When does AI Health Software Become an FDA Regulated Medicare Device? Lessons from WHOOP and UpDoc
Key Takeaways:
Adding AI to health software does not, by itself, trigger FDA medical-device regulation. What matters is how the software functions, its intended use and claims, the risk it poses to users, and the controls built around it.
The FDA’s June 2026 WHOOP closeout, paired with its January 2026 General Wellness guidance, shows that low-risk wellness wearables can keep estimating physiologic metrics like blood pressure under enforcement discretion, provided they avoid diagnostic claims and clinical-style presentation.
UpDoc’s 510(k) clearance shows that clinically meaningful, AI-enabled software can still reach the market through existing pathways—here, by pairing a deterministic, validated dosing core with a conversational AI layer.
Barrier to Market Entry or On-Ramp? FDA Signals a Nuanced Path for Digital Health and Wellness
For years, digital health companies and their investors have grappled with a fundamental regulatory question:
When does software intended to support health and wellness become a medical device subject to FDA oversight?
The answer has become increasingly complex as companies incorporate generative AI into software platforms capable of influencing clinical care, whether offered directly to consumers or to providers. The Food and Drug Administration (FDA), meanwhile, has continued to refine its approach to this area through its Software as a Medical Device (SaMD) guidance, seeking to balance patient protection with innovation while applying decades-old statutory and regulatory concepts to rapidly evolving technologies.
Two recent developments provide useful insight into the FDA’s current thinking:
The WHOOP Closeout Letter: Formally concluding an enforcement action concerning its Blood Pressure Insights feature after product modifications and the release of updated General Wellness guidance.
The UpDoc 510(k) Clearance: A venture-backed startup achieving traditional FDA clearance for its AI-enabled prescription management platform for Type 2 diabetes patients.
Viewed together, these developments illustrate FDA's continued reliance on a risk-based regulatory framework. Lower-risk wellness technologies may qualify for enforcement discretion when appropriately designed and marketed, while software performing clinically significant functions (including AI-enabled software) can continue to reach the market through existing device pathways when supported by appropriate controls and validation.
For digital health companies that incorporate AI into their software the emerging lesson is that details on how the software functions and presents data to users all greatly matter when it comes to whether or not the FDA will enforce a particular regulatory pathway. This includes considerations of:
How the software functions and the role that AI plays in the various functions,
The intended use reflected in official claims or that are inherent to the way the data is presented,
The risks the software presents to users and the role it has in clinical decision making, and
The controls implemented to mitigate those risks.
WHOOP: How far does FDA’s Flexibility for Wellness software extend?
The WHOOP matter drew significant attention after the FDA issued a Warning Letter challenging the company's Blood Pressure Insights feature. Within that Warning Letter, the FDA took a somewhat surprising stance, stating that blood pressure data is inherently medical in nature due to its intrinsic role in diagnosing hypertension and hypotension. The Agency questioned whether the feature remained within the scope of a low-risk general wellness product or instead met the statutory definition of a medical device simply because of the type of data it was processing, regardless of WHOOP’s claims that the data was not intended for clinical purposes.
For many developers, this appeared like a warning shot that traditionally clinical physiologic metrics would be viewed as inherently “medical,” raising considerable concerns that the FDA might be narrowing the practical reach of its General Wellness Policy for wearables. However, the regulatory landscape shifted favorably in January 2026, when the FDA finalized revisions to its General Wellness guidance.
The Shift in Physiologic Data Regulation
In that updated guidance, the FDA reversed its position on physiologic data, indicating that products using non-invasive sensing to estimate, infer, or output physiologic parameters would still qualify as General Wellness products not requiring FDA oversight if they met certain criteria. Despite the updated guidance, WHOOP's conversations with the FDA continued. Months later, in June 2026, the FDA issued a Closeout Letter terminating the enforcement matter after the company modified aspects of the Blood Pressure Insights feature and related marketing claims. These changes were explicitly intended to reinforce to users that the data was not intended to be diagnostic in nature. Rather than requiring WHOOP to pursue premarket authorization, the FDA accepted these changes because they addressed the underlying regulatory concerns while allowing the core functionality of the feature to remain largely the same.
What can Software Developers take away from FDA’s General Wellness Guidance?
For developers, the increased flexibility provided under the updated General Wellness guidance is a valid pathway to market. However, a product’s intended use involves not just official statements, but also specific details of the data collected and, importantly, how that data is presented to users within the interface.
Neither the WHOOP matter nor the new guidance should not be interpreted as reduced scrutiny of software that performs medical device functions. Products intended to diagnose disease, direct treatment, recommend therapies, or otherwise inform clinical decision-making remain substantially more likely to fall within FDA's medical device jurisdiction.
UpDoc: A Regulatory Blueprint for Clinical AI?
At the other end of the regulatory spectrum, UpDoc publicly announced in June 2026 its FDA clearance K253281 of December 2025 through the traditional 510(k) pathway for a software platform that supports prescription management and medication monitoring for patients with Type 2 diabetes. While the cleared indication is relatively specific, the regulatory significance extends far beyond diabetes management.
