Competing in AI: Why Large Model Competition Must Go Beyond Price

Recently, leading AI companies in China have begun to reduce their product prices. On May 21, Alibaba Cloud announced cuts for its Tongyi Qianwen commercial and open-source models. That same day, Baidu made its two primary models, ENIRE Speed and ENIRE Lite, available for free. The following day, iFlytek also announced that its Xinghuo API capabilities would be offered at no cost.

Price mechanisms are fundamental to market dynamics, with price competition serving as a key indicator of industry rivalry. This competitive landscape drives companies to improve production efficiency, enhance product quality, differentiate their offerings, and deliver excellent customer service—essential components for maintaining a competitive advantage. Price competition is a pivotal phase in any industry’s evolution, leading to the rise of quality enterprises and contributing to overall growth.

Surviving the challenges of price competition is crucial for strengthening the industry, particularly in the large model sector. For AI model companies, acquiring users is vital; attracting users leads to increased demand, which translates into valuable data and developmental insights. While improving product quality and refining data processing technologies are important, the influx of user data and feedback is essential for fostering effective supply-demand interactions.

Lowering prices for large models encourages user growth and expands networks, enhancing product quality through valuable feedback from users. As the user base grows, the diversity and unpredictability of their needs open opportunities for validating market demand and exploring new applications. The interaction between technical advancements and user needs enriches application landscapes through ongoing engagement.

As of 2023, China’s core AI industry has reached a significant scale of 578.7 billion yuan, with over 100 large models exceeding one billion parameters. Predictions indicate that the market for large AI models in China will reach $21.1 billion by 2026. The increasing number of large model enterprises intensifies competition for users. Although price competition can be fierce, this relentless force spurs companies to continuously enhance their technological and service capabilities. In this environment, enterprises that provide superior value for money will thrive, while those that cannot adapt risk falling behind.

While competing on price can capture market share, accelerating the implementation of large model applications requires additional efforts. For instance, Chinese large model vendors can deepen collaboration with various industries to gain insights into specific needs, allowing them to deliver tailored AI solutions. If 2023 marks the beginning of China’s large model industry, 2024 is expected to usher in a phase of significant growth. With lessons learned from the current competition, the Chinese large model sector is poised for rapid and robust development.