.gif){kind=link}
Hi! 👋
This repo contains a strategy analysis of the AI in Healthcare industry. It was a team project for OMT 520, Strategic Management of Technology, at Arizona State University. The question we were trying to answer: Microsoft, Google, and IBM are all fighting for the same hospital and insurer budgets, using completely different playbooks. Who actually has a real, lasting edge, and why?
read time: 8-10 min
- What it is: A strategy case study on the AI in Healthcare industry, built using Porter's Five Forces, the Resource-Based View (RBV) and VRIO, Core Competencies, and Generic Business Strategy frameworks.
- Team: A 4-person group project (Team 5): Ankush Harishchandre, Swarda Bhandare, Jingxin Wang, and me. Advised by Dr. Andrea Cherman.
- My part: The market sizing, segmentation, and Five Forces sections (the Industry Case).
- Companies studied: Microsoft, Google, and IBM in depth, with Amazon/AWS for context.
- The market: Global AI health spend went from about $1.1B in 2016 to $22-32B in 2023, and is forecast to cross $180-190B by 2030.
- Bottom line: Microsoft currently holds the strongest, hardest-to-copy position, because of Nuance's clinical voice data, its Epic partnership, and Azure. But the bigger lesson is that nobody wins this market on model quality alone. It comes down to who can combine the tech with real clinical data access, workflow integration, and governance.
.gif?raw=true)
- Why This Industry
- How AI Got to Healthcare
- What AI Actually Does in Healthcare Today
- Frameworks Used
- The Market, By the Numbers
- Five Forces: Is This Industry Actually Attractive
- Microsoft vs Google vs IBM
- How Each One Got Here
- Results and Conclusion
- My Contribution
- Documents in This Repo
- References
- Team
AI in healthcare gets talked about a lot, but most of that talk is about the technology itself: better diagnosis, faster scans, fewer errors. This report looks at it as a business instead. Who is actually making money here, what does it take to compete, and what happens to companies that get it wrong.
That last part matters more than it sounds. IBM built an entire healthcare AI business, Watson Health, spent over $4 billion on it, and then sold it off. Understanding why that happened, and why Microsoft and Google took different paths, is the real point of this report.
AI in healthcare did not show up overnight. It took about 60 years to get here:
| Era | What happened |
|---|---|
| 1960s-70s | MYCIN, an early expert system, tried to help doctors pick antibiotics. It never really caught on, mostly because it was not accurate or well known enough. |
| 1990s-2000s | AI starts helping read mammograms. Around the same time, hospitals switch to electronic health records, which quietly creates the huge datasets AI would later need. |
| 2010s | Deep learning arrives. AI starts matching human specialists at spotting skin cancer and eye disease in images. |
| 2010s | Predictive AI starts forecasting which patients are likely to get sicker, not just diagnosing what is already wrong. |
| Late 2010s-early 2020s | COVID-19 hits. Digital health adoption speeds up fast, and AI starts handling everything from writing doctors' notes to tracking hospital schedules. |
| 2020s | The rulebook catches up. The FDA makes it easier to approve AI medical software, the ONC sets data-sharing standards, and the AMA creates billing codes so hospitals can actually get paid for using these tools. |
The report breaks this into three jobs AI is doing right now:
- Clinical decision support. Reading images and patient data to catch tumors, heart problems, and infections earlier and more consistently than manual review.
- Administrative automation. AI scribes that write clinical notes automatically, plus tools that handle medical coding, billing, and referrals. This matters because clinician burnout and paperwork are a real cost driver in healthcare.
- Operational optimization. Predicting patient flow, staffing needs, and supply levels so hospitals run fewer bottlenecks and use their space and staff better.
All three run on the same three technical building blocks: machine learning (spotting patterns in past data), deep learning (handling messy, high-dimensional data like scans and genomic sequences), and natural language processing (turning doctors' unstructured notes into structured, usable data). And all three share the same weak spots: algorithmic bias, data security, and models that do not generalize well across different patient populations. That is why hospitals keep a human in the loop, and why the FDA has a specific regulatory framework for AI-based medical devices.
The report leans on five classic strategy tools to make sense of the industry. In plain terms:
- Porter's Five Forces: how attractive is this industry to compete in, based on how easy it is for new players to enter, how much power suppliers and buyers hold, whether substitutes exist, and how intense the existing rivalry is.
- Resource-Based View (RBV) and VRIO: a company only gets a real, lasting edge from a resource that is Valuable, Rare, hard to Imitate, and that the company is Organized to actually use.
- Core Competencies: the deeper organizational skill sitting behind the resource. Not what a company has, but what it is quietly really good at doing, and how hard that is to copy.
- Generic Business Strategies: a company competes on being the cheapest (cost leadership), being different in a way people will pay more for (differentiation), or picking a narrow slice of the market (focus). Try to do a bit of everything and you end up "stuck in the middle" with no real edge.
- Corporate and Cooperative Strategy: how a company grows, through acquisitions, mergers, and diversification, and who it partners with to get there.
Global AI health spending
| Year | Spend |
|---|---|
| 2016 | about $1.1B |
| 2023 | $22-32B |
| 2030 (forecast) | $180-190B |
| 2032 (forecast) | $500B+ |
Forecast growth rate (CAGR), 2024-2030: 35-45%.
Where the spending happens
| Region | Share |
|---|---|
| North America | 49% |
| Europe | 27% |
| Asia-Pacific | 18% |
| Rest of world | 6% |
North America spends the most overall, but Asia-Pacific is growing the fastest, driven by China and India.
