The $28 Billion AI Tax: Is Oracle Trading the Hospital for the GPU?

If you listened to the keynote speakers at CES earlier this month, or if you have been following the latest soundbites from Silicon Valley, you have likely heard the term "Agentic AI." The promise is seductive. We are told that by the end of 2026, autonomous AI agents will handle 90% of your clinical documentation, schedule your patients, and battle insurance denials while you sleep. The "Digital Doctor," they say, is finally here to save us from administrative purgatory. It sounds like nirvana. But then, there is Elon Musk. You might have seen the clip circulating on X this week. During a conversation with Peter Diamandis, Musk made a prediction that stopped my scroll. When asked if young people should still pursue a medical degree, he didn't hesitate. "No. Pointless," he said. He went on to predict that within three years, AI-driven humanoids (like Optimus) will be better surgeons than humans, and that everyone will have access to medical care better than what the President receives right now. (You can watch the full clip here). It is a beautiful vision. It is also missing some key info...... Musk is making the classic futurist mistake because he is confusing Capability with Implementation. Sure, in a sterile lab with perfect 6G connectivity, a robot might be able to suture a grape better than a resident. But healthcare doesn't happen in a lab. It happens in a rural ER with spotty Wi-Fi, a patient screaming in a language the intake nurse doesn't speak, and a legacy EMR system that crashes if you look at it wrong. So yeah the tech will be ready, but the hospitals won't be! While Elon is dreaming of "Robo-Surgeons," I am looking at the actual AI crisis happening in our hospitals today. It isn't about robots taking jobs. It is about "Shadow AI" breaking the system. A recent report from Wolters Kluwer has confirmed what many of us have whispered in the breakroom for months: Clinicians are cheating on their IT departments. While Health Systems spend millions vetting "Enterprise Grade" AI tools (platforms that often take 14 clicks just to open and require a dual-factor authentication that fails in the basement), exhausted doctors and nurses are bypassing protocols. They are quietly pulling out their personal phones. They are opening consumer-grade apps like ChatGPT-6 to summarize charts, draft appeals, and write discharge notes. They aren't doing this because they are rebellious or because they don't care about security. They are doing it because they are drowning. When the "official" hospital tech stack fails to save time on a Tuesday night in the ER, clinicians don't wait for a Governance Committee meeting. They find a workaround. As a Clinical Realist, I see this not as an innovation problem, but as an infrastructure failure. And it carries three massive risks for 2026 that no robot can solve. First is the Liability Trap. This is the one that keeps me up at night. If a physician uses an enterprise-approved tool and that tool hallucinates a diagnosis, the liability is shared. The vendor, the hospital, and the indemnifiers are all at the table. But if a doctor uses a "Shadow" tool on their personal iPhone and it hallucinates? The physician is on an island. Courts are already sanctioning professionals for citing non-existent cases generated by AI. Do you think they will be lenient on a doctor who uses a hallucinated summary to treat a patient? The "Digital Doctor" doesn't lose its license. You do. Second is the Data Leak. "Shadow AI" means PHI (Protected Health Information) is leaving your secure firewall. When a clinician pastes a complex case study into a public Large Language Model to get a second opinion, that data often becomes training data for the model. We aren't just leaking patient privacy. We are feeding the competition. Third is the "Digital Twin" Delusion. Hospital executives are currently making multi-million dollar strategic decisions based on workflow data from their official EMRs. But if 30% of the actual cognitive work is happening on shadow apps that the CIO can't see, the C-Suite is flying blind. We are building "Digital Twins" of a hospital that doesn't exist. You cannot optimize a workflow you cannot see. Get the "Clinical Reality Check" Before Everyone Else. I send these briefings to my private list 24 hours before they hit social media. Join other healthcare leaders who get the raw, uncensored analysis first. [Join the Clinical Realist List] So, Mr. Musk, with all due respect: We don't need you to tell us medical school is pointless. The tech is hear, we get that loud and clear. We need you to build tools that actually work in the chaos of the real world. Until then, we need a defense protocol. We can't ban AI because the genie is out of the bottle. And we can't ignore the burnout that drives it. We need "Implementation over Innovation." If you are a hospital leader, stop punishing clinicians for using Shadow AI. Instead, declare a 30-day "Amnesty Period" to find out what tools they are actually using. You will be shocked, but you will finally have the truth. If your enterprise tool isn't as easy to use as the app on your doctor's iPhone, do not buy it. In 2026, User Experience is a safety metric. If a tool is hard to use, it induces fatigue, and fatigue causes errors. Ultimately, we don't need Digital Doctors (yet!). We need Digital Scribes. The data is clear: Ambient AI tools are reducing burnout by over 13% in just 30 days. Refocus your AI budget entirely on reducing the "Pajama Time" (the documentation done at home). If the tech doesn't give the clinician 60 minutes of their life back, it is not an asset. It is a liability. The future of healthcare isn't about the algorithm. It is about the access. If we burn out the human workforce waiting for Elon's robots to arrive, there won't be anyone left to turn on the computers. It is time to stop buying "Future Tech" and start building "Now Infrastructure." Dr. Sarah Matt P.S. Navigating the 2026 Healthcare Landscape If your leadership team is struggling to balance the new $50B CMS Rural Funding with the reality of workforce burnout, you don't need another futurist predicting 2040. You need a guide who speaks fluent "C-Suite" and fluent "Frontline." I am currently booking Keynotes & Advisory for Q3/Q4 2026. My signature keynote, The Clinical Realist: Innovation that Survives the Bedside, offers a brutal, honest, and hopeful look at how to scale access without scaling burnout. [Inquire About Speaking Availability Here]

