CVPR 2026 Program Committee Reveals Relationship Between GPU Access on Accepted Paper Presentations in New Research
Compute Report finds accepted paper authors have access to higher GPUs and more development time
LOS ALAMITOS, Calif., 28 October 2026 – Today, the Program Committee of the 2026 Conference on Computer Vision and Pattern Recognition (CVPR) unveiled its Compute Report, which investigated if and how access to graphics processing units (GPUs) and development time impacted conference paper acceptance. The report revealed that authors with accepted papers had increased access to more GPU computing power when compared to rejected papers, and geographies that had access to higher computing capacity also had a greater representation in the program.
“There’s a wide sense that GPU access is one of the driving factors behind what’s happening in AI today,” said Vladimir Pavlovic, CVPR 2026 Program Chair; professor, Rutgers University. “We wanted to collect concrete data that would tell us if infrastructure is a contributing factor and let the evidence speak for itself.”
Drawing on mandatory compute disclosure forms provided by authors of 16,034 submitted papers, the Compute Report maps the hardware landscape of modern AI research with unprecedented breadth. Key findings include:
- CVPR 2026 accepted papers demonstrated a two-fold GPU advantage over rejected papers, reporting a median of 4 GPUs versus a median of 2 GPUs for rejected papers.
- CVPR 2026 accepted papers touted 67% more development time, with a median of 2,000 development hours, compared to just 1,200 hours for rejected papers, despite both groups working on similar project timelines averaging seven months.
- Countries whose researchers deploy more GPUs per paper are accepted at higher rates, consistently across the full distribution. The data shows that it is not the generation of GPU that predicts a country’s acceptance rate, but how intensively researchers use the hardware they can access.
- When normalized by population, AI research productivity falls on small research hubs: Singapore submitted 64.6 papers per million people and Hong Kong 53.2, compared to just 6.6 for the United States and 6.1 for China.
“Our primary goal with this report was to bring transparency to the data,” explained Pavlovic. “The hope is that by sharing this information we can help encourage access to compute on a global scale and enable a level playing field in AI and computer vision research.”
While the majority of information related to public research in AI and computer vision is open, the CVPR Program Committee recognized a blind spot in the correlation between paper acceptances and authors’ access to modernGPU infrastructure. With that in mind, the committee, supported by conference co-sponsors the IEEE Computer Society (CS) and Computer Vision Foundation (CVF), developed the Compute Report, based on self-reported GPU access data, to deliver transparency around the effects of GPU infrastructure on paper presentations.
For more information and to access the detailed findings of the analysis, visit https://computereport.github.io/cvpr2026/.
About CVPR 2026
The Conference on Computer Vision and Pattern Recognition (CVPR) is the preeminent computer vision event for new research in support of artificial intelligence (AI), computer vision, multi-modal AI, wearable AI, novel AI architectures, spatial computing, agentic AI, AI for science, embodied AI and robotics, and much more. Sponsored by the IEEE Computer Society (CS) and the Computer Vision Foundation (CVF), CVPR delivers the important advances in computer vision and pattern recognition that are expanding the frontiers of computer vision and defining the next generation of intelligent, interactive technologies. With a first-in-class technical program, including tutorials and workshops, a leading-edge, immersive expo, and robust networking opportunities, CVPR, which is annually attended by more than 10,000 scientists and engineers from across the globe, creates a one-of-a-kind opportunity for networking, recruiting, inspiration, and motivation.
CVPR 2026 took place 3-7 June at Colorado Convention Center in Denver, Colo., U.S. CVPR 2027 will take place 19-26 June at the Seattle Convention Center in Seattle, Wash., U.S. For more information about CVPR, visit cvpr.thecvf.com.
About the IEEE Computer Society
Engaging computer engineers, scientists, academia, and industry professionals from all areas of computing, the IEEE Computer Society sets the standard of excellence and champions global advancements to benefit humanity.
As the IEEE Computer Society celebrates its 80th anniversary in 2026, it continues to build on its rich legacy. Through conferences, publications, and programs that bring together computer science and engineering leaders at every stage of their careers, the IEEE Computer Society empowers, shapes, and guides the future of its 375,000+ community members, as well as the greater computing community, enabling new opportunities to better serve our world. Visit computer.org for more information.
About the Computer Vision Foundation
The Computer Vision Foundation (CVF) is a non-profit organization whose purpose is to foster and support research on all aspects of computer vision. Together with the IEEE Computer Society, it co-sponsors the two largest computer vision conferences, CVPR and the International Conference on Computer Vision (ICCV). Visit thecvf.com for more information.