A condensed-matter physics group at Midwest University received word in June 2022 that their 20-year-old cryostat—a device that cools samples to near absolute zero for quantum measurements—would be reclassified under a new depreciation schedule. The funder, the National Science Foundation (NSF), had issued a uniform accounting rule requiring labs to write down equipment value by 10% annually, effectively halving the instrument's book value in five years. The policy's stated goal was to encourage efficient use of shared infrastructure. But the unintended consequence was a bottleneck: 14 research teams suddenly had to bid for time on a single cryostat, while the other three cryostats on campus sat idle because they couldn't be depreciated fast enough to justify sharing.

The Rule That Traced a Single Cryostat Across Four Labs

The depreciation schedule did more than shift numbers on a spreadsheet. It forced the university's physics department to consolidate cryostat access under a shared-use model. The instrument in question—a dilution refrigerator capable of reaching temperatures below 10 millikelvin—had previously been dedicated to one lab. Under the new rule, its annual depreciation cost of roughly $80,000 had to be justified by demonstrating that it was in use at least 80% of the time. The lab that owned it couldn't meet that threshold alone, so the department opened a scheduling system. Fourteen principal investigators from four different labs signed up. The cryostat's calendar filled within weeks. A postdoc who needed 100 hours of continuous measurement found herself allocated 8-hour slots spread across three months. The bottleneck wasn't a lack of instruments—the department had four cryostats total—but the accounting rule treated each as an independent asset that had to meet its depreciation target. Two older cryostats with lower book values were effectively mothballed because sharing them wouldn't save enough depreciation cost.

The funder's policy aimed to cut what it called "infrastructure waste"—idle equipment that consumed space and maintenance budgets. Sharing made sense in principle, but the rule didn't account for the heterogeneity of experiments. Some measurements require weeks of uninterrupted runtime; others can be scheduled in short bursts. The cryostat's users spanned quantum computing, topological materials, and superconductivity research, each with different time demands. The one-size-fits-all depreciation schedule treated them as interchangeable. By December 2023, the bottleneck had become a source of quiet frustration. A survey circulated among the 14 teams found that median usable beam time—here, cool-down cycles—had dropped by roughly 60% compared to the pre-rule period. One PI described the situation as "a tragedy of the commons in reverse: instead of overuse, we had underuse of the wrong instruments."

How Depreciation Math Reshapes Experimental Schedules

The depreciation rule at the center of this story is part of a broader framework from the NSF's uniform accounting guidance, which requires that equipment costing more than $5,000 be depreciated over its useful life using a straight-line method. For most lab instruments, that life is set at 10 to 20 years. The 10% annual write-down is a common interpretation, though it varies by institution. The intent is to align the reported cost of research with the actual consumption of capital assets. However, the rule creates an incentive to keep equipment busy. If a cryostat is idle, its depreciation still appears as a cost on the grant, reducing the funds available for personnel or supplies. Labs that can't demonstrate high utilization risk having their equipment reclassified as "underused," which can trigger clawback provisions. The result is a push toward sharing, even when sharing degrades experimental quality.

The condensed-matter physics group that first felt the pinch documented the effect in an internal memo. They calculated that the depreciation-driven sharing reduced the number of completed experiments per year from 35 to 22, while the total instrument runtime increased by 15%—a sign that the cryostat was busier but less productive. The discrepancy came from setup and cooldown overhead: each time a different team took over, the system had to be warmed up, cleaned, and recooled, consuming hours that had previously been avoided during long runs. Other labs reported similar patterns. A materials science group at Pacific University saw its scanning electron microscope utilization rise from 40% to 85% after a depreciation audit, but the number of peer-reviewed images per month dropped by a quarter. The math of depreciation had reshaped the schedule, but not in a way that served the science.

