NSF officials framed the investment as a direct attempt to close that gap and, in doing so, to protect the country’s competitive position in emerging technologies. “NSF invests in small businesses that bring the most innovative cutting-edge technologies to the market and society,” said Erwin Gianchandani, the foundation’s assistant director for Technology, Innovation and Partnerships. “By closing this gap that is known as the ‘valley of death,’ NSF aims to strengthen our commitment to translational research and further develop innovation ecosystems in the United States.” The valley of death, as the agency describes it, is the stage at which a new technological innovation has been substantially de-risked but has not yet been commercially proven, making it difficult to attract the funding or other support needed to achieve commercial viability. For deep-technology companies, whose products often require long development cycles, capital-intensive prototyping and regulatory navigation, that stage can be especially punishing. Unlike software ventures that can iterate quickly and cheaply, hardware-based and science-heavy startups frequently need years of sustained investment before their first meaningful revenue arrives.

The pilot will concentrate its support on companies that have already cleared significant federal hurdles. Its services are aimed at small businesses funded through NSF’s Small Business Innovation Research and Small Business Technology Transfer programs, specifically those that have completed Phase II or Phase IIB awards. These represent the later stages of the federal small-business pipeline, by which point a company has typically demonstrated the technical feasibility of its innovation and begun the hard work of translating it into a manufacturable, sellable product. Lesia Crumpton-Young, the TIP section head for small businesses, emphasized the economic stakes of that translation. “Many NSF-supported small businesses have the potential to become great commercial successes and contribute significantly to economic growth and job creation throughout the nation,” she said. “We want to support small businesses to overcome the hurdles common to commercializing deep technology innovations.” The program’s architects argue that without a deliberate bridge at this late stage, the substantial public investment already made in these companies through earlier SBIR and STTR awards risks being wasted.

At the helm of the new institute is Ivan Garibay, a professor of industrial engineering and management systems at UCF, who will serve as principal investigator. Garibay’s framing of the problem captures the technical distinction at the heart of the pilot’s mission. “A validated technology is a scientific achievement, but it isn’t yet a business,” he said. “Many NSF SBIR Phase II companies hold genuinely de-risked technologies, yet still lack the business expertise, infrastructure, and capital access to turn that validation into a viable company.” His description of NCTI’s role is deliberately complementary rather than duplicative. “NCTI completes that pipeline rather than duplicating any part of it: we coordinate the full continuum of federal resources, managing the handoffs from I-Corps customer discovery through to commercialization support, so each company knows where it stands and what comes next.” That coordination function, in which a single institute orchestrates the various federal programs a startup might otherwise have to navigate alone, is one of the pilot’s most distinctive design features.

The support package offered to participating companies is structured as what the program calls concierge-level individualized assistance. Each selected business will undergo a diagnostic triage to identify its specific commercialization bottlenecks, followed by the creation of an individualized commercialization plan tailored to its technology, market and funding situation. Companies will be paired with dedicated mentor teams and will receive curated introductions to potential customers, investors, industry partners and national laboratories, the latter offering access to specialized facilities and technical expertise that small startups could rarely afford on their own. Alongside this individualized support, the institute will provide direct commercialization funding, expert mentorship and connections to investors. The intent is to address simultaneously the three deficits that Garibay identified as the core of the valley of death: business expertise, infrastructure and access to capital. Peter Atherton, program director for NSF Small Business Innovation Research and Small Business Technology Transfer, described the approach as a bridge. “Providing a ‘bridge’ to support our late-stage NSF-funded companies will move their impactful innovation to the market more quickly and thereby benefit the national economy,” he said.

