World Artificial Intelligence Conference 2025: A Landmark in AI’s Global Ascendancy

From 26–28 July 2025, Shanghai hosted the eighth edition of the World Artificial Intelligence Conference (WAIC), a premier international AI summit held at the Shanghai World Expo Centre and Expo Exhibition & Convention Centre. This year’s theme, “Global Solidarity in the AI Era”, reflected a deliberate emphasis on international cooperation amid intensifying geopolitical competition in AI development. 

Spanning 70,000 m² of exhibition space, WAIC 2025 broke all previous records, hosting over 800 exhibitors, including 50 percent international or non‑Shanghai Chinese firms, showcasing more than 3,000 frontier products, among them 40 large language models (LLMs), 50 AI-powered terminal devices, and 60 intelligent robots, with over 100 products making their debut globally or in China. 

Beyond the technology showcase, the conference included a High‑level Meeting on Global AI Governance, bringing together academicians, award‑winning scientists, policymakers, industry leaders, startups, investors, and youth innovators to debate regulation, applications, ethics, and the future of AI collaboration. 

Scale, Ambition and the “AI Capability Corridor”

This year’s WAIC set new benchmarks:

1. An unprecedented exhibition footprint of 70,000 m², a 34.6 percent expansion over the previous year.

2. More than 800 companies: over 50 percent hailing from outside Shanghai, including global heavyweights like Siemens and Schneider Electric, along with emerging domestic leaders such as Unitree Robotics, iFlytek, AgiBot, and MiniMax.

3. A record 3,000 cutting-edge exhibits, categorised into four thematic halls—Core Technologies, Industrial Applications, Smart Devices, and Ecosystem Connectivity—forming an integrated “AI Capability Corridor” that spans foundational research to commercialisation.

4. Over 100 world‑first or China‑first debuts, underscoring the event’s reputation as a launchpad for innovation.

This ambitious scale highlights China’s drive to position itself at the centre of global AI innovation and commercial development.

Thematic Structure and Focus Areas

WAIC 2025 featured 100 forums and panel discussions across five major intersections: innovation incubation, academic thought leadership, industry application, governance, and youth development. 

Key themes included:

1. AI safety governance and ethical standards in the context of rapid technology deployment.

2. Large model innovation and application, including vertical (domain‑specific) LLMs tailored for healthcare, finance, legal, manufacturing, e‑commerce, among others.

3. The development of AI infrastructure, especially computing power supply, chips, and dual-track advancement of capacity and governance.

4. Embodied intelligence, featuring an 80‑company robotics showcase—including humanoid robots, AI pets, robotic calligraphers and drummers—and real‑world installations like the immersive AI‑powered City Walk.

5. Youth engagement, via the SAIL Awards, Youth Paper Awards, Algorithm competitions, and Frontier Questions tracks focused on science, mathematics and AI model innovation.

Governance & Global Cooperation: A Strategic Pillar

A defining thread running through WAIC 2025 was the push for global AI governance. In his opening address, Chinese Premier Li Qiang called for a global AI cooperation organisation, warning that AI could risk becoming “an exclusive game for a few countries and firms”  . He proposed creating an inclusive framework and action plan involving governments, companies, researchers and international bodies, explicitly incorporating engagement with the Global South. 

China also released a China AI Safety Commitment Framework, aimed at balancing innovation with risk management and positioning itself as a champion of open-source collaboration and multilateral governance dialogue. The proposals stood in sharp contrast to the contemporaneous US low‑regulation AI strategy, which prioritises competitive primacy and deregulation.

Innovation & Domestic Ecosystem Resilience

WAIC 2025 marked a clear emphasis on Chinese technological self‑sufficiency in response to intensifying U.S. export controls:

1. The formation of the Model‑Chip Ecosystem Innovation Alliance, uniting chip manufacturers like Huawei, Biren, Moore Threads and LLM developers such as StepFun, to align model formats, APIs and supporting infrastructure. The goal: enable models to run across diverse Chinese accelerator hardware, overcoming fragmentation in processor architectures.

2. The launch of the Shanghai General Chamber of Commerce AI Committee, aimed at integrating AI into industrial transformation by bridging developers (SenseTime, MetaX, MiniMax, Iluvatar CoreX, etc.) with industrial operators.

3. Display of flagship products like Huawei’s CloudMatrix 384, purportedly matching high-end Nvidia GPU performance via clustered domestic hardware; Tencent’s 3D World Model, Baidu’s digital human livestream technology; Alibaba’s Quark AI Glasses—integrating navigation and mobile payments—indicating maturity in consumer and enterprise AI hardware capabilities.

Moreover, Shanghai’s local AI industry, already contributing a third of China’s AI talent and boasting annual revenue exceeding 118 billion yuan in Q1 2025, formed a potent backdrop to the national strategy, with city computing supply projected to exceed 100 EFlops by year‑end, approximating 40% of China’s national total at end‑2024.

