Following the fierce delivery war, the dining sector still faces many hurdles. Xiaocaiyuan, a chain restaurant operator, is restructuring its business model through proactive price cuts, digital tools, and AI, seeking a path forward in an era of market saturation.

After the intense battle for delivery market share, many food companies are now reassessing what delivery truly brings and re-evaluating the value of different consumer scenarios and sales channels. Around August 2025, the delivery war hit its peak, pushing daily order volumes across the market from roughly 80-90 million to over 200 million. For restaurants, it seemed like a feast: orders surged and revenue figures doubled. However, many brands were dragged into price wars, and the in-store dining experience was squeezed by platform delivery orders. The China Cuisine Association later described this as "inverted delivery pricing," where subsidies made delivery prices lower than dine-in prices, pushing consumers online while restaurants still bore rent, labor, and high platform fees.

Now, brands like Luckin Coffee and Xiaocaiyuan have responded with action. For instance, Luckin's 2025 delivery war results saw significant costs flowing into logistics, causing net profit growth to lag at half the pace of revenue growth. Its 2026 strategy involves accelerating store expansion, using higher store density to guide consumers toward "pick-up" scenarios, hedging against over-reliance on delivery channels. Xiaocaiyuan (0999.HK), an Anhui cuisine brand, is adopting a similar approach. This chain, with over 800 directly operated stores, has initiated a deep adjustment: scaling back delivery, proactively lowering menu prices, guiding customers back to dine-in, while advancing digital and AI initiatives.

Wang Shugao, Chairman of Xiaocaiyuan, has defined the period from late 2025 to mid-2026 as a "year of reform." On August 14, the company released its first interim report following the reforms. The results show first-half revenue of RMB 2.903 billion, up 7.0% year-on-year; net profit of RMB 289 million, down 24.3%; and adjusted net profit of RMB 263 million, a decline of 31.3%. Beneath the "revenue growth without profit growth" headline, the financials reveal a structural shift: dine-in is up, delivery is down. According to the report, dine-in revenue grew 18.1% year-on-year, with its share of total revenue rising from 60.7% to 67.0%. Delivery revenue fell 10.6%, with its share dropping from 39.0% to 32.6%. Table turnover improved from 3.1 times per day to 3.5, and inventory turnover days shortened from 25.4 to 19.7 days. Meanwhile, due to proactive price cuts, average spending per customer decreased from RMB 57.1 to RMB 50.5.

During the delivery war, a surge of delivery orders at peak meal times disrupted Xiaocaiyuan's dine-in service. By braking on delivery, the company aims to reduce dependence on it and strengthen the dine-in experience. Concurrently, digital and AI initiatives are reaching operational endpoints, enhancing supply chain and store efficiency. In early August, Wang Shugao told media that due to the deliberate reduction of delivery and price cuts, profitability in the first half of 2026 was poor. However, starting in June and July, as membership grew and repeat purchases improved, same-store sales turned positive, and the effects of the reforms are becoming visible. Xiaocaiyuan's approach may serve as a reference model for self-rescue and transformation in the restaurant industry.

Breaking Free from Delivery Dependence

A study by Fudan University, based on transaction data from over 40,000 restaurant merchants, showed that as subsidies increased, average daily order volume grew by 7%, but daily actual revenue fell by about 4%, with total profits dropping an average of 8.9% during intense competition. In August 2025, Wang Shugao noted that during the July "delivery war," Xiaocaiyuan's delivery orders surged, with over 100 orders arriving at 11 AM on some days, diverting management attention and hurting the dine-in experience. Shifts in the industry and consumer environment made management realize the old growth model needed to change. Under the platform's algorithmic traffic and subsidy rules, stores were losing control over their operations. So, amidst the subsidy frenzy, Xiaocaiyuan decided to slow down and regain control. From August 2025, it announced it would no longer participate in any platform delivery discounts. This immediately impacted business. At the time, delivery accounted for about 40% of revenue. After exiting discount activities, delivery order volumes dropped in the short term, with the monthly share falling below 30% at its lowest. Wang Shugao said cutting delivery initially cost the company about RMB 50 million in revenue. However, Xiaocaiyuan did not abandon delivery entirely. Instead, it later split delivery into "idle time" and "peak time" modes. During peak hours, dine-in experience takes priority; during idle times, more delivery orders are accepted to boost sales.

