Edge compute and robot semiconductors are the processors, accelerators, controllers, analog chips, sensors, power devices, and connectivity chips that sit inside a robot or next to a robot cell. They let the machine read sensors, run local inference, close motor-control loops, protect power rails, communicate with the factory network, and keep operating when cloud latency or bandwidth is not acceptable. The node matters when robot, warehouse, medical, defense, auto, and factory platforms move from pilots to qualified production, because each shipped endpoint then turns silicon content into revenue, margin, inventory drawdown, and cash conversion.
What the stack is: endpoint compute, embedded control, signal chain, sensing, power management, connectivity, safety, and support chips placed on robot control boards, sensor modules, motor drives, battery and power boards, edge gateways, and cell controllers.
What it does in the robotics chain: converts sensor input into low-latency decisions, executes motion commands, manages power and heat, links the robot to local networks, and keeps safety functions close to the machine.
Main pieces: CPUs and GPUs for compute, AI accelerators for inference, microcontrollers (MCUs) for deterministic control, digital signal processors, analog front ends, power-management integrated circuits (PMICs), motor drivers, image and position sensors, memory interfaces, wireless or Ethernet connectivity, and safety controllers.
Where it sits: physically inside the robot body, arm, camera pod, motor assembly, battery pack, charging dock, or local edge server; operationally between the sensor/actuator layer and the robot software stack.
How the theme uses it: robotics demand shows up as design wins that survive qualification, higher semiconductor content per robot, production shipments, industrial or automotive revenue, better utilization, lower inventory days, and margin stability.
Terms used later: a design win is a selected chip socket before full production; a socket is the role a chip fills on a board; content per robot is the dollar value of semiconductors in each endpoint; edge AI is inference run on or near the machine; qualification is the customer test process before volume shipments.
Report boundary: this is a tactical report-layer node. Durable company research, claim IDs, raw source maintenance, and missing security lanes stay in the linked knowledge pages. Price and volume context comes from read-only discovery daily_ohlc through 2026-06-12.
Current Setup
Endpoint silicon gateRobot chips need production sockets, not only platform relevance.
The semiconductor stack lets robots decide, sense, move, power, and connect locally. The investable test is whether robot and industrial endpoint programs create revenue that can be separated from data-center AI, handset, automotive, and broad analog cycles.
Best sourced exposureTXN has the cleanest analog and embedded-control route.
TXN, NVDA, MPWR, QCOM, NXPI, STM, AMD, and AVGO have local security lanes; NVDA, AMD, and AVGO need extra attribution checks because data-center AI can dominate.
Source gapADI and ON lack durable local security lanes.
They have market data and official filing extracts. QCOM, NXPI, and STM have lanes, but robot-specific revenue attribution is still incomplete.
Trading caveatRight-rail charts use live API data through 2026-06-12.
Static May 29 setup levels are stale and should not be used as current trading evidence without a fresh technical pass.
Design winChip selected for robot board
QualificationSafety, thermal, reliability tested
ProductionEndpoint shipments convert sockets
AttributionFilings separate robot demand
Robots need onboard compute, analog interfaces, power conversion, connectivity, sensing, safety control, and motion-control silicon when latency, reliability, privacy, or operating cost makes cloud-only inference impractical. Semiconductor design wins matter after a robot, machine, vehicle, warehouse cell, or surgical platform reaches qualified production. The useful proof is shipment volume, industrial or auto revenue, margin, customer breadth, inventory conversion, and production-program disclosure.
All ten ranked tickers have local daily_ohlc coverage through 2026-06-12. TXN, NVDA, MPWR, QCOM, NXPI, STM, AMD, and AVGO have local source-backed knowledge lanes, while ADI and ON were checked against current official filings or releases for latest-quarter revenue. The strongest theme tailwind is rising endpoint silicon content when robots need local inference, deterministic control, sensing, and power density instead of cloud-only operation.
Attribution is the main weakness. NVDA, AMD, AVGO, and MPWR can move on data-center AI, custom silicon, networking, or AI power demand; QCOM can move on handset and connectivity cycles; NXPI, STM, and TXN can move on broad auto and industrial analog cycles. ADI and ON still lack durable local security lanes, QCOM/NXPI/STM need robot-specific attribution checks despite having lanes, and current next earnings dates were not confirmed in this pass.
Source setup fields are preserved in hidden structured form in the required order: Why It Matters To The Theme, Tailwinds, Headwinds.
Basket
The ranked basket keeps the parent node order and sorts by robot-semiconductor economics, source support, and attribution risk before technical timing. Rank buttons and metric strips open the selected-security right rail; ticker links go to knowledge lanes only where those lanes exist.
QCOM belongs in this node because endpoint AI, connectivity, wireless, and embedded compute can sit in mobile robots, vehicles, gateways, and industrial devices.
