For a hundred and forty years, from the Crystal Palace onward, glass in facades has had one job: let light through while controlling heat. Over the last twenty years, two new jobs have joined it, technically unrelated to each other but often blurred together by marketing language under the generic label "smart glass": generating electricity from the solar radiation passing through it (photovoltaic glass, or BIPV) and changing its own light transmission on command, with no curtains or mechanical shading (dynamically controlled glass, electrochromic or thermochromic). These are two product families with very different physics, costs and commercial maturity, and it's a mistake to treat them as variants of the same idea.
Building-integrated photovoltaic glass (BIPV)
BIPV (Building Integrated Photovoltaics) doesn't add solar panels on top of an already-finished building, like a conventional photovoltaic roof: it integrates the photovoltaic cells directly into the building component — the facade, the skylight, the canopy — that this component replaces. In the most common version, crystalline silicon cells are laminated between two panes of glass exactly as in ordinary laminated glass, but spaced apart to let some light through: the resulting transparency (10–40% light transmission, depending on the spacing chosen) is inversely proportional to the surface covered by cells, and therefore to the electrical output. The typical generation of these semi-transparent modules sits between 60 and 130 Wp/m², much less than the 200–220 Wp/m² of a standard opaque photovoltaic panel, because every gap left open for transparency is surface that generates no current. Thin-film technologies also exist — amorphous silicon, dye-sensitised cells (DSSC) — which spread absorption more evenly across the pane, with a resulting colour tint that's often quite pronounced, less controllable aesthetically but potentially cheaper over large surfaces.
The most cited application case in Europe is the Copenhagen International School in Nordhavn (C.F. Møller Architects, 2017): over 12,000 coloured solar panels (blue-green shades) clad the entire facade of the building, in a project explicitly designed to make energy production visible as part of the school's teaching, generating a significant share of the building's electricity needs. The Cité du Vin in Bordeaux (XTU Architects, 2016) instead integrates BIPV panels into its curved glazed roof. In both cases, BIPV doesn't fully replace the building's electrical system: it supplements it, with a contribution that has to be assessed realistically based on orientation, shading and the transparency chosen — not on the rated output of opaque panels.
Electrochromic glass: electrically controlled tinting
Electrochromic glass changes colour and transparency by applying a small electrical potential difference to a stack of thin films deposited on the pane: an active layer, typically tungsten oxide (WO₃), changes structure as ions (usually lithium) migrate through it under the electric field, switching reversibly from a transparent state to a tinted one (dark blue), and back again by reversing the polarity. The complete transition from one state to the other takes anywhere from a few minutes to about a quarter of an hour depending on the pane's size, and covers a typical light-transmission range running from about 60% (clear state) down to residual values of 1–2% (dark state) — continuous control, not just a light/dark switch. The market's two reference manufacturers are SageGlass (now part of Saint-Gobain) and View Inc. (California): both have been selling the product commercially for over a decade, with installations in office buildings, airports and healthcare facilities, where automatic glare and solar-gain control removes the need for curtains or mechanical blinds, while keeping the outside view even in the darkest state. The same technology, developed in parallel for a very different sector, equips the windows of the Boeing 787 Dreamliner in place of traditional mechanical shades — proof that the principle, over thirty years after the first patents, is by now mature and reliable even in extreme operating conditions.
Thermochromic glass: the same goal, without electricity
Thermochromic glass pursues the same goal — modulating solar transmission — but passively, with no power supply and no control system: materials such as vanadium dioxide (VO₂) change crystalline phase at a critical temperature (around 68°C for the pure compound, which can be lowered with suitable dopants), changing their infrared transmission as a function of the pane's own temperature, and so indirectly of solar radiation. The advantage is the complete absence of wiring, power supply and electronic control, and so, in theory, zero maintenance cost; the disadvantage is lower precision — the response depends on the pane's temperature, not on a direct command tied to the actual use of the space — and commercial maturity still lower than electrochromic glass, with most applications still at the research or demonstration-project stage rather than a widespread industrial product.
Real-world performance and the limits renderings don't show
Semi-transparent photovoltaic glass always involves an explicit trade-off between transparency and energy output: the more light you want to let through, the less usable surface is left for the cell, and generation drops proportionally — there's no version that gives "all the transparency and all the energy." Cost per square metre remains significantly higher than a standard double glazing unit, and the payback period has to be calculated over the entire life cycle, not just the savings on the energy bill. Electrochromic glass, for its part, typically costs five to ten times an equivalent Low-E glazing unit, requires a control system (light sensors, control unit, wiring) that has to be designed and maintained like any of the building's electrical systems, and its uptake today remains concentrated on high-end office buildings or demonstration projects, more than on mainstream construction. Neither technology, as things stand, is meant to fully replace solar shading or a conventional photovoltaic system: both work best as an additional performance layer on a facade already properly designed according to the principles set out in the general guide to glass.
Where research is heading
The most promising development directions concern integration: glazed units that combine a BIPV pane and a dynamic-control layer in the same package, so they generate energy when maximising solar gain is needed and darken when the opposite is needed, without two separate systems to coordinate. In parallel, research into dye-sensitised cells (DSSC) and organic cells is trying to reduce the transparency-versus-output trade-off that currently penalises crystalline silicon in semi-transparent applications. But the factor that will really decide adoption at scale, more than the technology itself, remains economic: as long as the cost per square metre of these products stays a multiple of conventional Low-E glass, their adoption will follow the same trajectory as any flagship technology — first the manifesto buildings, then, slowly, everything else.