The typical MRI scanner demands a lot. Reinforced floors to bear its weight. Liquid helium to cool its superconducting magnets. Electromagnetic shielding thick enough to block radio interference. Dedicated HVAC systems to manage the heat. Construction timelines stretching across months, sometimes nearly a year. And that's before you factor in the equipment itself—anywhere from $1 million to $3 million for a standard system, with total project costs commonly reaching several million depending on site requirements.
What you get, in the end, is a diagnostic tool trapped by its own infrastructure. Powerful, yes. Clinically essential, certainly. But inaccessible to most clinics and far removed from the moments when early detection matters most.
That calculus is starting to shift. A cohort of companies—armed with new physics, sealed magnets, and AI reconstruction algorithms—is pulling MRI out of hospital basements and into mobile units, specialty clinics, even physicians' offices. The newest entrant is Adialante, a Redwood City startup fresh from Y Combinator's Summer 2026 batch. The company claims it has redesigned MRI "from the ground up" using what's known as RF-encoding physics. According to Adialante, its system weighs 80% less than conventional machines, runs 40% shorter, and operates quietly enough that you wouldn't need the soundproofing standard in traditional MRI suites. No specialized infrastructure required.
Six clinics have signed on. The company says it has secured $12.75 million in letters of intent, along with a $30,000 deposit from what it describes as a top-10 urology clinic. The first system is scheduled to scan by June 16, 2026—Adialante's demo day.
Whether it delivers remains an open question. But the company's pitch lands at a moment when the constraints that have defined MRI for decades are finally starting to give way.
The Infrastructure Problem
MRI technology has advanced incrementally over the years, but the fundamental bottlenecks have proven stubborn. A conventional installation still demands four to nine months from purchase order to first scan—factoring in design, construction, RF shielding (which alone can stretch across several days or weeks), and vendor installation. Helium, a non-renewable resource prone to supply shocks, has long been essential for cooling the superconducting magnets that generate MRI's powerful magnetic fields. Qatar outages in 2026 heightened interest in sealed and low-helium MRI platforms, raising operational risks for facilities dependent on legacy systems. The US Federal Helium System assets were privatized in June 2024, layering on further uncertainty.
Meanwhile, diagnostic backlogs persist. Scotland recorded a record number of patients waiting for CT and MRI scans at the end of 2025. NHS England missed interim diagnostic targets in September 2025, with the Nuffield Trust tracking roughly 1.8 million people waiting for diagnostic tests in January 2026—about 17% of that activity being MRI. In Canada, wait times for diagnostic imaging remained longer than pre-pandemic levels through 2024, despite progress on backlogs. A pattern of systemic capacity strain, in other words. The OECD's most recent Health at a Glance report, published in November 2023, underscored persistent variation in MRI unit density and exam volumes across countries—a gap that newer technologies are now trying to close.
Three Converging Forces
Three trends are reshaping the landscape, each reinforcing the others: helium-free magnet designs, AI-powered image reconstruction, and a push toward point-of-care deployment.
Start with helium. The major equipment manufacturers have all committed to reducing or eliminating it. Philips has installed more than 2,000 of its BlueSeal 1.5T systems worldwide, which use just 7 liters of sealed helium compared to the 1,500-plus liters in traditional magnets. At RSNA 2024, the company unveiled BlueSeal Horizon, billing it as "the world's first helium-free 3.0T platform." Siemens has deployed its MAGNETOM Free.Max 0.55T system, which uses less than 1 liter of helium, eliminates the quench pipe, and simplifies siting. GE HealthCare announced its "Freelium" sealed-magnet platform in 2025, touting less than 1% of conventional helium use. Fujifilm launched its ECHELON ZeroHelium 1.5T systems in Japan and the UK during 2024 and 2025.
These designs address operational risk and sustainability. But they haven't fundamentally changed the weight, footprint, or installation requirements.
That's where AI reconstruction comes in. Philips received FDA 510(k) clearance in 2025 for its SmartSpeed Precise "dual-AI" reconstruction software, which the company says can triple scan speed while sharpening images. GE has expanded its AIR Recon DL platform across its fleet. Clinical commentators at RSNA have noted that deep learning substantially improves speed and detail in musculoskeletal MRI, though whether it generalizes across all anatomies remains an open question. The critical point: faster, higher-quality scans from lower signal-to-noise regimes—whether from low-field magnets or novel encoding schemes—become clinically viable when AI can fill in the gaps.
The third force is decentralization. Hyperfine's portable 0.064T brain scanner, FDA-cleared and marketed under the Swoop brand, has expanded into neurology offices with its Optive AI software. The system received CDSCO approval in India in December 2025 and secured CE and UKCA marks in April 2026. Promaxo obtained 510(k) clearance in March 2021 for an office-based system used for transperineal prostate biopsy guidance at sites like Mississippi Urology.
Mobile fleets are proliferating, too. Siemens' Viato.Mobile units have been deployed in UK NHS trusts to reduce appointment backlogs during Community Diagnostic Centre construction. Mobile imaging service providers in the US—such as AIMI and Alliance Medical—cite asset-light capacity strategies and backlog reduction as core offerings. Third-party market estimates peg the US mobile medical imaging services market at around $425 million in 2023, projected to reach $549 million by 2029, though such figures warrant triangulation.
The RF-Encoding Bet

Adialante represents a different wager: not incremental improvement on existing architectures, but a redesign using RF-encoding physics.
