NSF 26-501: Collaboratory to Advance Mathematics Education and Learning for K-12 (CAMEL)

NSF 26-501: Collaboratory to Advance Mathematics Education and Learning for K-12 (CAMEL)

The U.S. National Science Foundation (NSF) CAMEL solicitation supports collaboration among researchers who study learning, K-12 education practitioners, and data scientists. Its purpose is to improve mathematics learning and education by connecting science of learning, classroom practice, data science, artificial intelligence, and technology. The solicitation was posted on December 10, 2025.

This page has a continuing but narrow application path. The Phase I deadlines have passed: the required Letter of Intent was due January 9, 2026, and the Phase I full proposal was due March 11, 2026. NSF still lists Phase II full proposals as accepted anytime, but Phase II is open only to awardees of Phase I. The rolling deadline in this page’s metadata therefore describes the continuing Phase II route; it does not mean that a new applicant can submit a late Phase I proposal or that a new Phase I round has been announced.

Key details at a glance

Why this opportunity is a good match for serious applicants

CAMEL is not simply a “produce a tool” grant. Its funding logic expects evidence-based collaboration and measurable outputs in teaching, learning, and data generation. The official solicitation states that CAMEL is intended to connect multiple disciplines: science of learning, cognitive and social science, computer science, machine learning, engineering, and education research, while also pulling in educational practice. That means proposals are rewarded when they treat mathematics education as an ecosystem problem rather than a one-off research intervention.

The opportunity is particularly useful for teams that can do both research and implementation-oriented work. NSF calls out the need for collaborative networks that combine theory and practice. A grant team that includes data scientists with practical classroom links and a team of researchers who can tie interventions to outcomes is generally more competitive than a grant that only proposes models or only does pilot studies. The call is explicit about producing high-value datasets that are usable by the broader field and likely to generate cumulative value.

The same structure also helps with long-term impact. Phase I is where networks and datasets are formed; Phase II is where outcomes are institutionalized through a national collaborators model, with a “collaboratory” built for continued coordination and training. Even if your main ask is to run one project, you should design as if the work can outlive the initial grant period.

What exactly Phase I and Phase II mean in practice

Many teams misread Phase I and Phase II as separate programs with independent applicant pools. In this solicitation, Phase II is only open to Phase I awardees. That has two consequences:

1. Every successful Phase I application should be built with a plausible handoff plan to a Phase II collaborative model.
2. Institutions applying only for Phase II are out; they must secure Phase I outcomes first.

Phase I itself has two pathways:

• 1a Novel Dataset Pathway: create new high-value K-12 math datasets where current collections are insufficient.
• 1b Existing Dataset Pathway: repurpose and increase value of existing K-12 math datasets with support for curation and safe sharing.

Both pathways are judged on quality of the research-to-practice loop and on technical rigor in data handling. The solicitation lists concrete expectations for each dataset: representativeness of learners, privacy controls, AI-readiness, size/quality sufficient for analysis, and FAIR principles (Findable, Accessible, Interoperable, Reusable). If your team cannot explain how data is collected, governed, annotated, and reused beyond your own project boundary, this is a strong reason to defer.

The Phase II concept is intentionally different: it is not just a “Phase I extension.” NSF describes Phase II as a collaboratory with a national coordinator, collaborative infrastructure, training in data science, and outcomes tracking. A project that gets into Phase II should already show that it can act as part of a connected network with shared standards and reproducible practice.

Eligibility and proposal constraints you must design around early

The eligibility section is broad in organizational type but strict in role and submission structure.

For proposal submission, the solicitation allows:

• Two- and four-year U.S. institutions of higher education (including community colleges), acting on behalf of their faculty.
• Non-profit, non-academic U.S. educational/research organizations.
• Federally recognized American Indian and Alaska Native tribal entities.

This combination means many organizations that do not hold university status can participate, but proposal ownership still follows NSF submission rules. Programmatic collaboration with museums, labs, or similar organizations can be strong if they are aligned with educational and research activity.

Other constraints with practical impact:

• LOI is mandatory for Phase I.
• PI or co-PI may submit only one Phase I proposal.
• Partners with Walton Family Foundation staff cannot be included as collaborators in proposals.
• Cost sharing is specifically prohibited.
• Collaborative proposals are restricted in system logistics; multiple organization submissions must go through Research.gov.

