Engineering Got $310 Million and the Only Three NSF Core Solicitations With No Proposal Cap At All. It Has Also Been Living Without Deadlines Since 2018 — and the Data on What That Does Is Not What You Think.

There is a quiet asymmetry inside NSF's August 17 funding release that nobody has priced correctly yet.

Twelve notices of funding opportunity went out that day, carrying more than $1.5 billion for foundational research and abolishing the fixed deadline across most of NSF's core research portfolio. That much has been covered — including here. What has not been covered is that the twelve solicitations do not treat researchers equally, and the gap between the most-restricted and least-restricted discipline is now enormous.

In chemistry, under NSF 26-519, you get two proposals a year — and being listed as senior personnel on a colleague's grant burns one. In materials research, under NSF 26-521, you get one proposal per fiscal year across all eight consolidated programs.

In engineering, you get as many as you can write.

The three solicitations

The Directorate for Engineering released three core NOFOs on August 17, and all three carry the same permissive terms:

That is $310 million against 695 to 850 awards — the largest directorate block in the release and, by award count, close to a third of everything NSF put on the table that day. Add the separately structured EFRI Wave-Based Computing solicitation at $30 million and engineering's August total reaches $340 million.

All three solicitations state plainly that there are no restrictions or limits on the number of proposals per organization or per PI/co-PI. All three prohibit voluntary committed cost sharing — you cannot buy goodwill with institutional funds. None require a letter of intent or a preliminary proposal. Unaffiliated individuals are not eligible; everything else in NSF's standard eligibility table is.

Underneath, each solicitation is a consolidation. CBET folds fourteen prior program solicitations into four clusters: Chemical Process Systems, Engineering Biological and Biomedical Systems, Energy Water and Resource Engineering, and Transport Phenomena. ECCS retires three long-standing program descriptions — PD 18-1517, PD 18-7564 and PD 18-7607 — into three programs: Electronic Photonic Magnetic and Quantum Devices; Circuits and Systems for Communications and Sensing; and Energy, Power, Control and Learning. CMMI runs four: Advanced Manufacturing Design and Operations, Engineering for the Built Environment, Intelligent and Interactive Dynamic Systems, and Mechanics of Advanced Materials Systems.

Engineering already did this, in 2018

Here is the part that changes how you should read the whole package.

The Engineering directorate did not adopt rolling submission on August 17, 2026. It adopted rolling submission on August 15, 2018, when a Dear Colleague Letter removed deadlines for unsolicited proposals to all core programs in CBET, CMMI, ECCS and EEC. The stated rationale then was the same one NSF used this month: give PIs flexibility, enable interdisciplinary collaboration, reduce institutional burden, let researchers "think more creatively, build strong collaborations, converse with Program Directors and carefully prepare proposals."

So engineering researchers have eight years of lived experience with the regime the rest of NSF just entered. The rest of NSF is catching up to engineering, not the other way around.

That matters because there is real data on what happens next — and it is not the flood everyone assumes.

The 59 percent

NSF has run this experiment before, deliberately, with measurement.

In 2015 the Division of Earth Sciences eliminated deadlines across four programs: geobiology and low-temperature geochemistry, geomorphology and land-use dynamics, hydrological sciences, and sedimentary geology and paleobiology. NSF announced the change in early 2015, paused submissions for three months, and switched to anytime submission in April.

Proposal volume did not rise. It collapsed. The four programs took 804 proposals in 2014 and 327 in the eleven months from April 2015 to March 2016 — a 59 percent drop.

It was not a fluke. A smaller instruments-and-facilities program in the same division had made the switch in 2011 and seen submissions fall by more than half, where they stayed.

The mechanism is unglamorous and entirely human. A deadline is a forcing function. It generates proposals that exist because the date arrived, not because the idea was ready — the half-formed resubmission, the defensive second shot, the "we may as well" from a lab with a spare week in October. Remove the date and those proposals never get written. What remains is the work someone chose to write on its own merits.

For anyone deciding where to spend the fall, the arithmetic follows directly. If engineering's core divisions behave the way Earth Sciences did, a submitted proposal in a no-deadline division is competing against a materially thinner field than the same proposal would have faced under a deadline. NSF's assistant director for geosciences framed the Earth Sciences result as improved odds for the proposals that do get submitted; some observers found the drop in applications troubling for exactly that reason — fewer ideas reaching review is not obviously a public good.