The technology traces back to research at the University of Minnesota's Center for Magnetic Resonance Research, where co-founder Efraín Torres completed his PhD and published a series of papers on what's called "frequency-modulated Rabi-encoded echoes" (FREE) and related methods. The technique aims to reduce dependence on conventional B0 gradient coils—the heavy, noisy, power-hungry components that create spatial encoding in standard MRI. By using radiofrequency pulses to encode spatial information instead, the system can theoretically shrink hardware, cut noise, and eliminate specialized infrastructure.
Adialante's two issued US patents—12,372,595 B2 (July 29, 2025) on multi-echo RF-based spatial encoding, and 12,625,209 B2 (published May 12, 2026) on RF-phase-encoded MRI with FM pulses—form the IP backbone. The company says its scanner requires 50% less hardware than conventional systems and runs 40% shorter and 80% lighter. Co-founder Parker Jenkins, who holds a master's degree from UMN and was featured in a March 2026 university news article on translating RF-encoding to accessible MRI, leads imaging hardware and AI development.
The business model is mobile, clinic-operated service. Adialante owns and operates the scanners, charging clinics a flat per-scan fee. Clinics keep the insurance reimbursement; Adialante captures recurring service revenue without requiring upfront capital from the clinic.
The company is starting with prostate imaging—a logical entry point. The American Urological Association and Society of Urologic Oncology's 2023 Early Detection guidelines endorse pre-biopsy MRI in biopsy-naïve patients to increase detection of clinically significant cancer, with pooled sensitivity around 0.91 and specificity around 0.37. NICE guidance in the UK similarly recommends multiparametric MRI as first-line for suspected localized prostate cancer. These are diagnostic indications, not population screening, which means reimbursement pathways exist.
Adialante's roadmap extends to breast MRI for screening, orthopedic and musculoskeletal imaging, renal, and brain. The breast screening market is more complex, though. The US Preventive Services Task Force, in its April 30, 2024 final recommendation, maintains an "I" (insufficient evidence) statement for supplemental MRI in women with dense breasts at average risk. However, the MQSA density notification rule, which began enforcement September 10, 2024, requires all US mammography facilities to inform patients and providers about breast density—potentially increasing demand for supplemental imaging discussions. For high-risk women, annual MRI is recommended by the American College of Radiology (May 2023), and trial data like EA1141, with results presented in 2023 and summary updated in 2025, showed higher cancer detection rates with MRI than digital breast tomosynthesis in women with dense breasts. The evidence base is still maturing. Reimbursement remains variable.
Precedents and Parallels
Hyperfine's trajectory offers a useful comparison. The company has focused on portable brain imaging at 0.064T—far below the field strength of diagnostic systems—positioning its Swoop device for ICU, ER, and now neurology office use. It has carved out a niche in point-of-care neuroimaging but hasn't yet challenged the diagnostic throughput of radiology departments. Promaxo, with its office-based prostate system, similarly targets a specific procedural use case rather than general diagnostic imaging.
Adialante's pitch is broader: diagnostic-grade imaging across multiple anatomies, delivered in a mobile system light enough to bypass the infrastructure barriers that have confined MRI to hospital basements.
The company has secured early traction—six clinics signed, $12.75 million in LOIs, and a $30,000 deposit from a major urology clinic. But it has yet to publicly detail clinical performance benchmarks, regulatory submissions, or human diagnostic validation at scale. The first system is set to scan by mid-June 2026. That will provide an initial read on image quality and workflow feasibility, perhaps more than the founders anticipated.
What Comes Next

The push toward accessible MRI is no longer speculative. Multiple pathways are converging: sealed magnets that cut helium dependence, AI reconstruction that enables faster scans from lower-field or gradient-light systems, and a regulatory and reimbursement landscape that increasingly supports point-of-care and mobile deployment.
The question is which architectures win in which clinical contexts.
For radiology departments managing high volumes, incumbent vendors' helium-free platforms at 1.5T and 3.0T—Philips BlueSeal Horizon, GE Freelium, Siemens' expanded mid-field offerings—will likely dominate. These systems maintain field strength and signal-to-noise ratio while reducing operational risk.
For specialty clinics and screening programs where capital and space are constraints, lighter, mobile systems like Adialante's could unlock new workflows. If the RF-encoding physics deliver on diagnostic quality, the "imaging clinic on wheels" model could change where and how early detection happens—particularly for cancers like prostate, where MRI-first pathways are already guideline-supported.
There are risks, naturally. Clinical validation remains the central hurdle. RF-encoded MRI has a peer-reviewed pedigree stretching back to 2018, but no system built on these principles has yet achieved widespread diagnostic deployment. Regulatory approval—likely via FDA's 510(k) or De Novo pathway—will require demonstrated substantial equivalence or novel safety and efficacy data. Reimbursement must follow; while prostate diagnostic MRI has established CPT codes, breast screening MRI for average-risk women does not yet have USPSTF endorsement, which payers typically follow.
And incumbents are not standing still. Philips, GE, Siemens, and Fujifilm are all rolling out designs that address siting and helium concerns, even if they don't match Adialante's claimed weight and footprint reductions.
Yet the direction is clear. MRI is moving out of hospital basements. Whether it's mobile fleets serving NHS trusts, portable scanners in neurology offices, or RF-encoded systems targeting urology clinics, the technology is being pulled toward the point of care.
Adialante's demo-day scans in mid-June will offer an early signal of whether physics-driven redesign can deliver on the promise of "diagnostic-grade MRI, within reach." If it can, the implications extend beyond a single startup—pointing to a future where early cancer detection no longer depends on infrastructure that most clinics can't afford.
