At JD’s 618 launch event on May 18, the spotlight wasn’t solely on the usual array of discounted prices, subsidies, and billion-yuan coupon giveaways—it was on how JD positioned AI as the centerpiece of this year’s 618 campaign.

JD disclosed that its AI-related R&D investment has soared by over 200% year-on-year. In the first quarter, nearly 80 million users interacted with “Jingyan,” JD’s AI-powered shopping assistant, representing a year-on-year surge of over 200%. JoyInside, a subsidiary of JD Tech, has forged partnerships with nearly 200 brands across home appliances, robotics, health, and toys, with plans to deploy over 10 million AI-enabled devices this year. Additionally, JD Health is set to launch multiple AI medical devices during 618, aiming to connect 1 million devices to JoyInside within a year.

Taken together, this data suggests that JD is moving beyond mere talk of “AI-driven shopping experiences.”

Historically, the market has evaluated JD based on tangible metrics: retail gross merchandise volume (GMV), supply chain efficiency, fulfillment costs, logistics assets, and profit margin recovery. JD is undoubtedly a formidable retail player—but also a weighty one. The upside lies in its strong competitive moats; the downside is a perceived lack of market imagination. Discussions about growth invariably revolve around questions like: Will price wars erode profits? Will logistics investments strain cash flow? Is user growth still sustainable?

With JoyInside, JD has an opportunity to rewrite its narrative: AI hardware as terminal gateways.

This, I believe, represents its greatest potential for a valuation re-rating.

I. AI Gateways Are Expanding Beyond Apps, and JD Is Positioned Between Hardware and Supply Chain

Over the past year, discussions about AI gateways have primarily focused on Alibaba, Baidu, and Tencent.

Alibaba emphasizes AI-driven e-commerce, focusing on redefining search, recommendations, shopping guides, and advertising. Baidu highlights its AI cloud and intelligent agents, aiming to drive new growth beyond search advertising. Tencent discusses Yuanbao, Hunyuan, and the WeChat ecosystem, with markets watching to see whether social and content scenarios can achieve scale.

JD’s approach is distinct. Rather than confining AI to apps, it is integrating AI into hardware—smart lamps, mattresses, cooking robots, wheelchairs, AI toys, and health monitors. While these categories may seem disparate, they share a common trend: AI is transitioning from “you ask questions in an app” to “you access capabilities directly in real-life scenarios.”

AI in mobile apps competes on entry points, usage time, account systems, and content ecosystems. AI in hardware competes not only on model capabilities but also on channels, categories, supply chains, after-sales, installation, services, and repurchases. For example, selling an AI study lamp is just the first step; content, tutoring, eye protection, data security, and user retention follow. An AI health device may lead to consultations, medication, follow-ups, deliveries, and in-home nursing. A smart wheelchair ties into elderly care, repairs, insurance, and family nursing.

This is JD’s sweet spot. Its strength has never been in creating the most buzzworthy traffic hubs but rather in building robust chains for goods, warehousing, delivery, after-sales, and services. Previously, this supported home appliances, 3C products, and daily essentials. Now, if AI hardware scales, it could become an AI terminal distribution capability.

Here’s where JD becomes truly intriguing: It may not build China’s most powerful large language model, but it could emerge as one of the most critical sales, delivery, and service platforms for AI hardware. For capital markets, that’s a game-changer.

If JD merely offers AI customer service or shopping guides, the market will view it as an efficiency improvement—lower operating expense ratio, higher conversion rates, reduced customer service costs. But if JoyInside succeeds, the market must reevaluate JD’s valuation anchor: Is it a retail platform or an AI hardware distribution network for household scenarios?

The distinction is significant. Retail platforms are valued on profit margins; AI terminal platforms are valued on gateway significance. The former faces valuation pressure from price wars; the latter can argue for asset revaluation.

II. JoyInside’s True Value Lies Not in Selling More AI Devices—But in Extending JD’s Reach

JoyInside shouldn’t be reduced to “JD selling AI hardware too.”

JD already sells hardware—home appliances, 3C, smart homes—as a core battleground. But JoyInside adds a layer: It wants JD to evolve from “selling others’ hardware” to “participating in defining hardware intelligence.”

This step is subtle but valuable. Historically, JD’s role in hardware was channel-driven: brands made products, JD handled sales, fulfillment, after-sales, and user operations. With JoyInside embedded in hardware, JD moves beyond shelves into product interaction layers. How devices wake up, converse, access services, and connect to JD Health, JD Home Services, JD Daojia, or JD After-Sales becomes new revenue opportunities.

This is JD’s most compelling AI narrative: shifting from one-time transactions to recurring services.

