The AI adoption boom is showing little sign of going bust as a new report by venture capital firm Menlo Ventures reveals that healthcare organizations are adopting AI 2.2 times faster than the broader economy.
“For years, AI in healthcare felt like a promise waiting for proof,” said Greg Yap, Partner at Menlo Ventures, in a statement. “Now we’re seeing evidence at scale that AI tools can actually fit into clinical workflows, make work easier, and improve outcomes.”
The numbers back it up: the findings from the State of AI in Healthcare Report show a record $1.4 billion in AI spending in 2025 — nearly triple 2024’s total — underscoring the growing role of AI in transforming clinical workflows, patient engagement, and drug discovery.
Providers drive majority of AI investment
Healthcare providers account for roughly 75% of AI investment, or about $1 billion of the total spend, according to the report. The majority of that funding has gone toward easing administrative headaches, including clinical documentation ($600 million) and coding and billing automation ($450 million). Patient engagement and prior authorization tools have shown the fastest growth rates at 20x and 10x year over year, respectively.
The report also notes that buying cycles for AI tools in the sector are shortening as providers race to capitalize on operational gains. Buying cycles for health systems are now 18% faster, and for outpatient providers, there has been a 22% acceleration. However, payers remain cautious, maintaining slower adoption cycles as they continue to test AI applications.
The biggest winners of this AI spending spree are startups, capturing 85% of the total spend. However, established electronic health records (EHR) providers remain the preferred vendors for many health organizations looking to invest in AI tools, indicating that brand familiarity and integration ease continue to be major factors in purchasing decisions.
Meanwhile, 67% of outpatient providers using ambient scribe tools, which are natural language AI systems that run in the background during doctor-patient consultations, are expected to switch vendors within three years, highlighting the ongoing battle for customer retention.
Finally, 66% of pharma and biotech companies surveyed are leading the charge in experimental AI across the drug development lifecycle by zeroing in on building or fine-tuning proprietary models, as well as developing data analysis tools that will ultimately design or conduct medical experiments.
Workflow integration emerges as the new challenge
The report comes at an inflection point in healthcare AI. Just this week, medical AI platform OpenEvidence, dubbed “ChatGPT for doctors,” raised $200 million in a funding round led by Google Ventures, pushing its value to an attention-grabbing $6 billion — signaling that the sector is moving past pilot models to more closely focus on specialized, domain-trained AI models.
Menlo’s findings suggest that AI integration, where product design and workflow alignment can create bottlenecks, is the next growth phase for the sector. As organizations move from pilot to production, the challenge is no longer proving the value of AI, but ensuring it works seamlessly within the complex workflows that keep hospitals, labs, and payer systems humming.
“The interesting thing about this phase of AI in healthcare is that the hard problems aren’t model problems anymore,” said Derek Xiao, Principal at Menlo Ventures, in a statement. “They’re product and workflow problems, whether that’s integrating AI into a busy hospital or embedding agents in wet lab processes for drug discovery. The companies that figure that out are poised to redefine how care and science move forward.”
Back in March, Microsoft unveiled its new AI voice assistant Dragon Copilot to help healthcare professionals with the burden of paperwork.