Either way, the strategic read is unambiguous: rolling submission rewards the people who keep submitting after everyone else stops.

Why engineering has no cap and chemistry does

The quota patchwork across the twelve NOFOs looks arbitrary until you line it up against directorate history.

MPS divisions imposed caps — two per individual per year in chemistry, one per PI per fiscal year in materials — because they are moving from deadlines to rolling submission for the first time, and they are bracing for a surge that, on the Earth Sciences evidence, may never arrive. The cap is insurance against a failure mode NSF has not actually observed.

ENG imposed no caps because ENG has eight years of post-2018 operating data and evidently does not need the insurance. Its program officers have already absorbed whatever behavior change rolling submission produces, and its divisions have already rebuilt review around continuous intake rather than panel season.

One structural detail from 2018 is worth carrying forward: the ENG deadline-removal letter paired anytime submission with a one-year moratorium on resubmitting declined proposals. Rolling submission does not mean instant reattempt. If your ECCS proposal is declined in November, the calendar constraint you thought you escaped reappears as a resubmission clock. Read the current solicitation text on this point before assuming you can turn a decline around in a quarter.

What $310 million against 850 awards actually buys

Divide it out. Across the three ENG core solicitations, the implied average award is roughly $365,000 to $445,000 total — spread across a three-to-five-year project, somewhere between $75,000 and $150,000 a year.

That is not a lab-sustaining grant. It is thread funding: a graduate student, part of a postdoc, some materials and travel. The same pattern holds across the entire August package, and it is the honest frame for what the $1.5 billion headline represents. NSF is buying breadth — a very large number of modest, foundational bets — not a smaller number of transformative ones.

CMMI's solicitation is unusually direct about what it wants in exchange. NSF states that proposals need not address near-term product, deployment, or commercial outcome to be competitive, while simultaneously aligning the division's portfolio with the Administration's FY 2028 R&D Budget Priorities around advanced manufacturing, autonomous systems, robotics and resilient infrastructure. Both signals are real, and they are not in tension: NSF wants foundational work in nationally prioritized areas, and it is telling you not to over-engineer a commercialization narrative you cannot support.

How to actually work a directorate that is not counting

Stop treating submission as an annual event. The single largest behavioral advantage available in a no-deadline division goes to the group that maintains a pipeline — one proposal in review, one in revision, one in drafting — rather than one that clears its desk each fall. With no cap and no deadline, throughput is a choice.

Submit when the queue is thin, not when the idea is old. Program officers in rolling divisions process continuously, but community submission behavior still clusters around residual habits: post-semester, post-holiday, post-conference. The weeks nobody submits are the weeks your proposal gets undivided attention.

Talk to the program officer before you write, not after you are declined. This advice is repeated to the point of cliché, and in a deadline regime it was frequently ignorable because the document did the work. In a rolling regime with four consolidated clusters per division, the single highest-leverage question — which of CBET's four clusters, or ECCS's three programs, will actually review this — is answerable in a fifteen-minute conversation and effectively unanswerable from the solicitation text alone.

Do not import an MPS mental model. If you are a chemist or materials scientist with an engineering-adjacent project, the constraint math is completely different on the two sides of the wall. A project that could plausibly land in CBET's Chemical Process Systems cluster costs you nothing from your ENG allowance, because there is no ENG allowance. The same project routed to MPS Chemistry spends half your annual quota. That is not a reason to force a bad fit — reviewers will see it — but it is a reason to have the conversation with both program officers before committing.

Ask about the resubmission moratorium explicitly. It is the one place where the "no constraints" story has historically had a constraint, and it is the difference between a decline costing you a quarter and costing you a year.

Budget for five years, not three. All three solicitations encourage proposals requesting up to five years of support for ambitious work, and all three state that budgets and durations must be commensurate with scope. In a portfolio where average award size is modest, duration is the variable that turns thread funding into something a student can actually graduate on.

The deadline that organized the academic year is gone across most of NSF. Engineering deleted its version of that deadline in 2018, and the field's experience since — plus the hard 59 percent from Earth Sciences — suggests that what replaces it is not chaos but a quieter, thinner, more continuous competition that rewards persistence over timing. Working out which of a dozen newly consolidated clusters actually fits a project, and what each one costs you to enter, is exactly the kind of triage Granted is built to shorten.