Yesterday I stood on stage at the SUNY UPSTATE UNIVERSITY HOSPITAL Health Justice Conference and looked out at a room full of brilliant and well-meaning professionals. These are people who have dedicated their lives to fixing the broken machinery of American healthcare. The energy was palpable. The slides were slick. The buzzwords were flying. Everyone was talking about "additive" solutions. We need new pathways, apps, AI, engagement. In between talks I was speaking to a group and asked a question that sucked the air out of the room. "How many of you have a Chief Innovation Officer?" About half the hands went up. "How many of you have a Chief Friction Officer ?" Silence. Not a single hand. We have a fundamental problem in how we approach health technology. We are obsessed with addition. We treat healthcare strategy like a hoarding situation where we keep piling new solutions on top of old problems without ever clearing the clutter. But as a surgeon, I can tell you that addition is rarely what saves a life. Extraction does. Removing the infection. Removing the tumor. Removing the blockage. It became the backbone of my talk yesterday. It is the only narrative that matters for 2026. We do not 'need' more innovation. We need a Friction Audit. The "Addition Bias" Trap Psychologists call this "addition bias." When humans are asked to improve an object or a situation, they overwhelmingly prefer to add something rather than subtract something. In healthcare, this bias is fatal. We add a digital check-in tool to "streamline" the front desk. But we forget to remove the clipboard of paper forms. Now the patient has to do both. That is not innovation. That is friction. We add an AI scribe to the exam room to capture the conversation. But we do not remove the requirement for the physician to click through forty-seven tabs to bill for the visit. That is not progress. That is administrative burden. Innovation without subtraction is just suffocation. The Friction Audit: A New Standard Yesterday in my talk I proposed a radical shift in strategy. I challenged every health system leader, startup founder, and trainee in the room to pause their product roadmap and conduct a Friction Audit . This is not a bug hunt. This is a strategic review of the "invisible walls" we build around care. A true Friction Audit looks at three specific layers. 1. Digital Friction This is the most obvious and the most ignored. I am not talking about a clunky user interface. I am talking about the "Login Wall." If a patient needs three different passwords to view their labs, message their doctor, and pay their bill, you have failed. The most advanced algorithm in the world is useless if the patient cannot remember their username. Read: The Cost of Fragmented Patient Portals 2. Physical Friction We often design digital tools for a world that does not exist. I shared the story of the "unplowed driveway" in rural New York. You can have the most advanced "Hospital at Home" monitoring kit in the world. It can be powered by the most sophisticated AI. But if that patient lives up a steep hill in a snowstorm and their driveway is unplowed, the ambulance cannot get to them when the AI detects a crisis. Friction is not just in the code. It is in the infrastructure. Data: The Broadband Gap in Rural America 3. Cognitive Friction This is the silent killer. We are asking sick, scared, and stressed patients to act as their own project managers. We ask them to coordinate between the specialist and the primary care doctor. We ask them to chase down prior authorizations. Every time we force a patient to understand the complexity of our org chart, we are introducing cognitive friction that leads to burnout and dropout. The Job Description: Chief Friction Officer If I were running a major health system today, I would hire a CFO . But not the financial kind. I would hire a Chief Friction Officer . Their job description would be simple. They would have zero budget for buying new technology. Their only power would be the power of the veto and the power of the delete key. Their primary Key Performance Indicator (KPI) would be Time to Care . Their job is to ruthlessly hunt down every policy, every click, every form, and every redundant question that slows the patient down or burns the clinician out. They are the guardian of flow. If we had this role in place five years ago, we would not be facing the crisis staring us down next week. The Ultimate Friction: The January 30 Telehealth Cliff We have a real-time example of "Government-Mandated Friction" happening right now. In just 10 days, on January 30, 2026 , the Medicare telehealth flexibilities regarding the "originating site" are set to expire. We are about to revert to a rule (unless Congress acts) that says a Medicare patient cannot receive telehealth services from their home. They must travel to a designated "originating site" (usually a clinic or hospital) to get on a video call with a specialist. Think about the absurdity of this friction. We are legally forcing an eighty-year-old woman with mobility issues to drive forty minutes on icy roads just to sit in a room and look at a screen. We have the technology to treat her in her living room. We have the connection. But we have a policy wall that says "No." This is the "Telehealth Cliff." It puts over a third of rural Medicare beneficiaries at risk of losing access to virtual care overnight. A Chief Innovation Officer would try to solve this by building a better scheduling app for the clinic, or a new patient engagement tool. A Chief Friction Officer would recognize that the drive is the barrier and fight to remove the rule. Or better yet develop a community strategy to bring transportation into the care plan, and make it a 'clinical necessity'. Read: January 30th Cliff The Bottom Line Stop looking for the next "Agentic AI" to add to your tech stack. Stop looking for the next shiny object at CES. Look for the friction you can subtract. If your algorithm can diagnose a rare disease with 99% accuracy, but your patient cannot log in to see the result, you have not innovated anything. You have just built a very expensive wall. -Dr. Matt\ Dr. Matt believes technology can erase the borders that limit access to care. This vision is the heart of her National Best Seller, The Borderless Healthcare Revolution . If your organization is drowning in "solutions" but starving for results, you do not need another consultant to tell you what to buy. You need someone to tell you what to delete. I am opening two spots next month to serve as a Fractional Chief Friction Officer for growth-stage health tech companies or progressive health systems. We will identify your three biggest barriers to care and dismantle them. And if you are new here, join thousands of others who get this newsletter every Tuesday. Subscribe to The Borderless Healthcare Revolution.