Rad AI – Weekly Recap

Rad AI is a technology-focused company operating in the photon-counting X-ray and CT imaging ecosystem, and this weekly summary reviews its latest market analysis and technical disclosures. Over the past week, the company emphasized both its role as an independent analytics provider on photon-counting CT (PCCT) competition and its deepening capabilities in detector-to-ASIC co-design, reinforcing its positioning in next-generation medical imaging.

Rad AI published an AI-generated, in-depth report on the global PCCT competitive landscape, shared via LinkedIn. The analysis highlights Siemens’ dominant position, with more than 80% market share and over 1,000 NAEOTOM Alpha installations, while noting that GE Healthcare’s Photonova Spectra system is still awaiting FDA clearance and faces silicon physics-related challenges. Canon’s PCCT solution is characterized as being under development, and Philips is described as having pivoted toward its Verida spectral CT platform instead of PCCT, based on concerns over complexity and the pace of adoption into routine clinical practice. The report also flags emerging competition from Chinese vendors Neusoft (NeuViz P10) and United Imaging (uCT Ultima), indicating that Siemens’ near-monopoly may be starting to weaken.

The underlying report is said to cover detector material trade-offs, FDA status by vendor, market projections, bonding technologies, and strategic implications for equipment buyers and investors. Key themes include Siemens’ entrenched installed base and clinical lead, regulatory and technical execution risk for GE and Canon as they scale PCCT offerings, and Philips’ choice to prioritize spectral energy-integrating detector CT as a way to reduce near-term technology complexity while potentially limiting its upside in PCCT. The presence of Chinese PCCT systems suggests intensifying price and technology competition, especially as these platforms seek approvals beyond China.

In a separate technical update, Rad AI disclosed that it is using Monte Carlo simulations to translate CdZnTe detector physics into concrete ASIC performance targets for photon-counting CT. The company reported quantitative timing metrics, such as 7 ns rise times and 3.2 ns to 50% charge, indicating electron-dominated fast signals and suppressed hole contributions. These results were linked to specific ASIC design guidance, including recommended shaping times of 15–25 ns, a minimum dead time of approximately 20 ns, and sub-800-electron noise requirements, providing clearer design guardrails for readout electronics.

Taken together, Rad AI’s publication of a detailed, AI-generated market report and its simulation-driven ASIC design guidance underscores a strategy centered on data-driven analysis and technical rigor. While the updates do not reference specific commercial contracts, they strengthen the company’s profile as a specialized analytics and co-design partner for medical imaging OEMs, which could support future collaborations and design-in opportunities as photon-counting CT adoption evolves. Overall, the week underscored Rad AI’s dual role as both a technical thought leader and an emerging decision-support provider in the advanced CT imaging market.