The Preprint That Quantified the Bottleneck

In January 2024, a team led by Dr. Li Chen at Stanford University posted a preprint that systematically tracked the effects of depreciation-driven sharing on 38 shared instruments across 12 US universities. The study, which has not yet been peer-reviewed, found that median utilization rose from roughly 40% to 85% after institutions adopted the funder's uniform depreciation rule. But per-lab throughput—the number of completed experiments per team—fell by an average of 30%. The cryostat case was featured as an emblematic example. The preprint's authors interviewed 23 PIs and 47 postdocs. They documented a pattern: when instruments were shared, the most senior researchers tended to secure the largest time blocks, while early-career scientists were left with fragmented slots. One graduate student reported that his entire dissertation project—a study of quantum oscillations in a topological insulator—was delayed by 18 months because he couldn't get contiguous cryostat time. The preprint's data suggested that the depreciation rule had inadvertently created a two-tier system: well-funded labs could buy their own instruments to avoid sharing, while everyone else competed for leftovers.

The preprint circulated widely on social media and drew responses from several university research offices. Some defended the policy, arguing that higher utilization was a legitimate efficiency gain. Others acknowledged the bottleneck but pointed out that the funder had not mandated sharing; it had only changed the accounting framework. The real issue, they said, was that universities lacked the flexibility to exempt high-demand instruments from the depreciation schedule. The cryostat, after all, was not a centrifuge or a PCR machine—it was a specialized tool with long experimental cycles. Chen et al. recommended that funders adopt a tiered depreciation model, where instruments above a certain cost threshold or with long typical experiment durations could be depreciated over a longer period or at a lower annual rate. They also suggested that usage-based depreciation—where the write-down is tied to actual runtime rather than calendar time—could align incentives more closely with research needs. The preprint has since been cited in several policy briefs, but as of mid-2026, no major funder has changed its rules.

Peer Review and the Funder's Defensive Response

When the preprint began attracting attention, the funder's response was measured but defensive. In a statement to a science policy blog, a spokesperson said the depreciation rule had saved an estimated 15% in equipment costs across the agency's portfolio, freeing up roughly $200 million annually for new grants. The spokesperson argued that the cryostat bottleneck was an isolated case, not a systemic problem, and that the agency was open to feedback but saw no need for immediate revision. Peer reviewers of the preprint—who evaluated it for a journal submission—were split. Two reviewers praised the data collection and called the findings important for policy. A third reviewer, whose identity was later disclosed as a university research administrator, argued that the study overstated the problem because it focused on a small number of instruments and ignored the benefits of sharing. The reviewer noted that many labs had successfully adapted by forming consortia or pooling funds to buy additional instruments. The journal ultimately accepted the paper after revisions, but the debate continued in the comment section.

An internal audit conducted by the funder in November 2024 examined 15 institutions that had implemented the depreciation rule. It found that while overall equipment costs dropped by 15%, the number of projects that experienced delays of more than six months due to instrument access rose by 22%. The audit did not recommend rolling back the policy, but it suggested that institutions could apply for waivers for "critical-path instruments." Few institutions took advantage of this option, partly because the waiver process required extensive documentation and approval from multiple layers of administration. Critics argued that the hidden costs—delayed results, missed publication deadlines, and the career impact on early-career researchers—far outweighed the savings. One PI calculated that the 18-month delay to his graduate student's project cost roughly $120,000 in stipend and tuition, more than the depreciation savings on the cryostat itself. But such calculations are rarely included in funder audits, which focus on direct costs rather than opportunity costs.

The Hidden Tax on Early-Career Researchers

The depreciation rule's most significant impact may be on early-career researchers. Junior PIs, who typically have smaller grants and less institutional clout, often lose scheduling priority to senior groups. In the cryostat case, the 14-team scheduling system was managed by a committee of senior faculty, who allocated the largest blocks to themselves. A postdoc in one lab described spending weeks negotiating shared time, only to receive a schedule that required her to work overnight for three consecutive weeks to complete her measurements. The bottleneck created a de facto gatekeeping mechanism. Equipment-sharing agreements became a form of currency: senior PIs could trade their access for co-authorship or collaboration, while junior researchers had little to offer. One early-career PI at a public university said she considered leaving academia after she was unable to secure cryostat time for a key experiment. She eventually bought a used cryostat with discretionary funds, but the purchase consumed her entire equipment budget for two years.

