You point an iPhone at a robot arm. You walk a slow loop around the cell. Minutes later, the arm moves in a browser, bumping the conveyor, tripping the cable.
That is Parallax Worlds' demo pitch to robotics engineers. Skip CAD-heavy twins. Hand video, get high-fidelity sim to break things in [StartupSeeker; citybiz].
Tanmay Agarwal founded the company. It exited stealth in fall 2025 with a $4M seed led by Pear VC. Total pre-seed/seed: $4.9M [Finsmes, Nov 2025; USA Today].
Cap table mixes believers and strategics. GS Futures, Kakao Ventures, Lightscape Partners, Gaingels, Nova Threshold, and Mana Ventures joined [Pear VC; WOWTALE, Feb 2026].
Kakao's follow-on hit Korean press. This signals robotics supply chain watchers [Seoul Economic Daily, Feb 2026].
The bet
Parallax: a modern reliability platform for robots [Parallax Worlds].
Wedge: ingest video, reconstruct, drop robots in for stress tests versus Isaac Sim or Gazebo from scratch [citybiz].
Pitch: validate in virtual replica pre-hardware. Shorten integrator weeks [Hiverlab].
Cost target: weeks from robot buy to reliable run. Reshoring, shortages, and humanoids backlog certainty [Hiverlab].
Why it could be big
Every 2026 shipper needs simulation hours. Most stitch open-source or build in-house.
Integrated phone-to-test compresses CAD/sim engineering/week job [citybiz; StartupSeeker].
Pear has a track record in robotics/infrastructure. $4M conviction eyes recon defensibility [Pear VC].
GS Futures has industrial fit. Kakao has a Korean ecosystem for heavy buyers [WOWTALE, Feb 2026].
The round in context
Pre-seed + Seed total | 4.9 | $M
Seed round (Nov 2025) | 4.0 | $M
$4.5M fits deeptech seed 2025. This provides runway for v1, partners, and numbers pre-A [Finsmes, Nov 2025].
The team and traction
Agarwal is named founder [Pear VC; Finsmes, Nov 2025].
Vertical stack: capture to simulation layer [citybiz].
Syndicate traction: Pear/GS/Kakao gets meetings.
The honest counterfactual
Bear: contested category. Isaac, Gazebo/MuJoCo, and twins circle buyers.
Fidelity is the game. Video struggles versus CAD physics for safety.
Bull: not raw physics versus Nvidia. Time-to-sim for teams without simulation engineers, iPhone for technicians [StartupSeeker; citybiz]. A roadmap fixes fidelity.
What to watch
Convert conviction to logos. Watch humanoid/arm partners, benchmarks versus twins, and GS/Kakao pilots.
2026 A hinges on video catching real failures.
Even partial success means the round writes.
Culture question: flooded robots from non-factory teams. Who ensures no room breaks?