Intelligence layer

AI Component Trends

AI view: 409 technologies over the last 90 days. Start with the ranking table below, then use the longer notes for context.

View ranking table

Tracked

409

Active

366

Peak

2

Primary view

Ranking table

Current signal board

A quick pass on which components are lifting, cooling, dominating footprint, or still building historical context.

Watch list

Cooling

4 items
Not Officially Disclosed

AI · 1 robots in the directory

-1
Cooling —
2× Intel i7 (8th gen, 6-core) edge computing

AI · 1 robots in the directory

-1
Cooling —
360° LiDAR plus dual-vision navigation with centimeter-level positioning, autonomous mapping, and AI obstacle detection for 1,000+ obstacle types

AI · 1 robots in the directory

-1
Cooling —
AgileCore platform; Google DeepMind Gemini Robotics integration (announced)

AI · 1 robots in the directory

-1
Cooling —

Coverage leaders

Broadest footprint

4 items
Reactive AI Obstacle Avoidance (200+ object types); SmartPlan 3.0

AI · 2 robots in the directory

2
2 verified in this window → Flat
10 TOPS AI platform with LiDAR and vision fusion for automatic mapping, path planning, 300+ obstacle recognition, and smart cliff protection

AI · 1 robots in the directory

1
1 verified in this window → Flat
128 TOPS onboard AI compute, self-developed spatial foundation model, on-device spatial intelligence, physical-space agent behavior, generative actions/dances, and large-language-model voice system

AI · 1 robots in the directory

1
1 verified in this window → Flat
1x Embodied Intelligence

AI · 1 robots in the directory

1
1 verified in this window —

Fresh data

Needs baseline

4 items
48 TOPS Dense (Geek), 117 TOPS (Education), or 200 TOPS (Professional); Doubao LLM included free for 6 months per official spec

AI · 1 robots in the directory

No baseline
No baseline —
AI neural path planning with IR obstacle detection and real-time route adjustment for pool floors, walls, steps, and irregular pool geometry.

AI · 1 robots in the directory

No baseline
No baseline —
AI-supported walking and dynamic balance reported by Humanoid.Guide; AKINROBOTICS has not published a detailed AKINCI-5 autonomy stack

AI · 1 robots in the directory

No baseline
No baseline —
Amoria Emotional AI System for voice interaction, contextual response, response style, memory-oriented conversation, and privacy-conscious local-first interaction principles

AI · 1 robots in the directory

No baseline
No baseline —

AI field notes

Keep the ranking table fast, then use this route-specific readout to understand what the lane is actually signaling.

Intelligence layer

What this lane is actually tracking

Use this filtered board to separate durable AI adoption from launch-week noise. It surfaces which reasoning, vision, and edge-processing stacks are becoming operational defaults across the robot catalog.

Look for technologies rising in both footprint and momentum, not just hype-heavy announcements.
The strongest signals usually come from stacks spreading across several categories, not one flagship demo robot.

Most used in the database

Reactive AI Obstacle Avoidance (200+ object types); SmartPlan 3.0

2 robots in the directory

#1
10 TOPS AI platform with LiDAR and vision fusion for automatic mapping, path planning, 300+ obstacle recognition, and smart cliff protection

1 robots in the directory

#2
128 TOPS onboard AI compute, self-developed spatial foundation model, on-device spatial intelligence, physical-space agent behavior, generative actions/dances, and large-language-model voice system

1 robots in the directory

#3
1x Embodied Intelligence

1 robots in the directory

#4

Cross-check next

Use the 30-day view to confirm whether today’s move is holding. Reactive AI Obstacle Avoidance (200+ object types); SmartPlan 3.0 currently leads this lane with 2 recent verifications.

Active now

366

Rows with fresh signal in this window

Sustained

308

Rows with confirmed direction

Read the trend correctly

Use signal for footprint, delta for immediate change, momentum for confirmation, and reliability to judge how much trust to place in the pattern.

Signal

How many robots carrying the component were verified in the last 90 days. Treat it as current footprint, not install base.

Delta

Change against the last stored snapshot. Positive means more recent verification activity, negative means cooling, and a dash means the baseline is still forming.

Momentum

Two consecutive moves in the same direction. Use it to separate one-off spikes from signals that are holding their shape.

Reliability

High reliability means multiple historical checkpoints, medium means limited history, low means the component still needs another capture before comparison becomes meaningful.

The 90-day window smooths out one-off updates. When you want the earliest hint of movement, cross-check against the 30-day view.

