Global AI in Medical Imaging and Radiology Market Outlook to 2032

Executive Summary

The global AI in medical imaging and radiology market — defined as the full value chain of AI algorithms applied to radiology, pathology, and emerging clinical imaging across diagnostic CT, MRI, X-ray, ultrasound, mammography, PET/SPECT, OCT, and emerging point-of-care imaging, plus AI orchestration platforms, foundation models for medical imaging, enterprise imaging informatics, integration services, plus first-year operations and support — is estimated at approximately US$5.5 billion in 2025 and is projected to reach approximately US$35 billion by 2032, expanding at a CAGR of 29–30 percent over the forecast period. The market sits at the intersection of healthcare, AI software, and enterprise imaging informatics — and represents the fastest-growing healthcare AI category with the strongest regulatory tailwind.

Three forces define the trajectory through 2032. First, FDA authorizations of AI-enabled medical devices crossed 1,100 in 2025 — with radiology accounting for approximately 76 percent of all AI medical device authorisations (1,104 radiology devices out of 1,451 total cumulative by end-2025 per FDA's AI Medical Device List). The FDA authorized 168 AI/ML-enabled devices in 2024 (a record), and 2025 Q4 saw 72 clearances of which 55 (76 percent) were radiology. The structural inflection event was Aidoc's February 2025 FDA clearance of a rib fracture triage solution built on its CARE1 Foundation Model — the first FDA clearance of a foundation-model-powered clinical AI device. In January 2026, Aidoc received FDA clearance for a tool detecting 14 critical findings in a single abdominal CT scan (liver injury, spleen injury, bowel obstruction, appendicitis, plus 10 others) — extending the foundation-model precedent. Second, the radiologist shortage crisis is the principal demand-side driver: imaging volumes are growing approximately 5 percent annually while radiology residency positions grow only 2 percent — radiology case loads grew 25 percent between 2018 and early 2025. The US is forecast to face a 17,000–42,000 shortage of radiologists, pathologists, and psychiatrists by 2033; UK consultant shortage is forecast at 40 percent by 2028; the average US radiologist salary reached US$571,000 in 2025 (up 9 percent year-on-year). AI deployment is the principal mitigation: 48 percent of radiologists report using AI in practice in 2024 surveys, with AI delivering 30–50 percent faster reporting, 30–75 percent scan time reductions, plus 40 percent radiology workflow step reductions in deployed platforms (INSTINX, DeepHealth, plus emerging enterprise platforms). Third, the big-tech-plus-medical-imaging-OEM partnership wave reached commercial scale at RSNA 2025: NVIDIA-GE HealthCare partnership (March 2025) targets autonomous X-ray and ultrasound powered by NVIDIA Isaac for Healthcare; NVIDIA-Philips partnership (May 2025) targets 2026 commercial introduction of AI Orchestrator generative AI capabilities; Microsoft-Siemens Healthineers collaboration for AI-based healthcare cloud infrastructure; plus emerging Anthropic, OpenAI, and Google DeepMind partnerships position big-tech AI infrastructure as the structural layer underneath enterprise imaging deployment.

For investors, hospital systems, radiology groups, medical imaging OEMs, and policymakers, the implication is that AI medical imaging has crossed from FDA-clearance theater to enterprise deployment in 2025 — but the structural reimbursement gap remains the principal commercial constraint. Only approximately 10 of the 1,100+ FDA-cleared AI medical devices have CMS payment coverage (as of January 2026); 26 CPT codes exist for clinical AI solutions but most remain Category III (experimental); the proposed Clinically Meaningful Algorithmic Analyses (CMAA) coding classification (under AMA discussion December 2025) plus CMS Transitional Coverage for Emerging Technologies (TCET, finalized 2024) provide the structural pathway. The 2026–2028 period is the decisive window for reimbursement framework maturation, foundation model commercialisation across CT-MRI-X-ray-ultrasound, plus the radiology AI workflow integration that determines whether AI compresses radiologist time-per-study or expands radiologist productive capacity.

Market Overview

Definition and Scope

This report scopes the global AI in medical imaging market as the full value chain of artificial intelligence applied to clinical imaging — AI algorithms for radiology (CT, MRI, X-ray, ultrasound, mammography, PET/SPECT, OCT, fluoroscopy), AI for pathology (digital pathology slide analysis), AI for dermatology (skin lesion analysis), AI for ophthalmology (retinal imaging, diabetic retinopathy screening), AI for cardiology imaging (echo, cardiac MRI/CT analysis), plus AI orchestration platforms (Aidoc aiOS, DeepHealth, Bayer Calantic, Blackford Analysis, plus emerging multi-vendor orchestrators), foundation models for medical imaging (CARE1, plus emerging Microsoft, Google, NVIDIA, and OEM foundation models), enterprise imaging informatics integration (PACS, VNA, RIS integration), AI cloud infrastructure (NVIDIA Isaac for Healthcare, Microsoft Azure AI Health, Google Cloud Vertex AI for Healthcare), plus integration services, training, support, and first-year operations.

The scope includes the AI software and platform value pool — algorithms, foundation models, orchestration platforms, integration services — but excludes the underlying imaging hardware (CT scanners, MRI systems, X-ray equipment, ultrasound machines) which is covered in the broader medical devices market. The scope captures both standalone AI vendor revenue (Aidoc, Viz.ai, Annalise.ai, Lunit, Qure.ai, plus emerging) and the AI software embedded in OEM platforms (GE HealthCare, Siemens Healthineers, Philips, Canon, United Imaging) plus emerging foundation model platform revenue (NVIDIA BioNeMo Healthcare, Microsoft, Google).

Evolution and Genesis

The AI medical imaging market evolved through four structurally distinct phases since 2015. The 2015–2018 academic-research-and-first-FDA-clearance phase was anchored by the Stanford-led CheXNet (2017) chest X-ray deep learning paper, Google DeepMind's diabetic retinopathy work, plus early FDA clearances (IDx-DR for diabetic retinopathy, the first AI-only diagnostic device, approved April 2018). Cumulative FDA AI/ML device authorisations remained under 50 through 2018, with the technology classified as experimental.

