This year marks the beginning of the 15th Five-Year Plan period, with a new blueprint already drawn. The "Focus Interview" program has launched a series titled "The Starting Year," focusing on key tasks to be steadily advanced over the next five years. Today's episode highlights the intelligent economy.

This year, terms like "small model," "intelligent agent," and "token" have increasingly entered public discourse, reflecting the growing influence of artificial intelligence across various sectors in China. The 15th Five-Year Plan outline proposes to further advance the development of Digital China and enhance digital-intelligent capabilities. The 2026 Government Work Report also introduced, for the first time, the goal of creating new forms of intelligent economy.

How can digital-intelligent levels be improved? What elements are required to build these new forms of intelligent economy? We visited the Yangtze River Delta region for insights.

With two months remaining until the start of the 2026 World Cup, related merchandise continues to sell well in Yiwu, Zhejiang. Factories in the area are operating at full capacity. In a knitting enterprise in Jinhua, Zhejiang, World Cup support scarves are being enthusiastically produced. This company, with only thirty employees, still uses handwritten work orders to record daily production. However, this year, the owner decided to break with this 22-year tradition by integrating each worker's tasks into a new system.

Here, the production organization methods of traditional small businesses are quietly transforming. By attaching a QR code, one can scan to view order quantities, production techniques, and track the real-time status of goods. Urgent and scattered orders previously deemed too risky can now be assessed and decided upon using system data. Whether small or large enterprises, connecting to digital-intelligent systems has become inevitable.

Last month, China's second domestically built large cruise ship, the "Ada·Flower City," was launched from the Shanghai Waigaoqiao Shipbuilding yard. In this mega-project involving over a thousand systems and more than 25 million parts, the application of AI technology significantly enhanced work speed and quality. At the shipyard, each production process is a complex system. In recent years, various systems have been developing their own AI models. Integrating these dispersed models into a more efficient, unified "brain" is a key focus for Waigaoqiao Shipbuilding.

In this starting year, regions across China are exploring how to leverage the trends of digitalization, networking, and intelligentization, using AI to empower numerous industries and foster new forms of intelligent economy. In Shanghai, practical efforts to deeply integrate AI with industry began several years ago.

Tang Wenkan, Deputy Secretary of the Party Committee and Director of the Shanghai Municipal Commission of Economy and Informatization, stated: "We must fully implement the 'AI+' initiative, use artificial intelligence to lead industrial paradigm shifts, and accelerate the creation of new forms of intelligent economy. The key lies in cultivating an innovative ecosystem and supplying new quality factors."

What kind of ecosystem and what elements are needed to build these new forms? This is a question continually pondered by AI expert Zhou Bowen. Zhou Bowen, Director and Chief Scientist of the Shanghai AI Laboratory, said: "We are at a historic inflection point. AI technology has matured to the point where it can genuinely transform industries, yet the full transformation hasn't entirely occurred. Infrastructure must transition from being merely adequate to being 'AI-ready.' Computing power, data, and models must be capable of coordinated scheduling."

Here, Zhou Bowen demonstrated the first crucial element for building the new intelligent economy: open-source, general-purpose large models. Zhou Bowen emphasized: "The value of an AI model isn't measured by test scores or correct answers, but by its ability to help scientists discover new scientific laws and assist engineers in solving practical industrial problems more efficiently."

In recent years, China has seen the emergence of several general-purpose open-source large models. In Shanghai alone, 150 large models have passed regulatory filing, providing significant便利 for AI development and spurring another new economic form: OPC, or "One Person + One Computer + AI tools," enabling individuals to start companies.

When reporters met Wang Wei, she was busy using AI software to adjust character designs. Two years ago, she was a deputy creative director at an advertising agency, collaborating with production companies and leading a team. Now, she has chosen to establish a "one-person company," independently handling the entire process of creating advertising short films.

