Supporters of Marcus Endicott’s Patreon can access weekly or monthly video consultations on this topic.
In March 2018, Tencent (腾讯), headquartered in Shenzhen, made its first confirmed contribution to a major digital human project by co-producing Siren, a photorealistic real-time digital human, alongside Epic Games, 3Lateral, Cubic Motion, and Vicon. Siren was modeled on Chinese actress Bingjie Jiang and driven live during the performance by performer Alexa Lee using full-body and facial motion capture, with Tencent providing high-level creative direction and look development for the character; Epic CTO Kim Libreri presented the demonstration at the Game Developers Conference State of Unreal keynote in San Francisco. Later that year, in August 2018, Tencent held its second Tencent Game Developers Conference in Shenzhen's Shekou district in Nanshan, where NExT Technology Center expert Xie Weibo delivered a keynote revealing the photorealistic rendering techniques behind Siren, covering approaches to skin, eyes, teeth, and hair and noting that the foundational rendering technology had precedents shown at SIGGRAPH 2016's Realtime Live segment. This conference session constituted the first in-Shenzhen public presentation dedicated to digital human technology.
In May 2019, the Shenzhen Municipal Government published its New Generation Artificial Intelligence Development Action Plan for the years 2019 through 2023, targeting an AI core industry scale of ten billion yuan by 2020 and aiming to drive three hundred billion yuan in related industries. Although the document did not specifically address digital humans, it established the foundational AI policy framework -- encompassing algorithms, chips, image recognition, and application scenarios -- that would later enable the city's digital human ecosystem to form. In December 2019, Shenzhen Houhai Hui (深圳后海汇), a commercial retail complex in Nanshan District, launched Xiang Xiang Hilda, a hyper-realistic virtual influencer designed as a Gen-Z trend ambassador, deployed on Xiaohongshu to build pre-opening buzz for the shopping center's planned 2022 opening; this made Houhai Hui the earliest known Shenzhen commercial real estate project to create its own virtual idol spokesperson and the earliest confirmed locally deployed virtual influencer in the city. Also in December 2019, Tencent's virtual idol Star Tong (星瞳), the QQ Dance virtual spokesperson developed by Tencent's CDD teams in Shenzhen, recorded a performance alongside national music master Fang Jinlong for the 2020 Sichuan TV Flowers Bloom Under Heaven New Year Concert, making her the first virtual artist to perform alongside a national music master, with the concert airing around the turn of the year.
In August 2020, Shenzhen Zhuiyi Technology (深圳追一科技有限公司), an AI company founded in Shenzhen in January 2016, publicly demonstrated its multimodal digital human Wukong at the first Shenzhen International Artificial Intelligence Exhibition. CTO Liu Yunfeng delivered a keynote on multimodal digital humans in the 5G era, introducing the company's Face digital human product, which integrates computer vision, speech recognition, and natural language processing. Zhuiyi would subsequently be ranked first nationally in digital human patent applications, with 67 patents filed by the end of 2021.
In June 2021, Xinhua News Agency and Tencent's NExT Studios jointly launched Xiao Zheng (小诤), described as the world's first digital astronaut and China's first digital journalist, coinciding with the Shenzhou-12 crew's arrival at the Chinese space station. A fifteen-person team based at NExT Studios in Shenzhen built the character over three months using Unreal Engine real-time rendering, and Xiao Zheng went on to broadcast the first episode of Tiangong Science Station before covering the second extravehicular activity in August and beginning a Mars reporting mission in October of that year, establishing herself as a mainstream-values virtual idol for state media. Also in June 2021, Star Tong began live streaming on Bilibili as a virtual UP, using a real-time driving technology pipeline developed by Tencent CDD in Shenzhen and built on Unreal Engine 4, shifting from pre-recorded stage performances to interactive fan engagement and marking Tencent's entry into the VTuber space. In September 2021, at Huawei Connect 2021, Huawei Cloud (华为云) publicly unveiled its first virtual digital employee Yun Sheng (云笙), introduced by Huawei Senior Vice President Zhang Ping'an. Created using Huawei's MetaStudio digital content production line with AI auto-modeling, AI speech-driven animation, and AI rendering acceleration, Yun Sheng became the public face of Huawei's digital human strategy; Huawei is headquartered in Shenzhen's Longgang District, and while the Huawei Connect 2021 event was held online, the company's Shenzhen base was central to Yun Sheng's development. In November 2021, Tencent AI Lab debuted the hyper-realistic 3D virtual human Xiao Zhi (小志) at Tencent's Digital Ecosystem Conference; built by the lab, which had been researching virtual humans since 2017 from its Shenzhen base, Xiao Zhi could synthesize speech from text input and automatically generate lip-sync and natural facial expressions through a multi-modal technology framework. That same month, Tencent's CDD department, in collaboration with Epic Games, completed construction of a 3,000-square-meter LED virtual production studio in Shenzhen's Pingshan District, featuring a forty-meter-by-six-meter curved 240-degree LED wall and a movable ceiling LED screen exceeding one hundred square meters supporting high-speed capture, giving Shenzhen a critical piece of physical infrastructure for virtual human and virtual production work. In December 2021, Shenzhen-headquartered Wondershare (万兴科技) launched a virtual human feature in its video presentation product DemoCreator, allowing users to transform their real appearance into digital virtual avatars via a standard webcam with eight virtual avatar options at launch, marking Wondershare's first entry into digital human technology as part of its metaverse strategy.
