Why does this AI unicorn have the confidence to make US hospitals' profits increase tenfold?

Recently, a super unicorn has emerged in the AI healthcare track.

On May 19th, the healthcare AI company Commure completed a $70 million financing round, reaching a valuation of $7 billion. The round was led by General Catalyst, with Sequoia and Morgan Stanley participating.

The reason why capital is willing to give such a high valuation is that Commure has targeted the most expensive and difficult - to - solve problem in the US healthcare system - back - end efficiency.

For a long time, the US healthcare system has been highly fragmented.

Electronic medical record giants lock data in their respective systems through closed interfaces and complex permission systems. The data flow between different hospitals, departments, and insurance institutions is extremely inefficient, ultimately trapping a large number of doctors in documentation, coding, reimbursement, and administrative processes.

As a result, the US healthcare system spends approximately $1 trillion on management and administrative costs each year. The operating profit margins of many hospitals are even only 2% - 3%.

What Commure wants to do is to use AI to connect the entire healthcare operation chain.

From front - end clinical document generation and environmental voice recording to back - end revenue cycle management (RCM), Commure hopes to transform the originally labor - intensive back - end system into an AI - driven automated network.

It has even set a very aggressive goal: to increase the operating profit margin of medical institutions from the current 2% - 3% to nearly 20%.

This potential is actually very huge.

For example, HCA, one of the largest healthcare groups in the US, has an annual revenue of over $100 billion. For a medical institution of this scale, even a 1 - percentage - point increase in the operating profit margin means billions of dollars in additional cash flow each year.

That's why Commure has quickly become one of the most capital - favored companies in the current AI healthcare field.

Today, Silicon - based Jun will take you to analyze what kind of company Commure is.

Aiming at the $1 - trillion healthcare administrative expenditure

Many people's imagination of AI healthcare still remains at the stage of "AI doctors diagnosing diseases" and "AI imaging assisting in reading films".

But the reality is just the opposite.

At least in the current US, the first - to - explode demand is actually AI helping doctors with miscellaneous tasks.

Because the most expensive and least efficient part of the US healthcare system is not in clinical practice but in the back - end.

In the US healthcare system, cumbersome administrative tasks such as appointments, coding, medical record writing, insurance claims, and appeal processing have almost entirely relied on manual labor in the past few decades.

According to data from the American Medical Association (AMA), for every hour an outpatient doctor in the US spends with a patient, they need to spend nearly 2 additional hours on electronic medical records and administrative work on average.

Some doctors have said bluntly that they work 14 hours a day, 8 of which are spent on writing medical records and handling insurance, and only 6 hours are spent on actually caring for patients.

In addition to the heavy workload, the long - term fragmentation of the US medical IT system has also seriously affected the operating efficiency of hospitals.

In the past few decades, a large number of point - based systems have emerged in the US healthcare IT industry. Some focus on electronic medical records, some on appointment systems, some on billing coding, and some on insurance claims and patient communication.

However, these systems are not interconnected. The data of a patient often requires a large amount of manual coordination when flowing between different departments, different software, and different insurance systems.

Because there has been a serious data blockade in the US healthcare industry for a long time. Electronic medical record giants lock hospital data in their respective systems through closed interfaces and complex permission systems.

This is why, although a large number of healthcare IT tools have emerged in the US in the past, it has been difficult to have a truly unified platform.

The inefficiency of the back - end system has also directly swallowed up hospital profits.

Currently, the operating profit margins of most medical institutions in the US are only 2% - 3%. According to the estimate of Commure's founder Kabir, the US healthcare system spends approximately $1 trillion on healthcare management and administrative processes each year.

Since the implementation of the "21st Century Cures Act" in 2020, the US government has started to force the interoperability of healthcare data, requiring healthcare IT manufacturers to open standardized APIs and clearly prohibiting "Information Blocking" behavior.

This is equivalent to removing the data barriers of traditional electronic medical record giants for the first time at the institutional level. It also gives opportunities to AI platforms like Commure.

To increase the hospital's profit margin by 10 times

Commure is targeting the back - end process that has been dragged down by the fragmented system.

Currently, Commure's AI products have covered most of the core processes of healthcare operations, from clinical document automation, patient interaction automation, to revenue cycle management automation, and back - end operation automation.

 Kabir believes that in the future, there won't be 100 point - based AI solutions. Eventually, they will converge into a unified platform.

The logic is similar to the 1980s when the spelling - check tool Grammatic was very popular and received a lot of VC attention, but was ultimately completely replaced by Microsoft Word as a built - in function.

Commure's goal is to use AI agents to reconnect these information islands and piece together the entire hospital's back - end system.

Among them, clinical document automation and revenue cycle management are the two most core links.

When doctors communicate with patients, the system will automatically monitor the conversation, extract key information, generate structured medical records, synchronize them to the electronic medical record system, and automatically generate the super - bill required for insurance claims.

In other words, in the past, after seeing a patient, doctors had to spend dozens of minutes or even an hour to supplement documents, record codes, and submit insurance materials. Now, these processes are completed in real - time by AI during the diagnosis and treatment process.

This completely liberates doctors from cumbersome administrative work, which is also the reason for its rapid growth.

Currently, Commure's entire platform supports approximately 200 million patient visits per year. Among them, the AI medical record product processes approximately 50 million visits per year; the patient interaction agent completes approximately 100 million communications such as appointments, rescheduling, and cancellations per year.

The company claims that these products can save at least 75 million hours of administrative time for doctors across the US each year, which is equivalent to giving each doctor 2 - 3 more hours per day to focus on patients.

