A national program that will connect innovators and large-scale operators to rapidly translate efficiency technologies into real-world AI systems.

STRIDE Ventures is launching the AI Efficiency Challenge, a program designed to dramatically improve the efficiency of at-scale AI/ML systems and data centers by accelerating the commercial adoption of translation-ready solutions. The initiative aims to strengthen the effective capacity and competitiveness of U.S.-based AI companies by reducing the cost of training and inference, addressing near-term limitations on data center capacity, and accelerating the time to market of new models.

The AI Efficiency Challenge is built on a central premise: researchers have already developed technologies capable of delivering substantial efficiency gains across AI/ML workloads. What’s missing is a structured pathway to bring these solutions into real operational environments – fast. The program will create a structured environment for collaboration among researchers, technology developers, and organizations running large-scale AI/ML workloads, with the goal of awarded teams achieving preliminary deployments at scale within one year of receiving funding.

“Efficiency improvements are the fastest path to expanding AI capacity without waiting for new infrastructure. We are proud to once again partner with the NSF – this time on the AI Efficiency Challenge. By focusing on translation-ready solutions and measurable, at-scale performance improvements, this initiative will deliver real, near-term impact across the industry and accelerate U.S. competitiveness in the global AI landscape.” – Annika Pierson, CEO of the Americas, Start2 Group

Three Participant Groups

• Innovators (“Pitchers”): Researchers, entrepreneurs, and innovators with translation-ready efficiency technologies
• Deployment Partners (“Catchers”): Organizations running AI/ML workloads at scale who commit to integrating new solutions and reporting efficiency gains
• Benchmarking Experts (“Umpires”): Teams developing industry benchmarks to assess the efficiency of the AI/ML software stack and validate measurable gains

This collaborative model ensures that every supported project is designed for implementation in real operational environments, not just demonstration settings.

The AI Efficiency Challenge will prioritize translation-ready, primarily software-based solutions that deliver efficiency improvements across the AI/ML pipeline — from data preparation to training through inference-serving and agentic architectures. Target areas include efficient AI/ML algorithms, automated model compression and distillation, MLOps and distributed system software (scheduling, placement, runtime, orchestration), efficient agentic orchestration, cloud-edge partitioning, and software-enabled energy and thermal management. Solutions must be deployable without long lead-time hardware or infrastructure dependencies.

Funding Awards

• Large Award: $3.5M
• Medium Award: $1.75M

Funding will be offered at two levels: a Large award of up to $3.5 million and a Medium award of up to $1.75 million per project. Awards will emphasize rapid team mobilization and close collaboration with deployment partners, and progress will be evaluated through a go/no-go framework directing continued resources to projects demonstrating measurable impact.

The AI Efficiency Challenge is administered by Start2 Group as the OT Contractor under an Other Transaction Agreement with the National Science Foundation.

Application Details

The call for submissions open May 18, 2026, with applications due July 13, 2026. Learn more about the AI Efficiency Challenge here: https://stride-ventures.com/ai-efficiency-challenge/.