From tap to doorstep: the tech powering physical AI

A single tap on a smartphone initiates a sequence of device, network and logistics systems that convert a digital order into a physical shipment. The process involves touch sensors, handset processors, network and cloud infrastructure, cybersecurity, warehouse automation and last‑mile sensing and control.

The process starts at the phone. A fingertip alters the screen’s electric field, a touch controller converts that disturbance into screen coordinates, and an application processor packages the action as a network request. MediaTek and Qualcomm supply the processors and modem platforms that make the input compute- and network-ready on the device. The request then leaves the handset over 5G or Wi‑Fi toward network and cloud layers.

At the network edge, traffic routing and filtering occur. Edge routing and denial-of-service protection route legitimate requests while blocking malicious traffic. Endpoint and cloud workload protection monitors runtime behavior to prevent breaches during processing.

Optical and electrical links carry the request into data centers. Suppliers of lasers, transceivers and signal‑processing components convert electrical signals to light for fiber networks. High‑speed switches route traffic between servers and accelerator clusters, and connectivity devices inside racks link processors, accelerators and memory.

Accelerated processors and GPUs run models that generate product recommendations, perform inference and produce fulfillment instructions. Custom AI silicon and network components move data between racks and sites. Data platforms store and serve governed enterprise data, while observability tools correlate metrics, logs and traces across the stack to detect latency or failures.

When compute confirms an order, software issues pick‑and‑place and packing instructions for automated warehouses. Robotic systems retrieve cases and build pallets, grid‑top robots fetch storage bins, and conveyors and sorters move goods between workstations. Barcode scanners, RFID readers and mobile terminals identify items and update inventory as goods move.

For the last mile, deliveries can be handled by people or autonomous systems. Lidar sensors map 3D surroundings, edge‑AI vision chips interpret camera feeds locally, and microcontrollers and power semiconductors manage vehicle control and energy. Those components are used in delivery robots, autonomous vans and drones.

For investors, the ROBO index focuses on physical automation and robotics hardware, while the THNQ index emphasizes cloud AI platforms and software. Some companies provide platforms used in both cloud services and edge or vehicle systems. As of July 2026, the 21 companies profiled across these stages are constituents of at least one of the ROBO or THNQ indices.

The U.S. ROBO ETF tracks the ROBO Global Robotics & Automation Index and the THNQ ETF tracks the ROBO Global Artificial Intelligence Index. L&G offers related UCITS ETFs, including the AIAI ticker for the AI index on the London Stock Exchange. VettaFi LLC is the index provider for ROBO and THNQ and receives an index licensing fee; it is not the issuer, sponsor or seller of those funds and has no obligation related to their issuance or administration.