In October 2025, Xiaocaiyuan brought in Gao Heng, a former technology expert from Alibaba, as Chief Information Officer (CIO) to lead digital and AI transformation at the store level. According to Gao, over the past two months, while delivery sales grew 22% month-on-month, average dine-in meal preparation time went from 8.06 minutes to 8.19 minutes, nearly unchanged, and delivery preparation time stayed between 7.13 and 7.19 minutes. This indicates that the 22% delivery growth did not compromise fulfillment capability. Wang Shugao commented on delivery: "Delivery is a necessity and a lifestyle. We're not abandoning it, but we need to control the degree. We cap delivery during peak times to protect dine-in, and push delivery during idle times." Data for the first half of 2026 shows delivery's share at 32.6%, a clear drop from its peak of 40% last year.

While shifting focus toward dine-in, Xiaocaiyuan also began proactive price reductions at the end of 2025. Wang Shugao explained the decision was based on his belief that a "13% net profit margin was too high," leaving room to lower prices to benefit consumers and achieve "volume for value." For example, the signature stinky mandarin fish was reduced from RMB 128 to RMB 108, and now to RMB 98. Prices for large and small portions of braised pork were also cut from RMB 65 and RMB 45 to RMB 49 and RMB 39. The financial report describes this as a strategy to drive dine-in traffic through price reductions.

Strategically returning to dine-in, Xiaocaiyuan realized it had long relied on platforms and struggled to directly control user assets. Wang Shugao decided to build a private domain presence, keeping dine-in traffic within its own ecosystem and converting in-store customers into long-term members. In late 2025, the company launched a membership program called "88VIP": for RMB 88, consumers receive benefits like a 10% discount on all items, free redemption of specific dishes, and points accumulation for menu items. These perks are exclusively for dine-in use, completely separate from delivery platforms. According to Xiaocaiyuan, this approach is showing early success. As of end-July 2026, paid members reached 1.51 million, exceeding initial expectations. Zuo Cheng, head of marketing, revealed that repeat purchase rates among paid members exceed 65%, compared to about 28% for non-paid members. Based on first-half data, paid members contribute 213% of the value of regular members. Within six months, the membership system drove approximately 1.89 million incremental customer visits.

However, the cost of reform is also visible in the financials. Sales data from 617 same-stores (defined as stores operating for at least 150 days in both the six months ending June 30, 2025, and June 30, 2026, excluding new stores opened after Q2 2025) shows first-half 2026 same-store sales of RMB 2.211 billion, down 12.57% from RMB 2.529 billion in the same period last year. Store expansion has also slowed, with net new openings of fewer than 20 stores in the first half, against an original target of 200 new stores for 2026. But by June and July, conditions began improving. Zuo Cheng noted that in June 2026, same-store dine-in revenue turned positive year-on-year, driven largely by the 88VIP program and price cuts. This combination of adjusting delivery, lowering average order value, and building a proprietary membership base has reshaped Xiaocaiyuan's business structure. Delivery revenue fell from RMB 1.057 billion in the first half of 2025 to RMB 945 million, while dine-in revenue grew 18.1% to RMB 1.945 billion, increasing its revenue share from 60.7% to 67.0%.

