JD Q2 2026 Earnings: 80 RoboBases in 5 Years, Zhiwolf in 60 Warehouses, Dual-Arm Robot in 10 Seconds — Embodied AI Gets Its Foundation

Forum topic · 小凯 · 2026-08-13

Summary

On August 13, JD.com released its Q2 2026 results: revenue of 346.4 billion yuan (down 2.9% YoY) but net profit up 14.5% to 7.1 billion yuan, with R&D spending accelerating for three consecutive quarters (+53.2% H1). The report highlights a large-scale embodied AI infrastructure push: the first RoboBase robotics base broke ground in Guangzhou, with plans for 80+ bases nationwide over five years covering data collection, testing, deployment, and maintenance. JD open-sourced EgoLive, claimed to be the industry's largest egocentric human first-view dataset, with 100+ universities applying for access, plus three open models (JoyAI-Echo, JoyAI-VL-Interaction, JoyAI-Video-Edit). Deployed systems include the Zhiwolf goods-to-person solution in 60+ warehouses globally, a dual-arm Yiwolf robotic arm completing pick-and-place in 10 seconds, thousands of autonomous delivery vehicles across 20+ provinces, drone delivery covering 78 villages, and the JoyAvatar digital human serving 80,000+ merchants. Unlike Tesla or Figure AI, JD is building robots as infrastructure rather than selling them. Management acknowledged these metrics don't yet translate to measurable revenue or cost benefits, with the next quarterly report as a key validation point.

On the evening of August 13, JD.com released its Q2 2026 and interim results. Taken together, several figures show that embodied intelligence in China is, for the first time, getting a "foundation-laying" effort.

The Financial Baseline

Q2 revenue: 346.4 billion yuan, down 2.9% YoY
Q2 net profit attributable to shareholders: 7.1 billion yuan, up 14.5% YoY
Services revenue: 79.3 billion yuan, rising to 22.9% of total
H1 system-wide R&D investment up 53.2% (accelerating for three consecutive quarters)
JD Logistics revenue +26.5%, JD Health +17.2%, JD Industrials +16.3%

Product revenue is declining while R&D and services revenue are rising — a traditional e-commerce company trading profit for the next technology cycle.

Three New-Infrastructure Commitments

1. RoboBase robotics bases — 80+ planned over five years

The first RoboBase project broke ground in Guangzhou in Q2. Per the plan, these bases will offer robot data collection, testing and validation, scenario application, and maintenance services. JD plans to deploy 80+ RoboBase bases nationwide over five years, aiming to build a complete physical-AI ecosystem loop of "R&D — manufacturing — application — service."

2. EgoLive egocentric dataset — 100+ university applications

JD launched China's first embodied data collection community in Suqian, Jiangsu, and open-sourced EgoLive, claimed to be the industry's largest human first-person-view dataset. Drawn from real retail, logistics, health, and industrial operations, every frame corresponds to real physical actions and industry needs. Over 100 universities and research institutes have applied for access.

3. Three open-source models — a physical-world model matrix

JoyAI-Echo: long-form audio-video generation framework
JoyAI-VL-Interaction: real-time video vision-language interaction model
JoyAI-Video-Edit: real-time streaming video editing model

JD claims this physical-world model matrix has reached a globally leading level overall.

Three Sets of Deployed Numbers

Warehousing

The Zhiwolf goods-to-person solution is live in 60+ warehouses globally (including the UK and Germany)
The Yiwolf robotic arm has been upgraded from single-arm to dual-arm, completing recognition, grasping, and stacking in 10 seconds

Last-mile delivery

Thousands of autonomous delivery vehicles operate routinely across 20+ provinces; Shenzhen launched the first nighttime autonomous delivery routes
A drone delivery network in Zizhong, Sichuan covers 78 administrative villages, crossing mountainous terrain in as fast as 7 minutes

Merchant services

JD's digital human JoyAvatar has served over 80,000 merchants
Q2 digital-human livestreaming accounts grew 3x YoY; daily active streaming accounts surged 6x
During 618, digital-human livestream hours and transaction volume both exceeded 4x YoY

A Different Path from Other Players

Tesla Optimus pursues a full-stack route: in-house robots, in-house data, in-house scenarios. Figure AI pursues single-point breakthroughs with a general-purpose humanoid deployed in BMW factories. Unitree pursues scale via cost-effective hardware across education and inspection scenarios.

JD's path is distinct: it is not selling robots — it is building robots as infrastructure. The 80+ RoboBases are essentially a hybrid of "robot 4S dealership + repair station + data center + training ground." JD opens up every real scenario in logistics, retail, industry, and health where robots can run, using them as training grounds. That sets it apart from US players.

The Real Questions

JD management explicitly acknowledged on the earnings call that these metrics do not yet directly show how much incremental revenue, profit, or cost reduction the AI investments have produced. Three validation points for the next phase:

1. Whether ad conversion rates improve as digital-human livestreaming scales 2. Whether the Yiwolf dual-arm robots actually lower per-unit warehouse fulfillment costs 3. Whether external technology services revenue reaches a quantifiable share

If any two of these three show positive signals within six months, JD's five-year, 80-RoboBase commitment stops being "foundation-laying" and becomes "runway-paving." The embodied-AI numbers in the 8-13 report are just the beginning — the 11-13 report will confirm whether this line of business is real.