In February 2022, primary and secondary school students in Shenzhen's Bao'an District experienced the first digital human school lesson in the city, titled 2022 New Spring First Lesson -- Button Up the First Button of Life, produced jointly by Huawei Cloud and Educator Magazine. The course featured Yun Sheng as the digital teacher and marked the first deployment of a digital human teacher in Shenzhen's primary and secondary schools. In March 2022, Global Digital Creations Holdings Limited (环球数码创意控股有限公司), founded in Shenzhen in 2000 and headquartered there, co-developed the virtual human Yuanxin (元心) with Tencent Cloud and Tencent AI Lab. Yuanxin, a kung-fu-style digital character, represented Global Digital Creations' entry into virtual digital human production, drawing on more than two decades of CG production and motion capture experience.
In June 2022, Shenzhen issued its Opinions on Developing Strategic Emerging Industry Clusters and Cultivating Future Industries, establishing the city's 20+8 industrial cluster framework, which explicitly identified virtual digital humans and virtual worlds as part of a strategic digital creative industry cluster and provided the foundational industrial policy for Shenzhen's digital human ecosystem. Later that same month, at the Huawei Partner and Developer Conference 2022 held in Shenzhen, Huawei Cloud announced a comprehensive upgrade to its MetaStudio digital content production line, introducing five digital human services -- styled digital human generation, realistic digital human creation, 3D video, digital human live streaming, and video production -- with key capabilities including digital human generation from a photograph in five seconds, 276 facial feature points, lip-sync accuracy above 95 percent, and motion capture latency below 100 milliseconds. In July 2022, Shenzhen-based FlashCut Smart Technology (深圳市闪剪智能科技有限公司), founded in June 2014 in the Qianhai area of Bao'an District and backed by Innovation Works, launched its digital clone and digital human features within the FlashCut AI intelligent video editing platform, offering over 200 digital human models, voice cloning, and unmanned live streaming capabilities, with product users eventually exceeding 260 million by 2024. In November 2022, several major institutional developments converged in rapid succession. The Shenzhen Special Economic Zone AI Industry Promotion Regulations took effect at the start of the month, making Shenzhen the site of China's first specialized local AI legislation and covering data governance, algorithm development, and industrial application in a way that provided the legal framework enabling the city's digital human ecosystem. Shortly thereafter, the 2022 Global Metaverse Conference Shenzhen opened under guidance from the Shenzhen Municipal Bureau of Industry and Information Technology, featuring a dedicated forum on digital human technology and application scenarios during which Vice Mayor Zhang Hua announced Shenzhen's intention to pursue virtual-real integration and platform-to-ecosystem development for metaverse applications. Mid-month, the Shenzhen Data Exchange (深圳数据交易所) was officially inaugurated with first-day cumulative trading exceeding 1.1 billion yuan across more than 106 application scenarios, laying the groundwork for digital human data asset registration, evaluation, and trading, with the exchange eventually listing over 2,000 data products by 2024. That same week, the 24th China Hi-Tech Fair was held in Shenzhen, where Mobvoi (出门问问) showcased its virtual digital human driving technology in the fair's metaverse exhibition area, presenting three metaverse software-as-a-service products: Yuanchuangdao, an AI short video platform; Qimiaoyuan (奇妙元), an AI digital avatar platform; and Moyin Gongfang, an AI voice tool. In December 2022, Shenzhen Yuanshi Technology (深圳市元师科技有限公司), based in Qianhai, organized a metaverse music class in which its virtual digital human teacher Mr. Yuan simultaneously connected students at schools in Hong Kong, Shenzhen, and Guangzhou, bringing together approximately 150 teachers and students across three Greater Bay Area cities; Mr. Yuan had first entered classrooms in September 2022 in Guangzhou, Yan'an, and Xinjiang before this cross-border deployment.
