Google utilized Veo and Gemini to recreate the most remarkable goal Pelé ever scored, a goal that was never captured on film.
TL;DR: Google DeepMind has recreated Pelé's "lost goal" from 1959 using Veo, Gemini Omni, and Nano Banana Pro, drawing on 2,000 records and eyewitness interviews. The reconstruction is displayed at the Pelé Museum.
On August 2, 1959, Pelé claimed to have scored the greatest goal of his career, performing three consecutive sombreros over defenders, a knee flick past the goalkeeper, and a header into the net, all without the ball touching the ground. This remarkable moment was never captured on film. For 67 years, the "Gol da Rua Javari" lived only in the memories of fans present. Now, Google DeepMind has successfully reconstructed it.
Collaborating with Pelé’s family and the Pelé Brand, the project blended traditional filmmaking techniques with three Google AI models: Veo 3, Gemini Omni, and Nano Banana Pro. Historian Anita Lucchesi collected nearly 2,000 historical documents, including stadium blueprints and family photos, and conducted interviews with surviving witnesses. A team filmed live-action footage at the original Rua Javari stadium using historically accurate leather balls and uniforms. This footage was processed through the AI models, which substituted a stunt double with Pelé’s likeness, altered the modern stadium to reflect the 1959 design and conditions, and created realistic crowd scenes from that era.
Flávia Kurtz, Pelé’s daughter, remarked, “He would be so proud to see all this happening. He’d often express that it was unfortunate the goal was never documented.” The reconstruction is currently on view at the Pelé Museum in Santos. Google unveiled Gemini Omni as a conversational video-generation model at I/O 2026, and the Pelé project is its most culturally significant application so far, showcasing how AI video generation can contribute to preservation rather than mere content creation.
The final result underwent processing with a filmout machine to achieve the aesthetic of 1950s cinema, followed by traditional VFX enhancements for ball compositing and color grading. Nano Banana Pro was engineered for precise image generation, and the Pelé reconstruction exemplifies the model's ambition: reconstructing a real event from existing fragments rather than merely producing content from scratch. In a year marked by AI inundating the internet with synthetic material, this project stands out as a rare example of technology being used to reclaim something genuinely lost.
Other articles
Google utilized Veo and Gemini to recreate the most remarkable goal Pelé ever scored, a goal that was never captured on film.
Pelé's "Gol da Rua Javari" from 1959 was never documented on film. Google DeepMind recreated it using 2,000 historical records, interviews with eyewitnesses, and AI models.
