Public Wireless Supply Chain Innovation Fund 2026: Up to $53 Million for AI-Native RAN Projects, Deadline September 9

Public Wireless Supply Chain Innovation Fund 2026: Up to $53 Million for AI-Native RAN Projects, Deadline September 9

The National Telecommunications and Information Administration (NTIA) has opened the fourth notice of funding opportunity under the Public Wireless Supply Chain Innovation Fund. The 2026 competition, titled Solutions for AI-Native RAN, is aimed at U.S.-based organizations that can build, deploy, and commercially validate a secure compute environment for native artificial intelligence in the Radio Access Network (RAN).

This is a large, technically demanding federal grant. It is not a general startup award, an exploratory research fellowship, or a grant for an early concept with no field path. The notice asks applicants to connect a working technical solution to a credible business case: operator revenue, lower operating costs, or innovative third-party services. Projects must also show how the proposed system can support open, interoperable, secure wireless equipment and move toward commercial deployment.

The opportunity is worth attention because it combines substantial federal support with a concrete demonstration requirement. NTIA may award up to $53 million in total, with an anticipated amount of up to $25 million per project. The tradeoff is equally important: applicants must provide at least 50% of total project cost as non-federal cost share, and the lead applicant must already have the organizational, technical, and commercial capacity described in the notice.

Key details

The Grants.gov record does not publish a standard award floor. Treat the $25 million figure as the anticipated amount per project, not as an automatic award entitlement. The full notice also says that requests above $25 million may be considered if the applicant explains the variance, while a request above $50 million in federal funding is disqualified under the statutory limitation described in the NOFO.

What the grant is designed to fund

The central idea is an AI-native RAN solution built around a secure compute environment. In ordinary terms, NTIA is looking for technology that puts useful AI capabilities into the radio access network while preserving security, interoperability, and a path to deployment. A promising application must explain the network fabric, the compute architecture, the AI functionality, the interfaces and standards involved, and the setting in which the technology will be tested.

The solution also needs a commercial reason to exist. The notice specifically points to revenue for operators, cost savings, and new third-party services. An application that describes an impressive model but cannot identify who would pay for the resulting capability will be a weak fit. The business case should connect technical milestones to measurable customer or operator value.

The funded work is expected to culminate in a field demonstration. Before the end of the performance period, a project must have at least a working hardware prototype and beta-level software, with the technology tested in a field setting appropriate to the project narrative. NTIA says the demonstration should generate evidence about functionality, security, network virtualization, competition, and the deployment of open and interoperable equipment.

The opportunity can support a project that begins at different technical maturity levels. That flexibility does not remove the exit requirement. Applicants should work backward from the final demonstration and show why the proposed work plan can reach that state within a period of no more than 18 months.

Who is eligible to lead an application

The lead applicant must satisfy two important tests. First, it must be capable of developing and commercially selling or deploying a solution that uses a secure compute environment to enable native AI capabilities in the RAN. Second, it must be organized under the laws of the United States or one of its states and have its principal place of business in the United States. The NOFO says that a subsidiary, affiliate, or branch whose ultimate parent is foreign or whose principal place of business is outside the United States is not eligible to apply as the lead entity.

The application must demonstrate this capability rather than merely assert it. Evidence may include existing products, deployments, field-test results, engineering staff, customer relationships, standards participation, relevant intellectual property, or prior work in RAN, cloud-native networking, secure systems, radio hardware, or applied AI. The right evidence depends on the proposed project, but it should make the applicant’s ability to execute easy to verify.

Partners can be involved. The notice permits entities to participate as subrecipients, contractors, consultants, or other allowable participants in multiple applications, subject to the detailed restrictions in Section 2.2.2. The lead applicant, however, may submit no more than one application under this notice and may propose only one project in that application. This makes internal coordination important for companies, universities, and consortiums that have several possible concepts.

The project location rules are also significant. Activities must take place in the United States, its territories, and possessions. The requirement covers design, development, prototyping, initial testing, and field demonstration, although the NOFO describes limited waiver authority for certain field demonstrations when compelling evidence shows that a waiver would materially advance the program objectives. Applicants should not build their core plan around an assumed waiver.

The NOFO contains additional restrictions involving countries of concern, covered nations, ownership and control, excluded entities, and untrusted vendors or equipment. Those provisions are detailed and should be reviewed directly before a consortium is finalized. A foreign participant may not automatically disqualify a project, but the lead applicant must check the exact participation and activity rules instead of relying on a general statement that international collaboration is allowed.