The 2019–2022 commercial-scaling phase saw rapid FDA clearance growth — from approximately 33 cumulative AI/ML devices through end-2015, to 95 by end-2019, to 343 by end-2022. The COVID-19 pandemic plus the resulting backlog in imaging studies plus AI-enabled triage for COVID-related lung pathology accelerated commercial deployment. Aidoc, Viz.ai, Annalise.ai, Lunit, Qure.ai, plus emerging vendors collectively raised over US$1.5 billion in cumulative venture funding through 2022. The Tempus AI IPO (June 2024, raising US$410.7 million at US$6.42 billion market cap) plus emerging public market entries validated the commercial pathway.

The 2023–2024 enterprise-deployment phase saw the structural shift from algorithm-only sales to enterprise platform deployment. Aidoc's aiOS platform unifying systems, teams, and data across radiology plus cardiology plus neurovascular plus vascular specialties reached deployment at over 1,300 hospitals globally. Viz.ai's stroke detection platform deployed across 1,600 hospitals with 13 cleared algorithms. The FDA authorized a record 168 AI/ML-enabled devices in 2024 — including 121 radiology devices. The big-tech entry began: NVIDIA Clara plus emerging Microsoft Azure Healthcare AI plus Google Health partnerships with major OEMs.

The 2025-onward foundation-model and integration phase is the current era. Three structural events define the new phase: (a) Aidoc's February 2025 FDA clearance of a rib fracture triage solution built on the CARE1 Foundation Model — the first FDA clearance of a foundation-model-powered clinical AI device, plus the subsequent January 2026 clearance of a 14-finding abdominal CT triage built on the same foundation model, (b) NVIDIA-GE HealthCare partnership announced March 2025 for autonomous X-ray and ultrasound powered by NVIDIA Isaac for Healthcare plus NVIDIA-Philips partnership announced May 2025 targeting 2026 commercial introduction, and (c) the FDA AI authorizations crossing 1,100 in 2025 with radiology maintaining 75–76 percent share. The foundation-model-and-integration phase is structurally consequential because it converts AI medical imaging from single-algorithm-per-clearance to multi-task-foundation-model deployment — fundamentally changing the commercial economics.

Key Market Drivers

• Radiologist shortage crisis driving structural AI deployment. Imaging volumes grow approximately 5 percent annually while radiology residency positions grow only 2 percent. Radiology case loads grew 25 percent between 2018 and early 2025. US is forecast to face 17,000–42,000 shortage of radiologists, pathologists, and psychiatrists by 2033. UK consultant shortage is forecast at 40 percent by 2028. Average US radiologist salary reached US$571,000 in 2025 (up 9 percent year-on-year). AI deployment is the principal mitigation — 48 percent of radiologists use AI in practice.
• FDA AI medical device authorizations crossing 1,100 in 2025. FDA has authorized approximately 1,451 cumulative AI-enabled medical devices through end-2025, with 1,104 radiology devices (76 percent of all AI medical authorisations). Annual authorisations reached 168 in 2024 (record). Radiology accounted for 75–76 percent of 2025 authorizations.
• Foundation model inflection via Aidoc CARE1 plus big-tech partnerships. Aidoc CARE1 Foundation Model received first-ever FDA clearance February 2025 (rib fracture triage), with subsequent January 2026 clearance of 14-finding abdominal CT triage. NVIDIA-GE HealthCare partnership (March 2025) plus NVIDIA-Philips (May 2025) plus Microsoft-Siemens Healthineers plus emerging Google DeepMind, Anthropic, and OpenAI partnerships position foundation models as the structural infrastructure layer.
• Big OEM AI suite consolidation at RSNA 2025. GE HealthCare, Siemens Healthineers, Philips, and Canon Medical Systems all launched enhanced AI suites at RSNA 2025 (Chicago, November 2025) — signalling accelerated R&D for radiology AI platforms. GE HealthCare retains the most FDA AI authorizations (120 including acquisitions), followed by Siemens Healthineers (89), Philips (50), Canon (45), United Imaging (38).

Macroeconomic and Regulatory Context

The market is operating against a structurally supportive but fragmented regulatory and reimbursement framework: FDA AI/ML medical device approval pathway (510(k), De Novo, Premarket Approval — with the FDA's Predetermined Change Control Plan (PCCP) guidance supporting iterative AI model updates without re-submission); EU CE marking under the Medical Device Regulation (MDR, in force May 2021) plus the EU AI Act (in force August 2024) classifying healthcare AI as "high-risk" requiring conformity assessment; UK MHRA AI Software as a Medical Device (SaMD) Change Programme; CMS Transitional Coverage for Emerging Technologies (TCET, finalized 2024) creating expedited Medicare coverage pathway for FDA Breakthrough Devices; emerging AMA Category I/Category III CPT codes for AI services (26 codes as of January 2026); plus the proposed AMA Clinically Meaningful Algorithmic Analyses (CMAA) coding classification (under discussion December 2025) for algorithm-based services.

The macroeconomic backdrop is structurally supportive but commercially fragmented. The radiologist shortage crisis provides demand-side support — 48 percent of radiologists use AI in practice but only 19 percent of pilot or deployed AI use cases reported "high" success per 2025 surveys. The reimbursement gap remains the principal commercial constraint — only approximately 10 of the 1,100+ FDA-cleared AI medical devices have CMS payment coverage. CMS issued AI-supplemented diagnostic codes (0877T-0880T) in the 2025 update — representing the most significant augmentation in CMS-certified health services codes (130 codes added, 23 deleted). The forward implication is that AI medical imaging commercial success increasingly depends on (a) enterprise platform deployment economics (vendor sells platform subscription, hospital realises productivity gains, reimbursement secondary), and (b) emerging Category I CPT code expansion plus CMS TCET pathway acceleration that brings reimbursement parity with FDA clearance.

