Medical Imaging Systems technology and investment research

Technical bottleneck

Why it's a moat — capital equipment cycles, installed-base service lock-in, and AI

Medical imaging is not a device business — it's an installed-base business. The initial scanner sale (MRI: $1–3M; CT: $500K–$2M) is a loss-leader or low-margin transaction. The real economics are in:

Service contracts: OEMs lock in multi-year service agreements covering preventive maintenance, parts, uptime guarantees, and performance optimization. Margins on service contracts are 40–60% — higher than the hardware itself. A hospital with 10 GE MRI scanners pays GE millions annually in service fees, regardless of whether they buy a new scanner. This is the installed-base moat: the machines stay in service for 7–15+ years, and every year they generate service revenue.

AI as a stickiness layer: GE's Edison platform, Siemens' AI-Rad Companion, and Philips' IntelliSpace embed AI into imaging workflows — automated protocol selection, image reconstruction, lesion detection, report generation. These AI tools are sold as software subscriptions on top of service contracts, making the relationship even stickier. Switching imaging vendors means losing access to the AI tools radiologists have integrated into their daily workflow.

Capital cycle dynamics: Hospital purchasing is slow and deliberate. A new MRI requires: budget approval (6–12 months), site preparation (construction, shielding, cooling — 6–18 months), installation and calibration (2–4 months), and staff training (ongoing). Total: 2–3 years from decision to operational. This creates inertia — hospitals don't switch vendors lightly because the switching cost includes physical infrastructure, not just the purchase order. GE's Allia upgrade pathways (2026 launch) explicitly target this by offering interventional suite modernization without full replacement.

The utilization threshold — why emerging markets matter: A traditional MRI needs 8–10 scans/day to break even in a typical hospital. This limits adoption in lower-resource settings. Helium-free MRI designs (reducing cryogen costs) and AI-accelerated protocols (reducing scan time) can lower the break-even to 3–5 scans/day — expanding the addressable market into mid-tier hospitals in India, Africa, and Southeast Asia. This is the next TAM layer.

The Midjourney disruption signal: In 2026, Midjourney (the AI company) unveiled a full-body ultrasound computational tomography (USCT) scanner — thousands of transducers, 17 GB/s data, 60-second 3D reconstruction via AI, no radiation, ~10× cheaper than MRI. This is not clinically validated yet, but it signals that AI-native imaging paradigms could eventually pressure traditional MRI/CT economics. The incumbents' defense is regulatory clearance, reimbursement pathways, clinician trust, and decades of clinical validation that new entrants lack.

Adoption

Why it matters now

GE HealthCare (GEHC):

• MRI (SIGNA), CT (Revolution), ultrasound (LOGIQ, Vivid, Vscan Air handheld)
• Interventional imaging: Allia platform — upgrade pathways launched 2026
• Edison AI ecosystem; emphasis on embedded AI throughout workflows
• Strong in service contracts and installed-base monetization

Siemens Healthineers (not in Healthspan universe):

• MRI (MAGNETOM), CT (SOMATOM), molecular imaging, AI-Rad Companion
• Photon-counting CT (NAEOTOM Alpha) — the most significant CT innovation in a decade, enabling spectral imaging at lower dose

Philips (not in current universe):

• MRI, CT, ultrasound, IntelliSpace AI platform, Azurion interventional

Key trends

• AI-native imaging emerging: FDA has cleared 1,500+ AI-enabled medical devices — imaging dominates. AI is shifting from "nice to have" to "must have" in RFPs.
• Photon-counting CT: Siemens' NAEOTOM Alpha enables spectral separation of tissues at the detector level — better tissue characterization, lower contrast dose. This could accelerate CT replacement cycles.
• Helium-free MRI: GE and Siemens developing low-helium or helium-free MRI designs to reduce operating costs and expand TAM.
• Site-neutral payment pressure: US hospital-based imaging costs 3–10× more than independent imaging centers for the same scan — Medicare site-neutral payment reforms could shift volume and pressure hospital capex.

Key players

Related Technologies

• Surgical Robotics — robotic surgery requires preoperative imaging for planning and intraoperative imaging for guidance
• Structural Heart & Transcatheter Valves — TAVR requires CT for annular sizing, fluoroscopy for deployment, and echo for post-procedure assessment
• Orthopedic Implants & Digital Surgery — orthopedic planning requires CT/MRI; Mako and ROSA integrate directly with imaging workflows