Market cap$230.9B
Next earningsNot confirmed
Latest qtr revenue$10.599B
Role in stack
Endpoint processors, modem and connectivity silicon, automotive compute, and local AI platforms for mobile and embedded edge devices.
Revenue mix
Q2 FY2026 official release and the local QCOM lane support handset, automotive, IoT, and licensing exposure; robot-specific mix is not durable until the lane separates endpoint or industrial robotics revenue.
Latest qtr revenue
Fiscal Q2 2026 revenue was $10.599B for the quarter ended March 29, 2026, sourced from Qualcomm's SEC-filed April 29, 2026 8-K earnings release.
NXPI belongs in this node because industrial and automotive processors, safety controllers, and connectivity chips can control robot and vehicle endpoints.
Market cap$74.4B
Next earningsNot confirmed
Latest qtr revenue$3.181B
Role in stack
Embedded processors, safety chips, secure connectivity, radar and edge-control silicon for vehicles, factories, and industrial endpoints.
Revenue mix
The Q1 2026 release and local NXPI lane point to automotive, industrial, software-defined vehicle, data-center control-plane, and physical-AI exposure; robot-specific mix still needs narrower source evidence.
Latest qtr revenue
Q1 2026 total revenue was $3.181B for the quarter ended March 29, 2026, sourced from NXP's SEC-filed April 28, 2026 8-K earnings release.
ADIAnalog Devices, Inc.
ADI belongs in this node because robots need signal-chain, sensing, power, and industrial analog devices to measure and control the physical world.
Market cap$203.3B
Next earningsNot confirmed
Latest qtr revenue$3.160B
Role in stack
Signal chain, sensing, power, isolation, motor-control interfaces, and industrial instrumentation for robots and factory endpoint systems.
Revenue mix
The fiscal Q1 2026 release describes growth across all end markets led by Industrial and Communications and record Data Center orders; robot mix remains a source gap.
Latest qtr revenue
Fiscal Q1 2026 revenue was $3.160B for the quarter ended January 31, 2026, sourced from ADI's SEC-filed February 18, 2026 8-K earnings release.
ONON Semiconductor Corp
ON belongs in this node because power switching, sensing, image sensing, and auto-industrial chips can convert robot electrification into semiconductor content.
Market cap$40.4B
Next earningsNot confirmed
Latest qtr revenue$1.513B
Role in stack
Power devices, sensing, image sensors, motor-control and auto-industrial semiconductors for electrified equipment and machine vision endpoints.
Revenue mix
The latest 10-Q supports total revenue, but ON has no maintained local security lane in this checkout; separate auto, industrial, power, and sensing claims need filing-level work.
Latest qtr revenue
Q1 2026 revenue was $1.513B for the quarter ended April 3, 2026, sourced from ON Semiconductor's May 4, 2026 10-Q.
STM belongs in this node because MCUs, motion-control chips, power devices, and sensors are core building blocks for embedded robot control loops.
Market cap$52.6B
Next earningsNot confirmed
Latest qtr revenue$3.1B
Role in stack
Microcontrollers, power semiconductors, sensors, motion-control chips, automotive devices, and industrial embedded-control components.
Revenue mix
The local STM lane supports MCU, power, automotive, industrial, RF/optical, and recovery context; robot-specific mix is still an attribution gap and Q1 revenue remains rounded from current source extraction.
Latest qtr revenue
Q1 2026 sales were about $3.1B, sourced from STMicroelectronics' April 23, 2026 6-K and a same-day earnings report summary.
AMD belongs in this node as an adjacent edge and inference-compute supplier, but current evidence is much stronger for data-center AI than robot endpoints.
Market cap$742.2B
Next earningsNot confirmed
Latest qtr revenue$10.253B
Role in stack
CPUs, GPUs, adaptive and embedded compute, and accelerator systems that can support edge inference and industrial control adjacencies.
Revenue mix
Q1 2026 Data Center revenue was $5.775B, or 56.3% of total revenue; Embedded was $873M and is the cleaner robot-adjacent segment.
Latest qtr revenue
Q1 2026 total revenue was $10.253B, sourced from the AMD knowledge lane and company quarterly materials.
AVGO belongs in this node as a custom silicon, networking, and connectivity adjacency, but robot endpoint attribution is the weakest in the basket.
Market cap$2,035.9B
Next earningsNot confirmed
Latest qtr revenue$19.311B
Role in stack
Custom silicon, Ethernet and connectivity, AI infrastructure silicon, and infrastructure software that can support robotics systems indirectly.
Revenue mix
Q1 FY2026 Semiconductor Solutions revenue was $12.515B, Infrastructure Software was $6.796B, and AI semiconductor revenue was $8.4B.
