NVIDIA Awards $60,000 Fellowships to PhD Students Pushing Boundaries of AI and Robotics

For a quarter-century, NVIDIA's Graduate Fellowship Program has been a cornerstone in nurturing the next generation of computing innovators. This year, the program announced its latest round of awards, offering up to $60,000 each to ten exceptional PhD students whose research spans autonomous systems, deep learning, robotics, and beyond. These fellows represent the vanguard of accelerated computing, tackling some of the most complex challenges in modern technology.

The selection process for this year's fellowship was notably rigorous, drawing from a global pool of applicants. Each awardee will embark on a summer internship at NVIDIA before their fellowship year begins, providing them with hands-on experience and direct exposure to cutting-edge research environments. This structured approach ensures that fellows not only contribute to NVIDIA's technological advancements but also gain invaluable insights into industry-leading practices.

The 2026-2027 Fellowship Recipients: A Diverse Range of Innovative Research

This year's cohort of fellows brings a diverse array of research focuses, each poised to make significant strides in their respective fields. Below are the ten recipients and their groundbreaking projects

  • Jiageng Mao (University of Southern California): Mao is delving into complex physical AI problems, aiming to leverage internet-scale data to develop robust, generalizable intelligence for embodied agents operating in real-world environments. His work promises to bridge the gap between theoretical AI models and practical applications, enabling more adaptable and reliable autonomous systems.
  • Liwen Wu (University of California San Diego): Wu's research is centered on enhancing realism and efficiency in physically based rendering through the use of neural materials and neural rendering techniques. This could revolutionize industries reliant on high-fidelity visual simulations, from gaming to architectural visualization.
  • Manya Bansal (Massachusetts Institute of Technology): Bansal is tackling a critical challenge in modern computing by designing programming languages tailored for accelerators. Her goal is to empower developers to write modular and reusable code without compromising the low-level control necessary for peak performance, thereby democratizing high-performance computing.
  • Sizhe Chen (University of California, Berkeley): Chen's focus on securing AI in real-world applications is particularly timely, given the increasing prevalence of AI systems. He is working on defenses against prompt injection attacks that are both general and practical, ensuring that AI agents remain useful while being resilient to adversarial inputs.
  • Yunfan Jiang (Stanford University): Jiang's project aims to develop scalable approaches for building generalist robots capable of handling everyday tasks. By integrating hybrid data sources—including real-world manipulation, large-scale simulations, and internet-scale multimodal supervision—his work could pave the way for more versatile and autonomous robotic systems.
  • Yijia Shao (Stanford University): Shao is exploring human-agent collaboration, focusing on AI agents that can communicate and coordinate with humans during task execution. Additionally, he is designing new interfaces for human-agent interaction, which could enhance productivity and efficiency in collaborative settings.
  • Shangbin Feng (University of Washington): Feng's research on model collaboration seeks to advance how multiple machine learning models—trained on different data by different people—can collaborate, compose, and complement each other. This work is crucial for building an open, decentralized, and collaborative AI ecosystem.
  • Shvetank Prakash (Harvard University): Prakash is pushing the boundaries of hardware architecture and systems design with AI agents built on new algorithms and curated datasets. His approach leverages agent-first infrastructure, which could redefine how we think about hardware-software co-design.
  • Irene Wang (Georgia Institute of Technology): Wang's project involves developing a holistic codesign framework that integrates accelerator architecture, network topology, and runtime scheduling. This framework aims to enable energy-efficient and sustainable AI training at scale, addressing critical challenges in large-scale AI deployment.
  • Chen Geng (Stanford University): Geng is focused on modeling 4D physical worlds using scalable data-driven algorithms and physics-inspired principles. His work on physically grounded 3D and 4D world models could have profound implications for robotics and scientific applications, enabling more accurate simulations of dynamic environments.

Acknowledging the Finalists: The Next Wave of Innovators

In addition to the ten fellowship recipients, NVIDIA also recognized five finalists who demonstrated exceptional potential in their research. These finalists include

Eigenmannia virescens 000
  • Zizheng Guo (Peking University)
  • Peter Holderrieth (Massachusetts Institute of Technology)
  • Xianghui Xie (Max Planck Institute for Informatics)
  • Alexander Root (Stanford University)
  • Daniel Palenicek (Technical University of Darmstadt)

While these finalists did not receive the fellowship awards, their innovative approaches and research potential were acknowledged as highly promising. Their work spans a range of disciplines, from advanced computing architectures to novel algorithms in AI, reflecting the breadth and depth of talent emerging in the field.

The Impact of NVIDIA's Graduate Fellowship Program

Since its inception 25 years ago, the NVIDIA Graduate Fellowship Program has played a pivotal role in shaping the future of computing. By providing financial support, mentorship, and industry exposure, the program empowers PhD students to pursue ambitious research projects that push the boundaries of what is possible in AI, robotics, and accelerated computing.

One of the standout features of this program is its global reach. Applicants from around the world are encouraged to apply, ensuring a diverse and inclusive cohort that reflects the international nature of modern technological innovation. This global perspective enriches the research landscape, fostering collaborations that transcend geographical boundaries.

The summer internship component of the fellowship is particularly noteworthy. It offers fellows a unique opportunity to work alongside NVIDIA's top researchers and engineers, gaining practical experience while contributing to high-impact projects. This hands-on approach not only accelerates learning but also provides a clear pathway for transitioning from academia to industry.

Looking Ahead: The Future of Accelerated Computing

The research areas pursued by this year's fellowship recipients are at the forefront of technological advancement. From securing AI systems against adversarial attacks to developing generalist robots capable of handling everyday tasks, these projects address some of the most pressing challenges in computing today.

For instance, the work on model collaboration and decentralized AI ecosystems could redefine how machine learning models interact and share knowledge, leading to more robust and adaptable systems. Similarly, advancements in physically based rendering and 4D world modeling have the potential to transform industries ranging from entertainment to scientific research.

As these fellows continue their work, they will undoubtedly contribute to NVIDIA's mission of accelerating AI and robotics innovation. Their research will not only shape the future of these fields but also inspire the next generation of innovators to tackle even more ambitious challenges. The 2026-2027 cohort represents a promising step forward in this journey, embodying the spirit of curiosity, collaboration, and technological breakthrough that defines NVIDIA's legacy.

Conclusion

The NVIDIA Graduate Fellowship Program remains a beacon for aspiring researchers and innovators worldwide. By investing in talent and providing the resources needed to explore bold ideas, NVIDIA continues to drive progress in accelerated computing. As the program enters its 25th year, it stands as a testament to the power of collaboration, innovation, and the relentless pursuit of excellence that defines the tech industry.