Market Size & Growth Outlook

The market grew from approximately US$0.8 billion in 2020 to approximately US$4.0 billion in 2024 — a 5× expansion in four years driven by FDA clearance acceleration (138 cumulative authorisations end-2020 to 950 cumulative end-2024), enterprise platform deployment (Aidoc reaching 1,300+ hospitals, Viz.ai reaching 1,600+), plus emerging big-tech partnerships. The 2023 acceleration (55.6 percent year-on-year value growth) reflects three forces: (a) FDA authorizations crossing 343 (2022) to 650 (2023) — nearly doubling in a single year, (b) Tempus AI IPO June 2024 raising US$410.7 million and validating the commercial AI healthcare pathway, and (c) the broader AI infrastructure investment wave benefiting healthcare AI categories.

The 2025 expansion to US$5.5 billion (37.5 percent year-on-year value growth) is anchored by three structural catalysts. First, the Aidoc CARE1 Foundation Model FDA clearance (February 2025) — the first FDA clearance of a foundation-model-powered clinical AI device, fundamentally changing the regulatory and commercial pathway for multi-task AI systems. Second, the big-tech-plus-OEM partnership wave reaching commercial scale: NVIDIA-GE HealthCare announcement March 2025, NVIDIA-Philips announcement May 2025, Microsoft-Siemens Healthineers collaboration deepening, plus emerging Google DeepMind and Anthropic partnerships. Third, the FDA authorizations crossing 1,100 in 2025 (1,451 cumulative including all medical devices) with radiology maintaining 75–76 percent share.

The forecast CAGR of 29–30 percent through 2032 anchors on three drivers. The first is continued FDA authorization growth: cumulative AI medical device authorizations are forecast to reach approximately 4,900 by 2032 (versus 1,451 at end-2025) — a 3.4× expansion driven by foundation model multi-task clearances plus emerging specialty expansion (cardiology, ophthalmology, pathology, dermatology). The second driver is enterprise platform deployment scaling: from approximately 30–35 percent of US hospitals using AI medical imaging in 2025 to approximately 75–85 percent by 2030 — driven by radiologist shortage urgency plus emerging reimbursement framework maturation plus enterprise platform value validation. The third driver is foundation model commercialisation across modalities — Aidoc CARE1 model expansion from rib fracture (2025) to 14-finding abdominal CT (January 2026) to broader CT-MRI-X-ray multi-task deployment through 2028 plus emerging OEM foundation models (GE HealthCare-NVIDIA, Philips-NVIDIA, Siemens Healthineers-Microsoft) collectively converts single-algorithm clearance economics into platform-multi-task economics.

Cumulative investment over the 2025–2032 window (AI biotech VC plus public-equity funding plus big-tech AI infrastructure investment plus OEM internal R&D plus enterprise deployment capex) is forecast at approximately US$140–175 billion across the AI medical imaging value chain — anchored by approximately US$75–95 billion in cumulative enterprise deployment value, plus approximately US$35–45 billion in cumulative AI biotech and platform vendor revenue, plus approximately US$25–35 billion in cumulative OEM AI R&D plus big-tech healthcare AI infrastructure investment. This investment magnitude reconciles to approximately 8× the average annual market size in the forecast window — consistent with high-growth platform infrastructure plus the parallel foundation model commercialisation. The implication for stakeholders is that AI medical imaging has structurally transitioned from "interesting technology" to "operationally-deployed enterprise infrastructure" — but reimbursement framework maturation plus enterprise platform consolidation will determine which AI biotechs plus OEMs plus big-tech platforms capture the value pool.

Market Segmentation

By Modality

CT dominates AI deployment value share (approximately 35 percent of 2025 deployment) because it is the highest-volume cross-sectional modality plus the principal trauma and emergency imaging modality. Aidoc's CARE1 Foundation Model rib fracture clearance (February 2025) plus the subsequent 14-finding abdominal CT triage clearance (January 2026) anchor the multi-task CT AI deployment. Viz.ai stroke detection plus pulmonary embolism plus intracranial hemorrhage AI plus emerging abdominal triage plus oncology staging collectively position CT as the platform modality for AI deployment. The share is forecast to grow slightly to 37 percent by 2030 — driven by foundation model multi-task expansion across CT applications.

X-ray (chest, musculoskeletal) represents the second-largest modality category (approximately 18 percent of 2025 deployment) — Annalise.ai's 124-findings chest X-ray model plus Aidoc CARE1 rib fracture plus Qure.ai tuberculosis screening (deployed across over 90 countries for global health applications) plus emerging fracture detection AI collectively dominate this modality. The relatively high X-ray AI deployment reflects the high volume of X-ray studies (approximately 4-5× the volume of CT) plus the lower cost-per-study making AI economics favorable. Mammography (12 percent of 2025 deployment) is anchored by DeepHealth 3D mammography AI (21 percent increase in cancer detection rate) plus Lunit Insight MMG plus iCAD ProFound AI plus emerging Mirai breast cancer risk prediction.

By AI Application Type

Triage and worklist prioritization dominates current AI application value share (approximately 32 percent in 2025) because triage AI addresses the highest clinical-value-per-deployed-algorithm: critical findings (stroke, hemorrhage, pulmonary embolism, aortic dissection) require expedited radiologist review, and AI-driven prioritization materially compresses time-to-treatment. Viz.ai's stroke platform showed 66-minute faster treatment with improved stroke detection. The share is forecast to compress modestly to 27 percent by 2032 as foundation model multi-task platforms grow.

The fastest-growing application category is foundation model multi-task platforms — forecast to grow from 2 percent share in 2025 to approximately 6 percent by 2032 (a 3× share expansion). Aidoc CARE1 Foundation Model (rib fracture February 2025, 14-finding abdominal CT January 2026, with broader CT-MRI-X-ray expansion roadmap) plus emerging GE HealthCare-NVIDIA Isaac for Healthcare plus Philips-NVIDIA AI Orchestrator generative AI plus Microsoft-Siemens Healthineers AI-Pathway Companion plus emerging Google DeepMind, Anthropic, and OpenAI healthcare partnerships collectively position foundation models as the structural multi-task replacement for single-algorithm clearances. The strategic implication: by 2030, approximately 40–50 percent of FDA AI medical imaging clearances are forecast to be foundation-model-derived (versus approximately 2 percent in 2025) — fundamentally changing the regulatory and commercial pathway.