The PI limit is often underestimated. Because the solicitation permits only one Phase I PI/co-PI per applicant identity, institutions should centralize role planning before concept notes are drafted. Running parallel internal projects in the same institution can run afoul of this if roles are reused without governance clarity.

Application timing and current status

Phase I is closed for the cycle described by NSF 26-501. The required LOI deadline was January 9, 2026, by 5 p.m. in the submitting organization’s local time. An LOI was required to be eligible to submit a Phase I proposal, and the LOI had to be submitted through Research.gov. The Phase I full proposal deadline was March 11, 2026, also by 5 p.m. in the submitting organization’s local time.

Phase II is the reason this page remains useful as a current opportunity record. NSF lists Phase II full proposals as accepted anytime, and the program page identifies this as the upcoming due date for Phase II proposals. That intake is restricted to Phase I awardees. A university, nonprofit research organization, tribal nation, or other eligible organization that did not receive a Phase I award should not interpret the rolling Phase II entry as permission to submit an independent Phase II application.

There is no confirmed new Phase I deadline in the official solicitation reviewed for this update. Applicants looking for a future Phase I round should monitor the CAMEL program page and contact NSF at [email protected] rather than treating the closed dates as an upcoming window. The current official solicitation remains the source of record for the Phase I requirements and the continuing Phase II route.

Required materials and proposal architecture that actually improve odds

For a Phase I submission in this solicitation, the LOI was required; it was not optional. It also defines a specific naming convention and technical materials expected from teams. Phase II proposals are accepted anytime only from Phase I awardees.

Core submission requirements

• LOI via Research.gov by January 9, 2026. This deadline has passed.

• Full proposal by March 11, 2026 (Phase I). This deadline has passed.

• Proposal title conventions:

  • LOI titles should begin with “CAMEL-CN:”
  • Phase I full proposal titles should begin with “Phase I CAMEL-CN”
  • Phase II titles should begin with “Phase II CAMEL”

• One-page synopsis requirements for LOI:

  • math learning and data science goals
  • network composition and interdisciplinary value
  • dataset description and justification
  • anticipated high-impact outcomes
  • training needs/challenges addressed
  • up to five comma-separated keywords, including learner age range, math area, and data format
  • pathway designation (Phase 1a, 1b, or both)

• Project description must address collaboration mechanisms, resource quality, alignment to standards, use cases, implementation plan, evaluation, and management approach.

• Collaboration and Management Plan (up to 2 pages) is required.

When teams over-prepare the science without showing collaboration mechanics, they often get low marks. This program is unusually explicit that collaboration is not optional. NSF is interested in how researchers and K-12 practitioners will co-design and co-produce knowledge over time.

Budget and compliance details that can quietly hurt applicants

The funding level is substantial per award, but not huge compared with broad national programs, so budget structure matters. The solicitation caps likely award size at $1.5M over up to three years and sets no voluntary cost-sharing.

Key points:

• no committed cost sharing required, and voluntary committed cost sharing is prohibited,
• standard NSF cost principles apply,
• data generation and sharing plans are central, not peripheral.

The most common budget issue in this kind of program is weak linkage between spending and collaborative outputs. NSF reviewers expect expenditure logic that supports sustained data curation, training use cases, community coordination, and evaluation. If costs are only tied to pilot content, proposals look unsustainable.

Who this is strongest for (and who should likely skip)

This is strongest for teams that can demonstrate these capabilities:

• proven collaboration across disciplines,
• access to K-12 practice settings or practitioner collaborators,
• clear plans for quality and reusable datasets,
• team members with experience in research methods plus implementation workflows,
• long-horizon planning for network building and training capacity.

Examples of strong fits include:

• university education or data science groups with established school partnerships,
• education labs with experience in machine learning for learning diagnostics,
• interdisciplinary consortia involving school districts, researchers, and technical specialists,
• institutions that can handle data governance and privacy for learner-linked resources.

If your current team is primarily theoretical and lacks either school-level practitioner ties or a data engineering backbone, consider applying only after building at least one pilot data governance framework and a shared evaluation protocol. The program is less suited for one-off classroom interventions without data reuse design.

