Back to feed
News Story
APriority76
机器之心
1 sources

Luming Launches NexCore: When Robot Skills Become Infrastructure

On the opening day of the 2026 World Robot Conference, Luming Robotics released Lumos NexCore, an embodied intelligence evolution engine designed to systematize the production and operation of robot skills, making them callable, trainable, and verifiable like cloud services. The platform integrates data, training, evaluation, skill management, and device operations to drive robots from single-point capabilities to large-scale applications.

SynthePulse Insight · AI deep readingMembers

Luming Unveils NexCore: How Robot Skills Become Reusable Infrastructure

Version 1 · 1 source

As robot demos become increasingly impressive, industrial deployment still relies on engineers crouching by production lines to reprogram. On the opening day of the 2026 World Robot Conference, Luming Robotics released the embodied intelligence evolution engine Lumos NexCore, aiming to connect data, models, skills, and devices into a closed loop, making robot capabilities callable, trainable, and verifiable like cloud services.

  • Luming Robotics released the embodied intelligence evolution engine Lumos NexCore on August 19, 2026, positioning it as skill infrastructure for industrial embodied intelligence.
  • NexCore integrates five capabilities: data assets, model training, evaluation and verification, skill packaging, and device operations, forming an observable and traceable closed loop.
  • Luming founder Yu Chao believes the second half of embodied intelligence is a competition of task closed loops, where victory depends on the product of data acquisition efficiency, model generalization capability, general-purpose hardware cost, and industrial delivery efficiency.
Open section navigationFrom 'Craft Workshop' to 'Skill Production Line'

From 'Craft Workshop' to 'Skill Production Line'

The current production method for robot skills remains 'craft workshop-style': for each new scenario, the team must complete the entire process of data collection, processing, training, parameter tuning, real-machine deployment, and failure rework, heavily relying on engineers' personal experience, and data and experience from one scenario are difficult to transfer to the next.

NexCore integrates data assets, model training, evaluation and verification, skill packaging, and device operations as platform services. The data asset platform solves the 'raw material' problem, the model training platform hosts the training process, evaluation provides 'acceptance criteria,' skill management packages models into reusable units, and robot operations connect to real devices and feed back high-value samples.

Users can even describe task objectives in natural language, and the platform orchestrates subsequent processes, enabling those who truly understand the business to participate in skill development without relying on a full technical team.

Free for now

Read the full analysis

4 more sections of analysis, plus the full takeaway

Loading

Credibility boundary

This article's information is primarily derived from Machine Intelligence's report on Luming Robotics' release of NexCore, which is an industry media retelling. Product features, founder views, and leaderboard results are all statements from Luming and have not been independently verified. Specific data such as dual-arm payload of 50 kg and MOS2.0 release timing are subject to the report.

Sources for this version

  1. 鹿明发布NexCore：当机器人技能成为一种基础设施机器之心

Primary report

机器之心

Primary source