Healthcare is entering an era of operational automation — and it is no longer hypothetical.
Aethon's TUG delivery robots have been deployed in more than 500 hospitals worldwide. Diligent Robotics' Moxi now operates in over 25 U.S. hospitals and has completed more than 1.25 million deliveries, including roughly 300,000 pharmacy runs. Health systems including Northwestern Medicine, Mount Sinai, AdventHealth, and Houston Methodist have moved from pilots to multi-campus deployments. Analysts project the hospital logistics robot market will roughly quadruple over the next decade, with most estimates putting growth at 10–17% annually.
The demand side is even clearer.
The Association of American Medical Colleges projects a shortage of up to 86,000 physicians by 2036, driven by a population aged 65+ growing 34% while 20% of practicing physicians are already at retirement age. The Bureau of Labor Statistics projects roughly 190,000 open RN positions every year through the early 2030s — while nursing schools turned away more than 65,000 qualified applicants in a single year due to faculty and clinical capacity constraints.
And the financial pressure is structural. According to the American Hospital Association, hospital labor costs reached $890 billion in 2024 — 56% of total hospital expenses.
These forces create a powerful, durable incentive for healthcare organizations to automate operational work wherever possible. The founders of Diligent Robotics built Moxi after observing that hospital staff spend roughly 30% of their time simply fetching and gathering items — work that requires no clinical license at all.
So discussions around hospital robotics naturally focus on the technology itself. Can robots navigate complex environments? Can they complete tasks reliably? Can they generate a positive return on investment?
These are important questions. But they may not be the most important ones.
The Technology Is Improving Faster Than the Organization
Robotics technology continues to advance at an extraordinary pace.
Autonomous mobile robots can already navigate hospital corridors, avoid obstacles, summon elevators, open doors, and complete delivery routes. Diligent's next-generation Moxi 2.0 — built on an AI foundation model trained on three years of real-world hospital deployment data — is designed to scale to 15+ units per site, and the company has stated plans to double its hospital footprint annually.
And the next wave is no longer a single-purpose delivery cart.
General-purpose humanoid platforms are moving from labs to commercial deployment. Figure's robots are working shifts in BMW's manufacturing facilities. Tesla has begun producing Optimus units at scale for its own factories, with external sales planned. Diligent itself now describes Moxi as a "humanoid platform." The vendors differ; the trajectory does not. The same economics pushing humanoids into factories and warehouses — labor scarcity, 24/7 operation, environments built for the human form factor — apply with even greater force to hospitals.
The technological barriers continue to shrink. The organizational barriers do not.
Hospitals were designed around human workers. Policies, workflows, escalation procedures, training programs, accountability structures, compliance processes, and operational management systems all assume that work is performed by people.
Consider what hospitals already do for humans: every clinician is credentialed. Every employee is badged. Access to medication rooms, supply chains, and patient areas is governed by identity, role, and context. Health systems spend enormous effort maintaining this infrastructure — because it is what makes a high-risk environment governable.
There is no equivalent institution-wide layer for autonomous systems.
As these systems become more capable, healthcare organizations face a fundamentally different challenge: not whether the robot works, but how it is integrated into existing operational and governance structures.
Robotics Creates Cross-Departmental Complexity
One of the most overlooked aspects of hospital robotics is that robots rarely fit neatly into a single department.
Consider a delivery robot transporting supplies. The robot may be purchased by Operations, supported by IT, maintained by Facilities, utilized by Nursing, managed through Supply Chain workflows, and subject to Security and Compliance requirements.
A single autonomous system touches multiple departments simultaneously. A fleet of them — sourced from multiple vendors — multiplies that complexity.
This creates new questions that traditional hospital governance structures were never designed to answer.
Who owns the robot? Who determines what tasks it is allowed to perform, and in which units, at which hours? Who is responsible when something goes wrong? Who approves expanded use cases as the vendor ships new capabilities?
Each vendor today answers these questions inside its own closed system. The hospital, as an institution, often cannot answer them at all — not across its full fleet, and not in real time.
These are organizational questions rather than technical ones. And they compound with every new system that enters the building.
Reliability Is Different in Healthcare
Many industries can tolerate occasional automation failures. Healthcare environments often cannot.
The data here is instructive. A 2025 feasibility study of autonomous medication delivery in a tertiary hospital found an overall delivery success rate of about 87% — but success dropped sharply as elevator congestion rose, with failure risk climbing meaningfully with each additional passenger competing for the same elevator. The robot's capability wasn't the constraint. The shared, contested, human-dense environment was.
