After the Cooling of Food Delivery Subsidies, JD.com Bets on an AI-Driven Efficiency Narrative

On August 13, JD.com released its second-quarter financial report, showing a divergent trend between revenue and profit.

Specifically, revenue decreased by 2.9% year-over-year to RMB 346.4 billion, while operating profit turned from a loss to a profit of RMB 4.5 billion. Adjusted net profit grew by 20% to RMB 8.9 billion.

In the second quarter of last year, JD.com was still competing for orders and users through food delivery subsidies and marketing investments. A year later, as subsidies tapered off and losses in new businesses narrowed, profits returned to the financial statements. However, the double-digit decline in home appliance and consumer electronics revenue indicates that growth pressure has not disappeared.

During the earnings call, JD.com CEO Xu Ran described this as a "clear turning point" in the profit trajectory. CFO Shan Su attributed the improvement to core retail gross profit, high-margin revenue streams such as advertising commissions, and narrowed losses in food delivery.

However, this appears more like a turning point for profit recovery rather than growth: JD.com Retail's revenue and operating profit are still declining. While food delivery losses narrowed significantly year-over-year, the quarter-on-quarter improvement was relatively limited. Management did not provide clear revenue guidance for the second half of the year.

More noteworthy is JD.com's changing approach to spending. In the second quarter, marketing expenses decreased by approximately RMB 6.7 billion year-over-year, while fulfillment and R&D expenses combined increased by about RMB 4.3 billion, with R&D spending rising by nearly 40%.

JD.com is shifting from purchasing traffic to building instant fulfillment capabilities and AI infrastructure. What will truly determine the quality of its growth going forward is whether investments in AI and supply chain can connect traffic, efficiency, and service revenue into a new growth loop.

After Narrowing Food Delivery Losses, Growth Pressure Returns to Core Retail Business

The decline in JD.com's second-quarter revenue was primarily driven by pressure on its traditional strong categories.

Revenue from electronics and home appliances fell 11.8% year-over-year, dragging down merchandise revenue by 5.4%. This is related to the high base in the same period of 2025, when trade-in policies were concentrated, leading to rapid growth in JD.com's home appliance and consumer electronics revenue. Entering 2026, the pace of subsidies, the high base, and price increases for some electronic products jointly suppressed sales volume.

During the call, management defined this as temporary pressure rather than a loss of competitiveness in the home appliance and consumer electronics sector. However, it is worth noting that JD.com did not provide guidance on the recovery magnitude of home appliance and consumer electronics revenue for the second half of the year, nor did it offer group revenue growth guidance.

It is evident that while the impact of the high base will gradually weaken, the timing of a return to positive revenue growth depends on consumer demand, not just financial base effects.

What truly helped JD.com offset the pressure from home appliances and consumer electronics was its general merchandise and platform businesses.

In the second quarter, general merchandise revenue grew by 5.6%, and service revenue increased by 6.8%, with platform and advertising service revenue rising by 8.3%. The proportion of service revenue in the group's total revenue rose from approximately 20.8% in the same period last year to 22.9%.

This has changed JD.com's profit structure. Self-operated merchandise revenue requires bearing procurement, inventory, warehousing, and distribution costs, whereas advertising and commissions do not occupy inventory and typically have higher gross margins. In the second quarter, JD.com's gross profit increased by 4.7% year-over-year to approximately RMB 59.3 billion, and the gross margin rose from 15.9% to 17.1%.

Shan Su emphasized during the call that the performance of platform and marketing service revenue was better than the overall average, which was a significant reason for the improvement in JD.com Retail's profit margin. In the second quarter, JD.com Retail's revenue decreased by 4.7% to RMB 295.4 billion, and operating profit fell by 3.3% to RMB 13.5 billion. However, the operating margin increased from 4.5% to 4.6%, reaching the highest level for JD.com during major promotional quarters.

Here, it is necessary to distinguish between "record-high profit margin" and "profit growth." The actual profit scale of JD.com Retail is still declining, but the decline in profit is smaller than the decline in revenue. Therefore, the 4.6% figure reflects the core business's ability to maintain profitability under revenue pressure, rather than indicating that the retail business has entered a period of rapid profit growth.

A larger source of the group's profit improvement was the narrowing of losses in new businesses such as food delivery.

In the second quarter, the operating loss of new businesses narrowed from RMB 14.8 billion in the same period last year to RMB 9.9 billion, a reduction of approximately RMB 4.9 billion. During the same period, JD.com Group's operating profit improved by about RMB 5.4 billion, meaning the narrowing of losses in new businesses contributed roughly 90% of this improvement.

Management attributed the narrowing of losses in new businesses mainly to improved operational efficiency in food delivery, increased revenue sources, and optimized marketing investments. They stated that investments in businesses such as Jingxi and Joybuy remain within the planned schedule.

However, compared to the second quarter of 2025, the narrowing of losses in new businesses was significant; but compared to the loss of approximately RMB 10.4 billion in the first quarter of 2026, the second quarter saw a reduction of only about RMB 500 million. In other words, food delivery has passed the most aggressive investment phase, but the speed of loss reduction has not continued to accelerate significantly.

Another easily overlooked issue is business reclassification.