Funding, cost share, and project length

NTIA may award up to $53 million across this notice and anticipates up to $25 million per project. Every application must include a minimum cost share equal to 50% of total project cost. The contribution may include cash or eligible in-kind support, but it must be real, documented, allowable, and available for the proposed work. Applicants should not treat the cost-share section as a bookkeeping attachment added at the end.

For example, a project with a total cost of $20 million must provide at least $10 million in non-federal cost share and may request up to $10 million in federal funding if the budget is structured that way. The correct ratio must be calculated from the total project cost, not by comparing the applicant contribution only with the federal request. The notice warns that applications below the 50% threshold are disqualified, and it explains that the percentage is rounded down to the nearest whole number.

NTIA encourages applicants to show how private-sector investment, cash contributions, in-kind resources, and other funding sources reduce the need for federal incentives. That does not mean the cheapest project automatically wins. It means the budget should make clear why each federal dollar is needed, what the applicant and partners are committing, and how the funding mix supports a credible route to deployment.

The expected performance period is up to 18 months from the date of award. Extensions may be possible at NTIA’s discretion, but applicants are assessed for the risk of exceeding the required period. A work plan with an 18-month schedule should therefore identify the critical path: architecture freeze, procurement, integration, security testing, field deployment, data collection, business-case validation, and final demonstration.

What to prepare for the application

The full NOFO is the controlling document, but its required materials include standard federal forms and a detailed project package. Applicants should expect to prepare the following:

• SF-424, Application for Federal Assistance;
• SF-424A, Budget Information for Non-Construction Programs;
• Project/Performance Site Locations Form;
• a project narrative with the executive summary and required technical sections;
• a detailed budget and budget narrative;
• cost-share documentation and the source of each contribution;
• a description of the AI-native network fabric;
• a field-demonstration plan;
• a business case for the proposed services or capabilities;
• a security plan;
• a work-plan summary with milestones and deliverables;
• qualifications of the applicant and participants;
• facilities and equipment information;
• an intellectual-property plan when applicable; and
• other required representations, certifications, and attachments.

The project narrative has to do more than describe the technology. The required sections are arranged to let reviewers judge whether the technical architecture, field demonstration, commercial plan, security approach, schedule, people, facilities, and partners fit together. Missing a required element can prevent the application from reaching merit review.

SAM registration deserves early attention. The NOFO specifically warns that SAM registration may take weeks and that NTIA will not provide an extension because registration was delayed. The lead entity should confirm its SAM status, Grants.gov access, authorized organization roles, and the unique registration that will be used to submit the package well before the deadline.

How the application and review process works

Applications must be submitted electronically through Grants.gov by 11:59 p.m. Eastern Time on September 9, 2026. NTIA will not accept a complete application or part of an application by mail, courier, email, fax, or another route. Build in time for Grants.gov validation, upload problems, and confirmation of timely receipt.

The review has several stages. An initial review checks whether the package is complete and responsive. Applications that omit required narrative elements may not advance. Merit review then uses at least three independent reviewers and a 100-point framework. The largest component is technical merit, worth 50 points: 30 points for the AI-native network fabric and 20 points for the security plan. Commercial viability is worth 35 points, including the business case. The remaining points address other factors in the notice, including the strength and feasibility of the proposed project.

The scoring structure gives applicants a practical writing guide. Technical novelty alone is not enough. The narrative should make the security architecture as concrete as the AI architecture, quantify the commercial hypothesis, and show why the field demonstration will produce useful evidence. Reviewers need to see what will be built, where it will run, how it will be tested, what success will look like, and who can deliver it.

After merit review, NTIA conducts programmatic review of the highest-scoring applications and then obtains selecting-official approval. The notice says awards are expected on a rolling basis. Unsuccessful applicants will be notified by email and may request a debriefing within ten business days of that notification.

How to judge whether this is a good fit

This grant is a plausible fit for a company, research-and-development organization, university-led consortium, or technology partnership that already understands the operational environment of wireless networks. The strongest prospective leads will be able to show a path from current capability to field-tested prototype, not just a research question.

