October 18, 1997. The opening of the Guggenheim Bilbao on the south bank of the Nervión River in the Basque Country of northern Spain. King Juan Carlos of Spain attended the inauguration. Frank Gehry, age sixty-eight, watched the press scrum from inside the building’s central atrium. The structure had cost $89 million USD to construct across four years from groundbreaking in October 1993. The Basque regional government had funded approximately 90 percent of the cost as a regeneration investment in a city of roughly 350,000 people that had lost its industrial base across the previous two decades. The Solomon R. Guggenheim Foundation, under director Thomas Krens, had contributed the brand and the curatorial framework.
The building was 24,000 square meters. The exterior was clad in approximately 33,000 individually-curved titanium panels. Each panel was approximately half a millimeter thick. Each panel’s geometry was unique. The titanium had been sourced from Russia at favorable post-Soviet-collapse pricing. The cladding was applied by local crews trained on shipyard welding methodology. Bilbao had been a shipbuilding city until the closure of the major yards through the 1980s. The displaced steelworkers built the building.
The construction had only been possible through CATIA, the computer-aided design software developed by the French aerospace firm Dassault Systèmes for the Mirage fighter jet program in the 1980s. Gehry’s office had licensed the software in 1991 and adapted it for architectural use. The 33,000 panels were each digitally modeled, fabricated from computer files, and installed in computer-specified positions. No two panels were identical. The building could not have been built in 1985. The combination of CATIA, post-Soviet titanium supply, and Basque regeneration funding had opened a window that closed again after Bilbao for buildings of that specific cost profile.
Frank Gehry was not just an architect. Frank Gehry was the moment late-twentieth-century architecture decoupled from the modernist geometric vocabulary I.M. Pei had codified at the Louvre and adopted a computer-enabled biomorphic vocabulary that could only be built with software originally developed for aircraft.
The Bilbao Commission
The site sat on the Nervión River’s south bank, on land that had been industrial waterfront through the twentieth century. Container yards, shipbuilding berths, steel-storage warehouses, and machine-tool shops had occupied the parcel until the early 1990s. The Basque regional government cleared the site as part of the Bilbao Ría 2000 plan, the broader urban-regeneration program initiated in 1992. The Guggenheim Foundation contracted directly with the regional government on terms that gave the Basque side ownership of the building and the Guggenheim brand-and-curatorial role for an initial twenty-year arrangement, since renewed.
The building’s program ran nineteen exhibition halls across three floors. The geometries varied. Some halls were orthogonal and conventional. Others followed the building’s curved exterior surfaces and presented galleries with no straight walls. The largest single space, the second-floor “fish” gallery, was 130 meters long and designed specifically to accommodate Richard Serra’s monumental steel sculptures, which would not have fit in the existing Guggenheim collection’s New York and Venice spaces. Serra’s “Snake,” completed in 1997, was installed as the building’s permanent anchor work. The central atrium rose 50 meters from the entry level. The titanium-clad exterior wrapped the program in computer-modeled curves whose interior expression read as a sequence of distinct spatial volumes rather than a unified plan.
Construction ran from October 1993 to summer 1997, with fit-out continuing through October. Edwin Chan, project architect in Gehry’s Santa Monica office, ran the CATIA digital modeling team that produced the construction documents. IDOM, a Spanish architectural practice based in Bilbao, partnered on site execution and local code compliance. The construction crews were drawn primarily from the closed Astilleros Españoles shipyards that had employed thousands of welders and steelworkers through the 1970s and 1980s. The shipyard crews applied ship-hull welding methodology to the titanium architectural cladding. The transfer of expertise from collapsed industry to live cultural-infrastructure construction was structural to the project’s local-employment economics.
The project came in roughly on budget at $89 million against an early estimate of approximately $100 million, a rare outcome for cultural buildings of that geometric complexity. Cost-overrun was the standard expectation for star-architect commissions through the 1990s. The Sydney Opera House had run roughly 1,400 percent over its original budget across a fourteen-year construction. Bilbao’s discipline was attributed to the CATIA workflow’s ability to define and price every panel in advance of fabrication, removing the field improvisation that drove overruns on traditionally-drafted complex buildings.
The CATIA Translation
The technological inversion behind Bilbao ran on aerospace software. CATIA had been developed by Dassault Systèmes through the 1970s and 1980s as a three-dimensional surface-modeling platform for the Mirage fighter jet and the subsequent Rafale program. The software’s specific capability was the precise definition of curved aluminum and titanium aircraft skins at tolerances aerospace required for aerodynamic performance. Each surface could be parameterized, modified, and re-extracted as a fabrication file readable by computer-numerical-control milling machines and cutting equipment.
