Cloudphysician Announces Nightingale Video Model for Hospital Monitoring

Clinical AI company Cloudphysician announced Nightingale on September 9, 2026, a domain-specific video foundation model built to monitor hospital patients around the clock. The company said the system is already deployed across more than 150 hospitals in India and is expanding into the US market.

The announcement was issued from Bangalore, India and San Francisco. Cloudphysician, founded by physicians, said its clinical AI is designed to expand bedside capacity while leaving clinicians and their existing workflows in place, and that Nightingale was trained on thousands of hours of clinical video. According to the company, the model watches for clinically relevant changes in a patient’s condition that would normally surface only when a physician makes periodic rounds.

Architecture and On-Premise Operation

The company contrasted Nightingale with general-purpose video AI systems, which it said identify basic activity such as whether a patient is present or moving. Nightingale, the company said, is engineered to recognize fine-grained clinical signals, including abnormal breathing patterns, nasal flaring, tube positioning that distinguishes feeding tubes from breathing tubes, signs of delirium, sedation levels, fall risk and pressure ulcer risk.

Cloudphysician described Nightingale as a compact foundation model developed from its proprietary clinical data with expert annotations, built from the ground up rather than adapted from existing open-weight models. According to the company, all processing occurs within the hospital: a single on-premise GPU ingests 100 concurrent camera streams, and patient video never leaves the building. The company called the result an efficient perception layer designed for edge and on-premise deployment, and said it reduces inference and bandwidth costs by roughly 100x compared with general-purpose, cloud-based, token-driven models.

CEO Statement and US Expansion Plans

Co-founder and CEO Dr. Dhruv Joshi said he spent years as a doctor watching patients whose conditions could turn within minutes, whether a shifting breathing pattern or an unattended patient trying to get out of bed. “Nightingale started in patient rooms, training to pick up the signals a bedside monitor misses and a physician would catch if they could be everywhere at once,” Joshi said in the announcement. He said the company is being built to serve as the eyes of a doctor or nurse, and that this kind of AI can mean the difference between life and death.

The announcement states that the United States has approximately 1 million hospital beds and that around 200,000 of them currently have cameras, a figure expected to grow to 500,000 by 2030. Cloudphysician said this installed camera base makes it possible to deploy AI that brings clinical-grade visual perception into care delivery around the clock. Its stated goal is to become the default visual perception layer for care delivery in what it characterized as a market that is becoming multibillion-dollar in size.

Company Background and Existing Products

Cloudphysician was founded in 2017 by Dr. Dhruv Joshi and Dr. Dileep Raman as an AI company building video-based solutions for hospitals, with a stated goal of enabling patient safety through autonomous care while patients are admitted. According to the release, the company has provided care to more than 200,000 patients and has helped save more than 10,000 lives across a network of over 200 hospitals.

On its website, Cloudphysician separately states it has treated more than 200,000 patients across more than 300 hospitals and cites up to 40% lower mortality rates. The company lists Smart-ICU, Smart-NICU, Smart-Dialysis and Smart-ER product lines, offices in San Francisco, Singapore and Bangalore, teams in both the US and India, and eight granted patents covering methods behind its AI models. It also says it has spent the past decade building one of the world’s largest multimodal datasets from high-acuity inpatient care, combining video, vitals and EMR data.

The company’s About page states that Cloudphysician was founded by two critical-care doctors who have spent thousands of hours in high-acuity care settings. Joshi’s background is in pulmonary and critical care medicine, with past associations at the Cleveland Clinic Foundation and MedStar Health. Raman’s background spans pulmonary, critical care and sleep medicine, with past associations at the Cleveland Clinic Foundation and Texas Tech University Health Sciences Center.

Cloudphysician also offers a product it calls its AI video co-pilot, which uses pre-trained, disease-specific AI agents trained on clinician-annotated video data. The company says an orchestrator layer coordinates specialized agents across multiple modalities so their insights remain contextual for caregivers, with listed agents covering respiratory rate and work-of-breathing detection, pain and sedation monitoring, and central line infection prevention. Demonstrated examples include detecting lowered bed rails and sending a fall-risk warning to the bedside team, flagging early risk of delirium, and visually identifying work of breathing.

The company states that its deployments are hardware-agnostic, run on-premises or in a customer’s cloud alongside existing monitors, cameras, tablets and phones, and that customer data stays within the hospital’s environment and is not used to train its models.