By Geography

The United States dominates global AI medical imaging deployment (approximately 51 percent of 2025 market value) because of the largest single radiology spend pool, the largest medical imaging install base, the FDA-leading authorization pathway, plus the deepest healthcare AI startup ecosystem. The share is forecast to compress modestly to 47 percent by 2032 as Chinese and Asia-Pacific deployment grows materially.

China is the fastest-growing region (forecast 8 percent to approximately 12 percent share by 2032 — a 1.5× share expansion). Chinese AI medical imaging is anchored by Infervision (founded 2015, NVIDIA-backed, multi-modality CT-X-ray-MRI platform deployed across 1,000+ Chinese hospitals), Yitu Healthcare (CT-based oncology and pulmonary AI), United Imaging AI (integrated imaging hardware + AI), plus emerging Tencent Healthcare AI plus Alibaba DAMO Academy healthcare AI initiatives. The Chinese NMPA AI medical device review framework (updated 2024) plus emerging dedicated state funding for healthcare AI plus the structural scale of Chinese healthcare (1.4 billion population with rapidly expanding imaging access) position China as the principal non-Western AI medical imaging growth opportunity.

By Vendor Archetype

The competitive landscape is structurally organised into seven archetypes. First, the Big OEM Imaging vendors (GE HealthCare, Siemens Healthineers, Philips, Canon Medical Systems, United Imaging, Fujifilm) collectively account for approximately 38 percent of 2025 AI deployment value — anchored by FDA AI clearance leadership (GE HealthCare 120, Siemens 89, Philips 50, Canon 45, United Imaging 38) plus bundling AI with hardware sales plus integrated workflow plus partnerships with big-tech AI infrastructure (NVIDIA-GE March 2025, NVIDIA-Philips May 2025, Microsoft-Siemens). Second, the Standalone AI Biotech vendors (Aidoc with US$420 million raised plus 1,300+ hospitals plus US$60 million 2025 revenue, Viz.ai with 1,600+ hospitals plus 13 stroke algorithms, Annalise.ai with 124-finding chest X-ray model, Tempus with NASDAQ June 2024 IPO at US$6.42 billion market cap) collectively account for approximately 27 percent — anchored by multi-vendor independence plus AI orchestration platforms plus foundation model leadership.

Aidoc's strategic posture as the leading standalone AI biotech combines aiOS platform deployment at 1,300+ hospitals globally with the CARE1 Foundation Model leadership. The February 2025 FDA clearance of a rib fracture triage solution built on CARE1 represented the first FDA clearance of a foundation-model-powered clinical AI device — a structural inflection event. The subsequent January 2026 clearance of a 14-finding abdominal CT triage built on the same foundation model extends the precedent. Aidoc has raised approximately US$420 million through 2025 (latest US$40 million Line of Credit July 2025); 2025 revenue reached approximately US$60 million. The strategic question: does Aidoc maintain foundation model leadership against big-tech-OEM partnerships (NVIDIA-GE-Philips collectively bringing structural scale) plus emerging entrants?

Viz.ai's strategic posture combines stroke and neurocritical care platform leadership (13 cleared algorithms plus 1,600+ hospital deployment) with emerging broader cardiovascular and oncology expansion. Viz.ai's stroke detection algorithm achieved AUC over 0.90 on retrospective datasets, with deployment showing 66-minute faster treatment. The forward question: does Viz.ai expand from stroke specialty to broader multi-specialty platform, or maintain stroke leadership while accepting broader-platform competition?

Tempus AI's strategic posture is uniquely positioned at the intersection of genomics and medical imaging AI. The June 2024 NASDAQ IPO raised US$410.7 million at US$6.42 billion market cap. Q2 2024 revenue of US$166 million (+25 percent YoY) with Genomics segment at US$112.3 million and Data & Services at US$53.6 million (+32.5 percent YoY) demonstrated the integrated genomics-imaging-AI business model. Tempus partners with Google DeepMind for emerging multi-modal AI plus extensive pharma partnerships for clinical trial enrichment.

The big-tech-healthcare-AI category (NVIDIA, Microsoft, Google, AWS, Anthropic) at approximately 6 percent of 2025 deployment value represents the structurally most-consequential emerging archetype. NVIDIA's Isaac for Healthcare plus BioNeMo Healthcare plus the GE HealthCare and Philips partnerships position NVIDIA as the foundational AI infrastructure layer. Microsoft's Siemens Healthineers cloud collaboration plus Precision Imaging (formerly Nuance acquired October 2021) plus emerging Azure AI Health integration provide cloud-anchored AI infrastructure. Google DeepMind's medical imaging research plus Verily plus Med-PaLM healthcare LLM provide differentiated AI capabilities. The structural implication: big-tech AI infrastructure grows from 6 percent to approximately 12 percent of 2032 deployment value — but primarily as infrastructure underneath OEM and standalone AI biotech platforms rather than as direct competitor.

The cautionary cases that anchor industry execution risk include: (a) only 19 percent of pilot or deployed AI use cases reporting "high" success per 2025 surveys, signalling that AI deployment lags initial expectations, (b) only approximately 10 of 1,100+ FDA-cleared AI medical devices having CMS payment coverage, creating structural reimbursement gap, (c) the radiologist salary increase to US$571,000 in 2025 (+9 percent YoY) signalling that AI deployment is not yet compressing radiologist compensation despite productivity gains, (d) multiple AI medical imaging companies (Caption Health, Subtle Medical, Imagia Cybernetics, plus others) facing financial stress or restructuring through 2023–2025, and (e) the historical Geoffrey Hinton 2016 prediction that radiologists would be obsolete by 2021 — which has demonstrably not materialised.

