Who Will Be Healthcare's DeepSeek?
Outsider threats to an entrenched industry
Last week, the artificial intelligence community was rocked by the sudden rise of DeepSeek, a scrappy Chinese upstart that seemingly came out of nowhere. Promising ChatGPT (or Claude/Gemini/Grok/Llama—pick your favorite LLM) level performance at a fraction of the cost, compute power, and energy consumption, DeepSeek’s R1 model quickly became the talk of tech circles.
Deep-pocketed American tech giants, long engaged in an expensive, resource-hungry AI arms race, saw their stocks sink and their methodologies questioned. Reactions varied, from Microsoft CEO Satya Nadella pulling out Jevons’ Paradox to Meta’s Mark Zuckerberg promising to stick to his playbook of “borrowing” from the competition. Public bravado aside, the tech industry was clearly put on notice.
How could such powerful incumbents be caught so off guard? How does a tiny AI outsider suddenly threaten some of the most well-resourced companies in history?
More importantly, what lessons can health care learn from AI’s David and Goliath moment?
Can a health care outsider challenge entrenched giants in an industry dominated by regulatory capture, red oceans, and manufactured opacity? Are AI and health care incumbents in similar positions, too big and dominant to face a credible threat?
Or will someone rise to become health care's DeepSeek?
Tools of Incumbency: Regulatory Capture
Artificial intelligence companies, much like health care incumbents, operate in a complex and evolving regulatory landscape. The relationship between AI firms and policy makers mirrors health care’s long-standing dynamic in several ways, most notably how dominant players use regulation to entrench their positions and stifle competition.
Elon Musk and Sam Altman have stoked fears of artificial general intelligence (AGI) to control the pace of AI development. Incumbents like OpenAI and Google lobby for strict AI regulations under the guise of safety, shaping government policy in ways meant to protect their dominance. (Look no further than the recently announced $500 billion Stargate project, a testament to how deeply big tech has embedded itself in national AI strategy.)
While these efforts may appear to be in the public interest, they strategically reinforce the position of large players while making things harder for smaller challengers.
Health care is no stranger to these tactics.
If AI incumbents wield the AGI boogeyman to justify regulatory barriers, hospitals and health systems invoke "misaligned incentives" to suppress competition from independent physicians and others.
Certificate of Need (CON) laws and the moratorium on physician-owned hospitals (POHs) make it nearly impossible to build competing facilities. Billing complexities, site-neutral payment disparities, and diverging reimbursement trends (up for hospitals, down for docs) create widening moats that favor legacy health systems.
Regulation also brings unintended consequences. (Ironically, sometimes to the detriment of the incumbents it was meant to protect.) Anti-China policies—including export controls—may have inadvertently fueled DeepSeek’s rise. By limiting access to American technology, restrictions forced DeepSeek to innovate by doing more with less: creating a leaner, more efficient model that caught incumbents flatfooted.
In health care, regulations have accelerated consolidation, paradoxically raising costs, stripping physician autonomy, and worsening patient care. Policy makers continually introduce legislation making it harder to compete, innovate, and challenge the status quo.
Such regulation is often framed as necessary to deter bad actors. In practice, it frequently serves as a shield for incumbents. But the shield isn’t impenetrable. DeepSeek figured out how to overcome policy restrictions.
Can the same be achieved in health care? Rather than be crushed by regulatory capture, can disruptors use it to their advantage?
The opportunity is there. The government is "all in" on value-based care, yet incumbents have been slow to abandon fee-for-service. Employers, desperate for relief from spiraling health care costs, are more willing than ever to engage outsiders through direct contracting.
The key is understanding and adapting to the regulatory landscape faster than incumbents. By following the DeepSeek playbook, regulatory capture can be turned into an advantage instead of a hinderance.
Red Oceans: Too Entrenched to Innovate?
It would be silly to suggest that artificial intelligence companies don’t innovate. Large language models like ChatGPT are technological marvels, improving with each iteration. But have AI companies become too enamored of their own hype cycles? With billions at stake, have they focused more on hyperscaling than on truly rethinking AI from first principles?
DeepSeek’s breakthrough came from learning to do more with less. Free from the red ocean of Silicon Valley’s AI arms race, the company was able to challenge fundamental assumptions. Big players bought more chips and vacuumed up more data. Meanwhile, DeepSeek rethought model training itself.
As it turns out, for all their brainpower, compute power, and financial power, AI incumbents may already be too bloated. Sound familiar?
Hospitals and health systems build sprawling medical campuses and skyscraping towers. They integrate horizontally and vertically, not to improve care, but to outmaneuver competitors and capture market share. The bigger they grow, the more capital investment and administrative complexity they require.
Academic medical centers churn out research papers but have little incentive to redesign how care is delivered. Traditional systems are optimized for billing, service utilization, and market expansion—not patient outcomes.
Health care red oceans create what Mark Cuban calls the CapEx Arms Race.
The lumbering, entrenched nature of health care incumbents has created an opportunity for willing innovators. Just as DeepSeek achieved its success by challenging existing AI models, health care disruptors can challenge existing care models.
Shedding the bloat of incumbency allows health care innovators to build efficient, patient- and physician-centric models that lower costs instead of inflating them. Entrenchment isn’t some brilliant strategic advantage—it’s an exploitable weakness inviting someone to rethink the system.