NCTI is not a sudden departure for UCF but rather the latest evolution of a commercialization pipeline the university has spent years constructing. Through initiatives including the Southeast Hub of NSF I-Corps, which trains researchers in customer discovery and lean startup methodology; NSF’s Accelerating Research Translation program, which builds institutional capacity to move academic discoveries toward practical use; and the Florida Semiconductor Engine, a regional coalition advancing the semiconductor industry, UCF has already helped researchers evaluate commercial opportunities, advance emerging technologies and strengthen regional innovation ecosystems. NCTI will build on that foundation while dramatically expanding its scope, extending UCF’s commercialization machinery to companies nationwide rather than confining it to Florida and the Southeast. University leadership has positioned the institute as a natural outgrowth of that trajectory. “NCTI builds on UCF’s growing national leadership in helping promising technologies become products, companies and economic growth,” said UCF President Alexander N. Cartwright, who added that Central Florida’s concentration of talent, industry and innovation positions the university to build a model with impact across the nation.

University research officials argue that the institute’s success will depend on the integration of the entrepreneurial infrastructure already in place at UCF. Winston Schoenfeld, the university’s vice president for research and innovation, described that integration as the anchor of the institution’s readiness. “UCF’s readiness to lead this national effort is anchored by our highly integrated entrepreneurial infrastructure critical in the ‘lab-to-market’ pipeline,” he said. “NCTI will ensure that federally funded technologies successfully transition into investor-ready commercial enterprises.” The phrase “investor-ready” is significant, because it identifies the precise handoff point where the valley of death typically opens. A company becomes investor-ready when its technology has been validated, its market has been defined, its team has been strengthened and its path to revenue is credible enough to attract private capital. Reaching that state requires exactly the combination of funding, mentorship and networking that the pilot intends to supply, and it is the state at which federal support has historically ended and private investment has historically been slow to begin.

Beyond its immediate services to individual companies, the pilot carries a broader experimental mandate: it is explicitly designed as a test of whether a scalable, long-term national model is feasible. Each stage of the program will be structured and independently evaluated, with the results determining whether NSF establishes a permanent program on the pilot’s foundation. That evaluative rigor reflects a growing recognition across the federal science establishment that commercialization outcomes can and should be measured as carefully as research outcomes. The factors, conditions and approaches that lead to the successful transition of deep-technology innovations into nationally impactful technologies will be systematically examined, generating evidence about which interventions, whether diagnostic triage, mentor teams, investor introductions or direct funding, actually move companies across the valley. If the pilot demonstrates measurable success, it could reshape how federal agencies support the late-stage development of science-based startups for years to come.

The stakes of the experiment extend well beyond the companies that participate. Deep technologies, which include advanced materials, quantum systems, biotechnology platforms, next-generation semiconductors and other science-intensive innovations, are widely viewed as critical to both economic competitiveness and national security, and competitors abroad have invested heavily in smoothing their own translation pipelines. When American startups fail in the valley of death, the underlying research, often funded by taxpayers through agencies like NSF, may be licensed abroad, shelved indefinitely or lost entirely, along with the jobs and industries it could have created. By intervening at the precise moment when public funding ends and private capital has not yet arrived, the NCTI pilot represents a deliberate attempt to keep that value inside the American innovation ecosystem. Whether a $20 million, two-year investment can meaningfully change the survival rates of deep-technology ventures remains to be seen, but the program’s designers are betting that coordinated, individualized, late-stage support can convert more laboratory-validated science into products, companies and economic growth, and that the lessons learned will justify a permanent national commitment.

Subject of Research: A $20 million NSF pilot program led by the University of Central Florida to help small businesses commercialize late-stage deep technologies and avoid the valley of death.

Article Title: NSF funds pilot program to support small businesses in accelerating commercialization of their late-stage deep-technology and avoid the “valley of death”

Article References: NSF funds pilot program to support small businesses in accelerating commercialization of their late-stage deep-technology and avoid the “valley of death”. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: NSF, valley of death, deep technology, commercialization, SBIR, STTR, University of Central Florida, NCTI, small business, technology transfer, innovation ecosystem, startup funding

The study’s central contribution is not a new algorithm. Khan openly states that every constituent method, from gradient-boosted forecasting to stochastic programming, is established. What is new is the explicit, testable coupling of those methods into a single auditable pipeline: a forecasting model produces a demand estimate, a statistical calibration layer converts that estimate into reliable uncertainty bounds, a leakage-safe classifier ranks logistics options by their risk of delay, and a risk-adjusted optimization model turns all of that evidence into concrete order quantities. Each handoff from prediction to prescription is separately verifiable, a design philosophy that stands in sharp contrast to the black-box dashboards many vendors sell to small firms.

