America’s innovative spirit has been a foundation of its prosperity for generations. Research and development drives productivity, new industry formation, national security, and improved health and living standards. Leadership in AI, quantum computing, robotics, biotech, advanced materials, and other frontier fields will be pivotal to maintaining our economic and strategic power for decades to come.

The United States begins with enormous advantages. Its research enterprise spans universities, federal laboratories, private companies, medical centers, and independent research organizations. U.S. R&D investment totaled around $993 billion in 2024¹—roughly three-quarters of it funded by business, with the federal government supplying 19 percent² and playing an outsized role in basic science, nearly half of which is performed by universities.³

This structure has created an unusually productive system. Federal support for investigator-driven research empowers scientists to pursue discoveries whose importance cannot be predicted in advance. Independent peer review and competition among researchers facilitates the exploration of thousands of ideas simultaneously, with common standards of rigor.

The U.S. also has a tradition of mobilizing public resources behind national objectives. The Manhattan Project, the Apollo program, the Human Genome Project, and Operation Warp Speed all demonstrated the value of aligning government, science, and industry. DARPA institutionalized a related model through empowered program managers, ambitious technical objectives, support for competing approaches, and finite programs focused on measurable milestones. ARPA-E and ARPA-H have adapted elements of that model to energy and health.

That leadership is now under pressure. The U.S. share of global R&D spending fell from 39 percent in 2000 to 29 percent in 2024, while China’s rose from 5 percent to 30 percent, narrowly surpassing the U.S. on a purchasing-power-adjusted basis.⁴ The U.S. faces increasingly capable competitors in AI, quantum science, and other fields. Capital-intensive frontier technologies also expose vulnerabilities in supply chains and in America’s capacity to turn discoveries into products at scale.

At the same time, America’s research system needs modernizing. Principal investigators now spend an estimated 44 percent of their federally funded research time on administrative tasks, and researchers collectively devote more than 10 million hours each year to grant applications.⁵ Onerous requirements have proliferated across agencies, often with inconsistent definitions and processes. Researchers receiving support from multiple agencies can face overlapping, inconsistent demands relating to security, disclosures, data management, reporting, and compliance.

Federal grantmaking methods also deserve scrutiny. Traditional peer review offers an unparalleled quality-control mechanism, but consensus-based review can favor proposals with a predictable path to success over novel ones. Some emerging scientific problems require a more comprehensive approach than grant-based financing allows—with multidisciplinary teams, expensive shared equipment, or long-term, sustained work. And until recently, federal research agencies devoted little systematic effort to testing which grantmaking and review methods yield the strongest scientific results.⁵

Frontier AI research presents a further constraint: it requires large amounts of specialized computing capacity, high-quality datasets, and the ability to test computational discoveries in physical laboratories. Similar bottlenecks of scale and cost arise in biotechnology, quantum science, and advanced materials. When equipment and pilot facilities cost more than an individual laboratory, startup, or university can afford, the range of researchers able to work at the frontier narrows.

America needs a research strategy that strengthens the underlying science base while increasing the productivity of the system itself—one that supports a wide field of researchers, eases the path toward breakthroughs, makes expensive infrastructure more accessible, and preserves the capacity to mobilize quickly behind urgent national objectives.

Leadership Now supports a national R&D strategy designed to keep the U.S. at the global scientific and technological frontier. Four principles guide our approach.

The overarching goal is a research enterprise capable of exploration and mobilization: thousands of independent sources of ideas, combined with institutional structures able to identify exceptional national challenges, assemble competing teams, provide shared infrastructure, commit resources, and move successful technologies toward rapid deployment.

