Healthcare is experiencing a transformation that seemed impossible just a few years ago. AI is detecting diseases earlier, personalizing treatments with unprecedented precision, and delivering complex care in more accessible settings. It’s driving measurable and meaningful improvements in patient outcomes, workflow efficiency, and care accessibility. It’s raising the standard of care around the world.
Yet, there’s another impact we need to discuss: AI is also resource-intensive. With $3 trillion projected to be invested in AI by 2028, its energy and water consumption has rightfully raised concerns.
Some estimates suggest that without new capacity, AI could consume over 9% of the U.S. electricity grid in the next decade, up from nearly 3% today.
It’d be easy to frame this as a binary choice — innovation or sustainability. But that’s not realistic, and it’s not responsible. Healthcare can’t afford to prioritize one at the expense of the other, even when weighing healthcare benefits against environmental costs can become impossibly complex. We need to be pragmatic about how we can advance both simultaneously without compromising on either.
So, let’s talk about what it will take to get this balance right.
Welcome to Healthcare 5.0
We’re entering an era defined by ultra-personalized, predictive, and preventive care — Healthcare 5.0. It’s a fundamental shift in how we deliver care that’s possible as we harness the power of AI.
Diagnostic tools leveraging AI are enabling earlier, more precise disease detection. The AI-powered GI Genius™ improves colonoscopy detection rates by up to 14%, helping physicians identify polyps they might otherwise miss. Aside from the direct patient health outcome benefit, better detection leads to earlier intervention, which can help prevent more resource-intensive treatments down the line.
Personalized treatment planning also shows real, measurable results. Spinal surgery using patient-specific alignment, such as the UNiD™ adaptive spine intelligence (ASI) system, has been shown to lower revision rates compared to traditional approaches. While this approach may require more resources initially, eliminating the need for a potential second surgery mitigates future consumption and delivers a superior patient experience. In addition, AI-enabled spinal surgery planning can provide predictable and reproducible outcomes for patients with spinal deformity.
AI and digital tools are also redefining where and how care is delivered. Digital capabilities, such as the Touch Surgery™ ecosystem, are being explored for ambulatory settings where using Live Stream enables access to experts from elsewhere. In an emergency or at a remote location, these tools support and augment on-site staff. Moving complex care from hospitals to outpatient centers also dramatically reduces resource use since hospitals generate an estimated over 29 pounds of waste per bed per day. Ambulatory care requires a fraction of that infrastructure while making care more accessible and affordable.
When it comes to workflow, AI is helping healthcare systems do more with what they have. Early data shows AI could be used to augment or automate 70% of healthcare workers’ tasks, allowing organizations to serve more patients without expanding infrastructure. As an example, Medtronic’s own LINQ II™ insertable cardiac monitor — deploying AccuRhythm™ AI algorithms — can save clinicians 400 hours annually of false alert reviews for every 200 patients they serve. Accomplishing more with existing resources is an efficiency that matters for both business and the planet.
It also matters for individual clinicians. By reducing administrative burden, AI can free up time for patient interactions and help address burnout.
The measurement challenge
Here’s where it gets complicated. Current sustainability metrics focus heavily on immediate environmental impacts — the energy consumed by a query or the water used to cool a data center. They don’t capture the full life cycle impacts of AI-enabled healthcare.
The industry needs frameworks that can meaningfully compare the environmental costs of AI against the healthcare value it creates. How do we measure the resources AI consumes against avoided hospital stays, prevented complications, or lives extended? AI’s ability to improve outcomes, efficiency, and accessibility creates value that must factor into sustainability discussions. And we need reliable and industry-wide measurement frameworks sophisticated enough to capture both sides of the equation.
It’s also worth noting that, from a resource-consumption perspective, not all AI is created equal. At Medtronic, we focus primarily on closed-loop AI models that operate within established knowledge paradigms — algorithmic AI that’s far less resource-intensive than large language models analyzing vast universes of open data. It’s an important distinction since generative AI applications require significantly more computational power, while the algorithmic AI embedded in some of our medical devices operates efficiently in more contained environments.
Healthcare companies can’t solve this alone. Enterprise AI providers and cloud computing partners need to measure AI efficiency using metrics that attempt to quantify AI’s performance against the resources consumed — something like “tokens per watt per gallon of water,” perhaps. The measurement challenge is as much about cross-industry collaboration as it is about developing the right metrics.
Practicing the balancing act
So, what can we do?
First, we need strategic AI deployment. Not every problem requires the most computationally intensive solution. The principle should be simple: Use the right AI tool for the right job, and only deploy AI when it demonstrably improves a healthcare success metric. This is what responsible AI should be about — particularly in healthcare, where, unlike other generative AI applications, there should be zero tolerance for hallucinations or clinically indefensible outputs. AI for AI’s sake is part of the problem.
This requires governance systems within companies that guide appropriate tool selection, helping teams match computational intensity to problem complexity while maximizing the benefit created.
We also need perspective on technology maturation. History shows us that emerging technologies become more efficient and cost-effective over time: Solar panels are 40% more efficient than in 2010, while electric vehicle battery energy density has boosted typical ranges from 100 miles per charge to 250–400 miles in around a decade. AI is following the same trajectory, with new chip architectures delivering more computational power while using less energy, and smaller models proving you don’t always need the biggest solution.
We should manage AI’s environmental impact early in its adoption, but we should also be realistic about giving it time to mature.
In the meantime, there are concrete steps we can take. Medtronic is investing in renewable energy at scale — our recent virtual power purchase agreement adds renewable energy equivalent to nearly all our North American energy needs. We’re modernizing facilities with more efficient technology and partnering with AI providers committed to sustainable practices. These aren’t perfect solutions, but they’re actions we can take today.
To learn more about our approach, see the Medtronic FY25 Impact Report.
The path forward
The path forward requires holding two truths simultaneously: AI is driving essential healthcare innovation that improves lives, and AI has environmental impacts we must actively manage. Healthcare leaders need to develop measurement frameworks that capture full life cycle value, implement mitigation strategies now, and collaborate across industries to create sustainable practices.
Healthcare has always faced a twin impact from climate change — the healthcare industry contributes roughly 5% of global greenhouse gas emissions while concurrently treating the adverse health effects caused by environmental pollution and climate change. That creates a compounding burden we can’t ignore.
This isn’t about choosing between innovation and sustainability. It’s about pursuing both with equal commitment, measuring honestly, implementing strategically, and collaborating broadly.
The healthiest future is one in which AI advancement and environmental sustainability aren’t opposing forces but integrated imperatives driving healthcare forward together.
