The Most Investible Layer Yet. Perception for Physical AI

Gaetano · Gaetano · June 13, 2026 at 13:41 · ⏱ 4 min read  | Read on Substack ↗
Summary
Gaetano argues that Layer 3 (Perception) is the most directly investible layer in Physical AI because it involves already-deployed technologies like machine vision, LiDAR, ADAS, and edge AI chips that turn raw sensor data into actionable intelligence. The article implies that the best opportunities are in companies that integrate software and hardware to make perception easier and more accurate, rather than just selling cheap sensors.
  • Perception converts physical world signals (from cameras, LiDAR, radar) into machine-usable information, enabling Physical AI systems to understand environment, distance, motion, and safety.
  • Companies like Ambarella (AMBA) and Ouster (OUST) are explicitly named as existing players in vision chips and LiDAR, respectively.
  • The author states that perception is already widely used in factories (machine vision), warehouses (cameras/scanners), and cars (ADAS), making it less abstract than earlier layers.
  • Value is expected to accrue to companies that make perception easier to deploy, more integrated, and more useful—through software attachment, edge processing, and workflow integration.
  • The author has taken positions in this layer (not disclosed in the article) and says these positions will have higher weightings than previous layers.
  • The article ends with 'The Companies I Am Focused On' but the text cuts off without listing them.
Read time 4 min
Length 4,743 chars
Category finance
Ideas
Gaetano Substack author, Gaetano
Author explicitly names Ambarella as a company in the perception layer (vision chips for edge AI). The article's thesis that value accrues to integrated perception solutions supports a bullish view on
Author explicitly names Ambarella as a company in the perception layer (vision chips for edge AI). The article's thesis that value accrues to integrated perception solutions supports a bullish view on Ambarella's camera processing chips. Risk: Competition from larger chipmakers (Qualcomm, Nvidia) could limit Ambarella's market share in automotive and robotics.
Gaetano Substack author, Gaetano
Author names Ouster as a LiDAR company in the perception layer. The article highlights LiDAR as providing depth and structural understanding, and Ouster is a direct play on this sensing technology for
Author names Ouster as a LiDAR company in the perception layer. The article highlights LiDAR as providing depth and structural understanding, and Ouster is a direct play on this sensing technology for autonomous vehicles and robotics. Risk: LiDAR adoption timelines remain uncertain; competition from cheaper solid-state LiDAR and Tesla's vision-only approach could pressure demand.
More from Gaetano

This newsletter, published June 13, 2026, features Gaetano discussing AMBA, OUST. 2 trade ideas extracted by AI with direction and confidence scoring.

Speakers: Gaetano  · Tickers: AMBA, OUST