Every year since roughly 2021, someone publishes an article titled some version of "n AI Trends Transforming Healthcare in [year]." Most of them could be dated to almost any year in that span and you wouldn't notice — the same three predictions about imaging, documentation burden, and "the future of personalized medicine," restated with a new hero image. I read AlphaSense's "7 Medtech Trends to Watch in 2026" expecting more of the same. The first two sections weren't. They were dense with specific, checkable claims — FDA clearance counts by specialty, named products with named regulatory milestones, named companies raising named funding rounds — the kind of texture that lets you actually go verify whether a "trend" is real or is three analysts citing each other in a circle.
So I did. This is what I found when I pulled the thread on section one — AI's move from standalone tool to embedded infrastructure — and section two — the surgical robotics market's transition from a near-monopoly to an actual contest. Some of AlphaSense's underlying figures sit behind its proprietary research platform and I can't independently verify them; where that's true, I say so explicitly and attribute the claim to its original reporting rather than presenting it as confirmed fact. Where I could get to primary sources — FDA press releases, SEC filings, company announcements — I went there directly, and in a few places the primary-source picture is a step behind, or more conditional than, the trend piece's framing.
Part I: AI Stops Being a Feature and Becomes the Foundation
The core claim of AlphaSense's first section is a shift in kind, not just degree: AI in medtech is moving "away from point solutions and standalone tools to full integration across clinical and device ecosystems." That is a meaningfully different claim than "hospitals are buying more AI products," and it's worth taking seriously on its own terms, because it changes what kind of company wins.
The adoption numbers, and what they actually rest on
AlphaSense cites industry survey data putting 70% of healthcare and life sciences organizations as "actively using AI," with broker research suggesting 88% of hospitals plan to leverage AI in 2026, primarily for medical record analysis, clinical imaging, and revenue cycle. Both figures sit behind AlphaSense's platform, sourced to a broker report and a survey I can't pull directly. I'd treat them as directionally credible but not independently confirmed — they're consistent with, if somewhat higher than, the public survey data I could reach. A 2026 HFMA survey of 95 healthcare finance professionals found 27% of organizations actively deploying AI at scale across multiple functions and another 53% running pilots in select areas — which is a different framing (deployment maturity within revenue cycle specifically) but points the same direction: AI adoption in the "boring," unglamorous parts of hospital operations is now the median case, not the exception. Separately, an NVIDIA-commissioned 2026 healthcare AI survey found 61% of medical-technology respondents already using AI for medical imaging, with 57% reporting measurable ROI from it. Different methodology, similar order of magnitude, same direction of travel.
The part of AlphaSense's claim I could verify most directly is the regulatory footprint, and it's the strongest piece of evidence for "AI has genuinely arrived" in this whole section. The article states 943 FDA-approved AI-enabled devices in radiology against just 109 in cardiovascular. I couldn't reproduce that exact split independently, but the aggregate figures I found from FDA-tracking outlets are consistent with the ratio: by the end of 2025, the FDA's cumulative AI/ML device authorization count had reached roughly 1,451, with radiology accounting for somewhere between 76% and 77% of the total — meaning radiology-specific clearances land in the 1,000-plus range against a low hundreds count for every other specialty combined. 2025 alone saw 331 new AI device authorizations, the most in the FDA's history for a single year. Radiology isn't just the leading specialty for AI adoption; it's disproportionately the specialty, by roughly an order of magnitude over its nearest competitor.
Why radiology specifically? The task structure is unusually favorable to current-generation AI: a bounded input (an image), a well-defined output (a finding, a measurement, a triage flag), enormous historical labeled datasets from decades of PACS archives, and a clear intermediate step — flagging for human review — that doesn't require the system to be autonomous to be valuable. AlphaSense's claim that embedded imaging AI can flag life-threatening anomalies "often reaching 95% accuracy before the radiologist even opens the file" is exactly this pattern: the AI isn't replacing the read, it's re-ordering the queue so the most urgent studies get seen first. That's a genuinely different, and much easier, integration problem than autonomous diagnosis — which is precisely why it's the specialty where "integration" claims are most credible right now.
The two problems nobody's marketing deck leads with
AlphaSense's own article, to its credit, doesn't stop at the adoption numbers — it flags two structural problems, and both are worth sitting with longer than the article does.