Notably, the FDA cleared the device as substantially equivalent to an existing drug-dose calculator (Hygieia’s d-Nav System), so the clinical decision-making core is deterministic in nature, vesting greater responsibility with the stewarding physician rather than delegating clinical judgment to a generative AI model. The patient-facing, large-language-model layer sits around that validated core rather than driving the dosing decision itself.
This distinction is consistent with the FDA's longstanding approach to software regulation. Historically, the agency has been more comfortable evaluating software that produces predictable, validated, and reproducible outputs than systems whose clinical conclusions depend on unconstrained generative responses that are more difficult to validate prospectively.
What is Architectural Decoupling as a Regulatory Strategy?
One emerging regulatory strategy involves separating the clinical decision-making engine from the conversational interface. From a regulatory perspective, this separation may simplify verification and validation by allowing developers to demonstrate the safety and effectiveness of the underlying clinical logic while limiting the AI's role to supporting user interaction and workflow.
UpDoc also reinforces a broader point: innovative AI-enabled software does not necessarily require entirely new regulatory pathways. Existing FDA frameworks, including the 510(k) process, remain well suited for appropriately designed software where intended use, clinical validation, risk controls, and human oversight are clearly established.
Looking Ahead: Product Design as Regulatory Strategy
Viewed together, the WHOOP and UpDoc developments provide a useful illustration of the FDA's current regulatory trajectory:
Low-Risk Wellness: The FDA appears willing to continue exercising enforcement discretion for appropriately designed, low-risk wellness technologies that avoid medical device claims.
Clinical AI & SaMD: The agency continues to apply established premarket pathways to software performing clinically significant functions, including software that incorporates AI.
Recent agency actions suggest the FDA is increasingly focused on intended use, the role software plays in clinical decision-making, the potential consequences of erroneous outputs, and the effectiveness of the controls implemented to mitigate those risks. For digital health companies, regulatory strategy is becoming inseparable from product design.
Key Action Items for Founders and Device Developers:
Incorporate Regulatory Considerations Early: Consider regulatory strategy alongside product design from the earliest stages of development, NOT after commercialization.
Audit Your User Interface: Ensure your UI and marketing claims align with your desired regulatory classification. Minor interface tweaks can mean the difference between enforcement discretion and a warning letter.
Isolate Generative AI: If utilizing generative AI, consider decoupling user workflows and communication modules from your core, deterministic clinical logic to simplify validation.
Frequently Asked Questions (FAQs)
Does adding AI to a digital health product automatically make it an FDA-regulated medical device?
No. The FDA’s analysis turns on what the software does—its intended use, the claims made for it, the risk it poses to users, and the controls that mitigate that risk—not on whether it uses AI. Software that diagnoses disease, directs treatment, or otherwise informs clinical decisions is far more likely to be regulated as a device than software that supports general wellness.
What changed in the FDA’s January 2026 General Wellness guidance?
The FDA clarified that products using non-invasive sensing to estimate or report physiologic parameters—including blood pressure—can remain low-risk general wellness products subject to enforcement discretion, provided they avoid diagnostic or “medical-grade” claims, do not characterize outputs as abnormal or pathological, and meet the guidance’s other criteria.
Why did the FDA close the WHOOP warning letter without requiring 510(k) clearance?
After the July 2025 warning letter, WHOOP modified its Blood Pressure Insights feature and labeling—including how the data is displayed to users—to reinforce that it is a wellness estimate rather than a clinical reading. In a June 2026 closeout letter, and consistent with the updated General Wellness guidance, the FDA said it did not intend to enforce device requirements against the modified feature.
What did the FDA actually clear in UpDoc’s 510(k)?
The cleared device (K253281, December 2025) is a prescription software product that helps adults with Type 2 diabetes manage and titrate basal insulin. The FDA cleared it as substantially equivalent to an existing drug-dose calculator, which means the clinical decision-making core is deterministic; the patient-facing large-language-model functions as a conversational layer around that validated core.
What is the “separate the engine from the interface” strategy for clinical AI?
It refers to designing a product so that the validated clinical decision-making logic is kept distinct from the generative, conversational AI that interacts with users. This can simplify verification and validation by letting developers demonstrate the safety and effectiveness of the deterministic core while limiting the generative AI’s role to communication and workflow.
How does Nixon Law Group help HealthTech Software Device developers?
As always when it comes to healthcare, the devil is in the details. At Nixon Law Group, we are laser-focused on healthcare innovation. We act as true partners, translating complex, evolving regulations into actionable corporate and product strategies built to accelerate your results. Our FDA and AI Governance attorneys take an integrated, comprehensive approach to your business, ensuring your product design aligns seamlessly with your regulatory, reimbursement, and data privacy goals.
Contact Nixon Law Group today to speak with our Digital Health and SaMD experts and ensure your regulatory strategy is built for success.