Security failures raise the stakes further. In 2022, researchers disclosed JekyllBot:5 — a set of five vulnerabilities in the servers controlling Aethon TUG robots that could have allowed attackers to remotely hijack robots operating inside hundreds of hospitals: machines with badge-level access to restricted areas, elevators, and medication transport. The flaws were patched. The lesson remains: an autonomous system inside a hospital is not just a productivity tool. It is a credentialed actor with physical access — and it must be governed like one.
A delayed supply delivery can impact patient care. A blocked corridor can affect patient movement. A compromised robot is a security incident with wheels.
In conversations with hospital innovation and strategy leaders, a recurring theme emerges: healthcare organizations are often less concerned about whether a robot can complete a task and more concerned about how the organization responds when it cannot — or when it does something it shouldn't.
Operational resilience becomes as important as technical capability.
The Governance Layer Is Already Forming — But Not for Robots
Here is the most telling signal that this is the real frontier: the institutions that govern healthcare are already moving.
In September 2025, the Joint Commission and the Coalition for Health AI released their first joint guidance on the Responsible Use of AI in Healthcare — calling on health systems to establish formal AI policies, governance committees spanning compliance, IT, cybersecurity, clinical, and operational stakeholders, ongoing quality monitoring, and reporting of AI safety events to the governing board. A voluntary certification program followed, along with detailed governance playbooks developed with more than 150 health AI leaders.
Healthcare's accreditation infrastructure has concluded that AI cannot be deployed tool-by-tool, vendor-by-vendor, without institutional oversight.
That conclusion currently applies to software. Physical autonomy is next — and it is harder. A clinical decision-support algorithm produces a recommendation. An autonomous robot produces actions in shared physical space: it moves through hallways, enters units, rides elevators, and handles medications. The governance questions are correspondingly more concrete:
Which systems are authorized to perform which tasks? Where can autonomous systems operate, and when? What level of human oversight is required, and how is it escalated? How are exceptions handled in real time — a code event, a locked-down unit, an outbreak protocol? How is accountability maintained across departments — and across vendors?
Historically, hospitals have developed governance frameworks for people (credentialing and privileging), devices (biomedical engineering and FDA oversight), software (IT governance, and now AI governance), and clinical processes (quality and safety programs). Autonomous physical systems will require the same evolution. Today, no equivalent exists.
The Multi-Vendor Problem
There is one more structural wrinkle worth naming.
The robot fleet of the 2030s hospital will not come from one company. It will be delivery robots from one vendor, disinfection robots from another, and general-purpose humanoids from a third — each with its own permissions model, its own dashboard, its own escalation logic.
No vendor's control system will span a rival's fleet. And the system being governed cannot credibly govern itself.
That means the governance layer — the institution-wide answer to "what is this system allowed to do, right now, in this context" — has to live with the hospital, not with any individual robot maker. It has to sit where credentialing, risk, and clinical governance already sit.
If every manufacturer builds its own governance silo, the fragmentation deepens with every deployment.
Looking Forward
The future of hospital robotics is often discussed as a technology problem. It is increasingly becoming an organizational problem.
The robots will continue to improve — the deployment numbers, the capital flowing into humanoid platforms, and the pace of capability gains all point one direction.
The more difficult challenge is helping healthcare organizations adapt their operational structures, governance models, and workflows to integrate autonomous systems safely at scale: a real-time, institution-wide answer to who is allowed to do what, where, and under whose authority.
Hospitals solved this problem for people. They are beginning to solve it for AI software. They have not yet solved it for the autonomous systems now moving through their hallways.
In the coming years, the hospitals that succeed with robotics may not be the ones with the most advanced technology. They may be the ones that are best prepared to govern it.
Sources: Diligent Robotics (Moxi 2.0 announcement, Oct 2025); ST Engineering Aethon deployment data; AAMC, "The Complexities of Physician Supply and Demand: Projections From 2021 to 2036" (2024); U.S. Bureau of Labor Statistics RN employment projections; American Association of Colleges of Nursing enrollment data; American Hospital Association "Cost of Caring" report (2024); feasibility study of autonomous medication delivery under elevator congestion, tertiary hospital, 2025 (PMC); Cynerio JekyllBot:5 disclosure (2022); Joint Commission & Coalition for Health AI, "Responsible Use of AI in Healthcare" guidance (Sept 2025).