AMI Raises $1.03B to Build World Models — Powering the Next Generation of Healthcare AI with Nabla
Our partners at Advanced Machine Intelligence (AMI) announced a $1.03 billion funding round to accelerate the development of a new class of AI systems known as world models— but what does this mean for healthcare?
Through our exclusive strategic partnership with AMI announced at the end of 2025, Nabla will gain first access to these emerging world model technologies, positioning us to help bring the next generation of safe, auditable agentic AI systems into healthcare.
Artificial intelligence has already begun transforming clinical workflows. Large language models have demonstrated remarkable capabilities in documentation, knowledge retrieval, and workflow support. But healthcare is not simply a language problem.
Clinical environments are dynamic systems involving multiple care teams, fragmented technology infrastructure, evolving patient conditions, and decisions that unfold over time. Supporting that complexity requires AI systems capable of more than generating text, systems that can understand situations, anticipate outcomes, and operate reliably within real-world workflows.
This is where world models will profoundly shift healthcare.
The Shift Beyond Language Models
Large language models have unlocked enormous value for healthcare. They have made it possible to automate documentation, extract insights from complex clinical records, and reduce administrative burden for clinicians.
But LLMs also face critical limitations. They are systems based on probability that can struggle with deterministic reasoning, adapting based on inputs like continuous multimodal data, and the ability to simulate how complex environments evolve over time.
World models represent a fundamentally different approach.
Rather than focusing solely on language, world models learn abstract representations of how environments function, more similar to how human beings reason about the world. These systems can simulate how situations evolve and which actions lead to which consequences, allowing them to plan sequences of actions while adapting to real-world constraints. This capability opens the door to use in complex environments like healthcare.
The Impact for Healthcare
Healthcare is one of the most operationally complex systems and has higher consequences for mistakes. Clinical care involves coordinated teams of physicians, nurses, medical assistants, educators, pharmacists, and care coordinators working across EHRs, scheduling systems, imaging platforms, diagnostic workflows, and patient communication tools.
Today’s AI tools often operate within specific tasks such as drafting notes or answering questions. These capabilities are valuable, but they only scratch the surface of what AI could bring to healthcare. The next phase of healthcare AI will require systems capable of understanding workflows, anticipating needs, and interacting safely with clinical systems.
World models provide a path toward that future.
Scaling Clinical AI Across Specialties
A central challenge in healthcare AI is the diversity of medical practice. While large language models perform well on general tasks, applying AI across specialized clinical domains is significantly more complex. Large language models depend on massive datasets, which can make adapting them to specialized environments more difficult.
Healthcare highlights this challenge. The U.S. system includes more than 40 core medical specialties and over 130 subspecialties, each with distinct workflows and core information.
Early adoption of ambient AI has been strongest in documentation-heavy specialties such as primary care, mental health, and emergency medicine, and is now expanding into fields like cardiology, oncology, and orthopedics as health systems scale deployment. Research published in NEJM AI evaluating Nabla in real-world clinical practice shows consistent benefits across specialties, including reduced documentation time, lower cognitive load, and improved clinician experience.
World models could accelerate this adoption. By learning more efficient representations of complex environments, similar to how humans build mental models, these systems may enable AI to perform reliably even in highly specialized clinical settings.
Nabla’s Vision for Clinical AI
At Nabla, our mission has always been to bring the most advanced and reliable AI technology into real clinical environments.
Our ambient AI assistant is already used across hundreds of health systems and provider organizations, helping clinicians generate high-quality documentation through ambient listening, dictation, and real-time coding support.
But documentation is only the beginning.
From the earliest days of the company, even before the emergence of GPT-3 — Nabla has invested in developing advanced machine learning systems designed specifically for healthcare workflows. We set out to provide next generation AI that will not simply automate tasks, but help clinicians navigate the complexity of care delivery, while restoring joy to practicing medicine.
Just as importantly, our technology is built in close collaboration with the clinicians and health systems who use it every day. Through our partnerships, we receive and utilize continuous feedback from physicians, nurses, and care teams. This ensures that our AI evolves alongside real clinical workflows and delivers meaningful value in practice.
Advancing Toward Agentic Clinical AI
Our partnership with Advanced Machine Intelligence reflects a shared belief that the future of AI will be defined by systems capable of understanding the environments in which they operate.
AMI’s research focuses on developing world models that can maintain persistent memory, reason about evolving situations, and plan actions under real-world constraints, with safety, reliability, and controllability as core design principles. Through this collaboration, Nabla has early access to these emerging technologies as we work toward building the next generation of healthcare AI systems.
This approach unlocks new capabilities for clinical AI, including:
- • Safe, deterministic, auditable decision-making
- • Simulation-based reasoning and “what-if” analysis
- • Handling of multimodal medical signals, including audio, imaging, and physiological data
- • A credible regulatory pathway for autonomous, agentic systems
These advances are particularly important as healthcare begins exploring agentic AI, or systems capable of performing actions within defined workflows on behalf of clinicians. For that vision to become reality, AI must operate with a deeper understanding of context, safety constraints, and system interactions.
By combining the strengths of large language models with world model architecture, we believe healthcare can move toward AI systems that are not only informative, but operationally useful, clinically trustworthy, and capable of supporting increasingly complex care environments while maintaining strong human oversight.
Building the Future of Healthcare AI
AMI’s $1.03B funding round reflects growing recognition that the next phase of AI will require new foundational architectures capable of operating in and navigating real-world environments.
For healthcare, this shift is imperative.
The systems that ultimately deliver the most value will be those capable of understanding clinical workflows, coordinating across fragmented infrastructure, and supporting clinicians safely and reliably.
At Nabla, we are proud to be working alongside the AMI team to bring these capabilities into healthcare.
Our mission remains unchanged: to restore the human connection at the heart of healthcare by building the most advanced and reliable AI assistant for clinicians.
By bringing the next generation of AI into healthcare, we’re helping build systems that understand not just language, but the realities of care itself.
Want to learn more? AMI Labs' Alex LeBrun and Yann LeCun share their thoughts on why healthcare needs to shift from LLMs to World Models on the Offcall podcast. Listen to the recording here.