In addition to clinical document automation, revenue cycle management (RCM) is also one of Commure's most core competitiveness.

In the US healthcare system, when a patient finishes seeing a doctor, it doesn't mean the hospital has earned money. There is a whole set of extremely complex processes afterwards - coding, insurance review, claim rejection handling, appeal, and payment collection... The entire process may last for weeks or even months.

Commure is comprehensively "agentizing" this process.

For example, in the past, a biller needed to spend several hours organizing materials and writing an appeal letter for an insurance claim rejection. Now, AI can automatically extract medical record information, generate an appeal document, and directly submit it to the insurance company.

The commercial value of this is actually very direct.

Originally, 100 people were needed to handle billing and insurance processes, but now 30 people may be enough; in the past, it took 45 days to receive payment, but now it may arrive in about 20 days; many projects that were previously rejected by insurance companies can also have their materials automatically completed by AI to increase the approval rate.

These changes will ultimately be truly reflected in the hospital's cash flow and profit statement. According to Commure's calculation, the full implementation of AI products can increase the operating profit margin of medical institutions from 2% - 3% to 20%.

For a large US healthcare group like HCA with an annual revenue of over $100 billion, even a 1 - percentage - point increase can create billions of dollars in additional cash flow.

Learning from Palantir's FDE model to shorten the implementation cycle

After targeting the core pain point of the inefficiency of the US healthcare system's back - end, Commure's expansion path is also very interesting.

As we all know, the healthcare industry itself has a strong inertia. Hospitals won't easily replace their core systems. The real difficulty is not just the product itself, but also channels, organizational trust, and workflow integration.

To quickly enter the hospital system, Commure has constructed a very unique expansion strategy.

On the one hand, it uses a product - led growth (PLG) strategy to start from small and medium - sized clinics.

Many doctors discover Commure's tools online, try them out, and then recommend them to their colleagues. The company then follows up and signs enterprise agreements. This allows Commure to quickly spread in the primary healthcare market at a relatively low cost.

On the other hand, it directly binds with the largest for - profit healthcare groups in the US.

Currently, Commure has established in - depth cooperation with large healthcare groups such as HCA and Tenet. Just one client, HCA, has 185 hospitals and approximately 2,000 outpatient care points, bringing a potential market of millions of users to Commure.

To make the AI tools truly integrate into the hospital's daily workflow, Commure has learned from Palantir's "front - line engineer" model and sent engineers directly into hospitals.

These front - line engineers work with doctors and nurses. Their first cooperation with HCA started with a small - scale pilot of 3 doctors. The engineers spent two months polishing the product in the most complex scenarios such as the emergency room until the product was good enough to be gradually promoted to the entire group.

This "pilot first, then replicate" model has significantly reduced the resistance to hospital deployment. The training time for medical staff has also been compressed to about 1 - 1.5 days.

More cleverly, Commure has not chosen to compete directly with traditional electronic medical record giants such as Epic and Meditech.

In the US healthcare industry, the electronic medical record system itself is already the most core infrastructure of hospitals, and the replacement cost is extremely high. Commure's strategy is to become an "AI toolbox" on top of them.

For example, doctors using Epic only need to press a button to activate Commure's environmental intelligent documentation and automatic coding functions. This means that Commure can quickly connect with millions of doctors in the Epic ecosystem without overthrowing the original system.

In addition, Commure also acquires companies to quickly obtain distribution channels, talent, and technology.

In 2024, Commure announced the acquisition of the listed company Augmedix at a valuation of $139 million.

Augmedix's core business is to help doctors automatically generate and manage electronic medical records. It has mature channels in approximately 40 healthcare systems across the US. However, its technical route is still mainly based on voice - to - text and traditional NLP. Although it has started to introduce large models, its overall iteration speed has clearly fallen behind the AI era.

Commure's advantage is precisely its AI technical ability. Thus, the two companies form a very typical complementarity: Augmedix has channels but weak AI; Commure has AI but lacks a mature distribution network.

And while expanding rapidly, Commure is also consciously controlling capital efficiency.

Different from most technology companies that burn money for growth, Commure attaches great importance to protecting shareholders' rights and interests. As of now, the company has raised approximately $750 million in total, but the dilution rate in the past two years has been very low.

This benefits from its use of a special financing model - CVF. In June 2025, Commure received $200 million in growth financing from General Catalyst's CVF.

CVF is essentially a financing tool that "uses customer payments to finance growth". The core logic is that the company has found the product - market fit (PMF), and the customer acquisition model is relatively stable, but the sales and marketing investment is very costly.

If the company continues to use equity financing to acquire customers, it will dilute too much equity. If it borrows money, it will increase the rigid repayment pressure. CVF pays the sales and marketing expenses for the company in advance and takes a portion of the payments from the revenue generated by these new customers in the future until it reaches the agreed upper limit. After that, the long - term value of these customers belongs entirely to the company.

It's easy to understand when it comes to Commure. According to the company's disclosure, currently Commure has served more than 130 health systems, and its annual recurring revenue (ARR) has reached hundreds of millions of dollars, with a continuous three - year doubling growth.

Overall, Commure is trying to integrate AI, payment, workflow, channels, and hospital operating systems into a new healthcare operating system. This also means that what Commure is competing for may not just be an AI application market, but the next - generation entrance to the entire US healthcare back - end infrastructure.

This article is from the WeChat official account "Silicon - based Observation Pro", author: Yuanyuan. Republished with permission from 36Kr.