NIA's FY 2026 Alzheimer's Research Apparatus Is Being Rebuilt Around AI Infrastructure and Single-Source Awards. What Investigators Need to Read Differently.

The National Institute on Aging spent the better part of 2025 reorganizing how it funds the data, biospecimen, and computational infrastructure that powers Alzheimer's and Alzheimer's-disease-related dementias research. The result is a portfolio of FY 2026 funding actions that look, on the surface, like a small set of routine cooperative-agreement renewals. Read more carefully, they describe a deliberate consolidation of the AD/ADRD research apparatus around a handful of high-value central nodes — and an explicit AI/data layer that did not exist as a coordinated funding mechanism eighteen months ago.

The Consolidated Appropriations Act, 2026 inserted a $100 million increase for Alzheimer's and dementia research at NIH, plus $41.5 million to implement BOLD public health infrastructure for dementia. The NIA Professional Judgment Budget for FY 2026 requested an additional $113 million in new resources, against a total resources estimate of $3.98 billion. The headline number is the appropriations bump. The operationally important number is how the institute is deploying it.

For investigators reading the FY 2026 NIA portfolio for the first time, three things have to be true simultaneously: there is more total money than there was in FY 2025, there are fewer competitive new awards in the highest-visibility infrastructure categories, and the AI/data layer of the research enterprise has been moved into its own coordinating-center structure. Each of those conditions changes the application calculus.

Three Single-Source and Limited-Competition Infrastructure Awards

The most consequential pieces of FY 2026 AD/ADRD funding are three RFAs that fund the central data and biospecimen infrastructure for the entire NIA portfolio. They are not open competitions in the conventional sense, but their structure is informative about where NIA is concentrating capacity.

RFA-AG-26-002 renews the National Centralized Repository for Alzheimer's Disease and Related Dementias (NCRAD), the biospecimen repository that currently supports 74 NIA-funded studies. NIA intends to commit $14.77 million in FY 2026 to fund one award. The mechanism is single-source — the cooperative agreement is being directed to Indiana University Indianapolis, which has held the NCRAD function since the repository's establishment, on the basis that no other institution has the equivalent infrastructure, sample chain-of-custody history, or integration with the existing study portfolio. For an individual investigator, NCRAD is not a funding target. It is the layer that determines which biospecimens are available to extramural studies and on what data-sharing terms — which is exactly the kind of infrastructural detail that has historically been invisible to first-time R01 applicants and that materially shapes which study designs are feasible.

RFA-AG-26-003 renews the National Alzheimer's Coordinating Center (NACC), the central data hub for the Alzheimer's Disease Research Centers (ADRCs). NACC consolidates clinical, neuroimaging, biomarker, and genetic data from the ADRC network into a single, harmonized resource that extramural investigators can request access to. The NACC renewal is the gravitational center of the AD/ADRD data ecosystem. Investigators planning secondary-analysis studies or pilot studies leveraging existing cohorts have to assume that the post-renewal NACC data structure, governance model, and access policies will be the de facto standard for the next five-plus years.

RFA-AG-26-007 is the new piece — the Artificial Intelligence and Technology Collaboratory Coordinating Center. NIA established a network of AI and Technology Collaboratories in 2022 to develop, validate, and deploy AI methods for dementia care and research. RFA-AG-26-007 funds the central coordinating center for that network — the entity that harmonizes methods, manages cross-site data sharing, runs the AI ethics and assurance functions, and coordinates the transition from research-grade models to deployable clinical tools. The coordinating center is the policy and governance layer for how NIA-funded AI work in dementia gets translated into care delivery. Its grantee will, in practical terms, set the terms on which extramural AI researchers gain access to the Collaboratory network's data and infrastructure.

The three RFAs together describe a vision of the AD/ADRD research enterprise as a structured stack: biospecimens (NCRAD), clinical and observational data (NACC), and AI/computational methods (the Collaboratory CC). Each layer is funded through a large, central, limited-competition cooperative agreement. The extramural investigator's job is to design studies that plug into the stack productively, not to compete for the infrastructure dollars themselves.

Where the New-Research Increment Actually Goes

The $113 million increment that NIA's professional judgment budget identified for new FY 2026 research does not flow into the infrastructure RFAs. It flows through the standing parent announcements that NIA has consolidated AD/ADRD research into over the last two cycles. PAR-25-332 is the R01 parent announcement for current topics in AD/ADRD. PAR-25-331 is the R21 parent announcement for exploratory and developmental work. The R03 parent and the K-series career awards round out the standing portfolio.

