Print

NSF's implementation of the revised 2 CFR

NSF Financial Assistance awards (grants and cooperative agreements) made on or after October 1, 2024, will be subject to the applicable set of award conditions, dated October 1, 2024, available on the NSF website. These terms and conditions are consistent with the revised guidance specified in the OMB Guidance for Federal Financial Assistance published in the Federal Register on April 22, 2024.

Important information for proposers

All proposals must be submitted in accordance with the requirements specified in this funding opportunity and in the NSF Proposal & Award Policies & Procedures Guide (PAPPG) that is in effect for the relevant due date to which the proposal is being submitted. It is the responsibility of the proposer to ensure that the proposal meets these requirements. Submitting a proposal prior to a specified deadline does not negate this requirement.

Updates to NSF Research Security Policies

On July 10, 2025, NSF issued an Important Notice providing updates to the agency's research security policies, including a research security training requirement, Malign Foreign Talent Recruitment Program annual certification requirement, prohibition on Confucius institutes and an updated FFDR reporting and submission timeline.

Dear Colleague Letter

Advancing Artificial Intelligence (AI) Agent Ecosystems through the National Science Foundation Pathways to Enable Secure Open-Source Ecosystems (NSF PESOSE) Program

Encourages NSF Pathways to Enable Secure Open-Source Ecosystems (NSF PESOSE) proposals that focus on the protocols enabling AI agent ecosystems.
Encourages NSF Pathways to Enable Secure Open-Source Ecosystems (NSF PESOSE) proposals that focus on the protocols enabling AI agent ecosystems.

Dear Colleague: 

The U.S. National Science Foundation (NSF) encourages proposals to the Pathways to Enable Secure Open-Source Ecosystems (PESOSE) program that focus on protocols enabling AI agent ecosystems. 

PESOSE program context

The PESOSE program strengthens open-source ecosystems that produce publicly accessible, modifiable, and shareable software and hardware, models and specifications, languages and measures for security and cybersecurity of artificial intelligence (AI) models and systems, metrology, data tools and benchmarking tools. These open ecosystems drive innovation across AI, cloud computing, finance, healthcare, education, manufacturing, mobility, and national security. 

PESOSE supports the maturation of the organizations that manage open-source ecosystems by helping them distribute development and grow broad user and contributor communities across academia, industry, and government. The program also advances security and privacy by addressing vulnerabilities that affect how open ecosystems evolve.

AI agent protocol ecosystems

AI agents are moving beyond simple chat interfaces. They can operate for long periods, write and debug code, search databases, manage email and calendars, buy goods, and operate laboratory equipment. These capabilities rely on open interfaces and shared standards that allow agents to communicate with tools, data sources, and each other. 

An AI agent ecosystem is a network of autonomous agents that communicate, coordinate, and work together using common protocols. These protocols let agents - built by different organizations or using different frameworks - smoothly share data, and perform tasks across different environments. They are essential for scalable, secure, and reproducible AI systems.

PESOSE and AI agent ecosystems

The U.S. has a long history of advancing open technology ecosystems and their enabling protocols. For example, early investments by the Defense Advanced Research Projects Agency (DARPA) and NSF enabled the Transmission Control Protocol/Internet Protocol (TCP/IP) which became the industry standard for addressing, transmitting, and routing data across interconnected networks. Today TCP/IP is the backbone of the Internet.

This Dear Colleague Letter (DCL) highlights NSF’s interest in PESOSE proposals that similarly advance the protocols for AI agent ecosystems, strengthen open‑source foundations and support growth across sectors.

PESOSE welcomes proposals that advance one or more of the following:

  • Interoperability. Develop open standards so agents can work across platforms and organizations without custom connectors.
  • Scalability. Create architectures that support large networks of agents, tools, and services while keeping integration simple.
  • Security. Design interoperable protocols with security features that enable agent operation across heterogeneous security regimes. Include:
    • Cross-domain data classification mapping,
    • Identity verification,
    • Role- and attribute-based access control enforcement,
    • Secure, policy-compliant communication channels,
    • Robust authentication and fine-grained authorization,
    • Zero knowledge proofs,
    • End-to-end auditability, and
    • Formally specified, safe message formats.
  • Open Science Applications. Demonstrate and evaluate security features in AI ecosystem protocols that enable open science.  
  • Partnerships. Build collaborations among academic, nonprofit, industry, and government partners to accelerate appropriate adoption.
  • Education and Training. Expand training for students and postdoctoral researchers to contribute to AI agent ecosystems, including security and reliability.

NSF will give strong consideration to proposals that address these priorities. Research must remain unclassified and publicly releasable.

Proposal Submission

When responding to this DCL, please title your proposal: “PESOSE / AI: Track #: (title)”.

Proposals must follow the requirements in the NSF Proposal & Award Policies & Procedures Guide (PAPPG) and the NSF PESOSE solicitation. NSF will evaluate proposals submitted in response to this DCL based on their alignment with the goals above.

For details, consult the full PESOSE solicitation and/or website.

Questions may be sent to: PESOSE@nsf.gov.

Sincerely,

Erwin Gianchandani
Assistant Director for Technology, Innovation and Partnerships

Related programs

Organizations

Top