Ask five questions before investing in a full proposal:

1. Can the lead applicant prove that it is a U.S.-organized entity with a U.S. principal place of business?
2. Does the team have a specific secure compute and AI-native RAN architecture rather than a broad AI-for-telecom theme?
3. Can the project reach a working hardware prototype and beta-level software within 18 months?
4. Is there a credible operator, service-provider, or third-party customer use case with measurable value?
5. Can the applicant document at least 50% eligible cost share without counting prohibited or uncertain contributions?

If any answer is no, the applicant may need a different funding route or a stronger partner before applying. A smaller, sharply defined demonstration is likely to be more persuasive than an oversized plan that depends on uncommitted infrastructure, unresolved vendor questions, or a future customer who has not agreed to participate.

Practical preparation strategy

Start with the exit demonstration and work backward. Define the exact hardware prototype, software maturity, test environment, network interfaces, security checks, and business evidence that will exist at project close. Then map every item to a milestone, responsible participant, budget line, and source of cost share.

Next, create a one-page risk register. Include supply-chain restrictions, vendor eligibility, foreign ownership or control, spectrum or site access, integration dependencies, data and privacy issues, cybersecurity testing, and the time needed for procurement. The purpose is not to make the project sound risk-free. It is to show that the team knows which risks could prevent the final demonstration and has credible mitigations.

The business case should be testable. State the baseline cost, revenue opportunity, service buyer, adoption barrier, and evidence that would change the deployment decision. Avoid broad claims that AI will make networks more efficient. Explain which workload changes, what measurement will show improvement, and how the result could be repeated by other operators or equipment providers.

Finally, ask an engineer, a security reviewer, a commercial operator, and a grants administrator to read separate parts of the draft. Their questions will expose different weaknesses. The best final proposal should be understandable to a technically sophisticated reviewer who is not already inside the applicant’s product team.

Common mistakes to avoid

The most serious mistake is missing the cost-share calculation. Recheck the ratio using the total project cost, document every contribution, and confirm that the proposed sources are eligible. A second mistake is submitting multiple concepts through one lead applicant; the notice permits only one application and one proposed project per eligible lead.

Do not describe a laboratory prototype as a field demonstration without identifying the field setting, participants, test conditions, security controls, and evidence to be collected. Do not leave the business case until the last page. The program is designed to support commercialization and deployment, so commercial viability is part of the core evaluation.

Applicants should also avoid vague partnership language. Name each participant’s role, facilities, equipment, technical contribution, and accountability for milestones. Check ownership, control, country, and vendor restrictions before writing a consortium into the budget. Lastly, do not assume a late SAM registration or Grants.gov problem will be excused.

Frequently asked questions

Can an international company apply as the lead?

No. The lead applicant must be organized under U.S. law and have its principal place of business in the United States. The rules for foreign participants in other roles are more detailed, so read Section 2.2.2 before including them.

Is this a grant for any artificial-intelligence wireless idea?

No. The proposal must address a secure compute environment that enables native AI capabilities in the RAN and must connect the technical work to operator revenue, cost savings, or innovative third-party services.

How much money can one project request?

NTIA anticipates up to $25 million per project. A request above that amount needs a reasonable explanation, and a request above $50 million in federal funding is disqualified under the notice’s statutory limitation.

Can cost share be in kind?

Yes, the notice permits cash or in-kind contributions, but the contributions must be eligible, documented, and sufficient to meet the minimum 50% of total project cost.

Can one organization join more than one proposal?

An eligible lead may submit only one application containing one project. The notice does not restrict an entity from participating in multiple applications as a subrecipient, contractor, consultant, or other allowable participant, subject to the detailed rules.

Where is the authoritative application guidance?

The Grants.gov listing and the attached FY26 PWSCIF NOFO control. Applicants should use the full notice for eligibility, technical requirements, required forms, formatting, submission, and review rules.

Official links and next steps

Review the official Grants.gov opportunity record, then download and read the FY26 PWSCIF NOFO – Solutions for AI-Native RAN. The official record identifies the opportunity number as NTIA-PWSCIF-26-01, the agency as NTIA, the funding instrument as a grant, and the deadline as September 9, 2026.

Before drafting, confirm SAM and Grants.gov readiness, validate the lead applicant’s U.S. status, assemble the cost-share evidence, and test the proposed schedule against the 18-month limit. Questions about access to the announcement can be directed through the contact information in the official record; the NOFO identifies Shanell Williams at [email protected] as a grantor contact. Submit the complete package through Grants.gov before the stated Eastern Time deadline and retain the electronic confirmation of timely submission.