Latest qtr revenue
Q1 FY2026 total revenue was $19.311B, sourced from the AVGO knowledge lane and company quarterly materials.
Basket note: market caps come from read-only discovery.instruments.market_cap queried on 2026-06-13. Next earnings dates were checked with bounded company IR and Nasdaq-style calendar searches on 2026-06-13; no company-confirmed dates were captured, so all cards use Not confirmed.
What Confirms Or Weakens
AreaWhat confirmsWhat weakens or invalidatesWatch next
01Node thesisEndpoints reach production
What confirms
Robot, warehouse, factory, medical, defense, or auto platforms move from pilots to qualified production and raise semiconductor content per endpoint.
What weakens or invalidates
Announcements stay at evaluation-kit, prototype, or platform-demo level without production shipments, program breadth, or customer revenue.
Watch next
Production programs
Shipment volume
Socket count
Qualification timing
02Economics and conversionRevenue carries margin
What confirms
Industrial, auto, embedded, power, or edge-compute revenue improves with gross margin, inventory days, utilization, and free cash flow.
What weakens or invalidates
Revenue growth comes with inventory builds, margin leakage, customer concentration, weak cash conversion, or price pressure.
Watch next
Segment revenue
Analog margins
Inventory days
Cash conversion
03Customer attributionRobot demand is separated
What confirms
Companies disclose endpoint, industrial edge, automotive, or robot-platform demand that is material enough to affect orders, revenue, or margin.
What weakens or invalidates
Price moves remain explainable by data-center AI, custom silicon, smartphones, broad auto recovery, or general semiconductor beta.
Watch next
Segment language
Product commentary
Backlog
Design wins
04Funding and channelBudgets become orders
What confirms
Robot OEMs, integrators, automotive customers, warehouses, medical platforms, and defense programs turn budgets into production purchase orders.
What weakens or invalidates
Customers delay capex, reduce inventory, extend qualification cycles, or keep orders in pilot lots.
Watch next
Distributor inventory
Book-to-bill
Customer capex
Defense procurement
05Policy and supplyAccess stays open
What confirms
Export controls, sourcing rules, and advanced-package supply do not interrupt robot-relevant chips or customer qualification timelines.
What weakens or invalidates
Tariffs, export limits, China restrictions, foundry bottlenecks, power device shortages, or memory constraints delay shipments or raise cost.
Watch next
SEC risk factors
Foundry capacity
Power devices
Memory availability
06Stale conditionRefresh trigger
What confirms
Discovery rows, API chart inputs, linked knowledge lanes, source-gap filings, revenue data, and metadata are refreshed after new market or operating events.
What weakens or invalidates
A new trading week, earnings release, 10-Q, product announcement, policy action, or supply-chain disclosure arrives before this node is refreshed.
Watch next
Discovery coverage
Source lanes
Setup levels
Metadata
Source Trail
Routes used for this node
Route
Use it for
Primary links and notes
Parent map
Robotics value-chain placement, parent return basket, and link-back target for this child node.
Revenue fields used in basket cards. Local lanes sourced TXN, NVDA, MPWR, QCOM, NXPI, STM, AMD, and AVGO; SEC-filed releases or filings cross-checked QCOM, NXPI, STM, and supplied the current ADI and ON revenue values.
Bounded web checks used company IR event pages, SEC filing context, and Nasdaq-style earnings calendar searches on 2026-06-13. No company-confirmed next earnings dates were captured, so all cards use Not confirmed. Do not treat estimated third-party dates as confirmed.
Discovery provenance
Read-only local coverage, market-cap fields, daily OHLC coverage, representative status, and API chart assumptions.
daily_ohlc latest as-of date is 2026-06-12 for ADI, AMD, AVGO, MPWR, NVDA, NXPI, ON, QCOM, STM, and TXN.
discovery.instruments.market_cap queried on 2026-06-13 for all ten tickers.
Right rail route: /api/securities/TICKER/chart?frequency=weekly&window=3y&as_of=latest.
Chart assumptions
How selected-security charts should be interpreted.
Weekly bars aggregate local daily rows by calendar week: first open, maximum high, minimum low, last close, and summed volume. Indicators are 20-week EMA, 100-week EMA, weekly volume, and 20-week average volume over a visible three-year horizon. Static May 29 setup thresholds are stale versus live API data through 2026-06-12 and must be refreshed before use as current trading evidence.
Data limitations
What should not be overclaimed from this node.
Price and volume context cannot prove robot demand. ADI and ON have no maintained local security lanes in this checkout. QCOM, NXPI, and STM have local lanes, but their robot-specific revenue mix remains an attribution gap. STM revenue is rounded from a same-day earnings summary plus the SEC-filed 6-K route because the downloaded 6-K HTML is fixed-position and did not yield a clean exact table extraction in this pass.