Highlights from Exhibits and Innovations

Large Language Models & Vertical AI

This sector-specific approach highlights a growing trend towards vertical AI solutions, where models are optimised for specialised tasks rather than general-purpose use. By tailoring LLMs to fields like healthcare or finance, companies can achieve greater accuracy, regulatory compliance, and operational efficiency. The presence of models designed for edge computing—capable of running on-device with lower latency and higher privacy—also signals a shift toward decentralised AI infrastructure, reducing reliance on cloud platforms. Notably, these innovations reflect a maturing domestic ecosystem in China that can now compete globally in both performance and cost. WAIC 2025 thus marked a major step forward in applied, scalable AI.

Robotics & Embodied Intelligence

This dramatic increase in robotics participation highlights the sector’s rapid commercialisation and growing integration into everyday life. Companies showcased not only technical advancements but also real-world usability, with robots performing complex, context-sensitive tasks in public environments. The “AI Shanghai” and City Walk experiences blurred the line between exhibition and urban reality, signalling a future where robots are not confined to factories but are embedded in society. From automated chefs preparing street food to AI calligraphers replicating traditional art forms, these displays emphasised both cultural sensitivity and technological innovation, reinforcing China’s ambition to make robotics a core part of smart city ecosystems.

Industrial Applications & Smart Devices

Siemens unveiled its Industrial Copilot AI assistant in China. Tesla introduced the Tesla Bot, while others exhibited AI glasses, transparent screens, AI pets and edge devices. These highlights marked the growing shift from academic AI theory to real-world product deployment across manufacturing, smart cities, autonomous vehicles and fintech sectors. 

Startups and Youth Engagement

WAIC 2025’s emphasis on nurturing young talent and cutting-edge innovation was not merely symbolic—it was a deliberate strategy to position the conference as a catalyst for the next wave of AI leadership. The Future Tech hatchery served as a dynamic launchpad for early-stage ventures, offering startups access to resources, mentorship, and, crucially, investors actively seeking breakthrough ideas. The atmosphere was one of experimentation and risk-taking, underscoring the belief that the future of AI will be defined not only by established players but also by nimble newcomers with disruptive potential.

The startup pavilion showcased innovations spanning AI for climate modelling, synthetic biology, real-time translation tools, neuro-symbolic AI architectures, and quantum-inspired machine learning. These projects reflected the interdisciplinary convergence that is increasingly shaping the future of artificial intelligence. WAIC’s ability to bring together technologists, academics, investors, and policymakers in one collaborative space helped bridge the often-disconnected worlds of research and commercialisation.

The SAIL Award (Super AI Leader), now a cornerstone of WAIC’s recognition programme, highlighted exemplary breakthroughs that combined scientific merit with real-world impact. Winning entries included AI models designed for early cancer detection and algorithms that optimise energy efficiency in large-scale urban grids. These awards not only elevated outstanding work but also encouraged a culture of research with societal relevance.

Perhaps the most intellectually ambitious addition was the Frontier Questions sessions—three tracks that invited thinkers from diverse disciplines to confront foundational challenges in science, mathematics, and AI theory. By tackling unresolved problems like explainability, the limits of algorithmic learning, or the integration of symbolic reasoning into deep learning systems, the sessions pushed attendees to think beyond commercial application and toward the philosophical and scientific core of AI.

Through these initiatives, WAIC 2025 firmly established itself not just as a showcase of present capabilities, but as a visionary forum for future breakthroughs.

Geopolitical Context & Contrasting Strategies

WAIC 2025 took place amid an intensifying US‑China tech competition, underpinned by export controls and diverging governance philosophies. While Chinese leadership championed open frameworks, inclusivity and cooperation, especially with developing nations, the US AI Action Plan adopted by the Trump administration emphasised minimal regulation to preserve American AI dominance. 

International figures at WAIC, such as Geoffrey Hinton and Eric Schmidt, echoed calls for global coordination on AI safety and ethics. Yet notable Western participants like Elon Musk were absent, reinforcing geopolitical friction themes. 

Through alliances such as the Model‑Chip Innovation Alliance and the Shanghai Commerce AI Committee, China aimed to reduce reliance on foreign hardware and assert industry standards domestically. These actions were presented as a strategic counterweight to US tech restrictions and emphasised the urgency of building resilient, self‑reliant AI infrastructure. 

Shanghai’s Strategic Role as an AI Hub

Hailed as a global AI centre, Shanghai leveraged WAIC to highlight its scale:

1. The city claims more than 300,000 AI professionals, roughly one-third of China’s total pool, and AI industry revenues of around 118 billion yuan in Q1 2025—a 29% increase year‑on‑year, with profits up 65%.

2. A local policy initiative supports high‑growth AI firms with computing power rental subsidies (up to 30% or 40 million yuan), boosting capacity and innovation.

3. Shanghai’s smart computing capacity is projected to exceed 100 EFlops by end‑2025, contributing a substantial share to China’s total of ~246 EFlops at the end of 2024.

