← Archive

The Pediatric Translation Gap: Why a $14B Sector Still Spends Single Digits on Children

US digital health closed 2025 with $14.2B in venture funding, yet pediatric companies still receive a single-digit share. Five structural mechanisms compound against pediatric pure-plays. The viable pediatric digital health business in 2026 is increasingly not a pediatric digital health business.

A $14 Billion Headline With a 5% Carve-Out

Children represent 21.7% of the US population, roughly 73 million people under 18. By every metric that matters to a venture-funded business they are a rounding error in healthcare innovation. Rock Health's 2023 pediatric market map, the most comprehensive public series, found that pediatric-focused startups averaged 1.7% of annual digital health funding from 2011 through 2019 and 5.1% from 2020 through Q3 2023 (see Chart 1 below). Through the first three quarters of 2023, pediatric companies raised $413.6 million across 20 deals, 4.7% of sector funding. Rock Health has not published a refreshed pediatric breakdown in its 2024 or 2025 year-end reports, and the firm's mapping of where 2024–2025 capital actually went — namely clinical workflow, AI infrastructure, and obesity — is consistent with pediatric exposure being incidental at best.

The same pattern shows up across the stack. A JAMA Pediatrics analysis of 553 new drugs and 918 indications between 2011 and 2023 found 25.2% carried pediatric labeling, but only 18.7% of nonorphan indications did. The FDA's own Report to Congress on Premarket Approval of Pediatric Uses of Devices found that across FY2008 through FY2023 only 29% of PMA and HDE approvals each year carried any pediatric indication, and CDRH's chief medical officer for pediatrics has publicly noted that fewer than 10% of approvals over the past decade have actually been labeled for under-18 populations, with only 3 to 4% for neonates.

AI is where this matters most going forward. Of 876 AI/ML-enabled devices authorized by the FDA from 1995 through March 2024, only about 17% were explicitly labeled for pediatric use, and only about 20% of those incorporated pediatric data into training or validation. In calendar 2024, 30 of 168 ML-enabled device authorizations carried a pediatric indication. NIH's RCDC system reports the "Pediatric" category at roughly $6.2 billion in FY2024, 12.8% of the $47.3 billion budget. Pediatric clinical trial registrations remain dominated by small, single-site, non-industry-funded studies. Across drugs, devices, AI/ML SaMD, NIH funding, and venture capital, pediatric representation runs roughly half to one-third of the population share, and AI deployment is widening the gap rather than narrowing it.

The variance in those statistics is itself the point (Chart 2 below). The 25% drug-labeling figure includes any indication touching pediatrics, often through extrapolation from adult studies under PREA, not products designed for children. The 29% PMA average masks the fact that many pediatric labels apply to 18-to-21-year-olds. The honest numbers are smaller than the headline numbers, in every direction.

Why the Gap Exists, and Why It Compounds

The argument is not that demand is missing. Pediatric clinical need is large, well-documented, and growing. The argument is that capital, regulation, evidence, reimbursement, and investor structure each independently select against pediatric work, and that they compound. Five mechanisms.

First, the payer mix. Per KFF, Medicaid and CHIP cover roughly 35.1 million children, about 48% of total Medicaid and CHIP enrollment, and Medicaid covers approximately 8 in 10 children living in poverty. Medicaid reimburses roughly 70 to 80% of Medicare rates and substantially less than commercial in most states. For an investor underwriting LTV from PMPM or fee-for-service economics, the same digital intervention generates materially less revenue when delivered to a child than to an adult, and the largest single payer of pediatric care has the least capacity to pay for innovation. This is the same dynamic our April 10 deep dive on OBBBA Medicaid exposure argued for the maternal cohort, applied one developmental stage further down.

Second, regulatory complexity. PREA can require pediatric studies for adult drug approvals, but the act's own deferral and waiver data show the majority of required pediatric studies get deferred until after adult approval, often for years. Pediatric trials require assent on top of parental consent, IRB oversight tuned to vulnerable populations, age-stratified PK/PD work, and pediatric-formulation development. For digital therapeutics and SaMD, age-of-dataset becomes a binding constraint: AI models validated on adult imaging, EHR, or behavioral data cannot be relabeled into pediatric indications without new clinical work.

