Expert advice for every grant — from EPA to DARPA

NSF 26-522 consolidates AAG, ATI and MSIP into a single astronomical sciences solicitation with $100 million and about 140 awards. Accomplishment-based renewals will be returned without review, voluntary cost sharing is prohibited, and routine instrument upgrades are explicitly deprioritized. The new Innovations track — one to three awards a year — is where facility operations and community observing time now live.

NSF 26-514, 26-515 and 26-518 put $310 million and up to 850 awards on rolling submission with zero per-PI limits, while chemistry rations two proposals a year and materials rations one. The Engineering directorate ran this experiment eight years ago. When NSF tried it in Earth Sciences, proposal volume fell 59 percent. Here is what that history predicts, and how to use a directorate that is not counting your submissions.

The OpenAI Foundation is putting $100 million behind Breakthroughs to Follow-Through, a Common Health Coalition initiative aiming to at least double hepatitis C cure rates in eight states and localities within two years. Alabama, Illinois, Louisiana and Massachusetts go first, money flows to local collaboratives and nonprofits, and the grants are deliberately not tied to any one AI vendor. Here is what makes an organization fundable under it.

DPA26BZ05-DV019 asks small businesses to reduce ISR video transmission from small drones by 90 to 99 percent while preserving what an operator actually needs, on a Jetson-class board at a 2-watt final power budget. It is Direct-to-Phase-II, up to $1.5 million base plus a $500,000 option, and its exclusion list — no weapon release, no lethal engagement, no named-person identification — is the most strategically revealing paragraph in DARPA's FY26 Release 5.

DPA26BZ05-DV018 is the largest award in DARPA's FY26 SBIR Release 5 — up to $3 million over 18 months, Direct-to-Phase-II only. It funds four subsystem tracks for an underwater 3D concrete printer that uses seafloor sediment and seawater with under 20 percent binder. Here is what the Trenton program already proved, why the track structure is really a supplier-qualification exercise, and who can clear the prior-feasibility bar in the four weeks that remain.

The Hydrocarbons and Geothermal Energy Office's University Training and Research program funds coal, oil and gas, and geothermal R&D at U.S. colleges and universities — but every proposal must include a non-academic partner and must build training modules that outlive the award. The LOI deadline is October 1, 2026, with full applications 15 days later. Here is what that compressed window means and why the workforce framing changes what a competitive proposal looks like.

The FY2026 MEP Center State Competition ran in two rounds: eight states worth $139.1 million, then fourteen more worth $232.4 million, both closing August 21, 2026. That's 43 percent of the national network recompeted in twelve months, under a 50 percent non-federal cost share, while the administration's own budget request proposed eliminating the program. Here's what the award tables reveal, why incumbents are genuinely at risk, and how to position for the FY2027 wave.

NSF 26-517 replaces the core solicitations of Environmental Biology, Integrative Organismal Systems and Molecular and Cellular Biosciences with a single directorate-wide opportunity. Proposals are accepted anytime, sorted centrally by NSF rather than targeted by the PI, and reorganized under three inquiry areas. Here is what the $280 million actually buys, why the loss of program targeting cuts both ways, and what to do before the September 16 webinar.

NSF 26-519 puts $150 million and 300 chemistry awards on a rolling submission schedule, then caps each individual at two proposals per year counting PI, co-PI and senior/key personnel roles. Materials caps at one. Physics, math, astronomy, geosciences and biology cap at none. Here is the full quota map across all seven new core solicitations, why the senior-personnel clause is the trap, and how to allocate a quota you did not know you had.

NSF 26-524 commits $30 million over FY2027 and FY2028 to wave-based computing — up to $2 million for four-year multidisciplinary awards, $100,000 to $400,000 for two-year exploratory ones. Eleven of the twelve August NOFOs consolidated existing programs. This one funds something NSF has never funded. Here are the three required thrusts, the tenure-track restriction, the one-proposal-per-person rule, and the spin-wave exclusion that disqualifies an entire subfield.

NSF 26-501 (CAMEL) commits $9 million to build AI-ready K-12 mathematics datasets across 6–7 awards of up to $1.5 million each. The money is co-supported by the Walton Family Foundation — and the solicitation bars WFF staff and partners from participating in any capacity, prohibits voluntary cost sharing, and requires reports be shared with the foundation after NSF review. Here's how the public-private structure works, and what Phase II means for teams already in.