Pei had drafted the Louvre Pyramid through hand-drawing and traditional engineering tabulation. The pyramid’s 673 glass panes were standardized rhombic and triangular units, manufactured through repetition of two basic shapes. The hand-drafting workflow was sufficient because the geometry was simple enough to define on paper. Gehry could not produce Bilbao through the same workflow. The building’s 33,000 panels were each unique. No two panels shared a geometry. The hand-drafting cost of defining 33,000 unique panel shapes would have made the project unbuildable inside any realistic budget.
Gehry’s office adapted CATIA for architectural use beginning in 1991. The adaptation work was led by Jim Glymph and Rick Smith, working with Dassault Systèmes engineers to translate aerospace workflows into building-design protocols. The Bilbao model was built in CATIA across 1993 and 1994. Each panel’s geometry was extracted as a fabrication file. The files went to titanium suppliers in the United States and to local Spanish fabricators. The installation drawings were generated from the same model and specified each panel’s exact position on the building’s structural frame. The construction sequence followed a pre-determined panel-installation order that the model had verified against structural and weather-sequencing constraints.
The CATIA workflow Gehry developed at Bilbao became the operational template for the next generation of complex-geometry architecture. Zaha Hadid’s office adopted similar parametric workflows for her work through the 2000s. The wider parametric-design movement that emerged in academic architecture through the late 1990s and early 2000s built on the same software stack. Gehry Technologies, a separate software company spun off from the practice in 2002, commercialized the workflow under the name Digital Project. The platform was eventually acquired by Trimble in 2014. The CATIA-to-architecture translation Gehry’s office had built at Bilbao became one of the structural shifts that distinguished 1990s architecture from everything that preceded it.
The Anti-Pei
The building’s lineage ran back eight years to a direct opposite. Pei had completed the Louvre Pyramid on March 30, 1989, the centerpiece of François Mitterrand’s Grands Projets program. The pyramid sat 21 meters tall at the center of the Cour Napoléon of the Louvre, the most-visited art museum on Earth. The structure was glass and steel, executed in 673 standardized panes across a four-sided form that was one of the five Platonic solids of classical geometry. Pei had been the obvious choice for the commission as the architect of the East Building of the National Gallery of Art in Washington (1978) and of a long catalog of late-modernist American institutional buildings. The Louvre Pyramid was the consolidation of a five-decade career inside the late-modernist register.
Gehry inverted every parameter. Pei’s geometry was Euclidean. Gehry’s was non-Euclidean. Pei’s structure was monolithic, executed as a single rigid form. Gehry’s was assembled from 33,000 panels that produced their visual coherence at the perceptual level rather than at the geometric level. Pei’s pyramid attached to a French royal palace at the symbolic center of Western institutional culture. Gehry’s Bilbao stood freely on a former industrial waterfront in a Basque port that the wider European cultural conversation had not engaged with since the Spanish Civil War of the 1930s. Pei’s commission was Mitterrand’s. Gehry’s commission was the Basque regional government’s.
The cultural register inverted at every level. Pei was a Chinese-American architect trained at MIT and Harvard, working at the institutional center of Western architecture from a Manhattan office. Gehry was a Canadian-American architect trained at USC, working from a Santa Monica office that was professionally located inside Los Angeles’s anti-Establishment architectural counter-current of the 1970s and 1980s. Pei represented institutional continuity. Gehry represented institutional disruption. Both architects were of the senior generation that had shaped postwar modernism, but their generational positioning inside the architectural profession was different. Pei was the senior modernist. Gehry was the senior disrupter.
The class read inverted. Pei said the geometry of antiquity will scale to the contemporary. Gehry said the geometry of antiquity cannot describe what computers can now build. Both were correct about their own time. The Pei pyramid was the last fully successful execution of the Euclidean institutional intervention. The Gehry Bilbao was the first fully successful execution of the computer-enabled biomorphic intervention. The two buildings define the architectural decade-boundary between 1980s late modernism and 1990s digital biomorphism.
The Bilbao Effect
The economic register was structural to the project’s lasting cultural impact. Bilbao had been an industrial port and shipbuilding city through the early twentieth century. The Astilleros Españoles yards and the Altos Hornos de Vizcaya steel works had employed roughly 40 percent of the metropolitan population at mid-century. The smaller machine-building plants ran parallel employment. Global shifts in shipbuilding to Korea and Japan, combined with Spain’s industrial restructuring under the post-Franco democratic governments of the 1980s, drove the closure of the major yards through the decade. Population declined. Per-capita income fell. The Bilbao metropolitan area entered the early 1990s with the economic profile of the depressed industrial cities of Northern England and the American Rust Belt.