Third, evidence economics. Pediatric populations are smaller, more developmentally heterogeneous, and fragmented across a relatively small set of academic medical centers. PEDSnet aggregates data across roughly a dozen children's hospitals: meaningful, but a fraction of the data scale available for adult populations through claims aggregators or national registries. A cross-sectional analysis of the FDA AI/ML device list found that age information was not reported at all in 81.6% of devices. The infrastructure that made adult AI possible does not yet exist at adult scale for pediatrics.

Fourth, reimbursement pathways. The CPT code system was constructed around adult primary and specialty care, and pediatric-specific codes are sparse outside developmental screening, vaccinations, and select procedural areas. Coverage of digital therapeutics is the cleanest illustration: when Akili Interactive's EndeavorRx received de novo authorization in 2020 as the first prescription video game, commercial coverage stayed patchy and Medicaid coverage was effectively nonexistent. Highmark's 2022 decision to cover prescription digital therapeutics was treated as industry-defining precisely because it was an outlier. The Medicaid and CHIP Access to Prescription Digital Therapeutics Act has been reintroduced in successive Congresses without passage.

Fifth, capital concentration. Generalist health tech VCs underweight pediatric, and there are very few pediatric-specialist funds at meaningful scale. The 2024 mega-fund concentration documented by Rock Health — with Andreessen Horowitz, General Catalyst, and a handful of others driving the bulk of late-stage rounds — has compounded the issue, since those funds are explicitly underwriting platform-scale workflow, AI infrastructure, and obesity care. The partial substitutes are children's-hospital innovation arms: Boston Children's IDHA (portfolio companies have raised over $100 million collectively), Cincinnati Children's Innovation Ventures (FY23 portfolio exit value of $293.7 million), and similar offices at CHOP, Texas Children's, Seattle Children's, and Nationwide, which PHD has been mapping in our Children's Hospital Innovation Directory. They produce commercializable assets but operate at a fraction of the capital scale of generalist health tech VC, and their incentives are mission-aligned rather than fund-return-aligned, which limits how aggressively they can lead later rounds.

What the Three Test Cases Actually Tell Us

Three companies illustrate how the mechanisms collide in practice, and they are the same three the field keeps coming back to.

Brightline is the cautionary tale for pediatric pure-plays underwritten on commercial PMPM. KKR led a $105 million Series C in 2021 at a reported $705 million valuation against $212 million raised across rounds. By November 2022 the company announced its first 20% workforce reduction. By May 2023 it cut another 20%. In September 2024 it exited 45 of the 50 states it operated in, halted new patient intake, and pivoted to a hybrid clinic model anchored on Medicaid contracts (including a $680 million California Medicaid arrangement with Kooth). Brightline did not fail because pediatric mental health demand was missing. It failed because an employer-channel virtual-care model premised on commercial PMPM economics did not generate enough referral density at scale, and because the Medicaid pivot it has now executed requires a fundamentally different operating model than the one the original capital structure was sized for. Mechanisms 1, 4, and 5, in compound.

Akili Interactive is the reimbursement story in its purest form. EndeavorRx received the first FDA de novo authorization for a prescription video-game digital therapeutic in 2020. Akili went public via SPAC in 2022 at a valuation reported near $1 billion, received a Nasdaq delisting notice in October 2023, cut roughly 40% of staff in September 2023, launched a non-prescription EndeavorOTC version to bypass payer dependence, and was acquired by Virtual Therapeutics in July 2024 for approximately $34 million, or about $0.43 per share. The science was not the problem. EndeavorRx priced at roughly $99 per month against fragmented commercial coverage and effectively no Medicaid coverage, which produced quarterly revenues in the low six figures against operating costs sized for a Nasdaq-listed pharma-style company. Mechanism 4, doing all the work.