In March 2023, Shenzhen University (深圳大学) approved China's first Virtual Digital Human micro-professional program, a collaboration between the university's College of Communication and Tencent Technology (Shenzhen) Co., Ltd., enrolling 40 students per year from computer science, media arts, and journalism backgrounds across eight courses including Digital Human Production, Metaverse and Media Philosophy, and Virtual IP Operations, with half the courses taught by Tencent's Content Development Division team. Later that month, Tencent launched Tencent Zhi Ying (腾讯智影), an AI creative assistant featuring a core Zhiying Digital Human function that enables users to generate digital human broadcasting videos from text or audio input, including an image-cloning capability requiring minimal source photos or video and targeting enterprise content creation at scale. In April 2023, Tencent Cloud launched its Intelligent Small-Sample Digital Human Production Platform, reducing digital human modeling time to three minutes and costs to several thousand yuan from the traditional standard of several hundred thousand, significantly lowering barriers for enterprise adoption. In May 2023, Shenzhen issued its Action Plan for Accelerating High-Quality AI Development and High-Level Application covering 2023 through 2024, simultaneously releasing the first batch of City+AI application scenario lists and establishing a one-hundred-billion-yuan AI fund cluster under a thousands-of-industries-plus-AI framework. Also in May 2023, Shenzhen Special Zone Daily (深圳特区报) debuted its AI digital human anchor Li Na on the Fei Yue Shenzhen column on the newspaper's 41st anniversary; Li Na was developed jointly with Tuzhan Intelligence (兔展智能) and can adjust tone, expressions, and actions to suit different news content. That same month, Wondershare launched Wondershare Virbo (万兴播爆), described as China's first cross-border full-scenario digital human AI video creative software, at the Global Short Video and Livestream E-commerce Summit held in Shenzhen, with the platform supporting digital human avatar customization, voice cloning, and video template customization, while Wondershare simultaneously launched a Digital Human Co-creation Business Partner Program nationwide. At the end of May 2023, the Several Measures on Accelerating the Cultivation of Digital Creative Industry Clusters took effect in Shenzhen; Section 12 explicitly supported metaverse, digital art, digital cultural tourism, digital performing arts, and AI art as new business formats, with specific encouragement for virtual digital human innovation within the creative industry cluster.
In June 2023, the 19th China (Shenzhen) International Cultural Industries Fair at the Shenzhen International Convention and Exhibition Center featured extensive digital human showcases, with Shenzhen Press Group's Digital Culture Pavilion presenting AI-generated content and digital human technology; Shenzhen-based Laihua Technology (来画科技) and Nanjing Guiji Intelligent (南京硅基智能) provided technical support, producing dozens of digital humans for various exhibiting units. In July 2023, Huawei Cloud comprehensively upgraded its MetaStudio digital content production line at the Huawei Developer Conference 2023 in Dongguan, near Shenzhen; based on the Pangu Large Model, MetaStudio gained digital human model generation and model driving services, enabling the generation of a digital human from a text description in ten seconds with over 50 customizable parameters and sub-100-millisecond latency for real-time meeting participation. In November 2023, the 25th China Hi-Tech Fair, the largest edition in the fair's history at approximately 500,000 square meters across two Shenzhen venues in Futian and Bao'an, drew 4,925 companies and saw more than 680 new results released, with AI large models, brain-computer interfaces, and AI-generated digital content among the prominent showcase categories. In December 2023, Aoto Electronics (奥拓电子) subsidiary Chuangxiang Shuwei (深圳市创想数维科技有限公司), a Shenzhen-based company founded in 2021, completed a Pre-A funding round of tens of millions of yuan from Shenzhen HTI Ventures and Shenzhen HTI Luohu Fund at a post-investment valuation of 145 million yuan, with funds directed toward R&D for the next-generation MetaBox XR virtual production platform and AIGC digital asset generation tools; by the end of 2023, Aoto had completed 67 XR and virtual production studio projects globally for clients including Nvidia, Tencent, and Amazon.
