Real problems worth solving

Helium-3 is required for the dilution refrigerators that cool superconducting quantum computers to operating temperatures below 10 millikelvin. But helium-3 does not exist in meaningful quantities in nature on Earth -- it comes almost entirely from the radioactive decay of tritium produced as a byproduct of nuclear weapons programs. Global supply is so constrained that in September 2025, Bluefors (the leading dilution refrigerator manufacturer) signed a contract with Interlune to purchase up to 10,000 liters of helium-3 per year extracted from the lunar surface, with deliveries starting in 2028. Why it matters: Because helium-3 supply depends on nuclear weapons program byproducts, the quantum computing industry's growth is literally coupled to geopolitics and nuclear policy, so any disruption -- sanctions, arms reduction treaties, or geopolitical conflicts -- could halt dilution refrigerator production, so new quantum computer installations would stop, so the entire superconducting quantum computing roadmap (IBM, Google, Rigetti, and dozens of startups) would be frozen in place, so the billions in quantum computing investment could be stranded by a supply chain vulnerability that has nothing to do with the technology itself. The structural root cause is that helium-3 production was never designed to serve the quantum computing industry -- it was an incidental byproduct of Cold War nuclear weapons manufacturing. As nuclear arsenals have been reduced and tritium production has declined, helium-3 supply has shrunk while quantum computing demand has grown. No commercially viable alternative production method exists at scale, and alternative cooling technologies (like Kiutra's magnetic cooling) are still in early development.

Dilution refrigerators -- the cryogenic systems that cool superconducting quantum processors to below 10 millikelvin (hundreds of times colder than outer space) -- have lead times of 6 to 9 months. But quantum hardware teams at companies like IBM, Google, and Rigetti iterate on chip designs every 12 to 18 months. This means the cooling infrastructure consumes 30-50% of the entire hardware development cycle just waiting for delivery. Why it matters: Because half the iteration cycle is spent waiting for refrigerators, hardware teams can only test 1-2 chip generations per year instead of 3-4, so the pace of qubit quality improvement is artificially throttled, so the gap between current noisy qubits and the fault-tolerant qubits needed for useful computation closes far more slowly than it should, so the timeline to quantum advantage stretches from 'years away' to 'decade away,' so classical computing continues to advance and narrow the window where quantum speedups would matter for real-world problems. The structural root cause is that the dilution refrigerator market is dominated by a handful of specialized manufacturers (primarily Bluefors in Finland and Oxford Instruments in the UK), and the manufacturing process requires highly specialized skills, custom components, and extensive testing at cryogenic temperatures. There is no mass-production process for these systems because annual global demand is only in the hundreds of units, which is too small to justify the capital investment needed to build high-volume production lines.

The global pool of professionals specializing in quantum error correction (QEC) is estimated at only 1,800 to 2,200 people, while industry projections indicate 5,000 to 16,000 QEC specialists will be needed by 2030. Companies like IBM, Google, Quantinuum, and Riverlane are competing for the same tiny talent pool, and university programs produce only a handful of QEC-focused graduates per year. Why it matters: Because QEC talent is so scarce, companies cannot staff the teams needed to develop real-time decoding systems, so fault-tolerant quantum computers remain stuck in the lab, so the entire industry's timeline to commercially useful quantum computing slips by years, so the billions of dollars invested by governments and private investors (over $4.2 billion in 2025 alone) risk generating no return, so quantum computing may lose credibility as a viable technology investment and funding could dry up before the field reaches its potential. The structural root cause is that quantum error correction sits at the intersection of quantum physics, coding theory, and real-time systems engineering -- a combination so rare that no single university department owns it. Physics departments do not teach real-time systems, computer science departments do not teach quantum mechanics at the required depth, and electrical engineering departments do not teach algebraic coding theory. This means QEC expertise can only be acquired through years of cross-disciplinary PhD research, creating a pipeline that produces dozens of qualified people per year when thousands are needed.

Chinese chipmakers including SMIC have learned to apply multipatterning techniques to deep ultraviolet immersion (DUVi) lithography systems, which are not restricted by export controls, to produce chips approaching 7nm and even 5nm-class performance. ASML sold 70% of its DUVi lithography systems to Chinese entities in 2024, and continues to service machines in China (potentially extending their lifespan to 30 years) despite U.S. pressure on the Dutch government. Meanwhile, enforcement gaps persist: in February 2026, Applied Materials was fined $252 million for illegally exporting ion implantation equipment to China. Why it matters: China can produce near-cutting-edge chips using unrestricted DUV equipment, so the stated U.S. policy goal of maintaining a multi-generational lead in semiconductor technology is being eroded despite increasingly complex export control regimes, so SMIC and other Chinese foundries can manufacture 7nm chips for military, surveillance, and AI applications that the controls were specifically designed to prevent, so the U.S. must continually tighten restrictions (creating a regulatory whack-a-mole dynamic) which increases compliance costs for American semiconductor equipment makers like Applied Materials, Lam Research, and KLA who lose billions in China revenue, so the diplomatic relationship with the Netherlands and Japan (whose cooperation is essential for controls to work) is strained by each new round of unilateral U.S. rule changes, so the export control regime risks becoming a worst-of-both-worlds outcome: not restrictive enough to actually prevent Chinese chip advancement, but restrictive enough to cost U.S. companies market share and push China toward full self-sufficiency. The structural root cause is that the physics of DUV multipatterning allows production of features smaller than the wavelength of light through repeated exposures, making technology-based export control thresholds inherently leaky. Export controls were designed around equipment capabilities (blocking EUV), but process innovation (multipatterning with DUV) can partially substitute for equipment capabilities. Additionally, ASML's DUVi machines require ongoing maintenance that only ASML can provide, and the Dutch government has been reluctant to restrict servicing of machines already sold, creating an installed base in China that will produce chips for decades.