For many restaurant brands trapped by platform traffic and price wars, Xiaocaiyuan's move is not easy, immediately costing it both revenue and profit. If key financial metrics show clear improvement in the next phase, this experiment could offer a replicable model for the industry. Yet, another industry example is Chagee (CHA.US), the listed tea chain. For much of 2025, it refused to join the delivery war, maintaining its own pricing. It ultimately reported 2025 revenue up just 4% and net profit down 53.5%. Founder Zhang Junjie reflected during an earnings call: "We took some detours" and "wasted half a year," admitting they "underestimated the impact of the delivery price war on the offline market."

Preparing for Large-Scale Expansion and Going Global

Another pillar of Xiaocaiyuan's transformation is the digitalization and AI adoption across the entire system, from stores to the supply chain. This is a path more restaurant companies are exploring. At Xiaocaiyuan's headquarters on the 33rd floor of the Jin Ying building in Nanjing's Jianye district, a massive screen displays real-time operational data. On August 9, key metrics for 833 operating stores—dine-in revenue, delivery revenue, membership sign-ups, repeat purchase rates—updated every minute. Gao Heng noticed that the day's real-time revenue was 10% lower than the same time the previous week. After a moment's thought, he realized: "Today is typhoon weather (landfall of 'Baihaitun'); last Sunday it wasn't."

In his 10-plus months at Xiaocaiyuan, Gao Heng has focused on building digital and AI capabilities at the store level. Each store has 13 to 16 screens across the front and back of house, with roughly 25 to 30 employees, meaning about two people share a screen. Since Gao arrived, every job's workflow has been digitized onto the screens: servers check table setups at 10 AM, bar staff have tasks at 10:20 AM, 2 PM, and 9 PM, and kitchen staff follow specific guidelines for cutting and preparing ingredients. All this information appears on screen at the designated times. "We want to minimize the skill requirements for frontline staff," Gao said. His goal is for any new graduate or even a part-time worker to know exactly what to do at any given time. The purpose is to prepare for larger-scale expansion and international ventures. "This is essential for scalable replication, allowing us to go from 800 stores to 3,000, or even overseas," Gao added. The digital system segments time and standardizes human actions into defined steps, replacing the traditional apprentice-style, experience-based training with "check the screen, follow the process."

AI extends into more nuanced areas. Gao revealed that in some stores, AI vision cameras monitor pass-through points and cooking stations, identifying whether dishes meet quality standards, whether staff are wearing masks, or if anyone is smoking in no-smoking zones. Currently, in nearly 300 stores, cooking robots have taken over repetitive tasks. Gao noted that stores using cooking robots have optimized kitchen staffing, with labor costs trending down and per-capita performance rising. In terms of operational efficiency, meal prep time and table turnover rates at robot-equipped stores outperform regular ones. On the supply chain side, digital and AI efficiency gains are also active. In May 2026, Xiaocaiyuan's new Ma'anshan factory began operations, designed with capacity to support 3,000 stores. Wang Ting, the factory manager, said the biggest change is the large-scale replacement of labor-intensive processes with robots and automated equipment. For example, braised pork cutting, previously done by hand, is now handled by automated cutting robots that precisely portion according to preset dimensions, ensuring consistency while freeing up manpower. The chicken cube production line, which once required seven people for filling, is now nearly fully automated, with workers only monitoring equipment status.

Diverse Solutions in an Era of Market Saturation

Xiaocaiyuan's situation and reforms hold relevance and reference value across the dining industry. In 2026, China's food service sector is slowly recovering from brutal price competition, but the overall market remains challenging. National Bureau of Statistics data shows that in the first half of 2026, national catering revenue was RMB 2.83 trillion, up 2.8% year-on-year, a 1.5 percentage point slowdown compared to the same period last year. Growth for enterprises above a designated size was just 1.8%. The China Cuisine Association notes that the industry has left the era of scale-driven growth and entered a phase of high-quality development focused on quality, efficiency, and value. Since the second half of 2025, a wave of transformation has spread among chain restaurants: Haidilao is contracting franchise expansion and optimizing store models, Yum China is enhancing its digital supply chain, and Laoxiangji has deployed cooking robots in hundreds of stores. The directions vary, but the underlying theme is consistent: as the dividends of extensive expansion fade, turning inward for efficiency has become a common choice for leading players.

