Artificial Intelligence continues to reshape global economies and industries, prompting institutions and individuals to rapidly adapt. The profound impact stems not merely from AI's ability to alter work and life, but from the underlying business models and investment strategies of its developers. This includes the concentration of capital, increasing demands for computational power, the competitive pursuit of specialized talent, and the significant infrastructural footprint required for sustained operation. This dynamic is particularly evident in the Greater Bay Area, which includes Guangzhou, Shenzhen, and Hong Kong. Government-led initiatives are actively promoting growth in this sector, translating intangible technological advancements into tangible physical forms through zoning adjustments, designated land allocation, and the emergence of new AI-focused building typologies, from advanced research labs to expansive data centers.
Hong Kong provides clear examples of this evolving landscape. Cyberport's transformation into an AI supercomputing hub, the introduction of a substantial Artificial Intelligence Subsidy Scheme, and the strategic development of San Tin Technopole as a key innovation and technology urbanism node—integrating data centers, prototyping workshops, and applied research—all demonstrate a proactive institutional agenda. Similarly, across the border, Shenzhen's Qianhai district is being developed to foster AI clustering and aggregation, while Guangdong's digital economy innovation and development policies further cement the region's ambitions. Together, these efforts illustrate how AI is not just an economic strategy, but a driving force behind a new urban architecture, manifesting in dedicated campuses, extensive server halls, complex logistics corridors, energy infrastructure, and the accompanying public discourse around 'innovation districts.'
The discourse surrounding AI's utility in architecture is well-established, covering aspects such as design assistance, visualization, construction optimization, streamlined manufacturing processes, and project management logistics. It is against this backdrop that the 2025 Hong Kong–Shenzhen Bi-City Biennale of Urbanism\Architecture (Hong Kong): Techformance (UABB HK 2025: Techformance) gains particular relevance. Through dedicated curatorial efforts, collaborating with architects and interdisciplinary practitioners, the exhibition transcends merely asking what AI can do for architecture. Instead, it critically examines AI's current influence on architecture: how it reconfigures spaces of production and culture, and how its 'performance' is presented, normalized, and debated within the built environment.
The UABB HK 2025 exhibition, 'Techformance,' critically assesses how AI is transforming the role of architects, shifting from individual authorship to a collaborative practice where machines contribute through generation, optimization, and fabrication. This raises fundamental questions about professional accountability: if design is increasingly influenced by algorithms and models, where does responsibility ultimately lie, and what role remains for human intuition, judgment, and cultural intent? The curators emphasize a crucial tension: while AI promises performance and efficiency in areas like sustainability modeling and robotic construction, this can lead to a reductive design logic unless architecture remains anchored by intangible values such as emotion, memory, ritual, and the social dimensions of space.
The exhibition's dual-venue structure serves as more than a practical arrangement; it acts as a curatorial framework that physically embodies the spectrum of ideas being presented. By staging 'Techformance' across Oi! and the East Kowloon Cultural Centre, UABB HK 2025 situates AI within two distinct architectural contexts deeply familiar to Hong Kong: the preservation and adaptive reuse of heritage, and the development of forward-thinking cultural infrastructure. Even before encountering the exhibits, visitors are confronted with the city itself performing the exhibition's central thesis: that AI functions not merely as a tool, but as an active agent mediating between historical continuity and novel creation.
Oi!, as a repurposed art space, highlights the enduring qualities of traditional construction, the constraints of existing structures, and the deliberate intelligence involved in working with what already exists. It frames AI within a narrative of continuity, where innovation is ethically linked to preservation. In contrast, the East Kowloon Cultural Centre functions as an institutional counterpart: a purpose-built performing arts complex designed for adaptability—spaces tailored for contemporary productions, advanced technical systems, and interdisciplinary arts and technology experiments. When viewed together, these venues sharpen UABB's central inquiry into an urban dilemma: in a future shaped by AI, will cultural creation be characterized by adaptive continuity or by new infrastructures designed to manifest 'the future' through sophisticated systems, grand scale, and highly functional spaces?
Within the diverse offerings of UABB, the installations Generative Futures: AI in Architectural Design by Hao Zheng and Does AI Dream by Urban Projection & Streetsignhk effectively illustrate the opposing effects of AI on architecture. Generative Futures presents AI as a practical extension of design methodologies, utilizing diffusion-based 3D form generation and neural networks for vectorized geometries. This approach propels architectural practice from speculative imagery to buildable, quantifiable outputs, re-positioning the architect as an orchestrator of parameters and an editor of results within an iterative process driven by metrics and optimization. Conversely, Does AI Dream champions elements that resist quantification: the illogical, the accidental, and the intuitively formed moments that infuse cities with vitality, posing a provocative question: if AI can 'dream,' what then remains of our own capacity for imagination? Together, these exhibits reinforce 'Techformance's' assertion that AI represents both a powerful creative force and a rationalizing system, and that the agency of architecture lies in its ability to navigate this inherent tension.
Yet, beyond AI merely serving as a design tool—generating forms, refining plans, or accelerating production—a crucial, yet often overlooked, question persists: what does it mean to design specifically for AI? The region's policy impetus is already driving the rapid construction of AI-dedicated environments, including data centers, supercomputing facilities, research laboratories, prototyping workshops, and the extensive infrastructural networks that support them. These structures are frequently perceived as mundane, efficient containers: insulated, secure, and service-intensive, where architecture is reduced to a neutral shell accommodating power, cooling, redundancy, and security. Constructed with urgency, they risk becoming the newest iteration of urban 'back-of-house' elements—vital to the city's economy yet spatially and socially inconspicuous, as if the digital realm must inherently manifest as an uncommunicative industrial black box.
If AI is now an extensive infrastructural system—an intricate web of energy, thermal management, land use, logistics, and human labor—then the architectural inquiry cannot be confined to user interfaces and workflows. It must also address how AI redefines urban typologies and politics: where these facilities are situated, what they displace, how visible or concealed they are permitted to be, and who ultimately benefits from their presence. Computational processes are far from weightless; they materialize as power substations, cooling plants, auxiliary power facilities, secure perimeters, and a land-use rationale that can easily transform significant urban areas into specialized zones with minimal public benefit. The true design challenge, therefore, is not merely to accommodate novel building types, but to mediate their integration into daily urban life—considering accessibility, permeability, public interaction, and crafting cultural narratives that portray 'innovation' as a shared future, rather than an exclusive industry.
However, the rapid pace and expansive scale of construction for these environments present a unique opportunity to shape them differently. Instead of concealing or isolating AI facilities, architecture can explore various forms of coexistence. This involves designing infrastructural perimeters that also foster public life, creating ground-level spaces that accommodate civic functions without compromising security, and integrating service systems—such as waste heat recovery, water management, and energy infrastructure—not just internally, but as visible and beneficial urban assets. The fundamental question is whether AI infrastructure can be thoughtfully embedded within the city as a new type of public utility, balancing functionality with civic reciprocity and public value, so that the power driving the future is seamlessly integrated into, rather than separated from, everyday life.