Trends & Developments

Aidoc CARE1 Foundation Model as Industry First (February 2025)

Aidoc's February 2025 FDA clearance of a rib fracture triage solution built on the CARE1 Foundation Model represented the first FDA clearance of a foundation-model-powered clinical AI device. The CARE1 architecture trains a single multi-modal foundation model on a broad range of radiology data, enabling task-specific fine-tuning across multiple clinical applications. The subsequent January 2026 clearance of a 14-finding abdominal CT triage solution built on the same CARE1 foundation model (detecting liver injury, spleen injury, bowel obstruction, appendicitis, plus 10 other critical findings) extends the precedent — demonstrating that foundation models can support multiple FDA-cleared task-specific applications without requiring full retraining for each. The structural implication: by 2030, approximately 40–50 percent of FDA AI medical imaging clearances are forecast to be foundation-model-derived (versus approximately 2 percent in 2025) — fundamentally changing the regulatory pathway, commercial economics, plus capability roadmap of the AI medical imaging category.

Radiologist Shortage Crisis Driving Structural Deployment

The radiologist shortage is the principal demand-side driver of AI medical imaging deployment. Imaging volumes grow approximately 5 percent annually while radiology residency positions grow only 2 percent — radiology case loads grew 25 percent between 2018 and early 2025. The US is forecast to face a 17,000–42,000 shortage of radiologists, pathologists, and psychiatrists by 2033; the UK consultant shortage is forecast at 40 percent by 2028. Average US radiologist salary reached US$571,000 in 2025 (up 9 percent year-on-year) — signalling that demand for radiologist time materially exceeds supply. AI deployment is the principal mitigation: 48 percent of radiologists report using AI in practice in 2024 surveys, with AI delivering 30–50 percent faster reporting, 30–75 percent scan time reductions, plus 40 percent radiology workflow step reductions in deployed platforms (INSTINX, DeepHealth, plus emerging enterprise platforms). The forward implication: AI medical imaging deployment continues structurally regardless of macroeconomic cycles because the radiologist shortage is itself structural.

Big-Tech-Plus-OEM Partnership Wave at RSNA 2025

The RSNA 2025 conference (Chicago, November 2025) showcased the structural convergence of big-tech AI infrastructure with OEM imaging platforms. NVIDIA-GE HealthCare partnership (announced March 2025) targets autonomous X-ray and ultrasound powered by NVIDIA Isaac for Healthcare plus synthetic data simulation — addressing radiology technologist shortages plus emerging global healthcare access expansion. NVIDIA-Philips partnership (announced May 2025) targets 2026 commercial introduction of AI Orchestrator generative AI capabilities. Microsoft-Siemens Healthineers collaboration for AI-based healthcare cloud infrastructure plus the Siemens AI-Pathway Companion decision-support system (launched 2024) connecting imaging data with clinical, molecular, and laboratory information. The collective implication: big-tech AI infrastructure is structurally embedding into OEM platforms — providing foundation model plus cloud plus simulation infrastructure underneath OEM hardware-plus-AI deployment.

FDA AI Authorizations Crossing 1,100 in 2025

FDA AI/ML-enabled medical device authorizations crossed 1,100 in 2025 (1,451 cumulative authorisations including all categories by end-2025), with radiology accounting for 76 percent of total. The FDA authorized 168 AI/ML-enabled devices in 2024 (a record), and 2025 Q4 saw 72 clearances of which 55 (76 percent) were radiology. Median approval time for radiology devices was 146 days — significantly shorter than devices reviewed under other panels. GE HealthCare retains the top spot with 120 radiology AI authorizations (including acquisitions), followed by Siemens Healthineers at 89, Philips at 50, Canon at 45, United Imaging at 38, Aidoc at 30. The structural implication: FDA pathway has reached operational maturity for AI medical imaging — clearance is no longer the binding constraint on deployment, replaced by reimbursement plus enterprise integration plus clinical workflow validation.

CMS Reimbursement Gap and CMAA Coding Proposal

Despite over 1,100 FDA-cleared AI medical devices, only approximately 10 have CMS payment coverage as of January 2026 — creating the structural reimbursement gap that constrains commercial deployment economics. CMS issued 26 CPT codes for clinical AI solutions as of January 2026, with three AI solutions receiving Category I CPT codes (FFR-CT for cardiovascular risk prediction, a diabetic retinopathy detection tool, plus one additional); all other AI solutions with designated CPT codes are Category III (experimental). The AMA discussed coding and payment for clinical AI tools in December 2025, considering a new coding classification tentatively titled Clinically Meaningful Algorithmic Analyses (CMAA) codes for algorithm-based services. CMS issued AI-supplemented diagnostic codes 0877T-0880T in the 2025 update — the most significant augmentation in CMS-certified health services codes (130 codes added, 23 deleted). CMS Transitional Coverage for Emerging Technologies (TCET, finalized 2024) creates expedited Medicare coverage pathway for FDA Breakthrough Devices, with CMS aiming to finalize coverage determinations within six months of FDA authorization. The forward implication: reimbursement framework maturation through 2026–2028 is the principal commercial unlock for AI medical imaging.

Foundation Model Multi-Task Platforms versus Single-Algorithm Clearances

The 2025–2032 structural shift from single-algorithm clearances toward foundation-model multi-task platforms fundamentally changes the AI medical imaging commercial economics. Aidoc CARE1 (rib fracture February 2025, 14-finding abdominal CT January 2026, with broader CT-MRI-X-ray expansion roadmap) plus emerging GE HealthCare-NVIDIA Isaac for Healthcare plus Philips-NVIDIA AI Orchestrator plus Microsoft-Siemens Healthineers plus emerging Google DeepMind Med-PaLM Imaging plus Anthropic healthcare-specific models collectively position foundation models as the multi-task replacement for single-algorithm deployment. The strategic implication: foundation model deployment compresses the per-algorithm marginal cost to near-zero (once the foundation model is trained, task-specific fine-tuning is materially cheaper) — fundamentally changing AI biotech business models toward platform subscription rather than per-algorithm licensing.

Competitive Landscape

The competitive landscape is structurally organised into seven archetypes (see segmentation §"By Vendor Archetype" above). The Big OEM Imaging vendors (GE HealthCare, Siemens Healthineers, Philips, Canon Medical Systems, United Imaging, Fujifilm) collectively control approximately 38 percent of 2025 AI deployment value — anchored by FDA AI clearance leadership plus bundling AI with hardware sales plus big-tech AI partnerships. Standalone AI biotechs (Aidoc, Viz.ai, Annalise.ai, Tempus) control approximately 27 percent. Specialty AI (Lunit, iCAD, Optellum, DeepHealth, Qure.ai) controls approximately 12 percent. Enterprise imaging orchestration platforms (Bayer Calantic, Blackford Analysis, Sectra AI Hub, Microsoft Precision Imaging) control approximately 11 percent. Big-tech healthcare AI infrastructure (NVIDIA, Microsoft, Google, AWS, Anthropic) controls approximately 6 percent.