Lessons from DeepSeek: Transparency, Consumerization, Commoditization
DeepSeek’s open-source approach challenges the black box nature of AI incumbents. By making its AI model publicly available, DeepSeek shifted the thinking around model development. Major players can no longer rely on data monopolies, proprietary models, or overwhelming (and expensive) compute power. Instead, now anyone can access cutting-edge AI.
DeepSeek’s move accelerates the consumerization and commoditization of AI, lowering the barrier of entry and forcing companies to compete on usability, accessibility, and cost.
This shift mirrors a transformation that is happening in health care, albeit slowly.
Just as AI incumbents thrive on closed models, health care incumbents benefit from a deliberate lack of transparency. Hospitals hide prices, quality metrics, and patient outcomes data behind opaque systems that prevent consumers (patients) from making informed decisions.
That information asymmetry is being challenged. Price transparency laws compel hospitals to disclose actual costs. Patient-reported outcome measures (PROMs) allow health care purchasers to make better-informed decisions. Technology tools make clinical expertise more accessible. Bringing it all together, health care marketplaces allow savvy consumers to find high value care.
Ultimately, greater transparency forces health care incumbents to compete on value rather than deliberate opacity.
DeepSeek’s emergence is expected to accelerate the commoditization of AI models. If high-quality AI is freely available, winners won’t simply have the biggest models, they’ll translate AI effectively into real-world use cases. The next era of AI competition will be about user experience, reliability, and adaptability—not who has the most powerful chips and big-name investors.
Health care is undergoing a similar transition to commoditization.
Unlike AI models, health care won’t become a commodity in the strictest sense; physicians aren’t interchangeable like software. But as transparency and price competition grow, health care services will become more consumer (patient)-driven. The winners won’t be those with the best brand. They’ll be those that offer the best value, experience, and outcomes.
The take home lesson from DeepSeek? The future belongs not to those who rely on legacy but to those who execute.
The DeepSeek of Healthcare
DeepSeek’s rise shows that innovation isn’t about having the most resources or highest degree of regulatory capture. It’s about being efficient and rethinking first principles. Instead of swimming with incumbents in a capital- and compute-intensive red ocean, DeepSeek found a unique way to challenge the status quo and threaten incumbents.
The DeepSeek of health care will follow the same blueprint.
It won’t be a sprawling health system trading on brand name and market power. It won’t surreptitiously influence policy while circumventing transparency. It won’t be a payor favoring stockholders over patients or a VC-backed health tech startup, Big Tech giant, or Big Retailer looking to extract value.
Health care’s DeepSeek will rethink care delivery itself and pose a real threat to incumbents. It will learn from how DeepSeek disrupted AI by exploiting regulatory contradictions, avoiding red oceans and the CapEx arms race, and competing where incumbents can’t: in experience, outcomes, and cost.
Embracing transparency, the DeepSeek of health care will force incumbents into marketplace competition. Instead of building billion-dollar campuses, it’ll develop efficient, engaging, high-quality models that deliver better outcomes at lower costs. While incumbents rely on branding, market dominance, and complex administrative structures, DeepSeek-like innovators will create physician-led, tech-enabled, patient-centric alternatives.
Succeeding in health care’s next era won’t be about entrenchment and incumbency. It will be about innovators capable of thinking differently and putting those thoughts into action.
Who Are the Early Contenders?
Can anyone rattle health care incumbents the way DeepSeek rattled the AI ecosystem?
Several companies are beginning to follow the DeepSeek playbook, rethinking care delivery and challenging health care’s establishment. Tech-enabled, value-based specialty care models like Commons Clinic, Maven Clinic, and Oshi Health are applying high-efficiency, data-driven care models to deliver better outcomes at lower cost.
Much like DeepSeek circumvented the AI arms race, Commons Clinic isn’t trying to outspend hospital systems—it’s rethinking how specialty care should be delivered.
Echoing DeepSeek’s approach, companies like Mishe Health, Health Rosetta, and Transcarent are “open sourcing” health care, navigating patients to transparent, cost-effective care. These models help employers and patients make informed health care decisions by surfacing high-value care.
Finally, groups like Surgery Center of Oklahoma and Mark Cuban’s Cost Plus Drugs leverage consumerization and commoditization to foster competition and lower costs while offering a superior patient experience. As regulations increasingly require transparency, these models will become even more attractive.
This is by no means an exhaustive list. Many more are joining the movement.
Can any of them threaten health care incumbents the way DeepSeek challenged the AI establishment?
The DeepSeek story is evolving. There are legitimate concerns about the company’s methods and motives. Still, DeepSeek proved that you don’t have to be massive or spend lots of money to threaten incumbents. You simply have to be driven, strategic, and nimble.
Organizations that embrace transparency, reject legacy inefficiencies, and prioritize patient experience and outcomes will be the ones to “DeepSeek” health care’s incumbents.
DeepSeek is making AI giants uncomfortable. Health care’s version of DeepSeek may be a matter of “when” not “if.” The only question is: Who will do it first?
Ben Schwartz, MD, MBA
Editor-in-Chief, The Surgeon’s Record
Commons Clinic Senior Clinical Fellow