Where the market breaks down by use case
| Segment | Share |
|---|---|
| Clinical Decision Support | 35% |
| Administrative Automation | 20% |
| Drug Discovery and Development | 18% |
| AI-Assisted Surgery | 15% |
| Remote Monitoring / Patient Engagement | 12% |
Who holds the market
| Company | Share |
|---|---|
| Microsoft | 18% |
| IBM | 18% |
| 16% | |
| AWS | 12% |
| Everyone else | 36% |
| Force | Rating | Why |
|---|---|---|
| Threat of new entrants | Low to moderate | Getting the clinical data, the EHR integrations, and the FDA and HIPAA approvals all take years. Microsoft's Epic relationship and Google's Mayo Clinic and NHS partnerships are hard for a newcomer to replicate quickly. |
| Buyer power | Moderate to high | Hospitals and insurers are sophisticated buyers. They run pilots, ask for proof of ROI and security, and hold off scaling until results show up. |
| Supplier power | High | Two groups hold real leverage here: GPU makers like Nvidia, and the hospitals and health systems that control the patient data needed to train these models. |
| Threat of substitution | Moderate | Manual workflows, older analytics tools, and consulting firms like McKinsey can still get a lot of this done without AI. |
| Competitive rivalry | High | Big Tech and smaller specialists are all chasing the same hospital and insurer budgets, each with a different angle. |
Supplier power and competitive rivalry are the two forces doing the most work here. Everything else lands somewhere in the low to moderate range.
VRIO compares whether each company's key resource actually holds up as a lasting advantage:
| Company | Key resource | Rare | Hard to copy | Organized to use it | Result |
|---|---|---|---|---|---|
| Microsoft | Nuance's clinical voice data, tied to Epic's EHR | Yes | Yes | Yes, through Azure | Sustained advantage |
| Med-PaLM 2 and DeepMind's research, tied to Mayo Clinic and NHS partnerships | Yes | Not really, R&D can be published and followed | Yes, through Cloud APIs | Temporary advantage, strong potential | |
| IBM | Watsonx, built on hybrid cloud with Red Hat | Limited, fewer EHR ties than the other two | No, built on open-source components others can match | Yes, through its consulting arm | Competitive parity |
And their core competencies, the thing each one is quietly best at:
| Company | What they are actually good at |
|---|---|
| Microsoft | Ecosystem embedding: hiding complex AI inside tools doctors already use every day, like Teams, Outlook, and Epic |
| Research-to-product: turning frontier science like DeepMind's work into cloud APIs fast | |
| IBM | Governance: designing for audits, compliance, and hybrid or on-prem deployment from the ground up |
This shows up in how each one competes. Microsoft runs a broad differentiation strategy, winning by embedding itself directly into clinical workflow. Google also differentiates, but through research and innovation leadership. IBM plays a focused differentiation strategy, going after the narrower slice of customers who care most about governance and compliance.
Microsoft acquired Nuance for $19.7 billion in 2022, which is where its clinical voice data came from. It launched Microsoft Cloud for Healthcare in 2020, deepened its Epic partnership in 2023 to bring Azure OpenAI into Epic's software, and has a long-running Walgreens partnership and OpenAI investment on top of that.
Google built its healthcare position through Alphabet subsidiaries (Verily in 2015, Calico in 2013), the DeepMind acquisition in 2014, and the Fitbit acquisition ($2.1 billion, 2021) for wearable health data. Its long-term research partnerships with Mayo Clinic (since 2019) and the NHS gave it real clinical testing ground.
IBM spent over $4 billion between 2015 and 2016 building Watson Health through acquisitions, then sold those assets off in 2022 when it could not turn that investment into solid growth. It has since refocused on Watsonx and Red Hat's hybrid cloud, leaning into a 10-year research partnership with Cleveland Clinic and work with Anthem and Pfizer.
Right now, Microsoft has the strongest position. The combination of Nuance's clinical language data, deep Epic integration, and Azure's healthcare-ready cloud is genuinely hard for a competitor to copy quickly. Google's advantage is real but earlier stage: its research is excellent, but it still has to prove it can turn that research into something that fits into everyday hospital workflow at scale. IBM's advantage is the narrowest of the three, strong for regulated, compliance-first customers, but limited by weaker access to clinical data and fewer front-line EHR relationships.
The bigger takeaway is that none of the three are winning purely on how good their AI models are. The market is shifting from "which model is best" to "who can actually deploy this inside a real hospital, prove it works, and pass a compliance review." The report's term for the winning position is an "ecosystem orchestrator": a company that combines the technology with deployment, governance, and long-term data access, rather than selling a single point solution.
This was a 4-person team project, and I want to be upfront about which parts were mine. My focus was the market sizing, segmentation, and competitive landscape work: the global spend numbers, the regional and segment breakdowns, the market share by firm, and the Five Forces summary. Ankush, Swarda, and Jingxin worked on the literature review, the VRIO and Core Competencies analysis, and the corporate strategy timelines.
| File | Description |
|---|---|
| Team5_FP8_11302025 (1).pdf | The full report: literature review, frameworks, industry case, strategy analysis, and conclusion |
Full citations are in the report above. Key sources include:
- Topol, E. J. (2019). High-performance medicine: The convergence of human and artificial intelligence. Nature Medicine, 25(1), 44-56.
- Porter, M. E. (1979). How competitive forces shape strategy. Harvard Business Review.
- Porter, M. E. (1985). Competitive advantage: Creating and sustaining superior performance. Free Press.
- Barney, J. B. (1991). Firm resources and sustained competitive advantage. Journal of Management, 17(1), 99-120.
- Prahalad, C. K., & Hamel, G. (1990). The core competence of the corporation. Harvard Business Review, 68(3), 79-91.
- Grand View Research. (2025). Artificial intelligence in healthcare market size, share and trends analysis report, 2025-2030.
- McKinsey & Company. (2023). The state of AI in healthcare: Moving from pilots to scale.