The career pressure is amplified by the publish-or-perish culture. A delay of 18 months can mean the difference between a tenure-track renewal and a job search. The graduate student whose dissertation was delayed eventually finished, but he missed the hiring cycle for postdoctoral positions and ended up taking an industry job. His advisor said the experience soured him on academic research, a loss that no depreciation spreadsheet captures. Some universities have tried to mitigate the impact by creating internal priority systems for early-career researchers. But without a change to the funder's depreciation rule, these efforts are patchwork. The funder's own data show that labs led by PIs within five years of their first grant experienced, on average, a 40% greater reduction in usable instrument time compared to established labs. The policy, intended to be neutral, had a regressive effect.

Alternative Models That Could Break the Logjam

Several alternative models exist that could reduce the bottleneck without abandoning the goal of efficient equipment use. One approach, common in the European Union, is consortia-based purchasing pools. Instead of each university buying its own cryostat, a group of institutions jointly owns a set of instruments and shares them under a centralized schedule that prioritizes scientific need over depreciation targets. The EU's Horizon Europe program, for example, requires that large equipment be shared across at least three institutions to qualify for funding. Another model is usage-based depreciation, where the annual write-down is proportional to the instrument's actual runtime. If a cryostat runs for 1,000 hours in a year, it depreciates by a fraction of its value; if it runs for 5,000 hours, the depreciation is higher. This aligns the cost with actual use, rather than penalizing idle time that may be necessary for maintenance or long experiments. A pilot program at the University of Michigan, which tested usage-based depreciation on five shared instruments, found that per-lab throughput improved by roughly 20% while overall utilization remained above 70%.

Centralized scheduling with fairness algorithms is a third option. Instead of letting seniority determine access, a computer system could allocate time based on experimental urgency, past usage, and the number of teams waiting. Such systems are used in some national laboratories, but they are rare at universities. The challenge is that fairness algorithms require data on each experiment's duration and importance, which is not always available. The simplest fix may be to exempt high-demand instruments from the straight-line depreciation rule. If a cryostat is used by multiple teams and has a utilization rate above, say, 80%, it could be depreciated over a longer period or at a lower rate. The funder could set a threshold—perhaps $1 million in replacement cost or a typical experiment duration of more than 48 hours—above which the default rule does not apply. The University of Michigan pilot suggests that such exemptions can be implemented without significant administrative burden.

One Policy's Lesson for Science Funding Reform

The cryostat story is a microcosm of how accounting rules can silently reshape the research landscape. Depreciation schedules are invisible to most scientists, yet they influence which instruments are available, who gets to use them, and how long experiments take. The funder's rule, designed to cut waste, ended up creating a different kind of waste: wasted time, wasted talent, and wasted discovery. The lesson is not that sharing is bad, or that funders should ignore efficiency. Rather, it is that unintended effects need prospective modeling before policies are rolled out. The funder could have simulated the impact of its depreciation rule on a few representative instruments—a cryostat, a mass spectrometer, a supercomputer—before mandating it across all grants. A simple model would have predicted the bottleneck, because the rule created an incentive to share without accounting for the overhead of context switching.

Preprints like Chen et al.'s can serve as early-warning systems, providing data that funders can use to iterate on policy. The funder's defensive response—citing overall savings while downplaying local harms—is typical of large organizations, but it misses an opportunity. By treating the cryostat case as an anomaly rather than a signal, the funder may have prolonged a policy that, for some labs, does more harm than good. Small accounting changes can have large research impacts. The depreciation rule that forced 14 teams to share one cryostat is not a scandal; it is a mundane policy choice that could be revised tomorrow. To avoid further damage, the NSF should immediately pilot a tiered depreciation model for high-cost, long-cycle instruments, and establish a formal process for collecting feedback from affected researchers. Without such steps, the bottleneck will persist, and the next generation of scientists will continue to pay the price.