Wang Wei's company is based in a "one-person company" startup community in Shanghai's Xuhui District, where many policies are tailored for "AI+" entrepreneurs. Lin Yue, Deputy Director of the Xuhui District Science and Technology Commission, explained: "We provide OPC enterprises with support such as model vouchers, computing power vouchers, and data corpus vouchers. We also offer a 'use first, pay later' model for computing power. Furthermore, financially, Xuhui District established the Zaozao Fund, offering 1 million yuan in financing support for excellent OPC projects. We have already invested in nearly 100 projects."

In this starting year, just in Shanghai's Xuhui District, five such OPC communities will be established, providing 2,500 workstations and attracting numerous one-person companies. Li Miaoyi, a one-person company entrepreneur, noted: "Yesterday, the park surveyed us about our needs. Our inherent need is for computing power because with it, we can compute more."

This corporate demand highlights the second key element for the new intelligent economy: sufficient computing power. Shanghai has already built a supercomputing scale exceeding 140 Petaflops. However, when training some large models, multiple computing centers might be utilized. Coordinating them to provide stable data services is a challenge being addressed by Zhou Bowen and his colleagues.

Zhou Bowen stated: "The laboratory is researching DeepLink technology, which can integrate discrete computing power. You can see computing resources coming from different locations, unified for training, and it can also manage scheduling across the future eight major national computing hubs."

To better meet computing power demands in the intelligent economy era, a series of new infrastructure projects are under vigorous construction in this starting year. Around computing power construction, some regions are exploring new pathways for intelligent economic development.

In Jiangbei New District of Wuhu, Anhui province, there is a building resembling a Rubik's Cube. Its futuristic interior houses a computing center named "Silicon Cube," where each light point represents a domestically produced server performing high-speed calculations. Wuhu is one of the ten national data center clusters in China's "East Data West Computing" project. Including Silicon Cube, seven computing centers have been built here, making computing power a new engine for the city's development.

Just months before the reporter's visit, space-based computing power from satellites was integrated here. Since the Wuhu cluster's construction began in 2022 under the "East Data West Computing" initiative, it has attracted many enterprises to build data centers. But how can these centers promote local development after completion? Wuhu city established a Big Data Construction, Investment, and Operation Company, creating Anhui Province's computing power coordination and scheduling platform, and identified industries related to computing power.

With a public platform for computing services, data processing has become the first industrial cluster centered around computing power. Upstairs from the big data company, an AI incubation base called "Computing Cube" has attracted over 20 enterprises related to algorithms and data.

During the visit, a thematic symposium on authorizing public data for the healthcare sector was underway, with the meeting room nearly full. Through the security technology platform built by Wuhu's Big Data Company, the city is considering authorizing the development of certain medical public data. Enterprises and experts from healthcare, data operations, and model development sectors attended. At this symposium, the reporter identified the third element driving the new intelligent economy: circulating data.

Wuhu is a city with a concentration of traditional manufacturing, including automotive, small appliances, and shipbuilding. With the continuous aggregation of data and computing power, the transformation of traditional manufacturing is accelerating. At one robotics company, several different robot models are launched each year. Initially operating as a project department within an automotive company, it became an independent entity in 2025 based on industry trends and its own situation, steadily increasing its product variety and quantity.

On a farm, data from every sensor and each drone flight, previously used only for that year's production, is now becoming a crucial foundation for large agricultural planting models.

Whether in Wuhu or Shanghai, regarding large model applications, computing power scheduling, or data circulation, the reporter gained a clear sense of a fourth essential element: security governance. Public data circulation must occur on independent, encrypted platforms. AI usage requires strengthened oversight, and the development of national AI security standards has already begun.

In this starting year, exploration of new forms of intelligent economy is expanding to more regions. Beijing is establishing national AI application pilot bases in healthcare, manufacturing, and education. Dongguan, Guangdong, hosted an application scenario innovation conference, opening over a hundred scenarios. Shandong is implementing the AI "Double Hundred Project," aiming to cultivate 100 enterprises and 100 industry-specific large models.