4. The WAIC City Walk and AI Shanghai Tours bridged formal exhibitions with real‑world urban scenarios, demonstrating integrated applications across lifestyle, mobility, and services.

Taken together, these aspects reinforce Shanghai’s ambition to become not only a national AI engine but a global showcasing ground for end‑to‑end intelligent ecosystems.

Criticisms, Challenges and Broader Implications

Despite its scale and spectacle, WAIC 2025 raised several challenges and open questions:

1. Some critics point to the propagandistic tone of official messaging, interpreting emphasis on cooperation and open‑source as positioning against perceived U.S. dominance, a narrative that may not fully reflect the complexity of global AI governance.

2. U.S. non‑signatories to shared AI governance declarations (e.g., the Paris “Statement on Inclusive and Sustainable AI”) may limit the effectiveness of proposed multilateral frameworks.

3. The fragmented hardware environment within China, different architectures (Arm, custom, PowerVR, etc.), poses a technical hurdle to seamless model portability despite unity efforts.

4. Human‑rights considerations around data governance, privacy, academic freedom, and state control, though acknowledged in principle, are often underplayed within official discourse, raising broader questions about the feasibility of truly inclusive AI governance.

While youth and startups are highly promoted, the international diversity of participants remains limited; about 17 percent of young‑talent entries come from abroad, suggesting scope for broader global inclusion.

Conclusion and Outlook

WAIC 2025 was a watershed moment for Asia’s, and particularly China’s, AI ambition. By combining a massive showcase of frontier innovation with governance proposals and strategic alliances, it solidifies China’s bid to reshape global AI norms and infrastructure. The event represented a juncture where public policy, industrial ambition, and global discourse on ethics and regulation converge.

China used the platform not only to display its growing technical prowess but also to position itself as a thought leader in AI governance. By proposing the creation of a global AI cooperation organisation and unveiling new frameworks for AI safety, China signalled its intent to influence how AI is developed and regulated worldwide. This marked a significant pivot from being a technology adopter to a norm-setting power.

At the same time, the exhibition served to domestically reinforce confidence in China’s AI ecosystem, especially amid rising geopolitical tensions and export restrictions. The strong presence of domestic LLMs, robotics, and chip innovation highlighted a shift towards greater self-reliance.

WAIC 2025 thus wasn’t just a tech expo, it was a calculated move in the broader race to define the rules of the AI age. Its long-term impact will hinge on whether its proposals lead to real collaboration or deepen global divisions.

Looking forward:

1. Whether China’s alliance efforts can deliver true interoperability across diverse hardware and software stacks remains to be seen.

2. The proposed global AI cooperation organisation could begin a gradual shift towards multilateral governance—if enough states and companies engage meaningfully.

3. The continued absence of key Western AI powerhouses—such as Elon Musk, OpenAI, DeepMind—underscores persistent geopolitical division in approaches and participation.

Nevertheless, WAIC 2025 stood as both technology festival and strategic declaration: a global invitation for AI cooperation—or, at minimum, a challenge to existing power structures in the AI ecosystem.

Final Reflection

While WAIC 2025 was undeniably a showcase of technological might, its broader significance lies in the narrative it projected and the questions it provoked. It brought together the dual forces of geopolitical assertion and technological aspiration, combining hardware exhibitions with nuanced conversations on governance, ethics, and accessibility. The event’s overarching message was clear: China is not just a participant in the global AI race, it aims to be a rule-maker and ecosystem architect.

One of the most notable aspects of WAIC 2025 was its effort to reframe the AI discourse from one of unilateral dominance to one of shared responsibility and inclusive growth. China’s call for a global AI cooperation organisation echoes similar sentiments raised in Europe and parts of Asia, pointing toward a multipolar world order where technological leadership is not monopolised by a few Western giants. Whether such a coalition will materialise, and be recognised globally, is uncertain, but the mere proposal shifts the terms of debate.

Moreover, the event showed how AI is no longer confined to the labs of elite universities or the research divisions of tech giants. With humanoid robots performing calligraphy in public plazas, AI pets interacting with children, and smart glasses offering real-time navigation, WAIC made a powerful case for AI’s integration into daily life. This “embodied intelligence”, AI embedded in physical, visible, and emotional interfaces—represents a significant leap in public engagement with emerging technologies.

However, challenges persist. Despite the grandeur and optimism, concerns remain about transparency, data governance, and academic freedom in Chinese AI development. The push for domestic chip ecosystems, while technically commendable, also reflects a decoupling trend that may further splinter the global technology landscape. For all the talk of cooperation, trust deficits between East and West remain a formidable obstacle.

Finally, WAIC 2025 served as a mirror to the world, revealing not just what China wants to project, but also what the global AI community must grapple with: How do we govern a transformative technology whose implications are still unfolding? Can nations with starkly different values collaborate meaningfully on shared standards? Will AI evolve in the image of those who control it, or can its governance be democratised?

The answers are not yet clear. But as WAIC 2025 demonstrated, the conversations have begun in earnest, and they are global, urgent, and unavoidable.