The technical machinery is worth unpacking. Demand forecasts came from XGBoost, a gradient-boosted tree model trained on 350,000 records from the public FreshRetailNet-50K dataset covering 120 store-product series. Rather than trusting the point forecast alone, the system applies conformalized quantile regression, a technique that uses a separate calibration period to correct the coverage of prediction intervals. The result was striking: an interval nominally covering 80 percent of outcomes actually covered 90.42 percent in the main sample, meaning the system deliberately over-protects rather than under-protects. Those calibrated bounds, not the raw forecast, generate the low, upper, and extreme demand scenarios that feed the ordering model.

On the logistics side, Khan faced a subtle trap common in supply-chain analytics: data leakage. Records of actual shipping duration and delivery status are only known after the fact, so using them to predict delays would produce a model that looks brilliant in testing and fails in deployment. The study excluded all post-outcome variables and trained classifiers using only information available before fulfillment, such as shipping mode, scheduled days, order value, and region. A calibrated logistic regression achieved a future-test ROC-AUC of 0.7277, moderate but honest discrimination. Notably, a hard classification threshold proved useless as an alert, flagging nearly everything as delayed, so the system instead feeds continuous calibrated probabilities into the decision model.

The prescriptive heart of the system is a stochastic linear program minimizing a conditional-value-at-risk objective. In plain terms, the model does not just minimize average cost; it explicitly penalizes the worst-case tail of shortage losses, the scenario in which a small business runs out of stock during a disruption. Decision variables specify how much to order from each logistics option, subject to budget, storage capacity, option limits, and a 95 percent service target. When that target proved infeasible under the stated constraints, which happened in 18.21 percent of risk-adjusted decisions, the system relaxed only the service row and recorded an explicit audit flag, converting silent failure into a visible escalation for human review.

The benchmark results complicate the usual AI triumphalism. XGBoost improved mean absolute error by just 2.74 percent over a humble seven-day moving average, and several other machine learning models failed to beat that simple baseline at all. Yet the risk-adjusted optimizer still delivered meaningful resilience gains: 3.93 percent stockout and 95.53 percent aggregate service, cutting stockouts by 71.91 percent relative to a fixed reorder point policy and by 45.45 percent relative to risk-neutral optimization, at a cost increase of 16.95 percent over the fixed rule. After Holm correction for multiple comparisons, these improvements over traditional and risk-neutral alternatives were statistically significant.

The most provocative finding, however, is what did not differ. The risk-adjusted optimizer was statistically indistinguishable from a far simpler adaptive policy that assigns inventory strategies to products using only their consumption value and demand variability, with thresholds frozen during calibration. A rule requiring no optimization model per item matched the sophisticated LP under the baseline scenario weights. That equivalence, though, turned out to be conditional. When Khan re-solved every policy under alternative probability profiles, equal weights, a routine-heavy profile, and a disruption-heavy profile, the optimizer became significantly more protective than the adaptive rule under both routine-heavy and disruption-heavy beliefs. Underweighting adverse states, the routine-heavy profile showed, can nearly erase the rank correlations of policy performance with the baseline.

No policy dominated everywhere, and the extremes are instructive. Ordering to the AI upper bound achieved the highest service but the highest cost, triggering an emergency-order flag on every single order. The classical economic order quantity remained the cheapest but suffered the worst service. Sensitivity analysis identified logistics delay as the single most consequential stress factor: raising delay stress from zero to 0.45 drove stockouts from 3.63 percent to 28.48 percent and service from 96.00 percent down to 73.22 percent, dwarfing the effect of any forecasting improvement. On an independent dataset of 40 UCI Online Retail product series, the policy rankings shifted again, with risk-neutral optimization taking the top service spot, evidence of context dependence rather than universal AI superiority.