The Basque regional government’s Bilbao Ría 2000 regeneration plan, initiated in 1992, identified cultural infrastructure as one of the regeneration vectors alongside the Norman Foster-designed Bilbao Metro (opened 1995) and the Santiago Calatrava-designed Zubizuri footbridge (opened 1997). The Guggenheim was the headline project and the largest single capital commitment. The expected first-year visitor count had been approximately 500,000. The actual first-year count reached approximately 1.3 million. The Basque government recovered its $89 million investment in tax revenue from tourism within roughly three years.
The phrase “the Bilbao Effect” entered urban-planning vocabulary inside eighteen months of the opening. Cities globally commissioned star-architect cultural buildings on the Bilbao model through the 2000s and 2010s. Daniel Libeskind’s Denver Art Museum extension (2006), Zaha Hadid’s MAXXI in Rome (2010), Herzog & de Meuron’s CaixaForum Madrid (2008), Tadao Ando’s Modern Art Museum in Fort Worth (2002), and dozens of regional museums and concert halls followed the pattern. The model spread to Asian cities through the same period, with Beijing’s Bird’s Nest and Water Cube (both 2008 Olympics), Shanghai’s Oriental Art Center (2004), and Guangzhou Opera House (2010) representing the Asian iterations.
Most of these projects produced smaller regeneration returns than Bilbao itself. The Bilbao Effect was specific to a confluence of factors that were difficult to replicate. Bilbao had been a city actively seeking regeneration with regional-government backing and an existing tourism infrastructure that could absorb the projected visitor volume. The Guggenheim brand carried weight in the European tourism market. Gehry’s particular celebrity-architect moment in 1997 was the peak of the star-architect economy. Subsequent projects often replicated the formal moves without the underlying economic conditions. The cities that commissioned post-Bilbao starchitect buildings frequently absorbed the construction costs without realizing the regeneration returns. The Bilbao Effect was real and was widely misunderstood by the cities that attempted to imitate it.
The Aftermath
Gehry’s subsequent career compounded the Bilbao gain. The Walt Disney Concert Hall in Los Angeles, designed before Bilbao but constructed after, opened October 23, 2003 at a cost of approximately $274 million. The design had been won in a 1987 competition funded by a $50 million bequest from Lillian B. Disney for the Los Angeles Philharmonic. Construction had stalled in the early 1990s over cost concerns, then resumed in the late 1990s after Bilbao validated Gehry’s complex-geometry approach. The Marqués de Riscal Hotel in Elciego, Spain (2006), the Fondation Louis Vuitton in Paris (2014), and the Luma Arles arts complex (2021) followed at increasing project scales and increasing CATIA-derived geometric complexity. Gehry’s office became one of the dominant complex-geometry practices through the 2000s and 2010s.
The architect himself, born February 28, 1929 in Toronto as Frank Owen Goldberg before his family moved to Los Angeles in 1947, continued to lead the practice past his ninetieth birthday. The Pritzker Prize had recognized him in 1989, the same year Pei’s pyramid completed. The award-recognition timing was structural to the decade-boundary the two buildings define. Pei represented the dominant architectural register of the 1980s as it received its final institutional validation. Gehry represented the dominant architectural register of the 1990s as it received its first commercial test at scale.
The deeper inheritance of Bilbao was the workflow more than the formal language. The 2000s and 2010s parametric-design movement that produced Zaha Hadid’s Heydar Aliyev Center (2012), BIG’s 8 House (2010), Snøhetta’s Oslo Opera House (2008), and the wider catalog of curved-surface contemporary architecture all built on the CATIA-and-successor software workflows that Gehry’s office had pioneered at Bilbao. The 1997 building was the proof-of-concept that opened the workflow. The proliferation that followed was the workflow’s commercial validation.
By the 2010s, complex-geometry architecture had become technically routine. Buildings that would have been unbuildable in 1990 became standard graduate-thesis projects by 2015. The Bilbao moment had been the moment when the technological possibility, the cultural appetite, and the regeneration economics had aligned. Subsequent decades inherited the capability without the alignment. Pei died on May 16, 2019 at age 102. Gehry continued to practice through the 2020s. The Pei pyramid still stands at the Louvre. The Gehry Bilbao still stands on the Nervión. The decades they describe have both moved on.
· · ·
The drafting tables closed. The slide rules retired. The pyramid finished. The software took over.