Haidilao is among the more systematic transformers. After 2025 net profit fell 14% and table turnover dropped from 4.1 to 3.9, the hotpot leader began closing underperforming stores, slowing expansion, and launching community-based all-purpose stores to strengthen dine-in scenarios. On January 13, 2026, founder Zhang Yong returned as CEO after four years, with subsequent adjustments focused on stabilizing the core hotpot business, advancing the "Red Pomegranate Plan," and building a smart middle platform. Gou Yiqun, after stepping down as CEO, shifted to overseeing "intelligent and automated management processes" and the smart platform. Compared to Xiaocaiyuan, Haidilao's adjustments are more systemic, covering organization, stores, brands, and technology.

Yum China offers another approach. In the first half of 2026, the company behind KFC and Pizza Hut reported total revenue of $6.409 billion, up 11% year-on-year, with net profit up 9%, maintaining growth in both metrics. This performance was supported primarily by a "scale plus franchise" strategy, with net additions of 1,196 stores. Analyst firm Dolphin Investment Research suggests Yum's competitive moat lies in its self-built logistics parks, tightly integrated supply chain, and a digital ecosystem centered on its super app—using apps, mini-programs, memberships, and delivery for multi-channel customer reach, with "Yum Cloud" powering supply chain and store management. The digital advantage is creating a closed loop of "traffic acquisition - member retention - smart operations - repeat purchase conversion": the more proprietary the channels and the deeper the private domain, the more precise the operations and higher the repeat purchase rate, which in turn lowers customer acquisition costs. However, there are concerns: delivery revenue now accounts for over 50% for both KFC and Pizza Hut, and dine-in momentum is weakening.

AI is a key that many restaurant companies are trying to unlock the future with. Green Tea Group, a listed dining company, reached an AI strategic partnership with Ant Digital Technologies on July 27. The collaboration will focus on store inspections, business analysis, and membership operations, using AI agents to improve operational management, commercial decisions, and member growth. Laoxiangji, another brand from Anhui like Xiaocaiyuan, is partnering with Tencent, DingTalk, and Ant Group's Afu to use AI in store operations and franchisee management, while also building food processing and warehousing logistics bases. The company's chief economist noted that these bases are key to upgrading from a basic industry chain to a digitalized full-chain operation.

Whether selling hotpot, chicken soup, or home-style dishes, leading companies are investing resources in three areas: supply chain (self-built central kitchens, direct sourcing, cold chain logistics), digitalization (real-time data, smart scheduling, sales forecasting), and AI and automation (cooking robots, AI store inspections, smart ordering). Investments once viewed as "cost items" are now becoming "qualification items" for valuation and survival. Xiaocaiyuan's investor relations advisory team head mentioned that restaurant valuations have traditionally been limited by store count, average order value, and table turnover, which determine revenue growth ceilings and profit margins. However, the application of physical AI might offer a breakthrough and become a second growth curve. Specifically, cooking robots can close the loop on the one unstandardized step in stores: "cooking." The biggest challenge in scaling Chinese cuisine is the "stir-fry" step. Ingredient preparation can be handled by central kitchens, and cold chain delivery can be daily, but the final cooking step has long depended on a chef's personal skill. Cooking robots solve this bottleneck by turning cooking from a craft into a replicable set of data parameters—oil temperature, number of stir-fries, when to add seasonings, how to adjust heat—all standardized.

However, a seasoned consumer investor noted that in dining, AI and digitalization ultimately must demonstrate real utility for store models. These tools are means, not ends; the evaluation criteria must rest on whether key single-store metrics improve sustainably. Shen Meng, founder of Chanson Capital, believes that dining companies can only grow through expansion. "If you can't build a solid growth expectation and return potential for investors, just telling stories won't work," he said.