In March 2024, Shenzhen upgraded its strategic industry cluster framework to 20+8 version 2.0, elevating AI to an independent cluster and adding intelligent robotics as a new future industry. That same month, Huawei Cloud held its digital human sales partner conference, at which MetaStudio director Huang Chao declared that 2024 would be the inaugural year for digital humans to accelerate large-scale application across industries, showcasing deployments in retail, hospitality, government, e-commerce, and education. In May 2024, at the 20th China (Shenzhen) International Cultural Industry Fair, Sichuan Cultural Industry Investment Group (四川文化产业投资集团) exhibited a Su Shi Q&A digital human that attracted over 3,000 attendees with a response accuracy rate of 98 percent, later improved above 99 percent, allowing visitors to converse with a virtual recreation of the Song Dynasty poet and demonstrating the application of AI digital humans to cultural heritage interpretation at a Shenzhen-hosted venue. In June 2024, Xiaoice (小冰公司), a Beijing-based company, and ShiMeta (深圳市视美泰技术股份有限公司), a Shenzhen-based company, announced a strategic partnership to launch the ShiMeta Digital Human Solution, a hardware-software integrated intelligent interaction terminal with digital human capabilities targeting enterprise offices, government services, tourism, and education. Later that month, at HDC 2024 in Dongguan, Huawei Cloud MetaStudio featured developer workshops and demonstrations of the interactive digital human Xiao Mei (小美), co-created with convenience store chain Meiyijia, alongside CodeLab training sessions for digital human product development; according to a Sullivan and Head Leopard Research report released in early 2025, MetaStudio achieved the top market share position in China's 2D AI digital human market in 2024, within a segment worth 2.89 billion yuan and growing at 101.2 percent year-over-year.
In July 2024, Shenzhen released its Action Plan for Building an AI Pioneer City, proposing 22 concrete measures across six areas and pursuing goals described as five pioneers in full-stack innovation, intelligent products, cross-border data, scenario applications, and intelligent driving. In August 2024, MoreFun Studios (魔方工作室群), a Tencent Games subsidiary based in Shenzhen, revealed F.A.C.U.L. -- an acronym for First-ever FPS AI Companion Who Understands Human Language -- at Gamescom 2024 in Cologne; F.A.C.U.L. is an AI companion integrated into the tactical first-person shooter game Arena Breakout: Infinite, capable of understanding and responding to player natural language commands in real time. Also in late August 2024, Yuanxiang Technology (深圳元象信息科技有限公司), known as XVERSE, founded in Shenzhen in November 2020 by former Tencent Vice President and AI Lab founder Yao Xing, released MotionGen, described as China's first physics-based 3D motion generation model, converting text prompts into realistic 3D character animations using a proprietary algorithm combining reinforcement learning, physical simulation, VQ-VAE, and Transformer architectures. In September 2024, Shenzhen's Futian District issued China's first-ever interim measures for the management of government auxiliary intelligent robots, requiring that each AI digital employee in government have a designated human supervisor and covering ethics, accountability, technical standards, application scope, and safety management. In October 2024, Natural Selection (Shenzhen) Intelligent Co., Ltd. (自然选择(深圳)智能有限公司), founded in February 2024 by Zhang Xiaofan with investment from Kingnet Network -- which holds a 17.2 percent stake -- released the promotional video for EVE, a 3D AI companion application focused on immersive virtual companionship, with an iOS launch planned for 2026.
In November 2024, the Shenzhen Municipal Bureau of Culture, Radio, Television, Tourism, and Sports issued a public tender for the Digital Tourism Shenzhen AI Digital Human Design and Production Service Project, seeking to create a digital human for tourism services covering dining, accommodation, entertainment, and shopping in Shenzhen. That same month, the 26th China Hi-Tech Fair was held across 400,000 square meters with more than 5,000 exhibitors from over 100 countries, marking the first edition operated by market-oriented professional organizations, with digital employee and AI showcases prominent throughout the floor; Wondershare exhibited AI and video creativity tools, and Tanji Technology (探迹科技) demonstrated digital employees for manufacturing, with intent transaction value exceeding 120 billion yuan. In December 2024, at the Shenzhen Global Investment Conference financial industry session, a digital human representing Deputy Director Zhu Jiang of the Shenzhen Municipal Financial Affairs Office presented the city's financial industry overview and investment policies, delivering remarks in Mandarin before switching to Spanish and English to welcome specific international banking partners -- an innovative deployment of digital human technology for official government investment promotion at a conference that drew more than 700 enterprises from 33 countries and saw over 900 billion yuan in signed project investments. Also in December 2024, Shenzhen published its Several Measures for Building Shenzhen into an AI Pioneer City, featuring 18 policy measures including computing power vouchers and model vouchers to subsidize AI development, while the Shenzhen Data Exchange separately launched its Data Property Rights Registration Platform with the first batch of data property registration certificates issued for multilingual speech data using blockchain for integrity verification, a platform directly relevant to digital human data asset management.
