Metabolist Architecture: Learning from a Vision of Buildings That Evolve

A building can outlast the way of life for which it was designed. Households change, equipment ages and cities acquire new needs. Must the whole building become obsolete when only parts of it need to change? The Japanese Metabolists explored that question in the late 1950s and 1960s, imagining architecture capable of growth, renewal and regeneration.

This investigation began with Nicholas Keng’s video, What Is Metabolist Architecture? It prompted a closer look at an intellectual inheritance that matters to ExoSett. The Wikipedia overview of Metabolism provides an accessible introduction; the research below examines what its ideas demand in practice.

The movement’s 1960 manifesto, Metabolism 1960: Proposals for a New Urbanism, brought together architectural and urban visions. Its biological metaphor concerned cycles of growth and renewal: relatively enduring infrastructure could support shorter-lived elements, allowing change without replacing everything at once.

This ambition extended beyond prefabrication. In Meike Schalk’s historical analysis, Metabolism belongs to Japan’s post-war reconstruction, rapid urbanisation and search for cultural resilience. Questions about collective organisation and individual freedom accompanied the technical proposals. Buildings, neighbourhoods and cities were all part of the inquiry.

The enduring insight is that different parts of a building need not share one lifespan. The difficult question is how those different cycles can coexist in an inhabited place.

Kisho Kurokawa’s Nakagin Capsule Tower, completed in Tokyo in 1972, made that ambition tangible. Its 140 accommodation capsules attached to two structural and service cores. The cores were intended to endure while capsules could be renewed. The planned replacement cycle never happened.

Tatsuyuki Maeda and Yuka Yoshida’s 2021 account explains a crucial physical obstacle: capsules could not simply slide out horizontally. Brackets and restricted clearance meant that removing one required removing capsules above it. A detachable connection did not provide independent access.

The same account describes expensive replacement proposals and the difficulty of securing agreement among separate owners. Investment in renewal competed with redevelopment interests. These constraints interacted: a technically demanding operation became a collective financial decision, rather than an individual household’s straightforward purchase.

The English abstract of Osaki, Otsuki and Lee’s study similarly identifies a growing separation between use, ownership and management, alongside technical and legal difficulties. Their full Japanese-language paper has not been reviewed here.

Maintenance and deterioration also matter. SFMOMA’s account connects declining upkeep with the pressure to sell valuable land. The tower was demolished in 2022, although individual capsules were saved. Its history cannot fairly be explained by one faulty detail or one missed decision. Nor does dismantling capsules for preservation demonstrate that routine replacement in an occupied tower would have been feasible.

Kiyonori Kikutake’s Sky House, built in 1958, offers a more intimate example. An elevated concrete structure supports an open living space, with service and private accommodation conceived as changeable elements. DOCOMOMO Japan describes the house as a working experiment that underwent alterations as the family’s needs changed.

Its significance lies in accommodating a household over time. That is encouraging evidence of lived adaptation, although it is a different test from replacing standard units across a building with many owners.

The connection with ExoSett’s stories about a first home, extending a building and downsizing is direct: useful architecture should respond to people’s changing circumstances. Those stories describe intended possibilities, rather than completed demonstrations.

Other projects explored the relationship between lasting structure and changeable accommodation. Kurokawa’s Takara Beautilion at Osaka’s Expo ’70 assembled capsules within a framework; Kenzo Tange’s broadcasting centres explored cores, working spaces and provision for growth. Their different scales and purposes caution against treating Metabolism as one uniform construction method.

Kikutake’s unbuilt Tower-shaped City is particularly intriguing. As described in the historical overview, it incorporated manufacturing infrastructure for prefabricated homes. Making replacement accommodation was part of the imagined urban system.

There is a conceptual parallel with ExoSett’s proposed meta pod: a pod equipped to help manufacture other pods. The proposals are different, and the meta pod remains an idea requiring engineering and validation. The shared question is valuable: where will a building’s future components come from?

ExoSett acknowledges its intellectual debt to this earlier thinking about adaptable architecture. Its proposed contribution lies in the mechanisms, interfaces and processes through which reconfiguration might become repeatable throughout a building’s life.