TSMC's fabs in Taiwan consume up to 99,000 tonnes of water daily at the Southern Taiwan Science Park alone, and Taiwan's overall water consumption is projected to exceed supply by 680,000 cubic meters per day by 2036. Simultaneously, Taiwan's semiconductor and AI industries are expected to add 5+ gigawatts of electricity demand by 2030 (enough to power 3.75 million homes), while the island relies on imported LNG for over 50% of its power generation and has limited renewable energy infrastructure. Taiwan has experienced drought conditions since 2021. Why it matters: TSMC produces 90%+ of the world's advanced semiconductors on an island facing both water and power deficits, so TSMC's planned capacity expansions (N2, A16 nodes) in Chiayi and Kaohsiung may face resource constraints that limit actual production volumes below nameplate capacity, so the global chip supply that powers smartphones, data centers, automobiles, and AI systems depends on resource availability on a water-stressed, energy-importing island 100 miles from a geopolitical adversary, so chip production interruptions during drought or typhoon seasons (which paradoxically bring floods, not usable reservoir water) become an annual recurring risk, so Taiwan's government faces an impossible trilemma between semiconductor industry growth, residential/agricultural water needs, and carbon emission reduction targets, so the world's most critical manufacturing capability is concentrated in one of the most resource-constrained and geopolitically exposed locations on Earth. The structural root cause is that semiconductor manufacturing requires ultrapure water (UPW) at volumes that scale linearly with wafer production, and producing 1,000 gallons of UPW requires 1,400-1,600 gallons of municipal water. Taiwan's geography (steep mountains, short rivers, limited reservoir capacity) makes water storage inherently difficult, while its lack of domestic fossil fuels or nuclear baseload (after phasing out nuclear plants) means electricity supply depends on LNG imports through the same shipping lanes that would be disrupted in a cross-strait conflict.

Intel Foundry Services (IFS), positioned as America's answer to TSMC dependence, posted an operating loss of roughly $7 billion in 2023 and continued multi-billion-dollar losses through 2024-2025, including a $2.3 billion operating loss in Q3 alone on $4.2 billion revenue. As of mid-2025, IFS had not secured a single major external foundry customer for its existing nodes, and key engineering talent has been departing for competitors. The U.S. government's strategy of having Intel serve as a trusted domestic foundry alternative is at risk of failure. Why it matters: IFS cannot attract major customers like Qualcomm, NVIDIA, or Apple, so Intel's multi-billion-dollar foundry pivot fails to generate the revenue needed to fund continued R&D and next-generation process development, so the U.S. loses its only potential domestic alternative to TSMC for leading-edge chip manufacturing, so the CHIPS Act's implicit assumption of a two-foundry domestic ecosystem (TSMC Arizona + Intel) collapses to a single-company dependency, so the national security goal of having a trusted U.S.-owned foundry for defense and intelligence chips remains unmet, so if Intel does not secure an anchor customer by late 2026 it will scale back from aggressive 14A (1.4nm) investment to a slower roadmap, permanently ceding the leading-edge foundry market to TSMC and Samsung. The structural root cause is that fabless chip companies (NVIDIA, Qualcomm, Apple, AMD) view Intel as a competitor in the chip design market and are reluctant to share proprietary designs with a foundry that also competes against them. Additionally, Intel's manufacturing process track record includes multiple delays and yield problems on 10nm/7nm (Intel 4/Intel 3), giving potential customers insufficient confidence to risk their product roadmaps on an unproven foundry relationship. The foundry business requires a trust relationship that takes years to build, and Intel is attempting to compress a decade of relationship-building into 2-3 years.