GE HealthCare's strategic posture is the most consequential among Big OEM Imaging vendors. The company retains the top spot by FDA AI authorizations (120 cumulative, including acquisitions of Caption Health 2023 plus emerging consolidation). The NVIDIA partnership (announced March 2025) targets autonomous X-ray and ultrasound systems powered by NVIDIA Isaac for Healthcare plus synthetic data simulation — addressing radiology technologist shortages plus emerging global healthcare access. GE HealthCare's RSNA 2025 platform launch (November 2025) showcased AI-forward imaging across CT, MRI, X-ray, ultrasound, and PET modalities. The strategic implication: GE HealthCare positions itself as the integrated AI-plus-hardware leader, leveraging NVIDIA AI infrastructure to differentiate from competitors.

Aidoc's strategic posture as the leading standalone AI biotech combines aiOS platform deployment at 1,300+ hospitals globally with the CARE1 Foundation Model leadership. The February 2025 FDA clearance of a rib fracture triage solution built on CARE1 was the first FDA clearance of a foundation-model-powered clinical AI device — a structural inflection event. The subsequent January 2026 clearance of a 14-finding abdominal CT triage built on the same foundation model demonstrates the multi-task expansion roadmap. Aidoc has raised approximately US$420 million through 2025 (latest US$40 million Line of Credit July 2025); 2025 revenue reached approximately US$60 million. The strategic question: does Aidoc maintain foundation model leadership against big-tech-OEM partnerships (NVIDIA-GE-Philips collectively bringing structural scale) plus emerging entrants? The 2026–2028 period is decisive.

Viz.ai's strategic posture combines stroke and neurocritical care platform leadership (13 cleared algorithms plus 1,600+ hospital deployment) with emerging broader cardiovascular and oncology expansion. Tempus AI's strategic posture is uniquely positioned at the intersection of genomics and medical imaging AI — the June 2024 NASDAQ IPO raised US$410.7 million at US$6.42 billion market cap, with Q2 2024 revenue of US$166 million (+25 percent YoY) demonstrating the integrated genomics-imaging-AI business model. Tempus partners with Google DeepMind for emerging multi-modal AI plus extensive pharma partnerships for clinical trial enrichment.

NVIDIA's strategic posture as the leading big-tech healthcare AI infrastructure provider combines Isaac for Healthcare (humanoid robot platform extending to medical imaging applications) plus BioNeMo Healthcare (foundation model platform) plus the GE HealthCare and Philips partnerships. NVIDIA captures structural value in the foundation model infrastructure layer — providing platform plus services revenue that scales as AI medical imaging deployments expand. Microsoft's strategic posture combines Azure Health AI cloud infrastructure plus the Nuance/Precision Imaging acquisition (October 2021 for US$19.7 billion, with Precision Imaging spun out as Microsoft's healthcare AI subsidiary) plus the Siemens Healthineers cloud collaboration. Google DeepMind's medical imaging research plus Verily plus Med-PaLM healthcare LLM provide differentiated AI capabilities but commercial deployment lags Microsoft and NVIDIA.

The cautionary cases anchor industry execution risk. Only 19 percent of pilot or deployed AI use cases reported "high" success per 2025 surveys — signalling that AI deployment lags initial expectations across enterprise integration. Caption Health was acquired by GE HealthCare in 2023 after facing commercial-scaling challenges as standalone ultrasound AI. Subtle Medical, Imagia Cybernetics, plus multiple emerging AI medical imaging companies have faced financial stress or restructuring through 2023–2025. The structural cautionary signal: only approximately 10 of 1,100+ FDA-cleared AI medical devices have CMS payment coverage, meaning that FDA clearance does not translate to commercial revenue without reimbursement framework maturation.

Challenges & Opportunities

Key Challenges

CMS reimbursement gap and Category I CPT code scarcity

Only approximately 10 of 1,100+ FDA-cleared AI medical devices have CMS payment coverage as of January 2026. Of 26 CPT codes for clinical AI solutions, only three are Category I (permanent, billable); all others are Category III (experimental, time-limited). The proposed Clinically Meaningful Algorithmic Analyses (CMAA) coding classification (under AMA discussion December 2025) plus CMS Transitional Coverage for Emerging Technologies (TCET, finalized 2024) provide the structural pathway — but full reimbursement framework maturation may extend through 2027–2028. The structural risk: AI medical imaging commercial revenue lags FDA clearance volume by approximately 60–80 percent through 2027.

Foundation model regulatory pathway uncertainty

Aidoc CARE1 received the first FDA foundation-model clearance in February 2025 — but the structural regulatory framework for foundation models with iterative updates remains under development. The FDA's Predetermined Change Control Plan (PCCP) guidance supports iterative AI model updates without re-submission, but the operational application to foundation models is still maturing. Emerging EU AI Act classification of healthcare AI as "high-risk" requiring conformity assessment plus the broader US AI policy uncertainty (December 2025 federal preemption EO, plus emerging state-level AI legislation) create structural regulatory complexity for foundation model commercialisation.

Enterprise deployment execution and 19 percent high-success rate

Only 19 percent of pilot or deployed AI use cases reported "high" success per 2025 surveys. The deployment challenge spans technical integration (PACS, RIS, EHR, plus radiologist workflow integration), clinical validation (algorithm performance in local patient population versus training data), economic justification (productivity gains versus subscription cost), plus radiologist behavior change. The structural implication: AI medical imaging deployment success requires enterprise platform sophistication plus clinical change management plus radiologist training — not just FDA clearance plus algorithm accuracy.

