NSF Just Bet $1.5 Billion on a New Funding Model. The X-Labs Initiative Could Reshape How Federal Research Money Flows.

On May 14, 2026, the National Science Foundation announced an initiative that quietly upends decades of how the agency funds science. The headline number is $1.5 billion over the next decade. The real story is the funding mechanism: NSF X-Labs will award research dollars not through the familiar 15-page proposal, three-reviewer panel, and grants.gov pipeline, but through Other Transactions Authority with milestone-based payments to independent teams pursuing specific technical capabilities.

If you have spent a career writing NSF proposals, the X-Labs RFP will feel less like a grant solicitation and more like a defense industrial base solicitation. That is not an accident.

Tracking the rapidly shifting landscape of federal research funding? Browse our Federal Research Grants Hub for a continuously updated view of NSF, NIH, DOE, and DARPA opportunities — filtered by topic, deadline, and award size.

What "X-Labs" Actually Is

The official framing from NSF describes X-Labs as "independent teams of researchers, engineers and entrepreneurs pursuing milestone-based federal funding to solve specific scientific challenges." Strip away the press-release language and three structural features matter.

First, the team composition is mandated. NSF wants academic researchers, applied engineers, and entrepreneurs in the same team — not a PI with a few co-PIs from adjacent departments. This is closer to a DARPA performer team or an ARPA-E translational consortium than to a standard NSF Directorate for Mathematical and Physical Sciences award.

Second, the funding model is milestone-based. Money flows when the team hits a pre-negotiated technical checkpoint, not in annual installments tied to a fiscal calendar. Teams that miss milestones face renegotiation or termination. Teams that hit milestones early can pull future tranches forward. This is closer to a Small Business Innovation Research Phase II contract or a DARPA Strategic Technology Office award than to a grant.

Third, the legal vehicle is Other Transactions Agreements (OTAs), not grants or cooperative agreements. OTAs are a flexible contracting authority Congress granted to certain agencies for prototype, research, and production work. NSF received OTA authority in the CHIPS and Science Act of 2022, and X-Labs is one of the most visible deployments of that authority to date. Practically, OTAs let NSF skip much of the Uniform Guidance (2 CFR 200) compliance machinery, negotiate intellectual property terms case by case, and bring in non-traditional performers — including for-profit firms that have never engaged with NSF — without forcing them through the federal grants compliance gauntlet.

The combination matters. NSF is signaling that for "generational breakthrough science," it will pay for technical outcomes rather than research effort. That is a meaningful philosophical shift for an agency whose flagship Directorate for Engineering and Major Research Instrumentation programs reward proposed methodology and investigator credentials more than committed deliverables.

The First Two Topics

The May 14 announcement opens the first round with two technical topics. NSF has said additional topics will follow in the coming weeks, but these two are the proving ground.

Scientific Instrumentation for Sensing and Imaging

NSF is seeking X-Labs to build "the next generation of scientific instruments, drawing on quantum sensing, artificial intelligence-driven computational imaging and entirely new chemical modalities." The framing is broad on purpose. Reviewers will look for proposals that combine at least two of those three pillars — quantum sensing physics, AI-driven reconstruction or computational optics, and novel chemical contrast agents or sample preparation — in service of a specific instrument capability.

Teams positioned to win in this topic will typically include a physics or optics group with quantum sensor experience (nitrogen-vacancy diamond, atomic magnetometry, single-photon detection, or similar), a computational imaging or machine learning group with hands-on experience in inverse problems, and either a chemistry group developing molecular probes or an engineering group capable of integrating components into a working instrument. The entrepreneurial leg of the team needs to articulate a credible commercialization pathway — which instrument market, which customer, which alternative the X-Labs instrument outperforms, and on what metric.

Quantum Systems: Interconnects and Integrated Photonics

The second topic targets "novel components to transfer quantum information and integrate heterogeneous quantum systems." This is the photonic interconnect and quantum networking layer — the pieces that have to exist before quantum computers, quantum sensors, and classical hardware can talk to each other coherently. Practically, that means efficient single-photon sources, low-loss frequency conversion, fiber-coupled quantum memories, and the integrated photonic chips that host all of the above.