On March 2, 2026, Iranian drone strikes forced Qatar's 77-million-ton-per-annum helium production facility offline, removing approximately 30% of global helium supply. Helium is irreplaceable in semiconductor manufacturing for cooling silicon wafers during fabrication and for leak detection in vacuum systems. South Korea, home to SK hynix and Samsung's memory fabs, imported 64.7% of its helium from Qatar in 2025, and industry experts estimated the global semiconductor supply chain had only a two-week buffer before production impacts. Why it matters: 30% of global helium supply went offline in a single event, so semiconductor fabs that depend on helium for wafer cooling have no substitute material and face production halts within their 2-week inventory buffer, so memory chip production (DRAM, HBM, NAND) is disproportionately affected because memory fabs consume more helium per wafer than logic fabs, so HBM supply (already fully sold out through 2026) faces further tightening at exactly the moment AI demand requires exponential growth, so helium prices spike and the 4-6 month recovery timeline estimated by industry consultants means chip production costs rise for an extended period, so the semiconductor industry's $600+ billion annual revenue is vulnerable to a commodity gas supply disruption in a geopolitically unstable region (the Strait of Hormuz) that also affects aluminum and LNG supply. The structural root cause is that helium is a byproduct of natural gas extraction (not produced intentionally), global production is concentrated in Qatar (30%), the U.S. (30%), Algeria, Russia, and Australia, and there is no synthetic alternative. The U.S. Federal Helium Reserve was privatized and largely sold off, semiconductor demand for helium is growing 15-20% annually, and no fab operator has invested in long-term strategic helium reserves because the gas was historically cheap and abundant.

The U.S. semiconductor industry faces a projected workforce shortfall of 67,000 workers (per SIA) to 300,000 workers (per McKinsey) by 2030, with annual demand for engineers expected to nearly double from 9,000 to 17,000 and technician demand doubling from 7,000 to 14,000 as early as 2025. New fabs being built under the CHIPS Act in Arizona, Ohio, New York, and Texas are competing for the same limited pool of semiconductor process engineers, equipment technicians, and cleanroom operators that does not exist in sufficient numbers. Why it matters: There are not enough trained semiconductor workers in the U.S., so CHIPS Act-funded fabs cannot ramp to full production on schedule even after construction completes, so the billions in federal subsidies fail to achieve their stated goal of domestic chip production capacity by the target dates, so the U.S. remains dependent on Asian manufacturing during the extended ramp-up period, so companies like TSMC Arizona resort to flying in hundreds of Taiwanese engineers on temporary visas (creating cultural friction and unsustainable staffing models), so the entire premise of semiconductor reshoring is undermined not by technology or capital but by the absence of human capital that takes 4-6 years to develop through university and on-the-job training programs. The structural root cause is that U.S. universities produce approximately 40,000 engineering graduates per year across all disciplines, but semiconductor manufacturing competes with higher-paying software, finance, and AI/ML roles for the same STEM talent. Semiconductor process engineering requires specialized knowledge (plasma physics, photolithography chemistry, materials science) that is not taught in standard EE/CS curricula, and the U.S. dismantled most of its semiconductor manufacturing training infrastructure when production moved offshore in the 1990s-2000s.

China is projected to control nearly 40% of global wafer fabrication capacity for legacy chips (28nm and above) by 2032, while the U.S. and Europe will account for just 10% and 3% respectively. Chinese foundries like SMIC, Hua Hong, and dozens of state-subsidized fabs are aggressively expanding mature-node production at below-market prices, while the CHIPS Act and EU Chips Act focus almost entirely on leading-edge nodes, leaving legacy chip manufacturing strategically unaddressed. Why it matters: 95% of chips in automobiles and the vast majority of chips in defense systems, medical devices, and industrial equipment are legacy-node chips at 28nm and above, so China's dominance in this segment gives Beijing potential coercive leverage over every country and industry that depends on these chips, so U.S. automakers (GM, Ford, Stellantis) and defense contractors (Lockheed Martin, Raytheon) face growing supply chain dependence on Chinese fabs for components with no domestic alternative, so China could replicate Russia's energy leverage playbook by undercutting global pricing to drive out competitors (UMC, GlobalFoundries, PSMC) and then restricting supply during geopolitical disputes, so the U.S. military's readiness and the resilience of critical infrastructure sectors from power grids to telecommunications are undermined by a structural dependency that current policy does not address. The structural root cause is that the 28nm node (called the 'forever node' in industry) is extremely cost-effective and unencumbered by export controls, so China can freely purchase DUV lithography equipment from ASML and build unlimited capacity. The CHIPS Act's guardrails and investment focus on leading-edge nodes (sub-5nm) left a policy vacuum for legacy chips, and U.S./European foundries cannot compete with China's state-subsidized pricing without equivalent government support that does not currently exist.