The strategic implication for investigators is that the parent-announcement pathway is where the new-research dollars are being allocated. There is no topic-specific RFA absorbing the FY 2026 increment in the way that, ten years ago, a dedicated NOFO would have absorbed a budget increase tied to a specific disease subtype, methods area, or population. The investigator-initiated pathway is doing the work that topic-specific NOFOs used to do.

This is consistent with the broader pattern across NIH that has unfolded over the last twelve months — the consolidation of the agency's NOFO catalog from more than 800 to fewer than 500 and the explicit push toward parent announcements as the primary submission vehicle for investigator-initiated work. NIA is one of the institutes most heavily affected by that shift, partly because the volume of AD/ADRD-specific NOFOs grew faster than other portfolio areas during the bipartisan funding ramp-up of the late 2010s and early 2020s. The current consolidation is, in part, a return to the baseline.

For investigators, the operational change is that the visible programmatic signals about where NIA wants to spend new-research dollars no longer come from a thick stack of topic-specific RFAs. They come from a smaller set of sources: the NIA director's blog, the National Advisory Council on Aging summaries, program-officer pre-submission consultations, and the institute's strategic plan refresh cycle. Investigators who built their submission strategy around scanning the NIH Guide for AD/ADRD RFAs are operating with an increasingly thin signal layer.

The Pipeline Investments

RFA-AG-26-011, the Expanding Research in AD/ADRD (ERA) Summer Research Education Program (R25), funds institutions to provide structured summer research experiences to early-stage investigators new to the AD/ADRD field. The R25 mechanism is small in dollar terms but strategically important — it is one of the few mechanisms NIA is using to backfill the pipeline of investigators entering the field at a time when the existing AD/ADRD workforce is aging out and when adjacent fields (basic neuroscience, geriatrics, computational biology) have their own recruitment pressures.

For institutions, the R25 mechanism is a credentialing opportunity. An institution that runs an R25-funded summer program in AD/ADRD research builds a defensible track record of training contribution to the field, which strengthens subsequent applications for T32 institutional training grants, P30 Alzheimer's Disease Research Centers, and Career Development awards. The R25 is the cheapest entry point into the AD/ADRD funding ecosystem at the institutional level.

For individual investigators, the R25 is the cleanest signal about where NIA is looking for new investigator capacity. The topic areas that the ERA program prioritizes — health disparities in dementia, computational methods, social determinants, real-world data approaches — are the same topic areas that the parent-announcement pathway is most receptive to for early-stage R01 and R21 applications. The R25 priorities function as a programmatic signal in lieu of the topic-specific RFA signals that NIA used to publish.

How to Read the FY 2026 Portfolio as an Applicant

The strategic posture for investigators submitting to NIA in FY 2026 has three components. First, infrastructure is not your funding target. The $14.77M to NCRAD, the comparable amount to NACC, and the new Collaboratory CC award are infrastructure investments that you use, not investments you compete for. Designing studies that plug into the post-renewal versions of these systems — including the harmonized data structures and access policies they will publish — is the right place to invest planning time.

Second, new-research dollars come through the parent-announcement pathway. A FY 2026 R01 submission to NIA almost certainly goes through PAR-25-332, not through a topic-specific RFA. That changes how the application is positioned: it must compete on scientific merit against the full investigator-initiated pool at NIA, not against a curated topic-specific cohort. The implication for review preparation is that the application should be written for a study section that may not have deep AD/ADRD expertise — which means more attention to mechanism, methods rigor, and broad significance than would be required for a topic-specific NOFO review panel. Investigators who have not used the parent-announcement pathway recently should also revisit program officer outreach as a pre-submission discipline — the program-officer signal is doing more of the work that the NOFO catalog used to do.

Third, the AI/data layer is where the field is going. The Collaboratory Coordinating Center exists because NIA expects AI methods to become a core research modality in AD/ADRD, not a peripheral specialty. Investigators with computational backgrounds entering the field through the parent-announcement pathway in 2026 are arriving at the moment when the institute has explicitly funded the infrastructure to support that work. The applications most likely to score competitively over the next two cycles are the ones that design AI methods around the data structures that NACC and the Collaboratory CC will support — not the applications that propose to build their own parallel data infrastructure.

The headline NIA budget number for FY 2026 will be repeated in advocacy materials and in congressional appropriations summaries. The structural picture underneath that number — the three-layer infrastructure stack, the parent-announcement-dominated new-research portfolio, the R25 pipeline mechanism — is what determines who gets funded. Investigators who learn to read the layers individually will find the portfolio more navigable than those who treat the institute as a single monolithic funder.