The Somatic AI revolution: How TechMed is reshaping healthcare as we know it

In a previous article, Zachary Bogue and I outlined the emergence of a new category of startup at the inter­sec­tion of technology and medicine: TechMed. These companies leverage artificial intel­li­gence, machine learning, computer vision, and novel sensor tech­nolo­gies to deliver break­throughs in clinical insights and care. They start not with a particular device, but with unique, proprietary datasets and sophis­ti­cated algorithms that can transform every stage of the patient journey.

In this sequel to that article, I will double-click on one of the most exciting and trans­for­ma­tive aspects of the TechMed revolution: the rise of Somatic AI. I’ll explain what it is, how it differs from other buzzed-about forms of AI, and why it represents an unprece­dented opportunity for startups, investors, and the future of care delivery. 

Somatic AI: healthcare’s new ​‘little black bag’ 

Much of the fervor in tech investing over the past two years has centered on Generative AI — large language models like GPT‑4 that can produce text, images, and code at levels that closely approximate human capa­bil­i­ties. In parallel, a class of Enterprise AI appli­ca­tions has gained traction by stream­lining knowledge work in domains like law, finance, and customer service. Both approaches have made inroads into healthcare, powering use cases like clinical docu­men­ta­tion and revenue cycle management. 

However, the most profound wave of AI-driven trans­for­ma­tion in medicine has yet to hit the mainstream. I call it Somatic AI — a new generation of AI systems purpose-built to understand and interact with the human body. Somatic AI combines insights from medical imaging, digital biomarkers, real-time sensing, and other patient data modalities to build high-resolution compu­ta­tional models of individual anatomy and physiology. Fed by a continuous flow of multi-modal data — a data river — these models serve as the foundation for a wide array of break­through clinical appli­ca­tions, from AI-guided surgical navigation and autonomous robotic procedures to intelligent patient monitoring and real-time clinical decision support.

The core enabling tech­nolo­gies behind Somatic AI — computer vision, sensor fusion, GPU-accelerated computation — are frequently birthed not in the medical device industry, but in other domains like autonomous vehicles, industrial robotics, and consumer electronics. In that sense, Somatic AI represents a second wave of artificial intel­li­gence in healthcare, building on but distinct from the initial forays focused on digital health workflows and back-end operations. 

The potential of Somatic AI to democratize early disease detection and inter­ven­tion is already being realized through wearable devices like the Apple Watch. In one recent example, a friend of mine woke up to an alert from his Apple Watch indicating that his heart was in atrial fibril­la­tion (AFib). This early warning allowed him to seek timely medical attention and potentially avoid serious compli­ca­tions. As Somatic AI advances, we can expect even more proactive and person­al­ized insights to be delivered directly to patients and their care teams.

Just as the iconic ​“little black bag” equipped physicians in the mid-20^th century with core diagnostic and treatment tools for house calls, Somatic AI platforms will serve as the foundation for a new era of data-driven, algo­rith­mi­cally enabled care. Portable ultrasound devices enhanced with computer vision and machine learning, like those made by Caption Health (acquired by GE HealthCare in 2023), can guide frontline providers through complex imaging exams while auto­mat­i­cally identifying signs of disease. The smartphone in a clinician’s pocket will provide instant access to Somatic AI models of each patient, trans­forming care from episodic and reactive to continuous and predictive.

Enabling the surgical superhero

The impact of Somatic AI extends well beyond the primary care setting. It will funda­men­tally transform how specialists approach complex procedures in fields like surgery, inter­ven­tional radiology, and radiation oncology. The ability to navigate the physical side of procedures and exam­i­na­tions is a key opportunity for Somatic AI. Two DCVC-backed companies, Proprio and Remedy Robotics, exemplify this potential. 

Proprio is pioneering a real-time surgical navigation and alignment platform, initially targeting spine and cranial procedures. The company’s core technology, derived from research at the University of Washington’s Sensor Systems Lab, leverages Lightfield cameras to capture rich spatial and depth information from the surgical field. (Lightfield imaging contains the intensity and direc­tion­ality of all light rays passing through a scene — orders of magnitude more visual information than standard two-dimensional, or 2D, images.)

Proprio feeds these Lightfield streams into GPU-powered deep learning models that construct three-dimensional repre­sen­ta­tions of patient anatomy, track instruments, and enable real-time navigation and alignment. The platform acts as a kind of surgical GPS, guiding the surgeon’s tools and decision-making to sub-millimeter precision. Surgeons can effectively ​“see” through blood and tissue to visualize underlying anatomical structures, track the position and trajectory of implants, and receive AI-generated alerts about potential deviations or safety hazards. Proprio received FDA 510(k) clearance in 2023 for its technology and is rolling out to leading spine surgical practices, having performed over 50 live and successful surgeries in the first half of 2024.