ASML Holding NV is the world's only manufacturer of extreme ultraviolet (EUV) lithography systems, which are required to produce every chip at 7nm and below. Each High-NA EUV machine costs over $380 million, contains more than 100,000 parts sourced from approximately 5,000 suppliers (including sole-source components like Zeiss optics and Cymer/Trumpf laser sources), and ASML can only produce approximately 50-60 EUV systems per year. No competitor exists or is expected to emerge before 2030 at the earliest. Why it matters: ASML is the sole EUV supplier, so every advanced chip produced by TSMC, Samsung, and Intel depends on ASML's production capacity and delivery schedule, so a single bottleneck at any of the 5,000 suppliers (particularly Zeiss for optics or Trumpf for laser sources) can delay machine shipments worth hundreds of millions of euros, so the entire global advanced semiconductor roadmap is paced by one company headquartered in Veldhoven, Netherlands with a population of 45,000, so any disruption (fire, geopolitical sanctions, pandemic-related supply issues) to ASML's Eindhoven campus would halt the expansion of advanced chip capacity worldwide for 12-24 months, so the $700 billion annual AI chip boom projected by industry analysts is physically constrained by how fast a single Dutch company can assemble machines the size of a school bus. The structural root cause is that EUV lithography required $10+ billion and 30 years of R&D across optics, laser physics, and precision engineering. The technology is so complex that it represents one of the most sophisticated machines ever built by humans, and the knowledge required to manufacture it is deeply embedded in ASML's workforce and supplier network. China's attempts to develop domestic EUV (through companies like Huawei-backed efforts) are estimated to be 15-20 years behind, and no government or company has been willing to make the multi-decade, multi-billion-dollar bet required to create a second source.

Building a semiconductor fab in the United States takes approximately 38 months compared to 19 months in Taiwan, and costs roughly 4-5x more due to higher labor costs, regulatory permitting delays, and the inability to construct 24/7. Despite the CHIPS Act allocating $52.7 billion and the 2025 'Building Chips in America' Act exempting key projects from NEPA environmental reviews, the construction timeline gap remains a fundamental competitiveness problem for U.S. semiconductor manufacturing. Why it matters: U.S. fabs take twice as long to build, so the $52.7 billion CHIPS Act investment yields production capacity years later than equivalent spending would in Asia, so companies like Micron (which delayed its $100B New York megafab first production from 2028 to 2030) and Intel (which pushed Ohio fab completion to 2026-2027) face years of lost revenue and market share during construction, so the U.S. share of global semiconductor manufacturing continues to decline during the critical build-out period, so America's goal of producing 20% of the world's leading-edge chips by 2030 becomes mathematically nearly impossible given current construction velocities, so the national security rationale for domestic chip production is undermined by the very regulations meant to protect public interests. The structural root cause is that U.S. construction is governed by a patchwork of federal (NEPA, Clean Air Act, Clean Water Act), state, and local permitting requirements that were not designed for the speed and scale of semiconductor manufacturing. Taiwan's government provides fast-track approvals within months, allows 24/7 construction, and coordinates infrastructure (power, water, roads) ahead of fab site selection, while U.S. fabs must navigate each of these independently, often facing 4-6 year NEPA review timelines even after the 2025 regulatory reforms.

Tokyo Ohka Kogyo (TOK) and JSR together control approximately 91% of the global photoresist market, and for EUV photoresists used in sub-7nm chips, Japan provides 100% of the world's supply. In November 2025, Japan's Ministry of Economy, Trade and Industry (METI) placed 12 core semiconductor materials including high-end ArF/EUV photoresists on its export control list, restricting supply to 42 Chinese companies, demonstrating that this monopoly can be and is being weaponized. Why it matters: Every advanced chip below 7nm (including all AI accelerators, leading-edge mobile processors, and HPC chips) requires EUV photoresist from Japanese suppliers, so any disruption to Japan's photoresist supply immediately halts production at TSMC, Samsung, and Intel foundries worldwide, so the entire advanced semiconductor supply chain has a single-country dependency that is not addressed by any CHIPS Act or EU Chips Act diversification effort, so chipmakers like TSMC and Samsung have zero negotiating leverage on pricing or supply priority for this critical input, so the global semiconductor industry's $700+ billion annual revenue depends on a handful of chemical plants in Japan that could be disrupted by natural disaster (Japan sits on the Pacific Ring of Fire), trade disputes, or policy shifts. The structural root cause is that EUV photoresist requires extraordinarily precise chemical formulations at the molecular level, Japan holds 70% of related patents globally, and the development cycle for qualifying a new photoresist chemistry takes 5-7 years of co-development with lithography tool makers and foundries. No country or company outside Japan has successfully replicated production-grade EUV photoresist, and the capital investment plus IP barriers make new entry nearly impossible within this decade.

TSMC's Chip-on-Wafer-on-Substrate (CoWoS) advanced packaging capacity is fully sold out through 2026, with demand for CoWoS wafers projected to reach 1 million in 2026 against a maximum capacity of roughly 130,000 wafers per month (about 1.56 million annually). Every AI accelerator from NVIDIA, AMD, and custom ASICs from Google, Amazon, and Microsoft requires CoWoS packaging, making it the single tightest chokepoint in the semiconductor stack. Why it matters: CoWoS capacity is oversubscribed, so NVIDIA and AMD cannot ship enough GPUs and AI accelerators to meet hyperscaler demand, so cloud providers like AWS, Azure, and GCP face 12-18 month lead times for AI server deployments (up from 6-12 months in early 2025), so enterprises building AI applications face GPU cloud cost increases of 40-300% depending on region, so the pace of global AI adoption and innovation is physically gated by a single packaging technology controlled by one company in one country, so any disruption to TSMC's Chiayi and Kaohsiung packaging facilities (earthquake, cross-strait tensions, power outage) would halt AI chip production worldwide with no alternative supplier able to fill the gap for 18-24 months. The structural root cause is that advanced chip packaging was historically treated as a low-value commodity step in semiconductor manufacturing, so neither TSMC nor the OSAT industry invested in scaling it proportionally to front-end wafer fab capacity. When AI demand exploded in 2023-2024, CoWoS capacity was at roughly 35,000 wafers per month, and even with TSMC nearly quadrupling output by end of 2026, the 2-3 year construction timeline for new packaging lines means supply cannot catch up to demand that is doubling annually.