For managers, the practical translation is a segmented playbook rather than a single prescription. Stable, low-value products can stay under classical EOQ logic. Service-critical items with moderate uncertainty may justify upper-bound ordering. Volatile or high-value products warrant risk-adjusted optimization and human approval flags for high-value, high-risk, or emergency orders. Crucially, the study argues that managers must document their probability beliefs, whether they are planning for routine conditions or disruption-heavy ones, because those beliefs alone can change which decision tool is worth its complexity. Given that delay stress outweighed forecast accuracy in every test, investing in logistics reliability may deliver more resilience than any marginal gain in demand prediction.

The study is candid about its limits. It uses public benchmark datasets rather than proprietary small-business operations, proxy cost parameters are fully disclosed but not audited, and the scenario probabilities are experimental design weights rather than empirically estimated disruption frequencies. The system is a human-supervised computational benchmark, not a field-deployed autonomous ordering platform. Still, as a reproducible, leakage-audited demonstration of how calibrated uncertainty should actually change ordering decisions, the work sets a new standard for the field. Its headline lesson deserves to travel far beyond operations research: the value of AI in decision-making lies not in prediction alone, but in the auditable pathway from calibrated evidence to constrained, reviewable action.

Subject of Research: An uncertainty-aware AI decision support system for small-business supply chain replenishment under demand and logistics uncertainty

Article Title: Artificial intelligence decision support system for small business supply chain resilience under uncertainty

Article References: Khan, M. R. I. (2026). Artificial intelligence decision support system for small business supply chain resilience under uncertainty. Discover Artificial Intelligence, 6(1), Article 1298. https://doi.org/10.1007/s44163-026-02331-8

Image Credits: AI Generated

DOI: 10.1007/s44163-026-02331-8

Keywords: artificial intelligence, decision support system, supply chain resilience, stochastic optimization, conformal prediction, XGBoost, logistics risk, small business, inventory management, CVaR, demand forecasting, uncertainty

The research, led by Md. Salman Syed Sani of Shahjalal University of Science and Technology together with colleagues at the University of Dhaka, focused on Gulistan and Gulshan-2, two commercial hubs that could hardly be more different. Gulistan, under Dhaka South City Corporation, is a dense, chaotic transport node packed with roadside stalls and small-scale enterprises heavily dependent on public transport. Gulshan-2, under Dhaka North City Corporation, is a more planned commercial-residential district with better infrastructure and a more established business class. The stakes of the comparison are enormous: micro, small and medium-sized enterprises contribute roughly a quarter of Bangladesh’s GDP and employ close to eighty percent of its workforce, yet their owners’ and workers’ mobility has been almost entirely invisible in urban transport research.

To capture that hidden population, the team surveyed 384 respondents, split evenly between the two districts, using a sample size calculated with Cochran’s formula at a ninety-five percent confidence level. Between April and October 2024, researchers approached business owners and employees directly at their workplaces during business hours, gathering data on education, income, business structure, residential location, commuting distance, travel time, transport mode and expenditure. Two focus group discussions, one in each district, added qualitative depth. The analysis combined descriptive statistics, multiple regression, Principal Component Analysis and geospatial mapping in SPSS, MS Excel and ArcGIS, an integrated toolkit designed to connect individual circumstances with the spatial structure of the city.

The demographic contrasts were stark from the outset. Gulistan’s workforce is young, with forty-eight percent of respondents aged between twenty and thirty, and education concentrated at primary and secondary levels. Gulshan’s workers are older, more educated and far more settled: thirty percent reported more than ten years in the same business, and thirty-seven percent had never changed jobs, compared with just twenty percent in Gulistan. Gulistan workers, by contrast, are chronic movers, with seventy-two percent reporting job changes within the same zone and half reporting three job changes in the past decade. In Gulistan, relocation decisions were driven mainly by business location, financial improvement and workplace relationships; in Gulshan, business location dominated, but residential location and better job preferences played larger supporting roles.