The first is hallucination and safety. AlphaSense cites an AlphaSense-platform expert call in which a diagnostic radiologist warns that AI hallucinations can lead to incorrect reporting and adverse clinical interventions. That call is paywalled, but the underlying concern is well documented in the open literature. A 2026 review of agentic AI and large language models in radiology catalogs hallucination as a first-order barrier to clinical trust — fabricated anatomical structures, missed findings, incorrect laterality, invented measurements — the kind of error that is plausible-sounding precisely because the model has learned what a correct-sounding report looks like, independent of whether it's actually correct for this image. The mitigations being proposed in that literature — multi-agent cross-validation, retrieval-augmented generation grounded in verified sources, explicit uncertainty quantification — are architecturally similar to what Microsoft's MAI-DxO diagnostic orchestrator does for text-based diagnosis, which I wrote about in The Compressed Career: don't trust one model's raw output, orchestrate several and force them to reconcile.
The second problem is data fragmentation, and it's the less glamorous but arguably more important constraint: high-quality foundation models require access to exceptionally clean, diverse, and well-governed clinical data to train effectively, and most health systems simply don't have that internally. A recent narrative review of foundation models in U.S. radiology makes the same point from the research side — models like CheXzero and BioMedCLIP show strong benchmark performance but face real limitation from lack of FDA clearance, limited external validation, and integration barriers with the PACS and RIS systems that actually run a radiology department day to day. The bottleneck on "AI as infrastructure" isn't model capability anymore. It's the unglamorous work of data governance, interoperability, and validation pipelines — which is a much slower-moving, much less venture-fundable problem than training a better model, and it's exactly the kind of problem that determines which vendors actually survive the transition from pilot to plumbing.
The regulatory picture is more contested than the trend piece suggests
AlphaSense frames the EU AI Act as a fixed catalyst: "the majority of substantive requirements... take effect in August 2026." When I went looking for the primary timeline, the picture is messier and actively moving. The core high-risk obligations under the AI Act — conformity assessments, technical documentation, human oversight requirements — do become applicable from August 2, 2026 for standalone high-risk AI systems under Annex III. But AI embedded in already-regulated products, which is where most AI-enabled medical devices actually sit (Annex I, alongside the EU's Medical Device Regulation), has a different and later runway. As of early-to-mid 2026, the Council of the EU has been negotiating a further push — standalone high-risk systems potentially moving to December 2027, and product-embedded high-risk AI, including medical devices, potentially moving out to August 2028. That negotiation was still in trilogue as of this writing.
The practical takeaway for a medtech company isn't "the deadline is August 2026, be ready." It's closer to: the direction of travel is toward more stringent, continuously-monitored obligations, but the exact date at which any given device category must comply is itself a live regulatory question being renegotiated in real time. Building compliance infrastructure now is still the right call — technical documentation and post-market monitoring systems take years to stand up regardless of which exact date lands — but anyone budgeting around a hard August 2026 wall for embedded medical device AI specifically is probably budgeting against a deadline that's already shifted.
Part II: The Robotics Market Stops Being a Monopoly Story
AlphaSense's second section makes a claim that would have sounded implausible five years ago: surgical robotics, long understood as "the Intuitive Surgical market," is becoming an actual multi-vendor contest. The broader medical robotics market is cited at $13.7 billion in 2025 growing to $27.1 billion by 2030. I couldn't locate that exact figure from a public source — market-sizing estimates for this category vary enormously depending on scope (surgical robots alone versus medical robotics broadly, which includes rehabilitation and hospital logistics robots), with public estimates I found ranging from roughly $19 billion to $77 billion by decade's end depending on the research house and category definition. The number matters less than the direction every estimate agrees on: mid-teens-to-high-teens percentage compound annual growth, driven by the same two structural forces AlphaSense names — an aging population and a healthcare workforce shortage that makes throughput-per-clinician-hour an increasingly binding constraint.
Intuitive's lead is real, and it's also visibly under pressure
Intuitive Surgical's position is exactly as dominant as the trend piece suggests, with one important nuance on timing. The da Vinci 5 system received FDA clearance for cardiac procedures — mitral valve repair, IMA mobilization, atrial septal defect repair, and several other cardiac indications — in late January 2026, which is genuinely new and expands the platform into a specialty (cardiac surgery) where robotic assistance has historically lagged general and urologic surgery. Intuitive's own guidance for 2026 is worldwide procedure growth of 13% to 15%, a deceleration from 2025's 18% — still very strong growth off an enormous base, but the deceleration itself is a data point worth watching, and it's consistent with a market where new entrants are starting to take incremental share rather than Intuitive capturing all category growth by default. By the end of Q3 2025, Intuitive had placed 929 da Vinci 5 systems, up from 689 just one quarter earlier — the installed base is scaling fast, and each of those systems is a multi-decade annuity of instrument and service revenue, which is the real economic engine under the procedure-count headline.