LASIK surgery clinics routinely advertise prices as low as $299 or $695 per eye, but these advertised rates apply only to mild myopia corrections, and the actual cost for most patients reaches $2,000-$3,500 per eye after mandatory add-ons including wavefront-guided technology, femtosecond laser flap creation, post-operative medications, and enhancement procedures -- none of which are required to be disclosed in advertising. Why it matters: consumers make the decision to pursue LASIK based on deceptively low advertised prices, so they arrive for consultations already psychologically committed and are then upsold on 'premium' technology presented as medically necessary, so the actual average cost of $2,246 per eye ($4,492 for both eyes) is 3-7x the advertised price, so patients who cannot afford the real price either take on medical debt or proceed with the cheaper 'basic' option that may use older technology with higher complication rates, so the patients most harmed by bait-and-switch pricing are those least able to afford the premium procedure. The structural root cause is that LASIK is classified as an elective cosmetic procedure exempt from health insurance coverage, which means it operates in a cash-pay market with no insurer negotiating prices or requiring transparency, and the FTC has not applied its advertising substantiation doctrine specifically to LASIK pricing practices despite extensive consumer complaints about bait-and-switch tactics.

US school-based vision screening programs identify millions of children with potential vision problems annually, but fewer than half of children who fail these screenings ever receive a follow-up comprehensive eye exam -- and among Black and Hispanic students, follow-up rates are even lower, despite these groups having higher screening failure rates. Why it matters: children who fail vision screening but never receive follow-up care continue to sit in classrooms unable to see the board clearly, so their academic performance suffers measurably (studies show 0.1-0.2 standard deviation improvement in reading scores when children receive needed glasses), so teachers misidentify these children as having learning disabilities or behavioral problems, so these children are disproportionately tracked into remedial programs or disciplined for 'not paying attention,' so the educational achievement gap between low-income/minority students and their peers widens due to an entirely correctable physical condition. The structural root cause is that school vision screening mandates exist in most states but include no funding or infrastructure for follow-up care, transportation to appointments, or provision of glasses, creating a system that identifies problems but provides no pathway to solve them -- and schools with predominantly low-income students are least likely to have full-time nurses who can coordinate follow-up care.

Warby Parker's Virtual Vision Test -- which uses smartphone-based technology to renew existing eyeglass prescriptions for stable-vision adults -- is prohibited in 27+ US states including major markets like Arizona, Colorado, Connecticut, Georgia, Maryland, Michigan, Minnesota, New Jersey, Virginia, and Washington, because state optometry practice acts require in-person examinations for any prescription issuance or renewal. Why it matters: adults with stable prescriptions in these 27+ states cannot use convenient telehealth tools to renew prescriptions, so they must take time off work and travel to an in-person office for a visit that often confirms no change, so this creates a $100-$250 cost barrier plus 2-4 hours of lost productivity per renewal, so cost-sensitive consumers wear outdated prescriptions rather than paying for an unnecessary in-person visit, so the US lags behind countries like the UK in telehealth adoption for routine vision care. The structural root cause is that state optometry boards -- whose voting members are practicing optometrists who profit from in-person exams -- write the practice act regulations that define what constitutes a legally valid eye examination, and when companies like Eyebot (partnered with Zenni) introduced 90-second automated vision kiosks in late 2024, the American Optometric Association immediately filed complaints with the FTC and FDA to block the technology.

US optical retailers routinely upsell blue-light-filtering lens coatings at $50-$150 per pair, generating 75-85% gross margins on these add-ons, despite a 2023 Cochrane Database systematic review of 17 randomized controlled trials finding no evidence that blue light filtering reduces eye strain, improves sleep quality, or protects against macular degeneration. Why it matters: consumers trust their optometrist's recommendation and pay $50-$150 extra per pair for a coating that provides no proven clinical benefit, so with 164 million corrective lens wearers in the US purchasing glasses every 1-3 years, the aggregate consumer overspend on blue light coatings reaches hundreds of millions annually, so this erodes trust in legitimate optical recommendations when consumers eventually learn the evidence does not support the claims, so patients may become skeptical of genuinely beneficial coatings like anti-reflective treatment, so the broader optical industry's credibility as healthcare providers is undermined. The structural root cause is that the FDA does not regulate eyewear lens coatings or their marketing claims (unlike drugs or medical devices), allowing manufacturers and retailers to make vague wellness claims without clinical substantiation, and optometrists face no professional sanction for recommending unproven add-ons because blue light coatings are classified as consumer products rather than medical interventions.

