At the beginning of 2026, something happened that the music industry had been simultaneously expecting and fearing: a track created entirely by artificial intelligence topped the VK Music and Yandex Music charts at the same time. The song "Play, Accordion," based on the poetry of Sergei Yesenin, became the first instance in Russia where a neural network outperformed living artists on the country's main streaming platforms.
The Mysterious SDP Project
The author of the track turned out to be neither a famous producer nor a major label, but a project with the concise acronym SDP — "Verse Worthy of a Song" (Stikh, dostoyny pesni). Behind this name lies writer Kirill Zbrodov, who launched the neural network experiment at the end of 2025. His idea is simple yet radical: to take works by classic authors that have already entered the public domain and, with the help of neural networks, turn them into full-fledged musical compositions.
Zbrodov set an ambitious goal for himself — to reach one million streams by the end of 2026. Reality exceeded expectations: by December of 2025, the counter had surpassed 10 million streams, and about 10 thousand posts using the project's tracks had appeared on social media.
Authentic Yesenin Verses
Many listeners first ask themselves the question: did artificial intelligence distort the classic text? In the case of "Play, Accordion," the answer is unequivocal — no. This is an authentic poem by Sergei Yesenin, written by him in 1922 and included in the famous cycle "Verses of a Scoundrel." The poet created it after one of his turbulent nights in a Moscow tavern, during a period of inner breakdown and existential crisis.
Zbrodov did not change a single word — the neural network received the original text as input and turned it into a musical piece. The voice we hear in the recording is also generated by artificial intelligence: slightly hoarse, brutal, tired — it fits perfectly with the unhurried, suspense-building beat. Previously, the band "25/17" had experimented with this poem, but it was the neural network version that became mainstream. The music/arrangement was also created by a neural network.
Technologies Behind the Scenes
The main tool for creating the music was the Suno AI neural network — one of the most advanced platforms for song generation. It allows you to upload text and get a full-fledged composition with vocals, arrangement, and production as output. In parallel, a powerful visual trend emerged around the song: social media users began creating videos where static photographs would "come to life" and dance to Yesenin's accordion. For this, another neural network was used — Kling 2.6 Motion Control, which transfers movements from a video reference to a photograph.
It was the combination of strong emotion and the simplicity of content creation that made the trend viral. To film a clip, you don't need to know how to dance, shoot video, or edit — a single portrait and light animation are enough.
How the Song Conquered the Charts
The success of "Play, Accordion" is explained by several factors. Firstly, the composition integrated powerfully into the recommendation algorithms of streaming platforms — it began to be pushed to literally every user "to listen to." Secondly, the audience's interest in the new and mysterious played a role: people were curious about what kind of track this was that sounded so human but was created by a machine.
Thirdly, the originality of the format itself is important. In an era when new music often seems like chewed gum in a bright wrapper, a neural network interpretation of a classic became a breath of fresh air. There is a feeling that AI is learning from humans how to fake emotions — and it is doing so convincingly. People believe what they hear.
Why This Matters
"Play, Accordion" is not just a viral hit. It is a signal that neural networks have ceased to be just a tool for assistance and have begun to conquer niches where thought, soul, intelligence, and heart were previously required. Classic poetry is free, neural networks are cheap, fast, and trainable. It seems it is no longer so important to people who exactly is singing — the main thing is the atmosphere, the voice, the sincerity.
Kirill Zbrodov is already working on new tracks: his compositions based on the verses of other classic authors have hit the charts. It seems we are witnessing only the beginning of a major transformation of the music industry, where the boundary between human and machine is becoming increasingly blurred.
