RadNet's adjusted EBITDA rose 36.3% year-over-year in Q1 2026, outpacing its 22.1% revenue gain — a divergence that points to AI-generated operational leverage, not simple volume growth.1
The radiology network deploys two AI platforms across its centers: TechLive, which optimizes MRI scheduling, and DeepHealth, which supports diagnostic imaging workflows. TechLive is explicitly credited with driving MRI utilization improvements through smarter slot allocation.1
MRI same-center volume grew 10.1%, well above industry baseline rates.1 Advanced imaging climbed from 26.9% to 29.3% of total procedural volume year-over-year.1 PET/CT aggregate volume surged 35.2%.1 Imaging Center EBITDA margins expanded 188 basis points.1
The mechanics are straightforward. When AI scheduling fills more scanner slots per shift without adding proportional fixed costs, EBITDA grows faster than revenue. Higher-value scans — PET/CT, MRI — carry better margins than commodity imaging. An algorithm that steers available capacity toward those procedures compounds volume and margin simultaneously.
The causal hypothesis — that AI deployment, not volume alone, explains the margin gap — remains unconfirmed at the center level. RadNet has not published a controlled comparison between locations with full TechLive and DeepHealth rollout versus those still in transition. A margin differential above 150 basis points between deployed and non-deployed centers would confirm the link.1
What aggregate data does confirm: the leverage is real. Volume-driven radiology revenue scales linearly with scanner count and patient throughput. AI-optimized throughput is increasingly non-linear — the same physical infrastructure generates more billable high-value procedures per shift.
The competitive implication is direct. Rivals without equivalent AI tooling face margin compression as RadNet captures referrals through faster scheduling and diagnostic turnaround. Deployment velocity now determines the size of the moat.
Healthcare diagnostics is splitting into two tiers: AI-optimized networks extracting non-linear returns from fixed assets, and everyone else running a slower, thinner-margin operation.
Sources:
1 RadNet Q1 2026 Financial Results and AI Deployment Analysis, May 2026