Only 44-50% of the 37 million Americans with diabetes adhere to recommended annual dilated eye exams, despite diabetic retinopathy (DR) being the leading cause of blindness in working-age adults and despite treatments that can reduce severe vision loss by up to 94% when caught early. Why it matters: more than half of diabetic patients skip screening, so one quarter of diabetic inpatients have previously undiagnosed retinopathy discovered incidentally during hospitalization, so vision-threatening DR progresses silently until irreversible damage occurs, so 1.84 million Americans already have vision-threatening DR as of 2021, so the US healthcare system spends billions treating advanced DR that could have been prevented with a $50-$150 screening exam. The structural root cause is that diabetic eye screening requires a separate visit to an eye specialist (not the primary care physician managing the diabetes), creating a fragmented care pathway where the referring physician has no mechanism to ensure follow-through, and the populations most affected -- Black (38.8% DR prevalence), Hispanic (31.0%), and low-income communities -- face the greatest barriers to specialist access including transportation, time off work, and lack of vision insurance.

Global childhood myopia prevalence surged from 25.3% (2001-2010) to 35.8% (2020-2023) -- a 42% increase in roughly 15 years -- driven by increased screen time and reduced outdoor activity, yet the evidence-based treatments that can slow myopia progression in children (low-dose atropine drops, orthokeratology lenses, specialized multifocal contact lenses) cost $1,000-$2,250 per year and are categorically excluded from vision insurance coverage in the United States. Why it matters: parents must pay $1,000-$2,250 annually out of pocket per child for myopia management, so lower-income families cannot access these treatments, so their children's myopia progresses unchecked to high myopia (-6.00 diopters or worse), so high myopia dramatically increases lifetime risk of retinal detachment (5-10x), glaucoma (2-3x), and myopic maculopathy, so these children face preventable blindness as adults -- a burden that falls disproportionately on families who could not afford $100/month for atropine drops during childhood. The structural root cause is that myopia management is classified as 'elective' rather than 'medically necessary' by US insurers because the treatments prevent future pathology rather than treating current disease, and the FDA has been slow to approve myopia-specific indications for atropine (still used off-label) which gives insurers justification to exclude coverage.

Approximately 63% of Americans -- roughly 210 million people -- have no vision insurance, and among the 80+ million Medicaid enrollees, 14.6 million (27%) live in states that provide no coverage for eyeglasses, while 6.5 million (12%) have no coverage for even routine eye exams. Why it matters: uninsured adults pay $485+ out of pocket for a basic eye exam and glasses, so for a single adult at the federal poverty level ($15,060/year), one pair of glasses costs more than a third of one month's income, so 25% of adults report forgoing routine eye care due to cost, so preventable conditions like glaucoma and diabetic retinopathy go undiagnosed, so these individuals eventually present to emergency rooms with advanced vision loss that is far more expensive to treat and often irreversible. The structural root cause is that the Affordable Care Act mandated pediatric vision coverage as an Essential Health Benefit but explicitly excluded adult vision care, and vision has historically been siloed from medical insurance as a separate 'voluntary' benefit category, meaning employers can choose not to offer it and state Medicaid programs can exclude it without federal penalty.

The FDA classifies contact lenses as Class II or III medical devices, and 38 US states mandate contact lens prescriptions expire after just one year, forcing 45 million American contact lens wearers to pay $100-$250 annually for a mandatory exam before they can repurchase the same lenses they have been wearing without issues. Why it matters: 45 million wearers pay $100-$250 each per year for exams that frequently result in zero prescription changes, so the aggregate annual cost to consumers is $3.6-$7 billion including exam fees and lost work time, so cost-sensitive consumers stretch their lenses beyond recommended replacement schedules creating actual infection risk, so the mandatory exam becomes both the barrier that causes the very harm it claims to prevent and a profit center for optometrists, so patients in rural areas with long travel times to optometrists face disproportionate burdens. The structural root cause is that the Fairness to Contact Lens Consumers Act (15 U.S.C. ch. 102) sets a federal floor of one-year expiration and state optometry boards -- controlled by practicing optometrists who profit from mandatory exams -- have resisted extending prescription validity despite evidence from countries like the UK and Japan where longer prescription periods have not led to worse outcomes.

