Neural Networks Take the Stage: Experiments With GigaChat Are Becoming a Tradition
The Mariinsky Theatre’s historic stage hosted a concert titled Music of the Theatre, where world-renowned musical works were performed against a backdrop of AI-generated visual scenery created with Sber’s neural networks. The event was organized with support from Russia’s Ministry of Culture as part of the cultural program of the St. Petersburg International Economic Forum 2026.
The Mariinsky Theatre hosted an unusual concert marking the 150th anniversary of the Union of Theatre Workers of Russia. The program featured excerpts from landmark operas and ballets that showcased the scale and legacy of Russian musical theatre.
Singing, Dancing, and Digital Scenery
More than 200 performers appeared on stage, including opera and ballet soloists, the Mariinsky Theatre Chorus, and the Mariinsky Symphony Orchestra. Conducting the performance was Valery Gergiev, People’s Artist of Russia and the theatre’s artistic director.
Audiences heard works by Tchaikovsky, Borodin, Puccini, Verdi, Prokofiev, Bizet, Shchedrin, and other composers, with a total of 15 musical pieces in the program. The evening opened with the overture Journey Through the Universe and concluded with a scene from Rimsky-Korsakov’s opera The Legend of the Invisible City of Kitezh and the Maiden Fevroniya.
The cast included some of the biggest names in Russian classical performance. Hibla Gerzmava appeared as the Diva, performing her aria, while Olga Borodina sang an excerpt from The Queen of Spades. The finale featured Yuri Vorobyov, Maksim Daminov, Alexander Trofimov, and Darya Tereshchenko.
Visual Design by SberStudios
The visual concept for the production was developed by SberStudios. The entire performance revolved around two symbolic figures, the Chronicler and the Muse, representing the historical journey of Russian theatre. Each musical number was accompanied by unique stage imagery generated with GigaChat and Kandinsky. The visuals were displayed on a massive LED screen.
Audience members noted the restrained and thoughtful use of the technology. Artificial intelligence helped create an emotional atmosphere without undermining the authenticity of either the historic stage or the works being performed.
Organizers believe experiments of this kind can reshape audience perceptions and expand the boundaries of traditional theatre. The use of artificial intelligence in the concert’s visual design offered a new perspective on established artistic traditions. Projects like this create new possibilities for theatre by helping reveal the meanings embedded in classical works in richer and more nuanced ways.
Russian Neural Networks Earn Recognition
An experiment on such a prestigious stage can be viewed as a form of recognition for neural networks within academic and classical arts. It is particularly significant that the visual design was created using Russian-developed generative models.
The Mariinsky Theatre is no stranger to ambitious experiments involving artificial intelligence. Last summer, it staged Russia’s first opera completed in collaboration with AI, Mandragora. GigaChat and Symformer helped finish a musical work originally started by Pyotr Tchaikovsky, while Kandinsky contributed to the creation of visual scenery.
Across Russia and around the world, artificial intelligence is becoming an increasingly practical tool for creative professionals. Audiences, in turn, gain access to new forms of cultural experiences that are more visual, more multimedia-driven, and potentially more accessible to younger generations. Experts believe AI-powered scenography could make academic art more engaging and easier to understand for people raised on digital media, without replacing live performance itself.
The concert at the Mariinsky Theatre served as a striking demonstration that artificial intelligence is gradually entering academic culture as both a production technology and an artistic tool. Russian generative platforms are already being used to reinterpret cultural heritage in contexts where content quality and respect for tradition are especially important. Neural networks such as GigaChat and Kandinsky may also attract interest from other Russian and international theatres and cultural institutions as part of broader solutions for generating visual content and creating stage designs.