Importantly, Proprio’s platform is not just enhancing surgical precision and outcomes in the moment but also generating a wealth of data on patient anatomy, surgical techniques, and procedural workflows that can be used to drive continuous improvement across the entire cycle of care. By capturing and analyzing this data at an unprece­dented level of granularity and using it to power AI-driven insights and recom­men­da­tions, Proprio is building a foundation for optimizing everything from pre-operative planning and intra-operative guidance to post-operative recovery and long-term follow-up. 

Another key tenet of TechMed is to use Somatic AI to expand and improve care to unserved populations, as we have seen in DCVC portfolio company Remedy Robotics. Today, only 15% of the global population has access to time-critical cardio­vas­cular care. Unfor­tu­nately, this care is often marked by imprecision and unac­cept­ably high compli­ca­tion rates. Remedy Robotics, founded by Dr. David Bell, a cardio­tho­racic surgeon, and Dr. Jake Sganga, a PhD in surgical robotics, is on a mission to provide optimized, compli­ca­tion-free cardio­vas­cular care and make it accessible to all people around the world. Remedy Robotics achieves this by integrating cutting-edge machine learning into the tradi­tion­ally hardware-driven field of surgical robotics. This approach enables enhanced visu­al­iza­tion and precise control of endovas­cular tools, allowing for the complete remote treatment — and soon, supervised autonomous treatment — of a range of time-critical cardiovascular conditions.

In addition, beyond surgery, imaging, and medical devices, having the ability to contin­u­ously (24×7) and remotely monitor and manage patients with at-risk conditions becomes an enormous enabler of TechMed. In the area of Somatic AI, all.health has developed an AI-enabled wearable wristband with a range of sensors, allowing health systems and payors to better manage their patients at risk by detecting serious issues earlier, thus providing better care and lowering healthcare costs — at population scale.

A new paradigm for healthcare — and investors

The rise of Somatic AI represents a fundamental shift in how we approach healthcare innovation and delivery. It challenges us to move beyond the traditional MedTech playbook of incremental device enhance­ments and embrace a new paradigm centered around data, intel­li­gence, and continuous opti­miza­tion. New compu­ta­tional insights, built at the inter­sec­tion of AI and anatomy, will factor into how medicine is conducted. 

In this new era, success will require more than just bolting AI onto existing devices or sprinkling in a few software features. It demands a wholesale re-archi­tecting of R&D, commercial, and post-market strategies to put data at the core. The winners will be those who can construct the most clinically relevant and diverse real-world datasets, develop the most intelligent and adaptable algorithms, and deliver the most compelling value propo­si­tions aligned with real-world outcomes.

For investors, the rise of Somatic AI offers massive oppor­tu­ni­ties. But success will require following three general rules:

1. Bet on hetero­ge­neous founding teams with deep expertise across both clinical medicine and core technical domains like AI/ML, computer vision, robotics, and sensors. Many of the most promising Somatic AI startups will be led by tech­nol­o­gists from fields like autonomous vehicles, AR/VR, and computer science partnering with prac­ti­tioners frustrated by the limitations of traditional medical devices.
2. Avoid ​“raging incre­men­talism.” Somatic AI enables funda­men­tally new paradigms of care, not just better versions of existing tools and treatments. Investors should back founders with the ambition to transform huge swaths of medicine instead of chasing incremental improve­ments or niche use cases.
3. Think in terms of lifetime patient impact vs. siloed episodes and specialties. The most valuable Somatic AI platforms will be those that capture and oper­a­tionalize data across multiple stages of the care journey and disease progression. They will break down the barriers between specialties and redefine care around the longi­tu­dinal needs of each patient.

Of course, realizing the full potential of Somatic AI will require close collab­o­ra­tion with patients, providers, and regulators. Entre­pre­neurs and investors must work with all these groups to address critical issues around safety, efficacy, data rights, and health equity. Industry and academia should partner on new approaches to validation, from in-silico trials and digital twins to adaptive post-market surveil­lance. The FDA has taken encouraging steps by releasing its Artificial Intel­li­gence and Machine Learning Action Plan, but much work remains to align regulatory frameworks with the unique challenges and oppor­tu­ni­ties of AI in medicine.

Ultimately, the rise of Somatic AI represents an opportunity to reimagine our healthcare system around the principles of 21^st-century technology — ubiquitous sensing, large-scale data analytics, intelligent automation, and continuous opti­miza­tion. From wearable devices that detect atrial fibril­la­tion to AI-guided surgical navigation and autonomous catheter-based inter­ven­tions, it promises to usher in a new era of care that is more preventive than reactive, more person­al­ized than population-based, and more accessible than ever before. 

For entre­pre­neurs, investors, and healthcare profes­sionals alike, the message is clear: the future of medicine lies in harnessing Somatic AI to understand, predict, and optimize human health in unprece­dented ways. Armed with these smarter tools and keener insights, clinicians can focus on the most human elements of healing while ensuring every patient receives the highest standard of data-driven, evidence-based medicine.