Only four US states (Kansas, Massachusetts, Alaska, and New Mexico) mandate that optometrists include pupillary distance (PD) measurements on eyeglass prescriptions, meaning patients in the other 46 states cannot order glasses online without separately obtaining their PD -- a measurement their doctor already took but is not required to share. Why it matters: consumers who want to buy cheaper glasses online from Zenni ($69 median), Warby Parker, or other competitors cannot do so without PD, so they are effectively locked into buying from the prescribing optometrist's in-house optical shop at 3-5x higher prices, so the 164 million Americans who need corrective lenses overpay by hundreds of dollars per purchase, so low-income consumers either buy ill-fitting glasses with self-measured PD or forgo online savings entirely, so the online eyewear market that could provide real price competition remains artificially suppressed. The structural root cause is that optometrists derive 40-60% of practice revenue from optical sales, creating a financial incentive to withhold PD measurements, and state optometry boards -- staffed by practicing optometrists -- set the rules governing what must be included on prescriptions, creating a regulatory capture situation where the regulated parties write their own rules.

EssilorLuxottica controls the entire eyewear value chain -- manufacturing frames that cost $10-$30 to produce and selling them for $250-$500+, while simultaneously owning the retail stores (LensCrafters, Sunglass Hut, Pearle Vision, Target Optical), the insurance plan (EyeMed), and the brands (Ray-Ban, Oakley, Persol, Oliver Peoples, plus 20+ licensed brands including Prada, Chanel, Armani, and Burberry). Why it matters: consumers pay 1,000% markups on frames, so the average American spends $200-$600 per pair of glasses, so the 164 million Americans who wear corrective lenses face a cumulative cost burden of tens of billions annually, so low-income households forgo or delay corrective eyewear, so children and adults with uncorrected vision suffer in school performance, workplace productivity, and driving safety. The structural root cause is that EssilorLuxottica (EUR 26.5 billion in 2024 revenue) operates the only vertically integrated business model in eyewear, owning 600+ factories, 128 distribution centers, 13,500+ company-owned stores, and 4,100 franchise locations, which means no competitor can match their end-to-end control, and regulators have not intervened because eyewear is classified as a consumer product rather than a medical necessity subject to price controls.

Ohio's Public Library Fund (PLF), which distributes state tax revenue to 251 library systems, saw a $27 million shortfall in 2024, and the state Senate's FY2026-2027 proposal would reduce the PLF by approximately $25 million compared to the House plan. Simultaneously, in multiple communities across the United States where organized book challenge campaigns have been active, library levy and tax ballot initiatives -- which historically passed with strong majorities -- are being voted down, creating a compounding funding crisis where political controversy directly reduces the financial resources available for library operations. Why it matters: libraries that face both state funding cuts and failed local levies must reduce hours, lay off staff, and eliminate programming simultaneously, so communities lose critical services like children's story time, homework help, adult literacy programs, and summer reading, so library usage declines as services deteriorate, creating a downward spiral that makes future levies even harder to pass, so in the most affected communities the public library system faces potential closure or consolidation with neighboring districts, so entire neighborhoods -- disproportionately low-income -- lose their only free access point for books, internet, meeting space, and community programming. The structural root cause is that the U.S. public library funding model requires continuous political approval at both the state and local level with no guaranteed baseline, making libraries uniquely vulnerable to the interaction between culture-war politicization (which energizes opposition voters) and fiscal austerity (which reduces state revenue sharing), a combination that no other public institution -- not police, fire, or schools -- faces to the same degree.

Public libraries nationwide report that staff are increasingly overwhelmed by patrons experiencing homelessness, mental health crises, and substance use disorders -- needs that exceed librarians' training and professional scope. While some systems like the Salt Lake City Public Library have added full-time social workers, and two-thirds of urban library leaders surveyed in 2024 identified community partnerships as their primary strategy, the vast majority of the roughly 9,000 U.S. public library systems have no social worker on staff and no formal protocol for connecting patrons to housing, mental health, or addiction services. Why it matters: library frontline staff without social work training are forced to perform crisis intervention, de-escalation, and case management daily, so staff experience secondary trauma, compassion fatigue, and burnout at accelerating rates, so experienced staff leave the profession and new graduates avoid public-facing library roles, so remaining staff cannot maintain core library services while also managing social crises, so libraries in high-need urban areas become less welcoming to all patrons and the community perception shifts from 'learning center' to 'shelter of last resort.' The structural root cause is that the systematic defunding of U.S. public mental health infrastructure since the 1980s deinstitutionalization movement, combined with chronic underinvestment in affordable housing, has made public libraries one of the few remaining free, open-to-all indoor public spaces -- absorbing social service demand that no other institution will fund or staff.

According to NTIA data, 1 in 5 U.S. households lacks home internet access, with rural broadband adoption at 73% compared to 86% in suburban areas. Public libraries have become critical last-mile internet providers: 47% now operate hotspot lending programs (up significantly from 2020), and the FCC updated E-Rate in 2024 to cover off-site hotspot lending. However, the $42.45 billion BEAD broadband deployment program faces implementation delays, and the proposed elimination of IMLS threatens the federal grants that funded many library hotspot programs in the first place. Why it matters: libraries are asked to fill the broadband gap but receive no dedicated, sustained funding to do so, so hotspot lending programs depend on one-time grants that expire without renewal, so libraries in the communities with the worst connectivity gaps cannot sustain the very programs those communities need most, so residents without home internet cannot complete job applications, access telehealth, attend remote school, or participate in increasingly digital government services, so the digital divide becomes a permanent barrier to economic mobility for rural and low-income Americans. The structural root cause is that U.S. broadband policy treats internet access as a market commodity rather than a public utility, so when the market fails to serve low-density or low-income areas, the burden falls on libraries as an unfunded mandate with no permanent federal mechanism to sustain their role as public internet providers.