The Pax Silica Artificial Intelligence Assistance Project (DFOP0019182) offered awards from $1 million to $50 million to build a cryptographic credentialing and provenance system for semiconductors, AI infrastructure, and critical minerals moving through Panama's ports — with a second phase expanding to the rest of the Pax Silica bloc. For-profits were eligible as primes. Here's what the NOFO signals about where foreign assistance money is going.

Two weeks after the DARPA Lift Challenge ended with the winner at 3.84:1 — short of the 4:1 target — DARPA posted a Direct-to-Phase-II SBIR that requires empirical proof of a 4:1 payload-to-weight ratio to bid. DPA26BZ05-DV022 offers up to $1.5 million over 24 months across two tracks that map exactly onto the Challenge's two design prizes. Here is who can actually clear the entry bar, what the month-9, month-18, and month-24 gates demand, and why the certification language is the real story.

DARPA's DIAL program already had AI rediscover the Kalman filter and wavelets from scratch. SPEED DIAL (DPA26TZ05-DV003) funds the unglamorous next step: turning algorithm discovery into a tool that plugs into MATLAB, Simulink, and COMSOL. Up to $750,000 base plus a $1.25 million option, Direct to Phase II only, with a three-accomplishment feasibility gate and a mandatory university partner. Here is the milestone schedule, the interpretability requirement, and who is actually eligible.

The Fall 2026 EDGE Grant Competition closes at 4 p.m. on September 11, with $750,000 for STEM and $400,000 for Entrepreneur applicants, awards up to $100,000, and a 3-to-1 match. The two tracks are not equally competitive, and the asset test disqualifies more applicants than the revenue test ever will.

HRSA-27-017 closes September 29, 2026 and HRSA-27-084 closes November 16, with $20,960,000 and $16,000,000 respectively. The award floors and ceilings are not budget guesses — they are resident headcounts at the $160,000 interim per-resident rate. Read them that way and the eligibility question answers itself.

The NIH Extramural Loan Repayment Program pays up to $50,000 a year against educational debt at roughly a 50% success rate — better odds than almost any NIH research grant. For FY2027 the six extramural subcategories collapse into three: L30, L40, and L70. Applicants who would have used L32, L50, or L60 must now pick a new lane. Here is the crosswalk, the 20% debt-to-salary rule, and why the consolidation changes how you frame the research.

DRL's Global Rapid Response Fund, DFOP0019705, closes August 31, 2026 with a single award of $9,569,050 and subgrants of $10,000 to $100,000. Only 40 percent of the award can flow through as subgrants, which means the majority of a rapid-response fund is not rapid-response money at all. Here is how to read that, and who is actually built to win it.

HHS launched STREAM on August 17 through ARPA-H — four technical areas spanning AI-designed herbicides, laser weeders, low-cost chemical sensing, and soil remediation. It runs inside an existing Mission Office ISO, which changes who can apply and how fast the clock moves after you submit.

IDNR's OSLAD program closes at 5:00 p.m. on August 31, 2026, with $37.5 million available, a $1,125,000 ceiling for land acquisition and $600,000 for development. The standard deal is 50/50. The deal that matters for small towns is the full-funding provision for economically distressed communities — and the thing that disqualifies applicants is not the project, it is GATA prequalification.

Five NHLBI Catalyze Product Definition RFAs — HL-26-016 through HL-26-020 — carry an October 21, 2026 receipt date, budget caps of $300,000 to $400,000 per year, and roughly 8 new awards annually. The science is not the gate. An Accelerator Partner and a 0.25:1 non-federal cost match are, and both take longer to line up than the nine weeks you have left.

The 2026 NYS HOME RFA closes at 4:00 p.m. on September 17. Its 100-point rubric gives 45 points to organizational capacity, 11 in bonus points to rural applicants in underserved areas, and subtracts up to 20 points for slow spend-down on contracts you already hold. Here is how the math actually works.

The Senate passed a continuing resolution 90-6 on August 8 with Section 157, language barring OMB from finalizing its 2 CFR Part 200 rewrite through December 11, 2026. The House CR, passed July 21, contains no such provision and expires a week earlier. Here is what the rider actually blocks, what it conspicuously does not, and how grantees should use a delay that may only last 16 weeks.