• Increase and stabilize federal support for fundamental research. Congress should establish durable, multi-year funding paths for the principal federal research agencies, including NSF, NIH, DOE’s Office of Science, NIST, and the national laboratory system, so that top scientists and high-value long-term projects are not vulnerable to political winds. Funding should maintain a broad portfolio of disciplines and research questions. Strategic priorities in AI, quantum science, biotechnology, energy, defense, and other fields all rest on a strong underlying scientific foundation.Predictability is itself a competitiveness asset. Universities, laboratories, and companies make hiring and capital decisions over many years, while leading scientists can increasingly choose among research systems around the world. Funding reversals, cutoffs and suspensions in recent years have interrupted vital work, demoralized researchers, and ceded America’s advantage to competitors.
• Get more discovery from every federal R&D dollar. OMB and OSTP should lead a government-wide effort to simplify research applications and harmonize federal requirements. A two-stage application process—with a preliminary proposal followed by full compliance and administrative materials only for finalists—should become the norm. The Build America Caucus estimates that wider use of this model could return more than one million researcher hours each year to science.⁵Agencies should also review and harmonize requirements across related programs, potentially through a more comprehensive common application portal at Grants.gov with agency-specific supplements. Any proposed new rules should be subject to an analysis of benefits and burdens to the research ecosystem, and existing obligations should be periodically evaluated, consolidated, or retired.The federal government spends close to $200 billion each year on R&D,⁶ yet only a small share is allocated to learning how those resources can be directed more effectively. Major scientific agencies should establish metascience units—building on NSF’s recently announced Office of Metascience⁷—to apply the scientific method to the funding and conduct of research itself. These units should test grant structures, review methods, and administrative processes, publish evidence about which approaches produce better research at lower cost, and facilitate the spread of successful innovations across the federal research system. Ideas worth testing include person-based grants that flexibly fund the researcher rather than the project, golden tickets that allow an individual reviewer to advance promising but unconventional proposals,⁸ fast grants for time-sensitive opportunities,⁹ and AI-assisted funding processes. They should also cover the contested question of institutional overhead, ensuring that the system is efficient and fair to taxpayers while also paying for necessary overhead.
• Give ambitious researchers more ways to succeed. Federal grantmaking should retain rigorous peer review while allowing greater variety in how promising science is supported. Agencies should expand early-career awards that let strong researchers establish independent programs sooner and test mechanisms that support exceptional researchers for several years without requiring every line of inquiry to be specified in advance. They should also test rapid grants for time-sensitive opportunities and limited mechanisms that allow expert reviewers to advance unusually promising high-risk proposals.Congress should expand and evaluate NSF’s recently launched X-Labs initiative¹⁰ and test comparable research organizations within or alongside other scientific agencies. These institutions can bring multidisciplinary teams together for several years around important scientific problems, provide shared instruments and infrastructure, and allow research programs to evolve as results emerge—addressing problems poorly suited to either short-term individual grants or tightly managed mission-centric programs.
• Build shared national infrastructure for AI-enabled science. Decentralized discovery is challenged when only the largest, best-resourced institutions can afford the tools required to work at the scientific frontier. Federal policy should treat compute, major scientific instruments, high-value datasets, and certain experimental facilities as national research infrastructure.Congress should place the National AI Research Resource¹¹ on a durable footing and expand researcher access to advanced computing, models, software, and expertise—with access competitively awarded to reach promising efforts at smaller universities, startups, and researchers outside established technology centers. Similar shared-access models should be developed for quantum science, biotechnology, advanced materials, and other capital-intensive fields.The federal government should also make high-value scientific datasets AI-ready by default, subject to privacy, security, and intellectual-property safeguards, with common standards for documentation and interoperability. AI can generate candidate drugs, materials, and designs faster than physical laboratories can test them. NSF and other agencies should expand remotely accessible and partially automated laboratory capacity, and federal equipment purchases should favor open software and data standards. Connecting federal datasets, compute resources, instruments, and laboratories through interoperable platforms would allow researchers to treat them as a national scientific network rather than siloed assets.
• Build stronger pathways from discovery to commercial scale. Promising discoveries often stall between proof of concept and commercial viability—lacking prototype funding, advanced manufacturing facilities, entrepreneurial expertise, investors, or an initial customer. Federal research policy should provide a clearer path across this gap without requiring every scientist to become an entrepreneur.Congress should strengthen the Small Business Innovation Research and Small Business Technology Transfer programs and expand proof-of-concept and translational awards tied to federally funded research. Research universities should make greater use of entrepreneurs-in-residence and advisers with industry experience, simplify licensing procedures, and develop standardized deal terms.Federal programs should also build stronger links among universities, companies, investors, state governments, and workforce institutions. Programs such as NSF Regional Innovation Engines can help more regions develop the research capacity, skilled workforce, suppliers, and private investment needed to turn scientific strength into durable economic activity.
• Strengthen the ARPA model and use mission-driven R&D selectively. DARPA, ARPA-H and ARPA-E share distinctive strengths that should be preserved: empowered program managers, ambitious technical objectives, competition among different approaches, finite programs, and the freedom to terminate efforts that fail. Congress should protect their operating flexibility and their ability to recruit technical leaders from universities and industry for limited periods of public service.The model should also inform temporary national challenge programs when a defined objective has exceptional economic, health, or security value. Operation Warp Speed demonstrated how government can combine development funding, manufacturing support, procurement, and advance commitments to accelerate technologies and save lives.¹² Such programs should begin with a specific objective and a credible theory of the imperative for federal intervention, finance competing technical approaches, and establish measurable milestones. Furthermore, they should identify from the outset how successful technologies can move toward a federal customer, manufacturer, private investor, or commercial market with clear decision points for expansion, redirection, or termination.
• Use government demand and strategic finance to accelerate critical technologies. Some strategically important technologies face a persistent gap between successful research and viable commercial scale: technical risk remains high while capital requirements exceed what ordinary lenders or early-stage investors will support. Government can help close that gap without choosing permanent national champions or operating commercial enterprises.The Defense Department’s Office of Strategic Capital offers a promising model, providing loans and loan guarantees to qualifying companies in strategically important sectors including semiconductors, quantum technologies, and synthetic biology.¹³ Public support should target identifiable financing or demand gaps, require private risk-sharing, and provide taxpayers with returns commensurate with the risks undertaken—including, where appropriate, equity warrants, revenue-sharing arrangements, or other mechanisms that allow participation in the financial upside.
• Preserve openness while protecting genuinely sensitive research. The U.S. has benefited enormously from attracting scientists, exchanging ideas, and participating in international research networks. Research-security policy should protect sensitive technologies, data, facilities, and supply chains through clear and targeted safeguards—not broad restrictions that isolate American universities or discourage talented researchers from working here.Lawful international collaboration should remain eligible for federal support. Security requirements should correspond to defined risks and use consistent standards wherever possible.America’s long-term advantage comes from the interaction of public investment, private enterprise, academic excellence, talented people, and trusted international partnerships. Federal research policy should preserve that combination while making the system faster, more productive, and better able to translate American ingenuity into American prosperity.