On May 25, 2024, the Seattle Public Library suffered a ransomware attack that disabled interlibrary loans, in-building Wi-Fi, public computers, hold placements, and physical media returns for months. In April 2025, the Pierce County Library System notified over 340,000 patrons and employees that hackers had stolen names, dates of birth, Social Security numbers, driver's license numbers, passport numbers, and financial information. As of 2025, 60% of libraries report attempted breaches in the past year, yet more than two-thirds have no documented data breach response plan and publish no privacy-protection guidance for patrons. Why it matters: libraries collect and store sensitive patron data (borrowing histories, personal identification, payment information) but lack dedicated cybersecurity staff or budgets, so successful attacks expose hundreds of thousands of people's private information, so compromised patron trust discourages community members from using library services, so libraries that serve as de facto public internet providers for vulnerable populations become vectors for identity theft rather than digital safety nets, so the public trust model that allows libraries to function as community institutions is undermined. The structural root cause is that library IT budgets are typically a small fraction of overall operating costs with no dedicated cybersecurity line item, librarians receive minimal cybersecurity training, and the decentralized nature of the U.S. library system (16,000+ independent systems) means there is no centralized security infrastructure, threat intelligence sharing, or mandatory compliance standards comparable to what exists in healthcare (HIPAA) or finance.

The 2024 PLA Annual Survey found that library director and beginning librarian salaries have not kept pace with inflation, representing a real-wage decrease from 2021 to 2024. Nearly one in three city libraries (29.1%) reported losing staff positions in the 12 months prior to the survey. Meanwhile, the Bureau of Labor Statistics reports the median annual wage for librarians was $64,320 in May 2024 -- a position that typically requires a master's degree in library science (MLIS), a credential that costs $30,000-$80,000 and takes 1-2 years to complete. Why it matters: the compensation gap between what librarians earn and what comparably-credentialed professionals earn in other fields makes recruitment increasingly difficult, so libraries operate understaffed and remaining staff absorb additional workload, so burnout and emotional exhaustion drive further turnover in a self-reinforcing cycle, so service hours are cut, programs are cancelled, and patron wait times increase, so communities that rely most heavily on library services experience the steepest service degradation. The structural root cause is that public library compensation is set by municipal or county pay scales that do not distinguish between the specialized graduate education librarians require and other civil service positions, while library budgets -- overwhelmingly funded by local taxes -- cannot compete with private-sector salaries for information professionals.

In 2024, the ALA tracked 821 attempts to censor library materials involving 2,452 unique titles -- the third-highest year since tracking began in 1990. Critically, 72% of these challenges were initiated by organized pressure groups and government entities, not individual parents (who accounted for only 16%). Beyond formal challenges, a growing 'soft censorship' phenomenon sees librarians preemptively removing or restricting books, declining to purchase controversial titles, or hiding materials from displays to avoid confrontation. Why it matters: librarians operating under sustained political pressure begin self-censoring collection decisions out of fear for their professional livelihood and personal safety, so library collections silently narrow without any formal challenge or public process, so patrons -- especially LGBTQ+ youth and readers of color -- lose access to representative literature without knowing it was ever available, so the chilling effect propagates to publishers and authors who anticipate reduced library sales for certain topics, so the intellectual freedom that distinguishes public libraries from commercial bookstores is eroded from within. The structural root cause is that library funding in the United States ultimately depends on local political approval (ballot measures, city council votes, board appointments), giving organized political groups outsized leverage to threaten librarians' budgets, employment, and institutional survival if collections include materials those groups oppose.

A December 2025 GAO report surveying approximately 16,400 public libraries found that an estimated 6,200 libraries (39%) have deferred maintenance backlogs exceeding $100,000 each, with 70% of all libraries expecting their backlogs to persist or increase over the next three years. About 2,700 libraries operate in buildings constructed before 1924, and approximately 7,000 report physical barriers to disability access. Why it matters: deteriorating HVAC systems, asbestos, and structural failures create unsafe conditions for the roughly 1.3 billion annual library visits in the United States, so libraries must divert operating budgets to emergency repairs rather than collections and programming, so patron experience degrades and usage declines in precisely the communities that need library services most, so the resulting cycle of declining usage and declining investment makes it harder to pass bond measures or secure capital funding, so public libraries in low-income and rural areas face existential threats from infrastructure failure alone. The structural root cause is that public library capital funding depends primarily on local property taxes and bond measures, which systematically disadvantage communities with lower property values, while no federal capital program exists specifically for library construction or renovation.