This topic sits squarely between the National Quantum Initiative Act and the CHIPS and Science Act priorities. Teams that have worked on the National Quantum Information Science Research Centers funded by DOE, the NSF Quantum Leap Challenge Institutes, or the DARPA Quantum Benchmarking program are likely to recognize the technical vocabulary. The X-Labs angle is that NSF wants prototypes of components, not theoretical analyses or single-device demonstrations — and the milestone payments will reflect that.

Who Should Actually Compete

The trap with a new mechanism is that everyone assumes their lab fits. Most do not. Three filters help.

Filter one — can you assemble a team in 60 days? The required composition (researchers + engineers + entrepreneurs) is non-negotiable. If your institution does not already have a tech transfer office that can quickly connect you with a startup co-founder, or if you have not maintained an industry network of engineering firms or potential customers, the timeline will eat you alive. The teams that win will be the ones that walked in pre-assembled, often having worked together on prior SBIR Phase II, ARPA-E OPEN, or DARPA seedlings.

Filter two — do you have committed milestones? A good X-Labs proposal lists what the team will deliver, when, and how the deliverable will be measured. "We will publish a paper on the topology of single-photon detector arrays" is not a milestone. "We will demonstrate a 16-channel superconducting nanowire single-photon detector array with 90% detection efficiency at 1550 nm and dark count rate under 100 Hz per channel, by month 18" is a milestone. If you are uncomfortable putting numbers on outcomes 24-36 months out, X-Labs is not for you.

Filter three — are you willing to negotiate IP terms? Under OTA authority, intellectual property allocation between the team, NSF, and downstream commercialization partners is negotiated rather than dictated by Bayh-Dole. That flexibility is a feature for entrepreneurial teams who want to retain rights to spin out a company. It is a complication for university tech transfer offices that have decades of muscle memory built around the standard federal IP framework. Get your TTO in the room early.

How X-Labs Fits the Broader Funding Landscape

This announcement does not exist in a vacuum. NSF's traditional core programs — Faculty Early Career Development, individual investigator awards across each directorate, the Major Research Instrumentation program — remain. The X-Labs $1.5 billion is additive, not a reallocation away from the curiosity-driven research base. But the directional signal is unmistakable: NSF is building the institutional muscle to fund applied translation alongside fundamental science.

The closest sibling program in the federal portfolio is the DOE Advanced Research Projects Agency-Energy (ARPA-E), which has used cooperative agreements with aggressive milestone structures since 2009. The closest model in DOD is the DARPA Strategic Technology Office. X-Labs effectively pulls NSF — historically the home of basic research — closer to those translational-research peers, while preserving the agency's identity as a basic science funder through its core directorates.

For researchers who have spent careers in the standard NSF lane, the strategic question is whether to learn the X-Labs mechanism or stick with what works. The honest answer for most PIs is to stay in the lane that fits your research. If your work is genuinely curiosity-driven, write a CAREER or a Faculty Early Career Development Program proposal. If your work has clear milestone-able deliverables, a committed team that crosses academia and industry, and a defensible commercialization path, X-Labs is where the federal money is moving.

The Companion News Brief

Granted's news team covered the headline announcement in our news brief on NSF X-Labs. This deep dive expands on the mechanism, the eligibility filters, and how the program fits into the broader landscape of milestone-based federal research funding. For solicitation language, deadlines, and the X-Labs webinar registration, check the official NSF X-Labs page directly, since the per-topic instructions will be updated as additional topics are released over the coming weeks.

What to Do This Week

If the two opening topics fit your work, three concrete actions before the first formal deadline.

Audit your team gaps. Map your current collaborators against the researcher-engineer-entrepreneur triad. Identify which leg is missing and start the conversation now. Founders of seed-stage hardware startups, fractional CTOs of deep tech companies, and engineering directors at instrumentation firms are typical fits.

Build a milestone table. Sketch your proposed deliverables at 6, 12, 18, 24, and 36 months. If you cannot commit to specific quantitative targets at each checkpoint, the proposal will not be competitive — and you will discover this gap faster on a whiteboard than in red-team review.

Pre-clear the IP conversation with your institution. University tech transfer offices need lead time to negotiate non-standard IP terms. If you wait until the proposal is drafted, you will discover institutional friction that delays or kills the submission. A 30-minute meeting now is worth a month later.

NSF has put $1.5 billion on the table for a new model of federal research funding. The teams that internalize what milestone-based, OTA-funded research actually demands — and assemble accordingly — will be writing the playbook the next generation of NSF programs use. Everyone else will be writing standard proposals to a smaller share of the same pie.