Rakuten Launches AI Commerce Platform Across Asia

Rakuten, Japan's largest e-commerce conglomerate, has officially launched an AI-powered commerce platform designed to transform online shopping experiences across 6 major Asian markets. The platform, dubbed Rakuten AI Commerce Engine (R-AICE), leverages proprietary large language models and computer vision to deliver hyper-personalized product discovery, dynamic pricing, and automated merchant tools — positioning the company as a serious AI-native competitor to Amazon and Alibaba in the region.

The rollout, which began in late June 2025, targets Japan, Taiwan, South Korea, Thailand, Indonesia, and India. Rakuten CEO Hiroshi Mikitani described the launch as 'the most significant platform investment in the company's 28-year history,' with an estimated $1.2 billion earmarked for AI infrastructure development through 2027.

Key Facts at a Glance

• Markets covered: 6 Asian countries at launch, with plans to expand to 10 by Q2 2026
• Investment: $1.2 billion committed to AI infrastructure through 2027
• Core technology: Proprietary LLM built on Rakuten's 1.7 billion-parameter language model, fine-tuned for commerce
• Merchant tools: Automated product listing, AI-generated descriptions in 12 languages, and predictive inventory management
• Consumer features: Conversational shopping assistant, visual search, and real-time personalized pricing
• Target: 30% increase in merchant conversion rates within the first 12 months

R-AICE Brings Conversational Shopping to the Forefront

The centerpiece of Rakuten's new platform is a conversational shopping assistant that allows users to describe what they want in natural language. Unlike traditional keyword-based search, R-AICE interprets complex, multi-attribute queries — such as 'a lightweight laptop bag that fits a 15-inch MacBook and looks professional for meetings' — and returns curated results ranked by relevance, price, and user reviews.

This approach mirrors what companies like Amazon and Google have been experimenting with in Western markets. However, Rakuten claims its model has a distinct advantage: it was trained on over 4 billion transaction records from its existing marketplace ecosystem, giving it deep understanding of purchasing patterns specific to Asian consumers.

The assistant also supports voice input in 8 languages, including Japanese, Korean, Thai, Bahasa Indonesia, and Mandarin. Visual search capabilities let users snap a photo of any product and instantly find similar items across Rakuten's marketplace. Early beta testing in Japan reportedly showed a 42% increase in product discovery rates compared to traditional search.

AI-Powered Merchant Tools Aim to Level the Playing Field

For sellers, R-AICE introduces a suite of automated merchant tools that dramatically reduce the friction of managing cross-border e-commerce operations. Small and medium-sized businesses (SMBs) have historically struggled to compete with large retailers on platforms dominated by algorithmic visibility.

Rakuten's new tools address this gap directly with several key features:

• Auto-listing generator: Merchants upload product images, and the AI generates optimized titles, descriptions, and tags in up to 12 languages
• Dynamic pricing engine: Real-time price optimization based on competitor analysis, demand forecasting, and inventory levels
• Predictive inventory management: AI models forecast demand spikes 14 days in advance, reducing stockout rates
• Customer sentiment analysis: Natural language processing scans reviews and social media to flag product issues before they escalate
• Ad creative automation: Generative AI produces banner ads, social media posts, and email campaigns tailored to each market

Rakuten estimates these tools can save merchants an average of 15 hours per week on operational tasks. The company is offering the basic tier free for the first year to accelerate adoption among its existing base of over 56,000 active merchants.

How R-AICE Stacks Up Against Amazon and Alibaba

Amazon has been aggressively integrating AI into its marketplace through its Rufus shopping assistant and AI-powered listing tools launched in 2024. Alibaba similarly invested heavily in its Tongyi Qianwen model to power Taobao and Tmall search experiences. Rakuten enters this arena as a smaller but highly focused player.

The key differentiator lies in Rakuten's regional specialization. While Amazon and Alibaba deploy generalized AI models across global operations, Rakuten has built market-specific models that account for cultural nuances in shopping behavior. For example, Japanese consumers tend to prioritize detailed product specifications and packaging quality, while Indonesian shoppers are more price-sensitive and respond better to social proof elements like user-generated content.

Rakuten's proprietary LLM, based on its 1.7 billion-parameter architecture, is smaller than the frontier models used by its competitors. But the company argues that a smaller, purpose-built model fine-tuned on commerce-specific data outperforms larger general-purpose models in transaction-relevant tasks. Internal benchmarks reportedly show R-AICE achieving 18% higher accuracy in product recommendation relevance compared to GPT-4-based alternatives.

The $1.2 Billion Infrastructure Bet

Rakuten's investment extends well beyond software. The company is building 3 new AI-optimized data centers — in Tokyo, Jakarta, and Bangalore — specifically designed to support low-latency inference for real-time commerce applications. These facilities will run on Rakuten's own cloud infrastructure, reducing dependency on AWS and Microsoft Azure.

This infrastructure play aligns with a broader trend among Asian tech giants. SoftBank, Samsung, and Tencent have all announced major AI infrastructure investments in 2025, collectively pouring over $15 billion into regional data center capacity. Rakuten's $1.2 billion commitment, while smaller in absolute terms, represents a significant bet for a company with a market capitalization of approximately $12 billion.

The data center strategy also serves Rakuten's mobile network operations. As the operator of Japan's fourth-largest mobile carrier, Rakuten Mobile, the company can leverage shared infrastructure to reduce per-unit costs across both its telecom and e-commerce businesses.

What This Means for Businesses and Developers

For Western companies eyeing Asian market expansion, Rakuten's platform represents both an opportunity and a competitive challenge. The automated translation and localization tools lower the barrier to entry for brands looking to sell in markets like Japan and Southeast Asia. However, the AI-driven personalization also means that local merchants who adopt these tools early will have a significant advantage in visibility and conversion.

Developers and AI practitioners should pay attention to Rakuten's approach of using smaller, domain-specific models rather than relying on frontier LLMs. This strategy validates a growing industry consensus that fine-tuned, purpose-built models can outperform larger general-purpose systems in vertical applications — a trend that could reshape how enterprises think about AI deployment.

For consumers, the immediate impact will be felt most strongly in Japan, where Rakuten holds approximately 28% of the e-commerce market. Shoppers can expect more intuitive search experiences, better product recommendations, and potentially lower prices driven by the dynamic pricing engine's competitive optimization.

Looking Ahead: Expansion Plans and Industry Implications

Rakuten has outlined an aggressive timeline for R-AICE's evolution. By Q4 2025, the company plans to introduce augmented reality try-on features powered by its computer vision models, initially targeting fashion and cosmetics categories. A full API suite for third-party developers is expected in early 2026, enabling external apps to integrate R-AICE's recommendation and search capabilities.

The broader industry implications are significant. Rakuten's launch signals that the AI commerce race is no longer a two-horse competition between Amazon and Alibaba. Regional players with deep local data and cultural understanding can carve out defensible positions by building specialized AI systems rather than competing on model scale alone.

As AI-powered commerce becomes the norm rather than the exception, the winners will likely be those who combine strong foundational models with rich, proprietary transaction data. Rakuten's 28 years of marketplace data across Asia gives it a resource that even the largest Western tech companies cannot easily replicate. Whether that advantage translates into sustained market share gains will depend on execution — but the scale of investment suggests Rakuten is playing for keeps.