Rakuten Develops Japan’s Largest Energy-Efficient AI Language Model + Video

Rakuten Group has taken a bold step into the future of artificial intelligence, announcing the development of a new large-scale language model (LLM) specifically optimized for the Japanese language. As AI continues to redefine digital services worldwide, Rakuten’s innovation promises not only enhanced performance but also remarkable energy efficiency, positioning the company at the forefront of AI-driven solutions in Japan. This move could significantly impact e-commerce, customer engagement, and the broader AI ecosystem by setting new benchmarks for large-scale, sustainable AI technology.

Rakuten AI 3.0: A Breakthrough in Japanese LLMs

Rakuten’s newly unveiled model, Rakuten AI 3.0, boasts an unprecedented 700 billion parameters, making it the largest LLM developed in Japan to date. Unlike traditional models that activate all parameters simultaneously, Rakuten AI 3.0 dynamically engages 40 billion parameters depending on the specific information being processed. This selective activation reduces unnecessary computation, allowing the model to operate far more efficiently than previous systems. In tests, the model demonstrated a potential 90% reduction in electricity consumption compared to other large-scale AI models, highlighting its energy-conscious design.

The company plans to integrate this model into various Rakuten services, from e-commerce platforms to personalized user experiences, leveraging AI to understand customer preferences and enhance service quality. By 2026, Rakuten intends to offer this technology externally, opening opportunities for other companies to benefit from its performance and efficiency.

Evolution of Rakuten’s AI Capabilities

Before Rakuten AI 3.0, the company relied on smaller models of 1.5 billion parameters or combinations of eight 7-billion-parameter models. With support from Japan’s Ministry of Economy, Trade, and Industry, Rakuten accelerated development starting in August, pushing the boundaries of LLM technology in the country.

This strategic development is part of a broader trend among Japanese companies focusing on domestic, Japanese-language-specialized AI models. For instance, SoftBank completed a 460-billion-parameter research-focused model in 2024 and plans to develop even higher-performance models in the coming years. Rakuten’s focus on energy efficiency and real-world application represents a critical evolution in making AI both practical and sustainable for large-scale deployment.

What Undercode Say: Analytical Insights

Rakuten’s development of AI 3.0 represents a convergence of performance, scalability, and sustainability—a combination rarely achieved in AI infrastructure. By activating only a fraction of parameters based on immediate needs, the system not only conserves energy but also reduces the latency often associated with massive LLMs. This approach suggests a shift from “brute force” AI, where more parameters equal better performance, to intelligent parameter management, a model that balances cost, efficiency, and usability.

Integrating Rakuten AI 3.0 into e-commerce services can revolutionize personalized shopping experiences. With the model capable of analyzing consumer behavior in real time, Rakuten can offer highly targeted recommendations, dynamic pricing strategies, and even predictive inventory management. Over time, this could translate into measurable gains in customer retention, operational efficiency, and competitive advantage.

From a broader perspective, energy-efficient LLMs like Rakuten AI 3.0 address a growing concern in AI development: the environmental footprint of massive AI models. Globally, energy consumption for AI training and inference is under scrutiny, and Rakuten’s 90% reduction in power usage positions the company as a pioneer in sustainable AI innovation.

Additionally, by focusing exclusively on the Japanese language, Rakuten is capturing a niche often underserved by global LLMs. Many international models prioritize English, leaving localized AI applications with suboptimal performance. Rakuten AI 3.0’s linguistic specialization ensures higher accuracy and context understanding for Japanese consumers, giving domestic companies a clear advantage in AI-driven digital services.

This development also has implications for Japan’s tech industry and national competitiveness. By creating indigenous AI models, Japan reduces dependence on foreign technologies, supports domestic AI talent development, and fosters an ecosystem for startups and enterprises to leverage high-performance AI solutions tailored to local needs.

Finally, the planned external release in 2026 suggests a commercial strategy beyond internal deployment, positioning Rakuten not just as a service provider but also as an AI technology vendor. This dual focus—enhancing internal services while creating external revenue streams—could accelerate the adoption of Japanese-language LLMs both domestically and internationally.

Fact Checker Results

✅ Rakuten AI 3.0 has 700 billion parameters, making it Japan’s largest LLM.
✅ Selective activation of 40 billion parameters allows up to 90% energy savings.
❌ There is no confirmation that the model is currently available for external use; plans target 2026.

Prediction

📊 Rakuten AI 3.0 will likely redefine AI applications in Japan’s e-commerce sector, driving personalized consumer experiences and operational efficiency. Its energy-saving model could influence other tech companies to adopt sustainable AI practices. By 2026, the model’s external release may create a competitive market for Japanese-language LLMs, potentially sparking international interest and collaboration.

▶️ Related Video (92% Match):

🕵️‍📝✔️Let’s dive deep and fact‑check.

References:

Reported By: xtechnikkeicom_fe9355e58f5b1da5c27e6927
Extra Source Hub (Possible Sources for article):
https://www.reddit.com/r/AskReddit
Wikipedia
OpenAi & Undercode AI