The moment “Oye Hoye Kya Scene Hai, Teri Walk Mein Taan Hai” played on season 2 of Samay Raina’s India’s Got Latent, it was clear the song had struck a chord. The show returned on Netflix and YouTube simultaneously and instantly became the talk of the nation. On the premiere episode, Contestant Balram Vishwakarma, better known as Rocking Goli, brought the energy with his performance, while rapper Young AJ wrote and co-performed the track. The standout moment on the show quickly snowballed into an internet phenomenon, with everyone from fans to celebrities joining in. Even DJ Chetas was impressed enough to invite Balram to perform at his shows.
The song quickly outgrew the reality show and found a life of its own online. During the premiere episode, Alia Bhatt and Sharvari couldn’t resist grooving to its infectious beat, and before long, Instagram was flooded with videos featuring the track. From dance reels to meme edits, Chai Ki Tapri was everywhere. It has now crossed more than 500 million cumulative streams and has featured in over a million Instagram Reels, making it one of the biggest viral songs of the year.
What makes its success even more surprising is how it was created. While most people assumed it was recorded in a professional music studio, Chai Ki Tapri was actually generated using artificial intelligence by a creator from Nagpur.
The viral hit nobody saw coming
The success of Chai Ki Tapri surprised not just listeners but also the company behind the technology used to create it. “I had no idea that Kush used Soundverse to generate it,” admits Sourabh Pateriya, Co-founder of Soundverse AI, to Hindustan Times in an exclusive interview.
Considering the platform now has over 3.5 million users who have collectively created more than 15 million songs, it isn’t surprising that individual tracks often go unnoticed. The viral success of Chai Ki Tapri came as a surprise even to the people behind the technology. He adds, “I got a text from a friend of mine saying, ‘Hey, I feel this song could be AI, and it’s worth checking if it’s made in Soundverse.’ I saw the episode, heard the song—it’s a really fun track. It definitely caught me off guard.”
Curious, the team searched their internal database and confirmed that the song had indeed been generated on Soundverse in mid-May. For Sourabh , however, the bigger takeaway wasn’t just that the song had gone viral. It was that millions of listeners couldn’t tell it was AI-generated. Many were convinced that an artist named AJ had actually sung it.
“For the longest time, people were doubting that it was AI. They always thought that AJ actually sang on it, which is not true. It gives me confidence that AI music is here to stay, and it’s becoming good enough for people not to even recognize it,” he explains.
Can AI really replace emotion?
As AI-generated music becomes more common, one question refuses to go away: can a machine create songs with real emotion? Critics argue that AI can imitate melodies and lyrics, but it cannot recreate the lived experiences that shape great music.
Sourabh understands that argument well. Before building Soundverse, he spent years performing as a live metal musician and continues to release music under the name Blue Nucleus. For him, making music has always been a balance between creativity and technical execution. “I realized there are two components of making a song,” he explains. “One is the creative component—your taste, creativity, storytelling, and aesthetics. Then there is the technical component—the engineering, the DAW, programming, arrangement, and recording.”
He believes AI isn’t here to replace the creative mind. Instead, it can remove the time-consuming technical work that often slows artists down. He adds, “I fundamentally feel that creators are better off if they can focus on the creativity of the track instead of the technicality. The technicality is something that can be automated.”
The next evolution of music
Even as AI becomes more capable, Sourabh doesn’t believe it can replace what makes great music memorable. “When it comes to high-quality songs, it’s impossible for AI to match that right now. Humans are the source of musicality, story, and emotions. AI is simply learning from humans,” he asserts.
He believes the musicians who stand out in the coming years won’t necessarily be the ones with the most technical skills, but those with the strongest creative instincts. The ability to shape an idea, create fresh sounds and tell authentic stories will become more valuable than ever.
Cultural identity, he says, will also remain impossible to automate. Whether it’s the flavour of Punjabi music, the rhythm of Marathi folk or the nuances of Gujarati sounds, those lived experiences can only come from people.
The economics of music-making, however, are already changing rapidly. “If a human was making four songs for ₹40 lakh, now they will make 10 songs for ₹1 lakh,” Sourabh notes.
Lower production costs, he believes, will allow far more people to create music than ever before. Today, only around 10 million people worldwide actively produce music. Over the next decade, he expects that number to reach one billion.
For Sourabh, AI represents a shift similar to the arrival of synthesizers, which once sparked fears that traditional instruments would disappear. Instead, both found their place and expanded what music could become. JHe explains, “The role of the musician changed between the instrument era and the synthesizer era, which gave rise to new genres. They all co-exist now, and people don’t talk about ‘synth music’ versus ‘non-synth music’—it’s all just music. Similarly, people will stop caring about AI versus non-AI music. We shouldn’t see musicians as traditional or modern, but as the carriers of storytelling in a sonic format.”