Algorithm performance generalisation and bias

AI medical imaging algorithms face structural generalisation challenges — algorithms trained on one population may perform less accurately on different demographic groups (race, sex, age), different equipment vendors (GE versus Siemens versus Philips), different imaging protocols, plus different clinical practice settings. Emerging research documents performance variance of 5–25 percent across populations for the same algorithm. The structural risk: deployed algorithms may underperform in real-world settings versus initial validation, eroding clinician trust plus creating liability exposure.

Key Opportunities

Foundation model multi-task platform commercialisation

Foundation model multi-task platforms are forecast to grow from 2 percent of 2025 deployment value to approximately 6 percent by 2032 (a 3× share expansion). Aidoc CARE1 plus emerging GE HealthCare-NVIDIA plus Philips-NVIDIA plus Microsoft-Siemens Healthineers plus emerging Google DeepMind and Anthropic foundation models collectively position multi-task platforms as the structural replacement for single-algorithm deployment. The cumulative foundation model investment opportunity through 2032 is approximately US$15–25 billion across foundation model development, fine-tuning infrastructure, plus deployment platforms.

Radiologist shortage opportunity at US$571,000 average salary

The structural radiologist shortage (US 17,000–42,000 by 2033, UK 40 percent consultant shortage by 2028) creates structural demand-side support for AI deployment. Average US radiologist salary at US$571,000 plus expected continued growth materially supports AI productivity gain economics — even a 10 percent productivity gain per radiologist generates approximately US$57,000 per year per radiologist in deployment value. The cumulative AI deployment opportunity from radiologist productivity gains through 2032 is approximately US$45–65 billion.

Big-tech-plus-OEM partnership infrastructure layer

NVIDIA-GE HealthCare-Philips plus Microsoft-Siemens Healthineers plus emerging Google DeepMind, Anthropic, AWS, OpenAI healthcare partnerships position big-tech AI infrastructure as the foundational layer underneath OEM and standalone AI biotech platforms. The cumulative big-tech healthcare AI infrastructure revenue opportunity through 2032 is approximately US$25–40 billion — primarily from NVIDIA hardware-plus-platform plus Microsoft cloud plus emerging entrants.

Emerging-market deployment expansion

Chinese AI medical imaging market grows from approximately 8 percent of 2025 global value to approximately 12 percent by 2032 — driven by Infervision, Yitu Healthcare, United Imaging AI, plus emerging Tencent Healthcare AI plus Alibaba DAMO Academy plus structural Chinese healthcare expansion. Indian AI medical imaging market (Qure.ai TB screening across 90+ countries, plus emerging deployment) provides emerging-market opportunity. Saudi Arabia Vision 2030 healthcare digitalisation plus emerging Middle Eastern deployment add incremental opportunity.

Key Policies & Regulatory Environment

FDA AI/ML Medical Device Authorization Pathway

The FDA's AI/ML-enabled medical device authorization pathway (510(k), De Novo, Premarket Approval) provides the structural regulatory framework. Cumulative authorizations crossed 1,451 in 2025 (1,104 radiology, 76 percent share). Annual authorizations reached 168 in 2024 (record), with 2025 Q4 at 72 clearances. Median radiology AI approval time is 146 days — significantly shorter than other medical device categories. The FDA's Predetermined Change Control Plan (PCCP) guidance supports iterative AI model updates without re-submission. The Aidoc CARE1 Foundation Model clearance (February 2025) established the regulatory precedent for foundation-model-powered AI medical devices.

CMS Transitional Coverage for Emerging Technologies (TCET, Finalized 2024)

CMS TCET creates an expedited Medicare coverage pathway for FDA-designated Breakthrough Devices that fit into established benefit categories, with CMS aiming to finalize coverage determinations within six months of FDA authorization. The 2024 final notice addressed the structural gap between FDA clearance (1,100+ AI medical devices) and CMS coverage (approximately 10 covered devices). TCET implementation through 2026–2027 is the principal commercial unlock for AI medical imaging deployment economics.

EU AI Act and Medical Device Regulation

The EU AI Act (in force August 2024) classifies healthcare AI as "high-risk" requiring conformity assessment. The EU Medical Device Regulation (MDR, in force May 2021) governs medical device CE marking with the In Vitro Diagnostic Regulation (IVDR) for diagnostic devices. The combined AI Act-MDR-IVDR framework creates structural compliance complexity for AI medical imaging deployment in EU markets — but the structural deployment growth continues anchored by EU healthcare system AI adoption. The EU AI Office (established 2024) is the principal regulatory authority for AI Act enforcement.

AMA CPT Codes for Clinical AI

The AMA's CPT coding framework includes 26 codes for clinical AI solutions as of January 2026 — with three Category I codes (FFR-CT, diabetic retinopathy, plus one additional) and the remainder Category III. CMS issued AI-supplemented diagnostic codes 0877T-0880T in the 2025 update — the most significant augmentation in CMS-certified health services codes (130 codes added, 23 deleted). The proposed Clinically Meaningful Algorithmic Analyses (CMAA) coding classification (under AMA discussion December 2025) provides emerging framework for algorithm-based services. The forward implication: CPT code expansion plus Category I code conversion is the principal commercial unlock through 2028.

UK MHRA AI Software as a Medical Device (SaMD) Change Programme

The UK MHRA AI SaMD Change Programme (launched 2021, ongoing through 2026) provides regulatory framework for AI medical device deployment in UK markets. The MHRA AI Sandbox programme supports innovative AI testing prior to formal approval. The structural implication: UK is supportive but post-Brexit divergence from EU framework creates separate regulatory pathway.

China NMPA AI Medical Device Review

China NMPA AI medical device review framework (updated 2024) provides regulatory pathway for Chinese AI medical imaging deployment. Domestic Chinese AI medical imaging vendors (Infervision, Yitu Healthcare, United Imaging AI) plus emerging Tencent Healthcare AI plus Alibaba DAMO Academy operate under the NMPA framework. The 15th Five-Year Plan (2026-2030) is expected to expand healthcare AI policy support.

India CDSCO AI Medical Device Regulatory Framework

India's Central Drugs Standard Control Organisation (CDSCO) AI medical device regulatory framework (under development 2024-2026) provides emerging pathway for Indian AI medical imaging deployment. Qure.ai (Indian-founded, deployed across 90+ countries for TB screening) plus emerging Indian AI medical imaging vendors operate under the CDSCO framework.

