When the Australian Research Council publishes Discovery Project outcomes, research offices across the country stop doing almost everything else for a day. The 2026 results have now done their annual work: they tell some researchers their lab has a future, tell many others to start the rewrite, send professional staff into a scheduling scramble, and reignite a fight about what Australian research funding is actually for.
The most recent complete comparison is the 2025 round. In DP25, the ARC approved 421 projects from roughly 2,370 applications, a success rate near 18 per cent, and committed about $222 million over the life of the grants. The 2026 outcomes have the same structural shape: strong proposals outnumber available grants by a wide margin, and the word impact appears on losing applications as often as winning ones.
What changed in the conversation is not the scarcity. It's the growing tension between the ARC's National Interest Test and the longer, slower work of fundamental research. The 2026 list renews that argument every time a research office matches a rejected project against the benefit language that won.
The scheme behind the results
Discovery Projects are the ARC's main investigator-driven grants. A team led by an eligible Australian researcher can propose work in any discipline, from pure mathematics to archaeology. Projects usually run between three and five years, with budgets typically falling in a range from roughly $300,000 to $600,000. That range funds personnel, equipment, fieldwork and research costs; unlike many schemes, the money can support postdoctoral research associates, which makes the outcome list a de facto job forecast for early-career researchers.
The ARC accepts applications in an annual cycle and sends them to external expert reviewers in the relevant fields. Assessment panels then rank proposals against two criteria. One is the investigator's track record and the design of the project. The other is the National Interest Test statement, which must explain how the proposed research will deliver benefit to Australian communities beyond the academy. It is this second part that generates the loudest arguments.
Photo by Akeyodia - Business Coaching Firm on Unsplash
A stress test called the National Interest Test
The flashpoint that still shapes the debate arrived in 2021, when then acting education minister Stuart Robert vetoed six Discovery Project grants after they had passed peer review, all in the humanities and social sciences. The intervention outraged research leaders and produced a years-long argument about ministerial discretion. The ARC's National Interest Test page now codifies what was, in 2021, an emergency reading of public benefit: applicants are required to state simply what benefit the research will give Australia, and assessors are asked to weigh that benefit against the cost. The full Discovery Projects selection outcomes page lists funded projects by institution and discipline, and the institutional spread keeps the debate local.
Critics say the test is unworkable as a filter because benefits from basic research arrive years later and often through channels no one predicted. Supporters reply that a scheme spending public money should be able to say what the public gets back. The result is that every applicant now writes two proposals in one document: the scientific case for a specialist reviewer, and the public-benefit case for a generalist panel. The second case consumes an outsized share of preparation time in most research offices.
Impact metrics, from ERA to DORA and back
Australia has been trying to measure research benefit for a generation. The Excellence in Research for Australia (ERA) exercise rated university research quality by discipline; the Engagement and Impact assessment asked institutions to show non-academic returns. The Engagement and Impact process runs separately from Discovery Projects but shares the same philosophy: research should leave a trace outside the journal.
That philosophy comes with measurement problems. The ARC became a signatory to the San Francisco Declaration on Research Assessment (DORA), which pushes universities away from journal impact factors. A grant process can reject journal metrics and still demand evidence of impact, but the evidence itself is easier for applied fields than for cosmology, poetry or cell biology. An engineering team can point to a road trial; a philosopher may have a decade before her work changes a policy argument. Early-career researchers feel this gap most sharply because they have not yet accumulated the collaboration record that impact statements tend to reward.
In recent annual data, the Group of Eight universities have picked up roughly two-thirds of Discovery Project funding. Smaller institutions and regional universities are not absent from the lists, but the concentration remains stable year after year. The consequence for the impact debate is that the same institutions with the largest research offices also dominate the production of impact narratives, which then shape what the ARC expects from everyone.
Photo by Brett Jordan on Unsplash
What this means for your lab
Two years ago, a collaborator I'll call Dr R spent parts of six weeks reframing an impact statement for a project on enzyme structure. The experimental design did not change; the reframing was about pharmaceutical applications and Australian manufacturing. She won the grant, but she began logging every hour she spent converting basic findings into benefit language. By the end of the year the tally sat just under 160 hours. That is the same time as teaching a small postgraduate subject, and it was invisible to her department.
Multiply Dr R across the system and the impact assessment itself becomes a research cost. The base rate is already unforgiving: most field-based applicants face success rates between 15 and 20 per cent, and a borderline impact statement is rarely the only reason a grant fails. So the sensible response is not to treat the impact section as an afterthought, nor to let it swallow the project. Write it early, write it plainly, and ask one honest outsider to read it. If a high school science teacher can repeat your first sentence back to you, the panel can grade it quickly.
For researchers waiting on the 2026 outcome list, the practical filter is the same in labs and archives: the grants that fund postdoctoral and research assistant positions are already named, and the funded teams have begun advertising. Watch the outcome list, then watch the job boards.
The next round starts now
The ARC will open the next Discovery Projects round on its usual annual clock, and the best preparation is not a new reference manager. It is a one-page impact draft that names the public problem, explains the method with its limits, states the next realistic benefit, and says who would use it. If a draft cannot complete those jobs in a page, the project narrative may not be ready.
Before you open the application portal, print the National Interest Test page and keep it next to your draft. Mark every sentence that reads like a peer review paragraph and rewrite it in the voice you'd use to explain the work to a minister's adviser. Then submit the draft to a colleague in a different discipline. The grant you save may be your own.