At the end of October 2025, JD Logistics acquired the instant delivery business, which previously belonged to new businesses. Since 2026, part of the delivery revenue has been transferred from new businesses to JD Logistics. Therefore, the 47.6% year-over-year decline in new business revenue to RMB 7.3 billion cannot be directly interpreted as a decline in food delivery order volume.

Changes in expenses better illustrate JD.com's true choices. In the second quarter, marketing expenses decreased by approximately RMB 6.7 billion year-over-year, a drop of 24.8%; however, fulfillment expenses increased by about RMB 2.3 billion, a rise of 10.4%. JD.com is reducing front-end customer acquisition costs such as subsidies and promotions, but investments in riders, delivery networks, and instant retail infrastructure are still required.

This means that the next stage of food delivery competition will no longer be primarily about who offers larger subsidies, but rather who can retain users, increase delivery density, and generate revenue from merchant commissions, advertising, and cross-selling at lower subsidy levels.

However, JD.com did not disclose data on food delivery order volume, loss per order, commission rates, advertising revenue, or user retention during this call. Therefore, it remains uncertain whether the food delivery business model is nearing viability.

AI Has Entered the Expense Sheet, But Not Yet the Independent Revenue Sheet

If the core task of food delivery is to reduce losses, then AI bears another task: finding new sources of efficiency and growth for JD.com.

In the second quarter, JD.com's R&D expenses reached RMB 7.3 billion, a year-over-year increase of 37.7%, with the R&D expense ratio rising from 1.5% to 2.1%. Meanwhile, marketing expenses decreased by 24.8%. Looking at these expenses together, the flow of JD.com's resources is quite clear: marketing expenses decreased by about RMB 6.7 billion, fulfillment expenses increased by about RMB 2.3 billion, and R&D expenses increased by about RMB 2.0 billion.

This does not mean that every yuan saved from promotional expenses was directly invested in AI, but it confirms that the company's expense focus is shifting from traffic procurement to fulfillment and technology construction.

During the earnings call, management discussed AI within the framework of retail and supply chain efficiency, emphasizing that AI implementation is mainly focused on consumer terminals, advertising recommendations, customer service, procurement, logistics, healthcare, and industrial supply chains.

On the consumer side, JoyInside has been integrated with nearly 200 brands, and the cumulative scale of connected devices has more than tripled compared to the "Double 11" shopping festival in 2025. It aims to embed JD.com's voice interaction, product knowledge, and transaction services into AI toys, robots, and smart hardware.

For JD.com, if consumers can complete product inquiries, after-sales service, and even place orders through smart devices, these devices could become new transaction entry points outside the JD.com App.

Xu Ran stated during the earnings call that as of this year's "618" shopping festival, JD.com's JoyInside had partnered with nearly 200 brands, with the cumulative number of connected devices more than tripling compared to the 2025 Double 11 period.

In healthcare scenarios, during the "618" period, the number of users served by JD Health's AI doctor "Dawei" was nearly four times that of the same period last year. In industrial scenarios, JD Industry deployed over 70 AI agents in the first half of the year, covering procurement planning, sourcing, fulfillment, and after-sales service. In logistics, thousands of autonomous delivery vehicles have been in regular operation in more than 20 provinces.

These data prove that JD.com lacks no AI scenarios, but they do not yet prove that AI has become an independent growth curve.

In this financial report, JD.com did not separately disclose revenue from AI products. Furthermore, the efficiency improvements brought by AI have not yet formed a consistent signal in financial indicators.

In the second quarter, JD Logistics' revenue grew by 24.3%, and operating profit increased by 15.6%, but the operating margin decreased from 3.8% to 3.5%. JD.com's inventory turnover days also extended from 34.1 days in the same period last year to 40.5 days. This indicates that while automation and algorithms may have improved local efficiency, they have not fully offset the impacts of business expansion, inventory, and fulfillment investments.

This is also the most unique aspect of JD.com's AI strategy: the scenarios are specific, but they are not yet reflected in the financial statements.

Alibaba and Tencent can explain AI commercialization through cloud revenue, model invocation volumes, and capital expenditures; JD.com's AI value is more likely dispersed in advertising conversion rates, procurement costs, inventory turnover, fulfillment costs, and customer service efficiency.

The advantage is that the scenarios are real; once efficiency improves, it can directly impact a transaction scale of hundreds of billions of yuan. The disadvantage is that the effects are difficult to verify, making it easier for management to attribute overall efficiency improvements to technology investments.

Therefore, to judge whether JD.com's AI is effective, one should not only look at R&D expenses and the number of products, but also focus on three sets of indicators: whether platform advertising and commission revenue can continue to grow faster than merchandise revenue; whether inventory and fulfillment efficiency can improve with AI penetration; and whether AI products begin to disclose paying customers, revenue, and renewal rates.

From this financial report and call, it is evident that JD.com has achieved profit recovery, but has not yet formed a complete growth loop.

Food delivery proves that the company can reduce losses, platform business proves that the revenue structure can become asset-light, and AI proves that JD.com is willing to continue investing. What remains to be proven is whether these three elements can jointly generate new revenue growth, rather than just making a financial report with declining revenue appear more profitable.