About AI Components

409 AI components represent the intelligence layer of home robots — from navigation algorithms and computer vision to large language models and adaptive behavior systems. AI is the most transformative technology category in modern robotics, determining how autonomously a robot can operate, how naturally it interacts with humans, and how well it adapts to unique home environments. The 90-day trends track which AI technologies manufacturers are actually shipping in production robots.

Most used: Reactive AI Obstacle Avoidance (200+ object types); SmartPlan 3.0 (2 robots), 10 TOPS AI platform with LiDAR and vision fusion for automatic mapping, path planning, 300+ obstacle recognition, and smart cliff protection (1 robots), 128 TOPS onboard AI compute, self-developed spatial foundation model, on-device spatial intelligence, physical-space agent behavior, generative actions/dances, and large-language-model voice system (1 robots), 1x Embodied Intelligence (1 robots), 2× Intel i7 (8th gen, 6-core) edge computing (1 robots).

Using This Trend Data

Components with high signal values and rising deltas are gaining manufacturer adoption — these represent technologies the industry is converging around. Components with declining signals may indicate either a technology being phased out or simply a gap in recent verification activity. Pay attention to momentum alongside the delta: a component with sustained upward momentum across multiple snapshots is a stronger signal of genuine growth than one with a single positive delta. Reliability indicators tell you how much confidence to place in the trend — high reliability means the pattern is confirmed by multiple data points, while low reliability means the trend is based on limited historical data. For purchasing decisions, combine trend data with the individual component detail pages linked from the table, which provide deeper technical context and robot compatibility information.

Large language models — LLMs are rapidly entering companion and assistant robots, enabling natural conversation instead of keyword commands. This is the fastest-growing AI category in terms of manufacturer interest.
Edge AI processors — On-device chips (NVIDIA Jetson, Qualcomm RB5, custom silicon) enable real-time processing without cloud dependency, improving response time and privacy. Growing adoption signals manufacturer investment in capable standalone operation.
Computer vision models — Object recognition, scene understanding, and visual navigation are becoming standard equipment. Robots with dedicated vision AI can identify specific objects (shoes, cables, pet waste) rather than treating all obstacles generically.
Simultaneous localization and mapping (SLAM) — Advanced SLAM algorithms process sensor data to build and update home maps in real-time. Improvements in SLAM AI directly translate to better cleaning coverage and fewer stuck situations.

Buying Context

The gap between robots with modern AI and those without is significant and growing. A robot with LLM-powered voice interaction, vision-based object recognition, and edge AI processing can operate far more autonomously than one relying on preset patterns and basic sensors. However, AI capability is hard to evaluate from specs alone — look at actual user reviews for real-world performance. The trend data above shows which AI technologies are gaining broad manufacturer adoption, indicating maturity and reliability.

Compare with the 30-day ai trends for a broader adoption picture.

Related
Components Directory Components Glossary Robot Categories Compare Robots

Frequently Asked Questions

How often is the 90-day trend data updated?
Recalculated on every page load from current robot verification dates. Signal = robots verified in the last 90 days. Snapshots for delta/momentum are stored periodically.
What does "No baseline" mean for a ai component?
First measurement in this window — no previous snapshot to compare. Once a second snapshot is taken, the component gets a delta and eventually momentum data.
What is the difference between 30-day and 90-day trends?
30-day is volatile and sensitive to individual product launches. 90-day smooths noise and reveals sustained adoption. A component rising in both views is showing genuine growth.
How is momentum different from the trend delta?
Delta compares against the single last snapshot. Momentum requires two snapshots and checks for sustained direction — two consecutive increases or decreases. Momentum is a stronger signal of real change.
Can trend data predict which ai components will be popular?
Trends show historical adoption patterns, not predictions. But sustained upward momentum in the 90-day view tends to continue as it reflects manufacturer consensus. Cross-reference with industry news and announcements.
What is the difference between cloud AI and edge AI in robots?
Cloud AI sends data to remote servers for processing, enabling more powerful analysis but requiring internet connectivity. Edge AI processes data locally on the robot's own chip, providing faster response times and better privacy. Many modern robots use a hybrid approach — edge AI for real-time decisions (obstacle avoidance) and cloud AI for heavier tasks (map analysis, conversation).
How do I know if a robot's AI is actually good?
AI quality is difficult to assess from spec sheets alone. Look for: (1) specific capabilities like object recognition (can it identify shoes, cables, pet waste?), (2) independent reviews testing navigation in cluttered environments, (3) user reviews mentioning whether the robot learns and improves over time. The trend data here shows which AI platforms have broad manufacturer adoption — a proxy for maturity.