June 12, 2018 - China Announces Artificial Intelligence Development Programs

On June 12, 2018, you're looking at one of China's most consequential AI moves — the formal launch of programs built directly on the 2017 New Generation AI Development Plan. China committed funding, named national AI teams, and targeted industries like manufacturing, healthcare, and transport. It also unveiled its first cloud AI chip. These programs weren't incremental — they were designed to make China the world's dominant AI power by 2030, and the details reveal exactly how.

Key Takeaways

• China's Three-Year Action Plan (2018–2020) prioritized developing high-end smart products and enhancing intelligent manufacturing to transform the real economy.
• The plan built directly on the 2017 New Generation AI Development Plan, adopting a three-phase timeline targeting 2020, 2025, and 2030 milestones.
• Target sectors included intelligent vehicles, service robots, drones, and medical imaging as central application areas.
• The State Council designated fifteen Chinese companies as national AI teams, each leading development within specific specialized sectors.
• China's 1+N funding model blended public-private capital, leveraging government guidance funds alongside direct subsidies and social capital investment.

What China's 2018 AI Announcement Actually Covered

When China released its Three-Year Action Plan for Promoting Development of a New Generation Artificial Intelligence Industry (2018–2020), it wasn't starting from scratch—the plan built directly on the 2017 New Generation AI Development Plan and Made in China 2025.

You'll notice the announcement focused on four core priorities: developing high-end smart products, reinforcing foundational technologies, enhancing intelligent manufacturing, and improving public support systems.

Rather than addressing ethics governance or education workforce pipelines explicitly, China concentrated on industrializing AI applications and transforming its real economy. Much like how modern platforms rely on engagement signals rather than follower counts to determine distribution, China's AI strategy prioritized measurable industrial output over institutional prestige.

The plan targeted specific sectors—intelligent vehicles, service robots, drones, and medical imaging—while positioning China as both a science and technology superpower and a dominant cyber superpower by 2020. Complementing these national efforts, the China AI Development Report 2018 was formally launched at Tsinghua University on July 13, 2018, further documenting the country's trajectory in artificial intelligence.

The 2017 Blueprint That Made It Possible

Issued by China's State Council on July 20, 2017, the New Generation AI Development Plan served as the top-level blueprint that gave the 2018 Three-Year Action Plan its foundation and direction. It set three clear phases: basic progress by 2020, major breakthroughs by 2025, and global AI leadership by 2030. You can think of it as the strategic backbone from which all subsequent programs drew their structure.

The plan prioritized building open innovation systems, establishing ethics frameworks to govern responsible development, and constructing talent pipelines through education and workforce training. It also called for integrating AI across economic, social, and defense sectors. Without this 2017 foundation, the 2018 announcement wouldn't have had the policy infrastructure or strategic clarity needed to move forward effectively. The plan also identified key technological enablers driving this new stage of AI development, including mobile Internet, big data, supercomputing, sensor networks, and brain science as the critical forces enabling breakthroughs across priority research areas.

China's leadership explicitly viewed AI as core to international competition, recognizing that achieving first-mover advantage in the field was essential to the country's broader national strategy for innovation-driven development. This mirrors how foundational open-source efforts, such as Linus Torvalds' release of the Linux kernel in 1991, demonstrated that early strategic investment in a transformative technology can shape an entire field for decades.

How the 1+N Funding Strategy Actually Works

The "1" anchors forward-looking development of next-generation AI through dedicated major S&T programs, while the "N" draws on existing initiatives like National Key R&D Projects in high-performance computing, big data, robotics, and quantum information.

You'll find the financing model blends public-private capital through government guidance funds — totaling 2,107 by 2022 with a $1.86 trillion target size — alongside direct subsidies, competitions, and social capital.

Grassroots innovation gets support through low-cost hackerspaces and mass innovation bases, ensuring resources reach beyond elite institutions and into broader AI development ecosystems. Infrastructure for AI workloads increasingly relies on cloud computing platforms like AWS, which spans 39 geographic regions and 245 countries and territories to support global deployment at scale. By 2023, actual capital raised through these guidance funds reached $940 billion (6.51 trillion RMB) from combined private and public sources.

Inside China's New Generation AI Megaproject

China's New Generation AI Megaproject sets an ambitious course toward global AI leadership by 2030, targeting breakthroughs in theory, technology, and applications across every major sector. You'll see this initiative unfold across three critical milestones:

1. 2020: Advances in big data, swarm, and autonomous intelligence establish foundational progress.
2. 2025: World-leading breakthroughs push AI industry value beyond 400 billion RMB core and 5 trillion RMB related sectors.
3. 2030: A mature AI theory system transforms manufacturing, medicine, agriculture, cities, and national defense.

The megaproject cultivates domestic talent while embedding ethical standards into its innovation framework. It's not just technology-building — it's reshaping China's entire economic and social infrastructure through brain-inspired computing, hybrid intelligence, and autonomous systems deployed at unprecedented scale. Parallel to China's push, enterprise software leaders like SAP have already embedded machine learning and predictive analytics into core business modules, demonstrating how AI integration at scale can reshape industries from manufacturing to finance.

Which Industries Did China Target for AI Growth?

Backing its bold 2030 ambitions with sector-specific action, China's New Generation AI Megaproject didn't just set targets — it zeroed in on five industries as primary engines of AI-driven growth: manufacturing, healthcare, finance, retail, and transport.

In manufacturing, AI adoption surpassed 30% among larger firms by 2025, focusing on production optimization and quality management. Samsung's own AI-driven factory strategy similarly targets full automation by 2030, deploying AI agents for quality control, production, and logistics across its global manufacturing network.

Healthcare innovation drove AI into assisted diagnosis, drug R&D, and traditional Chinese medicine acceleration.

