NSF CAREER Award 2026: The Complete Guide for Early-Career Faculty

Eligibility Requirements

CAREER eligibility is stricter than most NSF programs. You must meet all of the following:

• First untenured faculty appointment: You must be in a tenure-track or tenure-like position. Research scientists, lecturers, and adjunct faculty are not eligible. Visiting assistant professors who are not on a tenure track are typically ineligible.
• Timing window: You must submit your first CAREER proposal by the deadline in the 5th year of your first tenure-track appointment (some programs allow the 5th year). You remain eligible until you receive tenure or complete 10 years in a tenure-track position, whichever comes first.
• U.S. institution: Your employer must be a U.S. college, university, or two-year institution.
• No prior CAREER award: You cannot hold or have held a previous NSF CAREER award.
• No concurrent CAREER submission: You can only submit one CAREER proposal in any given year. You cannot have a CAREER proposal simultaneously under review at two different NSF divisions.

Important edge cases: Time on leave (parental, medical, military) may not count toward the 10-year limit. Contact your NSF program officer to confirm your eligibility window if you've had any career interruptions.

Award Amounts by Directorate

Note: These are direct cost ranges. Facilities and Administrative (F&A / indirect) costs are added on top, subject to your institution's negotiated rate.

Success Rates: What the Data Shows

NSF publishes aggregate success rate data but doesn't break it out by CAREER specifically in public-facing reports. Based on data from division program officers and university grants offices:

• Overall CAREER success rate: 20–25% (comparable to R01 at NIH, but higher for resubmissions)
• First submission success: approximately 15–18%
• Second submission (after reviewer feedback): approximately 30–35%
• Third+ submission: approximately 35–40%, but applies to a self-selected population

The key implication: most CAREER awardees submitted more than once. Treating the first submission as a "practice round" with built-in reviewer feedback is a legitimate strategy, provided you have enough time in your eligibility window.

Intellectual Merit: What Review Panels Look For

NSF review panels evaluate Intellectual Merit using five criteria:

• Significance and novelty: Does this advance knowledge in the field? Is the question important? Is the approach original?
• Conceptual and technical approach: Is the methodology sound and appropriate? Are the risks acknowledged and mitigated?
• PI qualifications: Is the PI demonstrably prepared to lead this work? (Publications, preliminary data, training)
• Resources: Does the PI have access to the facilities, equipment, and collaborators needed?
• Preliminary data: Is there evidence that the approach is feasible? CAREER proposals need enough preliminary data to establish credibility, but less than a mature R01.

A critical distinction: CAREER proposals are compared against other CAREER proposals, not against all NSF proposals. Early-career-appropriate preliminary data is expected — a CAREER reviewer who complains that you don't have 10 years of data is reviewing the wrong program.

Broader Impacts: The Differentiator

Broader Impacts is where most CAREER proposals fail. NSF explicitly states that Broader Impacts receive the same weight as Intellectual Merit in review panels. In practice, many scientists spend 90% of their writing effort on science and 10% on Broader Impacts — and it shows.

NSF defines Broader Impacts broadly: contributions to STEM education, broadening participation of underrepresented groups, building research infrastructure, advancing national security and economic competitiveness, and general societal benefits.

What strong Broader Impacts look like:

• Integration with research: Educational activities that directly use the research — students analyzing real research data, course modules built around the funded project, K-12 partnerships where your lab's work is the curriculum content.
• Measurable outcomes: "I will mentor 2 PhD students and 4 undergraduates annually, including targeted recruitment from [specific HBCUs or community colleges]" is better than "I will mentor diverse students."
• Existing partnerships: Broader Impacts activities with existing community partners, schools, or organizations are more credible than new relationships you promise to establish after funding.

What weak Broader Impacts look like: "I will publish papers in top journals and present at conferences." This is what everyone does. It is not a Broader Impact — it is the minimum expectation for any funded researcher.

What NSF Won't Tell You About CAREER Awards

1. Program officer relationships matter before submission, not just after funding. NSF program officers have latitude in which proposals they advocate for at the panel stage. Reaching out before submission — to confirm fit, ask about panel trends, and discuss your proposed approach — creates awareness and goodwill. Do this 3–6 months before the deadline, not 3 weeks before.

2. The review panel composition varies year to year and significantly affects scores. A computer science proposal reviewed by a panel with two experts in your exact subfield will score differently than the same proposal reviewed by a panel with no such experts. You cannot control this, but you can write for a broader audience rather than assuming deep domain knowledge in reviewers.

3. NSF CAREER panels sometimes deliberately fund across subfields to maintain program diversity. If 80% of submissions in a panel are in one hot subfield, the panel may fund proportionally less from that subfield than the raw scores would suggest, to ensure CAREER awards go to a range of research areas. Niche but important subfields sometimes have structurally higher funding rates.

4. The "Fascination of Plants Day" problem: Many PIs include stock K-12 outreach activities (visiting schools, science fairs, public talks) as Broader Impacts without any institutional infrastructure. Reviewers are increasingly skeptical of these without evidence of existing school partnerships, established contacts, or a track record of educational engagement. If you don't currently do this, don't propose it as your main Broader Impact.

2026 Deadlines by Directorate

CAREER deadlines are annual and directorate-specific. The typical schedule for 2026 proposals:

Always verify current deadlines at nsf.gov/pubs/2021/nsf21553 (CAREER solicitation). Deadlines occasionally shift.

Resubmission Strategy

CAREER proposals can be resubmitted without limit within your eligibility window. Unlike NIH (which allows only one resubmission), NSF allows as many rounds as needed.

After receiving reviewer comments (summary statement), the standard resubmission process:

• Analyze the summary statement: Understand which critiques are substantive (need significant revision) vs. stylistic (easy to fix). Most summary statements have 2–3 core concerns that, if addressed, would make the proposal competitive.
• Call the program officer: Within 2 weeks of receiving your summary statement. Ask whether the score is competitive for the next cycle, which concerns were most impactful, and whether the panel was appropriate for your work.
• Generate additional preliminary data: If reviewers questioned feasibility, you have 12 months before the next deadline to generate data that directly addresses those concerns. This is often the single most impactful thing you can do.
• Revise, don't just respond: Resubmissions are not rebuttals. Don't write the same proposal with added defensive text. Actually restructure the affected sections to preemptively address the concerns without explicitly saying "reviewer X said..."

5 Common NSF CAREER Mistakes

Frequently Asked Questions

Last updated April 2026. NSF CAREER award amounts, deadlines, and eligibility requirements are subject to change with each annual solicitation. Verify current requirements at nsf.gov.