In February 2025, MicroCloud Hologram Inc. (微美全息), a Shenzhen-headquartered company listed on NASDAQ, announced the integration of DeepSeek R1 Model semantic understanding technology into its Holographic Digital Human GPT to enhance conversational capabilities for holographic digital human interactions. Later that month, Shenzhen's Futian District formally launched 70 AI digital smart employees built on DeepSeek via the district's Government Large Model 2.0 platform, covering 11 categories and 240 government scenarios including document processing, citizen services, and emergency management, and drawing on 120 million government data items accumulated over ten years; key performance metrics included document formatting accuracy above 95 percent, review time reduced by 90 percent, and citizen complaint routing accuracy rising from 70 to 95 percent, with each AI employee operating under the supervision of a designated human guardian per the September 2024 Futian regulation, and Shenzhen had by that point completed full DeepSeek R1 deployment across its entire government cloud, making it Guangdong Province's first city to do so. Also in February 2025, multiple Shenzhen-based companies showcased DeepSeek-powered digital humans at the Shenzhen International LED Exhibition: Mengpai Group (梦派集团) demonstrated DeepSeek-integrated 3D human cloning and AI digital employees for telecom showrooms; Liard (利亚德) subsidiary Virtual Motion Point (虚拟动点) merged its LYDIA motion model with DeepSeek; and Unilumin (洲明科技) connected DeepSeek to its multi-modal interactive Agent platform for holographic transparent display cabinets. Toward the end of February 2025, Shenzhen Technical University (深圳技术大学) launched its digital human Run Xiaozhi (润晓知), running on the full 671-billion-parameter DeepSeek R1 and V3 versions, integrated with the campus unified authentication system and an enrollment AI agent; the university simultaneously launched what was described as Shenzhen's first large models and DeepSeek applications practical general education course, with technical support from Volcano Engine, a ByteDance subsidiary.
In March 2025, Shenzhen published its Action Plan for Accelerating AI Terminal Industry Development covering 2025 through 2026, identifying nine priority product categories for AI terminal products and continuing the city's policy momentum in building infrastructure for digital human and AI applications. That same month, Shenzhen's Luohu District (罗湖区) launched holographic digital human customer service at the Luohu District Government Service Center as part of a government AI event, with approximately 80 AI application scenarios deployed across Luohu's government, education, healthcare, and city management domains. Also in March 2025, Global Digital Creations reaffirmed its commitment to virtual digital human technologies in its 2024 annual report, citing capabilities spanning virtual hosts, virtual tour guides, virtual brand ambassadors, and CG animation production built on over two decades of digital content creation experience from its Shenzhen headquarters. In April 2025, during Shenzhen Intellectual Property Promotion Week, an AI Industry IP Achievements Special Release Conference saw the issuance of China's first batch of AI digital human data intellectual property registration certificates; Shenzhen National South IP Agency (深圳国新南方知识产权代理有限公司) facilitated the first registration for Shenzhen Youmi Intelligent Technology (深圳优秘智能科技有限公司), covering its smart anchor Xi Xi (希希), with certificates spanning voice data, video training data, and portrait image data intellectual property. In July 2025, Bangyan Technology (邦彦技术股份有限公司), a Shenzhen-based company listed under stock code 688132, officially launched its Nuwaai Intelligent Agent Digital Human Platform at the 2025 World Artificial Intelligence Conference in Shanghai; the platform features a dual-brain architecture consisting of an emotional intelligence brain and an IQ brain, and its launch represents the most recent milestone in a decade of Shenzhen companies advancing digital human technology from early photorealistic demonstrations through policy-enabled commercial deployment to AI-native mass-market application.
[Mar 2026]