Finance leveraged AI for credit risk, fraud detection, and hyper-personalized engagement, with 90% of Chinese firms viewing generative AI as essential.

Retail transformation doubled AI-influenced firms by 2024, targeting 71% efficiency gains through personalized marketing and smart logistics.

Transport prioritized autonomous driving and intelligent cabins, reinforcing China's broader AI Plus initiative across interconnected sectors. By 2025, the number of AI companies in China had surpassed 6,200, with the core AI industry valued at more than 1.2 trillion yuan, underscoring the country's rapid expansion across all targeted sectors. These efforts were further supported by government-backed investment structures, as the State Council designated fifteen China-based companies as national AI teams, each assigned to lead development within specific specialized sectors.

What China's First Cloud AI Chip Revealed About Its Strategy

Releasing China's first cloud AI chip in May 2018, the Chinese Academy of Sciences signaled something bigger than a hardware milestone — it revealed a strategic pivot toward AI self-sufficiency.

Developed by Cambricon Technology and backed by Alibaba, the MLU100 processed over 1.66 terabytes per second, targeting servers and data centers directly. It told you three things about China's direction:

1. China was building domestic supply chains to reduce foreign tech dependence
2. China was pursuing cloud sovereignty through homegrown semiconductor infrastructure
3. China was expanding from edge devices into large-scale cloud AI environments

This chip wasn't just hardware — it was a declaration that China intended to compete globally on its own technological terms. The MLU100 was specifically engineered to support deep learning tasks, enabling capabilities such as data analysis, pattern recognition, and prediction across large-scale AI environments. The announcement was publicized through Xinhua and China Daily, reflecting the government's intent to broadcast this technological achievement to both domestic and international audiences. This drive for semiconductor independence mirrors how Qualcomm's patent and licensing model allowed it to embed its technology into global standards, demonstrating how controlling foundational intellectual property shapes long-term technological influence.

How China's AI Patent and Paper Output Shifted After 2018

Tracking China's AI patent output after 2018 tells you just how deliberately the country converted policy ambition into measurable innovation.

AI patents jumped from 3% of China's total patent output in 2014 to 20.5% by 2023, with China granting 183,302 AI patents that year compared to just 48,197 in the US.

You'll notice the growth wasn't uniform—computer vision and machine learning accelerated sharply, reflecting targeted industrial policy rather than broad experimentation.

China's organizational structure reinforced this momentum, with research collaboration spanning tech firms, universities, and government institutions simultaneously.

That distributed model contrasts directly with US concentration among large private incumbents. Despite China's commanding lead in total volume, the United States continues to produce more higher-impact patents by citation and commercial significance.

Patent quality also gained regulatory attention, with China revising examination guidelines in 2025 to tighten disclosure standards for AI model applications. The 2025 revisions, effective January 1, 2026, require AI model patents to disclose essential components, hierarchical structure, and the specific steps, data, and parameters needed for training. Baidu's ERNIE platform illustrates how domestic AI development translated into commercial deployment, with 300 million users and 14.65 million developers engaging with the model since its integration into Baidu's core search infrastructure.

Where China's AI Research Funding Actually Went

When China announced its New Generation AI Megaproject in October 2018, the Ministry of Science and Technology didn't just signal intent—it committed 870 million RMB across up to 39 projects targeting fundamental theories, key technologies, smart chips, and integrated systems.

You can trace where the funding actually landed across three channels:

1. The National Natural Science Fund directed nearly 700 million RMB toward ~500 projects covering algorithms, theory, and robotics research.
2. National Key R&D Programmes embedded AI into existing data pipelines across key industrial sectors.
3. Public investment reached a few billion dollars total, tilted heavily toward applied research and experimental development.

Funding transparency remained limited, though, since guidance funds operated more like venture capital than structured R&D allocations. China's public AI R&D spending was assessed to be on the same order of magnitude as the United States' planned FY2020 federal AI R&D spending, suggesting rough parity rather than dramatic outspending. Among the key hardware targets embedded in China's smart chip initiatives, processors built on ARM's licensing model attracted particular attention due to their widespread adoption across mobile and embedded applications, with ARM growing from 1 licensee in 1990 to over 280 licensees globally. By 2025, state capital had moved beyond infrastructure and research labs to invest directly in commercial frontier-AI companies, with Shanghai State-owned Capital Investment leading rounds in frontier-LLM startups such as MiniMax and StepFun.

Why China's 2018 AI Programs Are Built to Dominate by 2030

Beyond where the money went lies the more revealing question: what was it actually designed to achieve? China's 2018 AI programs aren't incremental investments—they're a sequenced blueprint targeting global dominance by 2030.

The three-tiered roadmap moves deliberately: optimize development conditions by 2020, secure world-leading technologies by 2025, then claim primary global innovation status by 2030. You're looking at compressed timelines that reflect serious national prioritization, not bureaucratic scheduling.

Every layer reinforces another. Talent pipelines feed innovation ecosystems. Industrial applications generate commercial dominance. Regulatory frameworks legitimize expansion. The geopolitical implications are unavoidable—this isn't just economic competition, it's structural repositioning.

The ethical tradeoffs embedded here deserve equal scrutiny. Governance frameworks are built incrementally, trailing technological deployment rather than guiding it, which tells you something critical about what's actually being prioritized. Underpinning the entire strategy is an explicit push for open-source collaboration, bridging industry, academia, and military sectors to accelerate innovation across every tier of the roadmap.

Hardware acceleration is central to realizing these ambitions, with China's AI ecosystem increasingly dependent on purpose-built chips that distribute workloads across CPU, GPU, and dedicated AI processing units to enable on-device intelligence without cloud dependency.