Review expectations and common mistakes

NSF uses standard merit criteria of Intellectual Merit and Broader Impacts, with program-specific filters layered on top. In CAMEL, successful proposals are usually the ones that can show:

• strong scientific and educational relevance,
• clear use-case path from data to outcomes,
• practical plan for collaboration and coordination,
• explicit handling of privacy and FAIR principles,
• credible plan for measuring impact (not just output metrics).

Common weaknesses that can weaken an application:

• treating the dataset as an internal product only,
• weak or generic collaboration plans,
• missing the required pathway clarity (1a vs 1b),
• insufficient justification for why a dataset is “high value,”
• ignoring PI limit constraints and filing multiple PI submissions,
• not documenting how outputs will transition into collaboration infrastructure.

A strong application usually reads like a design document, not just a grant proposal. It should explain not only what you will do, but how your team will ensure other organizations can actually use what you produce.

Practical preparation checklist before the LOI

Use this pre-submission order and keep a shared tracker:

1. Confirm organizational eligibility.
2. Assign one official PI and ensure no PI/co-PI duplicates an additional Phase I proposal.
3. Choose pathway(s): 1a and/or 1b, and test each against available datasets.
4. Draft one-page LOI with explicit outcomes and data value statement.
5. Prepare a 1-page keywords line (age ranges, math domains, data format).
6. Draft a collaboration and management section with named roles: principal investigators, practitioner liaisons, data governance lead, evaluation owner.
7. Confirm Research.gov account access and submission owner if using external collaborators.
8. Draft privacy and FAIR plan: de-identification strategy, metadata standards, sharing mechanism.
9. Build a review timeline backward from 2026-01-09 LOI and 2026-03-11 full proposal.

If you start after the year starts, this still works, but teams that do this in time generally submit more coherent docs because they can align the one-page LOI and full proposal requirements simultaneously rather than retrofitting later.

FAQ

Is this still open now?

Only the Phase II path remains available, and it is limited to Phase I awardees. The Phase I LOI and full proposal deadlines have passed, and no new Phase I cycle is announced in the official source reviewed here.

Can my institution submit both pathways?

Yes, proposals may be submitted through 1a and/or 1b, but the submission title must follow the required format and you must clearly justify each track and its distinct outcomes.

Can applicants without a traditional education department participate?

Yes, if they meet eligibility as an allowed organization type and are proposing a team with clearly relevant expertise. Many strong applications are interdisciplinary and do not come from a single education department.

Is there any funding year guarantee?

The amount listed is anticipated and contingent on available funds and competitive proposals. Expected FY2026 available amount is up to $9.0M, but NSF does not guarantee funding levels.

Do you need letters of intent?

For Phase I, the LOI is required and must be submitted before the full proposal.

Risks, limitations, and next steps for applicants

The biggest risk is treating CAMEL as just another dataset grant. It is designed as a collaborative network and ecosystem project. If your logic ends at data collection and publication, reviewers may view it as incomplete.

Second risk is submission mechanics. The required one-title convention, LOI timing, PI restrictions, and Research.gov requirements for collaborative submissions are straightforward but can create automatic disqualifications if missed.

Third risk is underestimating the transition from Phase I to Phase II. Even if Phase II has no fixed date, proposal quality should already include continuity planning.

Because this is a federal program with institutional-level governance, teams should assume a minimum of several internal approvals, legal/data governance checks, and review cycles before submission. Build that process as a project component from the start.

Official links

• Official solicitation (official source): https://www.nsf.gov/funding/opportunities/collaboratory-advance-mathematics-education-learning-k-12/nsf26-501/solicitation
• NSF program page: https://www.nsf.gov/funding/opportunities/collaboratory-advance-mathematics-education-learning-k-12
• NSF PAPPG page (proposal rules): https://www.nsf.gov/policies/pappg_23_1/nsf36/
• NSF Proposal & Award process: https://www.nsf.gov/awardsearch

Before submitting, keep your working notes focused on evidence and compliance, and avoid writing to fill pages. CAMEL rewards proposals that can be read as credible, data-responsible, and durable for K-12 learning ecosystems.