Hardware sales are the first layer. Strong 618 sales boost GMV but offer limited valuation upside. The second layer is brand partnerships—nearly 200 brands onboard JoyInside—suggesting JD can monetize AI capabilities as a platform service. The third layer is post-activation service revenue: health management, consumable repurchases, memberships, in-home care, maintenance, and content subscriptions. The fourth layer is long-term data loops and user relationship deepening.

If this chain works, JD’s business model will undergo marginal shifts. Users previously “bought” on JD, with weak post-purchase engagement unless repurchasing or needing after-sales. AI hardware changes this. Persistently online devices can transform transactional relationships into service-based ones. A JoyInside-connected health device might see weekly usage; an AI toy, daily interaction; a smart mattress, long-term sleep tracking; a cooking robot, ongoing ingredient, consumable, recipe, and after-sales needs. This is JD’s desired long-chain engagement.

Capital markets love long chains for stronger order visibility, higher user lifetime value, and profit potential beyond markup margins.

Of course, JD isn’t there yet. It’s still in the catalytic phase, with thematic hype and expectation repair. But it offers a new trading framework: past JD valuation focused on retail margin recovery; future valuation hinges on whether AI hardware drives new categories, services, and gateways.

Many still view JD through old e-commerce lenses, but JD is integrating AI into hardware, health, logistics, and services. JD Logistics has Chaonao (a large logistics model), JD Health has Jingyi Qianxun (a medical AI model), JD App has Jingyan, and hardware has JoyInside. Individually, none are revolutionary, but together, they resemble a JD-style “AI operating system”: front-end sales, mid-end fulfillment, back-end services, all powered by AI.

JD lacked a compelling capital story. JoyInside gives it one.

III. From Thematic Trading to Valuation Shift: Three Hurdles for JD

My stance: Pay attention, but don’t rush to conclusions.

AI hardware is hot, but the sector faces two risks: flashy launches with poor user retention; impressive shipment numbers but weak service revenue.

For JoyInside to drive valuation shifts, JD must clear three hurdles.

Hurdle 1: Real Usage Frequency

Ten million devices sound impressive, but capital markets won’t settle for installation numbers alone. Post-sale questions matter: How many devices are activated? Daily usage? Retention? Repurchases? Service conversions? If AI labels merely mask one-time sales, asset revaluation is unlikely.

AI hardware isn’t just about voice assistants. Consumers are pragmatic—if functions are clunky, slow, or lack real-world scenarios, devices gather dust. The industry has seen too many “smart hardware turning dumb.” JD must prove JoyInside isn’t marketing fluff with real usage data.

Hurdle 2: Standardized Brand Integration

Nearly 200 brand partnerships are positive but introduce complexity. Hardware categories vary in chips, sensors, interactions, and service chains. Custom projects for each brand raise costs and erode margins at scale. To scale as a platform, JD must standardize model capabilities, device integration, content security, service orchestration, and after-sales systems.

Standardization is key to marginal cost reductions.

Otherwise, JoyInside risks becoming an “AI project factory”—loud but unprofitable.

Hurdle 3: AI Investments Reflecting in Financials

JD’s 200%+ YoY AI R&D growth is a catalyst but also a pressure point. Investments boost risk appetite but disrupt short-term profits. Investors now demand more than R&D spending—they want revenue, gross margins, and cash flow improvements.

JD has spent years repairing profit margins, shifting market pricing logic from “money-burning retail” to “efficient supply chain platform.” Increased AI spending must clarify ROI paths. If JoyInside drives AI hardware GMV, brand service revenue, device activation income, and health service revenue, it becomes a mid-term narrative. If it’s just AI hype without earnings, it reverts to cost pressure—JD’s AI story must deliver results.

Alibaba’s AI narrative focuses on cloud revenue and e-commerce ads; Baidu’s on AI cloud, search redefinition, and Robotaxi; Tencent’s on WeChat, gaming, and ad monetization. JD’s AI story will ultimately hinge on whether it can embed AI into supply chains, hardware terminals, and service fulfillment—and translate that into visible revenue.

JD’s smartest move is avoiding claims of “strongest model.” It knows its strength lies in supply chains and service networks. JoyInside as merely an AI hardware brand has limited upside; as JD’s foundational capability to connect brands, devices, and household services, it reshapes valuation logic.

Short-term, this is a 618 AI catalyst; mid-term, it’s JD’s strategic shift from retail efficiency to AI terminal distribution; long-term, it’s JD’s chance to escape low-valuation retail frameworks.

I see JoyInside as JD’s new card for capital markets.

It hasn’t cashed in yet, but it lets JD move beyond price wars, margins, and fulfillment costs. JD’s old story was “selling goods faster, stabler, cheaper.” Now, it wants AI hardware to enter households via JD, then bring services back through those devices.

If this chain works, JD won’t just sell AI hardware.

It’ll sell the next generation of household gateways.