The challengers are further along than "entering the market" suggests, and further behind on approval than headlines imply
Medtronic's Hugo RAS system is the most concrete challenger story, and the public record backs AlphaSense's framing closely. The FDA cleared Hugo for urologic procedures — prostatectomy, nephrectomy, cystectomy — on December 3, 2025, a category covering roughly 230,000 U.S. procedures annually. The clearance rests on the Expand URO IDE study: 137 patients across the three procedure types, meeting its primary safety and efficacy endpoints. Medtronic has signaled it intends to file for general and gynecologic surgery indications next, which would meaningfully broaden Hugo's addressable procedure volume beyond urology alone. This is the first genuine soft-tissue robotic platform to clear FDA review as a direct alternative to da Vinci, not an adjacent or complementary system — which is precisely why it matters more than a typical competitive product launch.
Johnson & Johnson's OTTAVA tells a different story than the trend piece implies. AlphaSense describes it as "targeting approval this year." What I found in J&J's own press releases and SEC filings is a company that submitted OTTAVA to the FDA for de novo classification in January 2026 — covering upper-abdomen general surgery procedures like gastric bypass, gastric sleeve, and hiatal hernia repair — built on IDE study data including a completed Roux-en-Y gastric bypass trial, with a separate inguinal hernia trial still enrolling under a late-2025 IDE approval. As of this writing in July 2026, I found no public record of OTTAVA actually clearing FDA review; it remains in the submission-and-review stage. "Targeting approval this year" is a defensible characterization of J&J's stated ambition, but it reads differently once you know the system is still awaiting a first decision rather than moving toward launch — the gap between "submitted" and "cleared" has, for da Vinci 5's own cardiac indication, historically run several months to over a year, and de novo review (a novel-device pathway, since OTTAVA has no exact predicate) tends to run longer than a standard 510(k).
In orthopedics, Stryker's position is less a challenger story and more an entrenchment story. The Mako platform's installed base — over 3,000 systems, more than 2 million cumulative global procedures — plus a Morgan Stanley survey finding that 75% of hospital executives planning to purchase an orthopedic robot in the next 12 months intend to buy a Mako, describes a company with such a commanding position that the more interesting news isn't defense of the core platform but its extension downward. Stryker's February 2026 limited release of Mako RPS — a handheld, cutting-block-free robotic saw compatible with its existing Triathlon knee system — is a deliberate move into the segment of surgeons and sites who want some of the accuracy benefit of robotics without the capital cost, footprint, and console-based workflow disruption of a full Mako system. That's a strategically different bet than Medtronic's or J&J's: rather than a new full-scale platform competing head-to-head, it's a lower-cost, lower-friction product designed explicitly to widen the addressable market — including the Ambulatory Surgery Centers AlphaSense flags as a 2026 growth vector, where capital budgets and physical space are both tighter than in a hospital OR.
Telesurgery and training: the two capabilities that make the rest of this scale
Two smaller items in AlphaSense's robotics section are, in my reading, more structurally important than their brief mentions suggest, because they address the two things that actually gate how far robotic surgery can expand: geography and workforce.
Sovato's telesurgery platform closed a $26 million Series B in November 2025 (bringing total funding to $41 million), notably with Intuitive itself as an investor — a market leader backing a company whose stated purpose is to make robotic surgery system-agnostic and deliverable to rural or underserved locations regardless of which platform is physically present. That's a meaningfully different bet than a hospital simply buying more robots: it's an attempt to decouple surgical expertise from surgeon physical location entirely, using high-speed connectivity, real-time imaging, and haptic feedback. If it works at clinical scale, it addresses the workforce-shortage driver AlphaSense names for the whole robotics category more directly than any single new robot platform does, because it multiplies the effective reach of the specialists who already exist rather than requiring new ones to be trained.
Which is the other constraint: robotic platforms proliferating faster than the surgeons trained to use them creates its own bottleneck, and that's what Surgical Science's RobotiX Express, launched in October 2025, is built to relieve. It's a suitcase-portable, high-fidelity robotic surgery simulator, explicitly positioned against the current default of training being tied to OR availability — a scarce, expensive resource that most institutions ration tightly. Decoupling training from live-console time via a genuinely portable simulator is a small-sounding announcement that's actually load-bearing infrastructure for every other trend in this section: none of the new platforms from Medtronic, J&J, or Stryker matter commercially if there aren't enough credentialed surgeons to operate them, and simulation is the only scalable way to build that credentialed pipeline without tying up a $2 million OR console for training instead of surgery.