Maven Clinic is the structural answer rather than a pediatric pure-play. Maven's October 2024 Series F closed $125 million at a $1.7 billion valuation, taking total raised past $425 million, with about 17 million covered lives via employer and health-plan contracts. Founded as a women's health company, Maven layered fertility, maternity, postpartum, parenting, and pediatrics onto one continuous platform. It captures commercially-insured economics across a consumer's reproductive arc rather than relying on the lower-yield pediatric-only window. Lyra Health's July 2025 acquisition of Bend Health tells the same story from a different direction: bolt youth mental health onto an adult-platform parent valued north of $5 billion. The viable pediatric digital health business in 2026 is increasingly not a pediatric digital health business. It is a maternal-child-family business that includes pediatrics, or an adult-platform that buys pediatric capability when it needs it. Summer Health, with $19.1 million raised on a $45-per-month direct-pay subscription model, is the rare scaling pediatric pure-play, and it has gotten there by routing around the insurance problem entirely.

The comparison that sharpens the analysis is women's health. Per Rock Health, 2024 women's+ health funding reached $671 million, 6.6% of total digital health funding, the highest share since 2021. Women's+ health has had an inflection. Pediatric, on the same series, has not. The difference is instructive: women's+ benefited from a consumer-payer rail (employer fertility benefits, OOP wellness spend) that produced commercial unit economics, plus a recognizable workforce of women founders and women GPs whose deal flow concentrated capital. Pediatric has neither, and the consumer-payer rail is constrained because the consumer is a parent purchasing on behalf of a child.

What Actually Closes the Gap (and What Does Not)

A 24-to-36-month outlook follows from the structure rather than from anyone's preferred theory of change.

The AI gap is the most durable consequence and the one PHD readers should weight heaviest. With roughly 17% of FDA-cleared AI/ML devices carrying pediatric labels and only about 20% of those validated with pediatric data, the next generation of clinical decision support, ambient documentation, imaging interpretation, and risk stratification will be substantially better-validated for adults at the moment of mass deployment. Pediatric clinicians will face a familiar choice (off-label use of adult-validated tools, or no tool) precisely as those tools get integrated into health-system AI procurement. Closing this requires pediatric-specific datasets at adult scale, which neither PEDSnet, individual children's hospital data assets, nor existing claims aggregators currently provide. A federally-coordinated pediatric data infrastructure investment would move the needle here. It is conspicuously absent from the current AI policy conversation.

The children's-hospital-as-innovator model is likely to expand as a partial substitute for VC. Boston Children's IDHA, Cincinnati Children's Innovation Ventures, CHOP's Office of Entrepreneurship and Innovation, and Texas Children's commercialization arm are increasingly active in seeding spinouts and licensing IP. More pediatric IP will originate inside hospital walls, with hospital ownership stakes, hospital data dependencies, and exit dynamics that look more like university tech transfer than Sand Hill Road. Strategic acquirers (adult-platform incumbents like Lyra, maternal-platform parents like Maven) will increasingly serve as the realistic exit path, at modest multiples.

The policy interventions that have actually moved the needle are the ones that compelled adult-program sponsors to do pediatric work: PREA, BPCA, the RACE Act for pediatric oncology. The interventions that depended on independent pediatric demand-side economics (pediatric digital therapeutic Medicaid coverage, dedicated CMS pediatric innovation pathways) have not. Forecasted improvement in the next 24 to 36 months should weight toward extrapolation-and-mandate mechanisms. Policy in the Akili shape (reimbursement-floor legislation that depends on Medicaid agencies opting in) is unlikely to clear in the current Congress. Policy in the PREA shape (requiring adult-program sponsors to do pediatric work as a condition of approval) is the lever that has historically worked.

The maternal-child continuum bet is, on present evidence, the dominant strategy for capital that wants pediatric exposure without pediatric-only economics. Companies serving the full fertility-to-pediatric arc capture commercially-insured economics during reproductive years that subsidize lower-yield pediatric service, and they de-risk the small-addressable-market problem by bundling. PHD has been making this argument in pieces across the OBBBA exposure work (April 10, April 24). The translation gap is the bow on top of the same parcel.

The sharper version of the call: do not underwrite a pediatric pure-play on the assumption that capital will eventually correct the gap. The gap is the equilibrium. Underwrite either to a continuum business model that captures commercial economics elsewhere on the arc, to a direct-pay subscription that routes around the insurance problem, or to a children's-hospital-origination strategy that exits to an adult-platform parent. The companies that have scaled in 2025 and 2026 are doing one of those three things. The ones that aren't are restructuring.