Residential patterns told a similar story of divergence. Ninety percent of Gulshan workers live in central city locations close to their workplaces, clustered in neighbourhoods such as Natun Bazar, Badda, Banasree and Shahjadpur, and typically within five to ten kilometres of work. Gulistan workers are far more dispersed: fourteen percent live in outer suburban areas, and some commute from districts as distant as Narayanganj and Munshiganj. Fifteen percent of Gulistan workers travel more than thirty kilometres to work, and thirty percent commute between ten and twenty kilometres. The consequence is a transport profile dominated by buses, which carry fifty-six percent of Gulistan commuters, followed by informal leguna minibuses at eighteen percent and rickshaws at fifteen percent. Gulshan workers, living closer, rely far more on walking and non-motorized transport.

The financial burden follows the same gradient. Nearly a third of Gulistan workers spend between 1,000 and 1,500 taka per month on commuting, and a quarter spend between 2,000 and 3,000 taka, significant sums for low-income earners. Focus group participants described the trade-offs in blunt terms. One Gulistan employee explained that affordable housing near the commercial area is nearly impossible to find, forcing a daily long-distance commute. Another reported choosing slower, cheaper transport options to keep expenses down, while a third said peak-hour congestion often doubles travel time and makes a regular work schedule difficult to maintain. Gulshan participants, by contrast, described commutes that, while not congestion-free, are at least predictable, and some said paying more for reliable transport is a worthwhile investment in efficiency.

The statistical core of the study delivered its most striking finding. Principal Component Analysis extracted five components explaining ninety-two percent of the total variance in the data, and when those components were fed into a logistic regression classifier, the model distinguished Gulistan from Gulshan respondents with ninety-four percent accuracy. The first component, dominated by business scale and economic capacity, separated smaller, informal, foot-traffic-dependent enterprises in Gulistan from larger, more formal and better-connected businesses in Gulshan. In effect, the analysis showed that mobility inequality in Dhaka is not randomly distributed but structurally embedded in the geography of the two districts, so much so that a worker’s commuting and business profile alone can identify which district they work in.

Yet the regression models delivered a humbling lesson about what actually determines commuting time. Education, age, income and transport mode all failed to reach statistical significance once other factors were controlled. Only two variables mattered: sex, at the ten percent significance level, and commuting distance, which was highly significant at the one percent level. The models explained roughly eighteen and twenty-three percent of the variance in commuting time for Gulistan and Gulshan respectively. The authors interpret this through the lens of Hägerstrand’s time-geography and the Alonso-Muth urban location model: in a city as congested as Dhaka, congestion, road conditions and residential-workplace mismatch impose constraints that cut across socio-economic groups, while lower educational attainment can indirectly restrict housing choices, pushing workers to the periphery and lengthening their journeys. The researchers are careful to note that only the direct effects of distance and sex are statistically supported; the wider causal chain remains a theoretically informed interpretation.

Perhaps the most distinctive insight concerns entrepreneurs themselves. Business owners, especially those managing multiple premises, consistently prioritized commercial accessibility over residential convenience. Participants explained that they choose business locations based on customer demand and established supplier relationships, even when that means travelling farther from home, and that changing residence is easier than relocating a business whose customers and suppliers are already anchored in place. This inverts the assumptions of conventional commuting studies, which treat the workplace as fixed and the residence as the adjustable variable. For small business owners, the reverse is often true, and any transport policy that ignores the gravitational pull of customer networks will misread the mobility behaviour of a huge share of Dhaka’s workforce.

The policy implications are correspondingly place-specific. In Gulistan, the authors argue, priority should go to improving the reliability, affordability and coordination of buses, legunas and rickshaws, along with safer boarding areas and better connections from peripheral districts. In Gulshan, the challenge is protective: preserving continuous footpaths, safe crossings and first- and last-mile connections against the encroachment of parking and construction. At the metropolitan scale, the study calls for mixed-use development paired with affordable rental housing near commercial centres, since proximity alone helps no one who cannot afford to live there. Gender-disaggregated transport data, delivery access for multi-location businesses, and coordinated monitoring by the two city corporations round out the recommendations. What the study ultimately demonstrates is that mobility inequality is not only a problem between cities or social classes; it can be manufactured, block by block, within a single metropolis, and fixing it requires reading the fine grain of the urban fabric itself.