What This Actually Means, Taken Together
Line up Part I and Part II and a shared structural pattern emerges that's easy to miss reading them as separate "trends." In both AI software and surgical robotics, 2026's real story isn't a single breakthrough product — it's the maturation of the surrounding infrastructure that determines whether a capability that already works in a demo can actually run at scale: data governance and validation pipelines for AI, regulatory clearance pathways and reimbursement alignment for robotics, training and connectivity infrastructure for both. The technology curve, in both cases, has outrun the institutional and logistical curve, and 2026's actual news is the institutional curve visibly starting to catch up — an FDA that authorized more AI devices in 2025 than any prior year, an EU building (and still actively renegotiating) a genuinely comprehensive AI-specific device framework, a training-simulator company solving the unglamorous surgeon-pipeline problem, a telesurgery startup solving the geography problem.
None of that is as exciting as "AI beats doctors" or "autonomous surgical robot," which is probably why it gets less coverage than the headline-grabbing capability claims I've written about elsewhere on this site. But if you're trying to figure out which medtech bets actually compound over the next five years rather than making a good conference keynote, the plumbing is where I'd be looking.
References and Sources
- AlphaSense. "7 Medtech Trends to Watch in 2026." alpha-sense.com/resources/research-articles/medtech-trends
- The Imaging Wire. "FDA AI Approvals Surge Past 1k for Radiology" (Dec 2025). theimagingwire.com
- IntuitionLabs. "FDA's AI Medical Device List: Stats, Trends & Regulation." intuitionlabs.ai
- HFMA. "The Revenue Cycle of the Future: AI boom and workflow redesigns." hfma.org
- NVIDIA Blog. "From Radiology to Drug Discovery, Survey Reveals AI Is Delivering Clear Return on Investment in Healthcare" (2026). blogs.nvidia.com
- PMC. "Agentic AI and Large Language Models in Radiology: Opportunities and Hallucination Challenges." pmc.ncbi.nlm.nih.gov
- PubMed. "The Emergence of Foundation Models in U.S. Radiology: A Narrative Review of Clinical Utility, Safety, and Evaluation." pubmed.ncbi.nlm.nih.gov
- Certivo. "EU AI Act August 2026: Compliance Guide for Manufacturers Integrating AI Into Products." certivo.com
- MDXCRO. "EU AI Act for Medical Devices: SaMD Compliance Deadlines & Requirements." mdxcro.com
- Gardner Law. "The EU AI Act Has Arrived." gardner.law
- Intuitive Surgical. "Da Vinci 5 Cleared for Cardiac Procedures" (Jan 2026). isrg.intuitive.com
- MedTech Dive. "Intuitive says general surgery, acute care fuel US robot momentum" (2026). medtechdive.com
- MedTech Dive. "Intuitive readies da Vinci 5 for broader launch after placing 110 robots in Q3." medtechdive.com
- Intuitive Surgical 10-Q filings, Q2/Q3 2025. sec.gov
- Medtronic. "Medtronic announces FDA clearance of Hugo robotic-assisted surgery system for urologic surgical procedures" (Dec 3, 2025). news.medtronic.com
- Urology Times. "FDA grants clearance to Hugo robotic-assisted surgery system for urologic procedures." urologytimes.com
- Johnson & Johnson. "Johnson & Johnson Submits OTTAVA Robotic Surgical System to the U.S. FDA" (Jan 2026). jnj.com
- MedTech Dive. "J&J submits FDA de novo request for Ottava robot in general surgery." medtechdive.com
- Stryker. "Stryker introduces Mako Handheld Robotics with the limited market release of Mako RPS" (Feb 2026). stryker.com
- Medical Design & Outsourcing. "Stryker aims for ASCs and reluctant surgeons with its new Mako RPS robotic system." medicaldesignandoutsourcing.com
- Sovato. "Sovato Closes Series B to Advance the World's First and Only Remote Robotic Surgery & Procedure Platform" (Nov 2025). sovato.com
- Medical Design & Outsourcing. "Telesurgery software provider Sovato lands Intuitive as a new investor in $26M Series B round." medicaldesignandoutsourcing.com
- Surgical Science. "Surgical Science Launches RobotiX Express: A Breakthrough in Accessible Robotic Surgery Training" (Oct 2025). surgicalscience.com
- Grand View Research. "Medical Robotic Systems Market Size & Share Report, 2030." grandviewresearch.com
- Dr Neal Aggarwal. "The Compressed Career: What AI Means for Physicians and Surgeons," July 4, 2026. drnealaggarwal.info/post/2026-07-04-the-compressed-career
This essay reflects publicly available research and reporting as of July 2026. Figures attributed to the AlphaSense platform's proprietary research (broker reports, expert calls, and surveys behind its paywall) are presented as reported by AlphaSense, not independently verified by me, and are noted as such throughout. Nothing in this essay constitutes investment, medical, legal, or regulatory advice, and no part of it should be used as the basis for a clinical, purchasing, or investment decision without independent verification of the underlying primary sources.