Future Outlook

The global AI in medical imaging and radiology market is positioned for sustained 29–30 percent CAGR through 2032, reaching approximately US$35 billion in value with approximately 4,900 cumulative FDA AI medical device authorizations. The market has crossed from FDA-clearance theater to enterprise deployment in 2025 — anchored by FDA AI radiology authorizations crossing 1,100 (76 percent of all AI medical devices), Aidoc CARE1 foundation model FDA clearance February 2025, NVIDIA-GE HealthCare-Philips RSNA 2025 platform launches, and the structural radiologist shortage crisis driving deployment. The forecast structure is three-phased: a 2025–2027 acceleration phase (35–40 percent annual value growth) anchored by foundation model commercialisation plus CMS TCET reimbursement pathway plus enterprise deployment scaling, a 2028–2030 maturation phase (24–32 percent annual value growth) where deployment economics stabilise, and a 2031–2032 plateau phase (23 percent annual value growth) as cumulative deployment approaches structural saturation in mature markets.

The competitive structure is forecast to evolve from the current Big-OEM-led pattern (GE HealthCare-Siemens-Philips-Canon-United Imaging-Fujifilm collectively at 38 percent of 2025 deployment value) toward a more balanced structure with foundation model standalone AI biotechs plus big-tech infrastructure gaining share. Big OEMs maintain approximately 35 percent of 2032 deployment value (compressed from 38 percent) — anchored by FDA clearance leadership plus integrated hardware-plus-AI bundling plus big-tech AI partnerships. Standalone AI biotech grows from 27 percent to approximately 30 percent share — anchored by Aidoc foundation model leadership plus Viz.ai stroke specialty plus Tempus genomics-imaging integration plus emerging foundation-model entrants. Specialty AI (Lunit, iCAD, Optellum, DeepHealth, Qure.ai) maintains approximately 12 percent share. Big-tech healthcare AI infrastructure (NVIDIA, Microsoft, Google, AWS, Anthropic) grows from 6 percent to approximately 12 percent share — primarily as infrastructure underneath OEM and standalone platforms.

The modality mix shifts modestly. CT remains the dominant modality (35 percent of 2025, 37 percent of 2032) anchored by foundation model multi-task expansion. X-ray and mammography modestly compress as a share but grow in absolute terms. Ultrasound grows from 9 percent to approximately 10 percent driven by NVIDIA-GE autonomous ultrasound plus Butterfly Network POCUS plus emerging point-of-care applications. PET/SPECT plus pathology plus ophthalmology plus dermatology collectively maintain small but stable share.

The AI application type mix shifts toward foundation model multi-task platforms (forecast 6 percent of 2032 share, versus 2 percent in 2025). Triage and worklist prioritization compresses from 32 percent to 27 percent share (still largest single category). Detection plus quantification plus image acquisition plus workflow automation plus patient risk prediction collectively maintain stable share. The structural implication: foundation models do not replace single-task algorithms entirely but provide the platform layer underneath multiple task-specific deployments.

The geographic structure shifts modestly. The United States compresses from 51 percent of 2025 deployment to approximately 47 percent by 2032 — driven by emerging-market growth rather than US weakness. China grows from 8 percent to approximately 12 percent by 2032 — anchored by Infervision plus Yitu Healthcare plus United Imaging AI plus emerging Tencent and Alibaba healthcare AI. Europe maintains approximately 21–22 percent share. Other Asia-Pacific (Korea, India, ANZ, SEA) grows to approximately 8 percent by 2032. Middle East plus Latin America plus Africa collectively grow to approximately 4 percent.

The foundation model commercialisation trajectory is the most consequential 2025–2032 technology development. From approximately 2 percent of 2025 deployment value to approximately 6 percent by 2032. By 2030, approximately 40–50 percent of FDA AI medical imaging clearances are forecast to be foundation-model-derived (versus approximately 2 percent in 2025). Aidoc CARE1 (rib fracture February 2025, 14-finding abdominal CT January 2026, with broader CT-MRI-X-ray expansion roadmap) plus emerging GE HealthCare-NVIDIA Isaac for Healthcare plus Philips-NVIDIA AI Orchestrator plus Microsoft-Siemens Healthineers plus emerging Google DeepMind Med-PaLM Imaging plus Anthropic healthcare-specific models collectively dominate the foundation model category.

The CMS reimbursement framework maturation through 2026–2028 is the principal commercial unlock. The current 10 of 1,100+ FDA-cleared AI medical devices with CMS payment coverage creates structural reimbursement gap that constrains commercial deployment economics. The CMS TCET pathway (finalized 2024) plus emerging Category I CPT code conversion plus the proposed CMAA coding classification (AMA discussion December 2025) collectively position the reimbursement framework for structural improvement through 2028. The forward implication: AI medical imaging commercial revenue scales materially as reimbursement framework matures — potentially adding US$8–15 billion in incremental annual deployment value by 2030.

The principal risk to the outlook is sustained reimbursement gap that materially compresses commercial deployment economics. If CMS TCET implementation lags expectations, if Category I CPT code expansion stalls, or if commercial payers fail to follow CMS coverage decisions, AI medical imaging commercial revenue growth could plateau at approximately 18–22 percent CAGR (versus the 29–30 percent base-case forecast). The mitigation pathway requires sustained AMA-CMS coordination plus emerging Category I CPT code expansion plus commercial payer adoption following CMS decisions.

The secondary risk is foundation model regulatory uncertainty that constrains commercial scaling. The Aidoc CARE1 February 2025 FDA clearance established the precedent — but the structural regulatory framework for foundation models with iterative updates remains under development. If FDA imposes more restrictive requirements for foundation model updates, or if EU AI Act compliance creates structural friction, foundation model commercialisation could slow versus the base-case forecast. The mitigation pathway requires sustained FDA-vendor collaboration plus emerging international regulatory harmonisation.

For tailored support and detailed market analysis, see our offerings on Services or Contact Us.