Subject of Research: Intra-urban mobility and commuting patterns of small business owners and employees in Dhaka's commercial districts

Article Title: Assessing the influence of intra-urban mobility on small business dynamics in Gulistan and Gulshan commercial areas of Dhaka

Article References: Sani, M. S. S., Rahman, M. M., & Labib, M. I. H. (2026). Assessing the influence of intra-urban mobility on small business dynamics in Gulistan and Gulshan commercial areas of Dhaka. Discover Cities, 3(1), Article 196. https://doi.org/10.1007/s44327-026-00375-0

Image Credits: AI Generated

DOI: 10.1007/s44327-026-00375-0

Keywords: Dhaka, intra-urban mobility, small business, MSMEs, commuting behavior, urban planning, transport geography, residential location, principal component analysis, spatial inequality, public transport, Bangladesh

The technical logic behind Pearce’s objection is straightforward. In experimental science, hardware cost is a quantifiable variable, just like measurement precision, throughput, or error rate. When a paper describes an open-source instrument, the bill of materials is the specification that makes replication possible. Removing it does not merely soften a claim; it breaks the chain of reproducibility that peer review is supposed to guarantee. A peer-reviewed comparison of cost, reproducibility, and performance, Pearce writes, is no more an endorsement of one product than a methods-comparison paper is an endorsement of one technique. It is part of the scientific record needed for replication, evaluation, and efficient public spending. Demanding its removal forces researchers to publish incomplete research, which he characterizes as fundamentally anti-science and inconsistent with the United States government’s own public-access and open-science policies, which now emphasize free, immediate access to federally funded research outputs and the data underlying them.

The economics at stake are substantial. Peer-reviewed research has repeatedly demonstrated that open-source scientific hardware simply costs less than proprietary equivalents. A recent HardwareX review found average savings of 87 percent overall, rising to 92 percent for 3D-printed tools and 94 percent for tools that combine 3D printing with Arduino-class open-source electronics, while earlier case studies often reported savings in the 90 to 99 percent range. The United States currently spends billions of dollars annually on scientific equipment, much of it proprietary and often manufactured internationally. Pearce points to Finland as a cautionary benchmark: an analysis there found that Finnish science funders could save between 2.84 and 27.7 million euros per year simply by shifting to open-source purchasing. Scaled to the American research enterprise, he argues, a conservative estimate suggests federally funded science could run ten times faster, with ten times more equipment at the same cost, if open hardware became the default.

The letter offers a concrete illustration of what such savings look like in practice. An open-source syringe pump library, developed in the United States, documented a device that can be built for roughly 50 dollars in any American laboratory, while internationally manufactured proprietary equivalents cost 1,500 dollars and in some cases up to 2,500 dollars. Stating that price difference, Pearce insists, is not disparagement of the private sector; it is an empirical fact. That design library was downloaded thousands of times within its first months of publication, producing a calculated substitution valuation of between roughly 1 and 12 million dollars in savings to the scientific community in a single year. Since then, the designs have been downloaded tens of thousands of times, conservatively saving scientists tens of millions of dollars. Censoring such data, Pearce argues, prioritizes the commercial comfort of legacy vendors over the efficiency of government researchers and the success of American small businesses.

A central misunderstanding, according to the letter, is the assumption that open hardware is anti-commercial. By definition, open-source hardware licenses permit anyone, including private enterprises, to study, modify, distribute, make, and sell products based on a published design. Far from competing with industry, federal researchers who publish open designs are providing free, rigorously tested research and development to American small and medium-sized enterprises, enabling those firms to commercialize new products without the crippling overhead of initial prototyping and design. The Open Source Hardware Association, a United States-based organization composed primarily of such enterprises, exists precisely because open licensing can be a viable and profitable business model. Some authors in this field have gone so far as to call for all government-funded hardware research to be open source by default, on the grounds that the public has already paid for it.