The Rural Utilities Service is putting $410 million of Section 22001 budget authority back in the market through the Powering Affordable Reliable Technology program. Letters of Interest open September 8 and close October 9, 2026 — both at 11:59 a.m., not p.m. — and RUS evaluates them on a rolling basis in the order received.

The Energy Department's $65.5 million oil and gas production and delivery NOFO splits across four topic areas with wildly different odds. Topic Area 3 alone carries $28 million for Hydrocarbon Infrastructure Test Sites at roughly $7 million per award over five years — and it requires a defined physical site you already control. Here is how the money is actually allocated.

The Energy Department's first Genesis Mission SBIR/STTR Phase I solicitation offers roughly 40 awards of up to $250,000 against $10 million — but you cannot submit a full application anymore. A four-question, 700-word pitch now decides who gets to apply, there is no Direct to Phase II path for new entrants, and the $147.4 million Phase II pool is closed to everyone who has never won a DOE award.

NOT-OD-26-098 bars K award funds from supporting clinical trials, feasibility studies, and ancillary studies for due dates on or after October 12, 2027, and retires the Independent Clinical Trial Required K NOFOs. NIH has no replacement mechanism yet — only an RFI closing September 30, 2026. Here is what the policy actually says, who it hits, and how to sequence around it.

POSE became PESOSE — Pathways to Enable Secure Open-Source Ecosystems — and the added word is not cosmetic. NSF 26-506 brings a third track funding vulnerability remediation in existing ecosystems, keeps the $300,000 planning and $1.5 million build tracks, anticipates 40 to 60 awards against $40 million, and closes September 1, 2026. The I-Corps requirement and the PI eligibility rule are where most applicants will trip.

NSF 26-525 keeps the $15 million, seven-year headline that makes Expeditions the largest single investment CISE makes. What changed is underneath: funding is now split into three explicit phases, Years 6 and 7 are contingent on a Year 5 site visit, the Innovation Transition supplement is gone, and the competition calendar moved from June to December — pushing awards out past 2028.

On August 14, 2026, the California Arts Council approved nearly $24.6 million across 1,200 grant awards. Buried in the release is a rare thing: before-and-after data showing exactly how much the ranked-priority bonus moved the pool. Small organizations went from 53% of applicants to 72% of awardees. Geography moved 19 points. First-time status barely moved at all. Here is what that means for the 2027 cycle.

The Council on Foundations opened a Working Across Differences Accelerator for community foundations — a $20,000 planning grant, up to $45,000 in follow-on funding, and a November 2026 cohort start, backed by Mott and the Walmart Foundation. The dollar figures are small and the strategic value is not. Here is how to read a two-stage cohort program, and why the vocabulary this fund chose is itself the lesson.

HHS announced $102 million in New Access Points awards on August 13, 2026 — 158 health centers, more than 415 new sites, nearly 1 million more patients. That works out to roughly $246,000 per site, which does not build a clinic. Here is what the award actually buys, why the real selection happened before applications were scored, and what unfunded applicants should do next.

The Innovative Postsecondary Models Fund — Gates, Lumina, Strada, Ascendium, ECMC, College Futures, Haas Jr., and Raikes, administered by Panorama Global — will make 7 to 10 grants of $750,000 to $1,000,000. Eligibility is unusually wide and the gates are unusually specific. One of them eliminates most applicants outright, and it has nothing to do with the quality of the idea.

NIH proposed on August 14 to withhold overall impact scores and percentiles from applicants, replacing them with three bands: Most competitive, Competitive, Not discussed. Paylines already vanished in January. Here is what applicants actually lose, what survives in the summary statement, and how to rebuild a resubmission decision rule without a number.

The Cold Chain Grants for Emergency Food Assistance Program closes October 1, 2026, funding cold storage equipment at up to $200,000 per project with a 10% cash match. But USDA is not awarding to food banks — it is awarding to nonprofit intermediaries who will run competitive subaward programs. Two audiences, two entirely different jobs, and one of them has 43 days.

The FY 2026 Coordinated Tribal Assistance Solicitation lets federally recognized Tribes apply once for grant programs spread across BJA, COPS, OJJDP, OVC, and OVW — more than $107 million in a single competition. Purpose Area 7 is gone this cycle, the two-step Grants.gov and JustGrants deadlines are a week apart, and the applicants who lose usually lose on the structure of the request, not the need.