Phoenix. Cost is not Waymo’s advantage. Execution credibility is.
Thesis B: Tesla wins the long-term economics ramp
Tesla’s sensor cost advantage — approximately $650 per vehicle versus $4,000–6,000 for Waymo’s sensor stack (est.) — combined with an owner-enrollment network model and a 6M-vehicle training flywheel creates a structural cost moat that no purpose-built AV fleet can replicate. If Cybercab launches at scale, cost-per-mile could undercut Waymo by 40–60%. The key risk: Tesla’s Cybercab launch in Austin is still supervised (~50 vehicles), and converting that into a fully driverless commercial network requires a FMVSS waiver from NHTSA — regulatory approval that Waymo already holds.
Thesis C: China sets the global pace
Baidu is running driverless commercial rides in 10 cities. WeRide operates in 30+ cities across 7 countries. China’s government-mandate model deploys faster than US state-by-state regulation allows, and China has funded over 20 humanoid startups with factory deployment targets since 2023. The US is not “winning” globally — it leads in technology sophistication and safety-data transparency, but trails in commercial-scale deployment breadth. The competitive risk is narrative: if Baidu reaches 500,000 weekly rides before Waymo expands to its fifth major city, the global framing shifts regardless of underlying technology differences.
Section 2 — Benchmark article index
| # | Article | Key finding | Slug |
|---|---|---|---|
| 1 | Tesla/Waymo Ramp Index | Waymo 150K+ rides/week; Tesla Austin launch | physical-ai-tesla-waymo-ramp-index-jun2026 |
| 2 | Humanoid Robot Race | Figure $675M; Optimus 5K–10K units est. | humanoid-robot-race-index-jun2026 |
| 3 | AV Safety and Regulatory | Waymo 4-state approvals; Tesla needs FMVSS waiver | av-safety-regulatory-scoreboard-jun2026 |
| 4 | Investment Tracker | Bezos in 4 companies; OpenAI in 3 | physical-ai-investment-tracker-jun2026 |
| 5 | Compute Silicon | Thor bottleneck for humanoid startups; Tesla/Waymo isolated | physical-ai-compute-silicon-index-jun2026 |
| 6 | Sensor Stack | Tesla ~$650 vs Waymo ~$4–6K per vehicle (est.) | av-sensor-stack-perception-index-jun2026 |
| 7 | Robotaxi Unit Economics | Break-even: Waymo ~20K vehicles; Tesla network ~3.5K (est.) | robotaxi-unit-economics-index-jun2026 |
| 8 | Global Race | Baidu 10 cities; Wayve $1.05B; China domestic chip push | physical-ai-global-race-index-jun2026 |
Section 3 — H2 2026 watch list
The five events that will move the scorecard before year-end:
-
Tesla Cybercab FMVSS waiver decision — NHTSA must grant an exemption from federal safety standards written for human-driven vehicles before Cybercab can operate driverless commercially in the US. Timeline and conditions are the single largest regulatory gate for Tesla’s ramp.
-
Waymo Atlanta launch and fleet scale milestone — Waymo’s expansion beyond its current four markets is the next commercial proof point. A fifth city launch and trajectory toward 200,000 rides/week would confirm the network-effects thesis.
-
Optimus production number at Tesla Q3 2026 earnings — Tesla has guided for 5,000–10,000 Optimus units by end of 2026. The Q3 update will reveal whether the ramp is on track, ahead, or behind — and whether humanoid production is scaling to automotive-like volumes.
-
Baidu Apollo IPO or Waymo IPO signals — Both companies have been flagged as potential standalone public offerings. Any S-1 filing, roadshow, or formal IPO statement would force a public unit-economics disclosure and recalibrate the private valuations that currently anchor investor narratives.
-
NVIDIA Jetson Thor allocation relief for humanoid startups — Thor is the compute bottleneck for most non-Tesla humanoid programs. Any meaningful increase in chip allocation — whether from production ramp, new contract announcements, or alternative chip qualifications — would accelerate the entire second tier of the humanoid race.
How to read the scorecard
The master table above reflects the state as of mid-2026. No single company leads across all ten dimensions. Waymo leads on commercial deployment and near-term regulatory standing. Tesla leads on cost structure, data flywheel, fleet scale, and humanoid ambition. China leads on deployment breadth and government-enabled speed.
The race is not decided. The second half of 2026 will answer the questions that matter most: Can Tesla get its FMVSS waiver? Can Waymo cross 200K rides/week? Does China’s humanoid factory push produce verifiable output? Follow the H2 watch list — the scorecard will shift.