The commercial track record cited in the letter is extensive. Arduino, whose open-source microcontroller platform has become the de facto standard for embedded prototyping in laboratories worldwide and underpins countless HardwareX publications, is now owned by Qualcomm, a San Diego-based technology firm with roughly 44 billion dollars in annual revenue. SparkFun Electronics of Boulder, Colorado, grew into a 32-million-dollar-a-year business with 146 employees by selling open-source electronic components and educational kits. Adafruit Industries in New York City generates more than 45 million dollars annually selling open-source hardware. Backyard Brains of Ann Arbor, Michigan, commercialized open-source neuroscience tools first described in peer-reviewed open hardware literature, while Opentrons of Long Island City democratized laboratory automation with open liquid-handling robots that displaced proprietary systems costing an order of magnitude more.

The examples extend into heavy manufacturing and even the military. LulzBot, headquartered and manufactured in Fargo, North Dakota, is an entirely open-source 3D printer manufacturer that has competed directly with proprietary incumbents and is used by the United States Marines, whose motto of improvising, adapting, and overcoming is well served by hardware whose designs can be freely modified and repaired. Adjacent firms have flourished on the same ecosystem: re:3D of Austin, Texas, a spinoff founded by NASA engineers, and MatterHackers of Lake Forest, California, both built thriving businesses around the open-source RepRap 3D printing project, the self-replicating rapid prototyper that catalyzed the modern desktop printing movement. Collectively, Pearce writes, these open hardware firms employ thousands of Americans and generate hundreds of millions of dollars in annual revenue, all built on openly licensed designs. When a federal researcher publishes a low-cost environmental sensor or diagnostic tool, he argues, they are planting the seeds of the next generation of American hardware startups.

Open hardware also carries a strategic dimension that goes beyond price. Because designs can be produced with distributed manufacturing technologies such as 3D printing, CNC milling, laser cutters, and open-source microcontrollers, scientific equipment can be fabricated locally and on demand within the United States. This insulates research supply chains from global disruptions, from pandemics to trade conflicts, and repatriates manufacturing capability that has migrated overseas. A federal grant that once could afford only a single proprietary instrument can instead fund an entire laboratory of open-source equivalents, accelerating discovery across any scientific field the equipment touches. Pearce frames these efficiencies as exactly the kind of fiscal discipline that cost-conscious administrations should champion, suggesting that agencies celebrate open hardware breakthroughs with press releases and awards rather than suppress them.

The letter closes with concrete demands directed at the Bureau Approving Officials. Pearce urges them to permit empirically grounded bills of materials and cost comparisons in manuscripts, to issue written guidance distinguishing comparative scientific analysis from endorsement, and to align manuscript review with agency public-access plans and the broader open-science expectations of the Office of Science and Technology Policy. The underlying principle, he argues, is simple: federal researchers should be allowed to report proudly how they are saving taxpayer money and providing open-source research and development to the American public. Whether the agencies respond by revising their review guidance or by continuing to redact economic data will signal how seriously the United States takes both the openness of its science and the competitiveness of the small manufacturers who depend on it.

The dispute is likely to resonate well beyond the federal bureaucracy. As open-source instruments, from syringe pumps to bioprinters, become mainstream tools in laboratories worldwide, the question of who may publish what about them has become a genuine constraint on the scientific record. Pearce’s letter, published as an editorial in HardwareX, places the issue squarely before the officials who control the clearance of government manuscripts, and before a research community that has watched open hardware deliver documented savings measured in the tens of millions of dollars. The stakes, as he frames them, extend to the integrity of peer review itself: a literature stripped of cost data is a literature that can no longer fully support replication, evaluation, or the efficient use of public funds.

Subject of Research: Open-source scientific hardware and federal censorship of cost data in U.S. government research publications

Article Title: An open letter to the Bureau Approving Officials of the United States government

Article References: Pearce, J. M. (2026). An open letter to the Bureau Approving Officials of the United States government. HardwareX, 27, Article e00825. https://doi.org/10.1016/j.ohx.2026.e00825

Image Credits: AI Generated

DOI: 10.1016/j.ohx.2026.e00825

Keywords: open-source hardware, HardwareX, scientific equipment, censorship, bills of materials, 3D printing, federal research policy, open science, cost savings, distributed manufacturing, small business, peer review