The accommodation frame is intended to provide the principal structural load path, with pods occupying its bays. An associated service frame provides for access and utilities. The design objective is a structure that remains stable with empty bays, allowing pods to be installed, removed or replaced independently. Frame members meet at defined nodes; pod connections and utility interfaces must be specified as deliberately as the frame itself.

Repeatable installation methods, including jigs, are intended to make alignment and handling dependable. Defined interfaces should enable multiple manufacturers and suppliers to participate in a common architecture. These are proposals to be demonstrated through engineering, testing and commercial application.

The XML building specification is intended to hold the authoritative facts: geometry, materials, tolerances, connections and performance requirements. Drawings are derived representations of those facts. The fuller specification remains under development; a document passing schema validation does not establish that its engineering is sound.

The most useful test is concrete: imagine replacing one pod in an occupied building several decades after construction. What must still be true for that operation to succeed?

Engineering. The building must remain stable during every stage, including with a bay empty. Equipment needs a usable lifting and removal route. Utilities must be isolated and reconnected, while neighbouring occupants retain safe access and services. Weatherproofing, drainage, fire separation and inspection must work before, during and after the exchange. Water ingress and neglected maintenance can defeat a replacement strategy long before a crane arrives.

Economics. Replacement must offer sufficient value compared with refurbishment or demolition and rebuilding. The comparison includes temporary accommodation, disruption, lifting equipment, transport and approvals. A pod’s factory price alone cannot establish affordability. Sometimes repairing an existing pod will be the better choice.

Manufacturing. A compatible replacement must be procurable when needed. Common interfaces could widen the supplier base, but participation depends on demand, production capability and a dependable way to verify compatibility. An intention to welcome competing manufacturers does not itself create that market.

Ownership. The building owner should retain and maintain its XML specification: a housing authority for its housing, for example, or a hotel chain for its hotels. The record should reflect authorised changes and pass with the building on a sale. Where ownership is divided, arrangements must also establish who can approve work affecting shared structure and services, how costs are allocated and how occupants are protected.

Longevity. The owner’s specification should remain sufficiently complete to commission compatible replacements even if the original supplier disappears. New generations of pods must be assessed against the actual building’s interfaces and requirements. Updating the wider system must not silently invalidate an older building’s record. The enduring frame itself will also require inspection, maintenance and eventual decisions about renewal.

These are ExoSett’s practical conclusions from the history. They make reconfigurability a continuing responsibility, spanning design, ownership and manufacture.

The Metabolists deserve appreciation for asking how architecture could accommodate change, and for building experiments that exposed the difficulty of doing so. The fate of one tower cannot settle the value of that inquiry.

ExoSett seeks to develop an architectural system through which that vision might become technically practical, repeatable and commercially sustainable. For ExoSett, the Metabolist vision provides both inspiration and a practical test: can change become an ordinary, affordable part of a building’s life?

Accessible introductions and building histories

• Nicholas Keng, What Is Metabolist Architecture? Video; the starting point for this investigation.
• Wikipedia, Metabolism (architecture). An introductory overview and guide to projects.
• DOCOMOMO Japan, Sky House, selection no. 054. Building history and adaptation.
• Alexxa Gotthardt (2024), A Towering Acquisition, SFMOMA. Maintenance, demolition and preservation.

Academic literature

• Meike Schalk (2014), The Architecture of Metabolism. Inventing a Culture of Resilience, Arts 3(2), 279–297. Historical and social context.
• Tatsuyuki Maeda and Yuka Yoshida (2021), The real reason why Nakagin Capsule Tower was never metabolized, Docomomo Journal 65, 118–120. A preservation participant’s account of replacement barriers.
• Yuta Osaki, Toshio Otsuki and Yonggeun Lee (2021), A Study on the Transition of Nakagin Capsule Tower as One of the Earliest ‘One-Room Mansions’ in Japan, AIJ Journal of Technology and Design 27(66), 807–811. Japanese-language paper; only the English abstract was consulted.

The connections to ExoSett are our interpretation; these sources do not evaluate or endorse the system.

Explore Reconfigurability, or return to Research and Reading.