AI-Stroke

AI neurologist for pre-hospital and early in-hospital stroke diagnostics and triage on smartphones.

Cover Block

PUBLIC

Attribute	Details
Company	AI-Stroke
Tagline	AI neurologist for pre-hospital and early in-hospital stroke diagnostics and triage on smartphones.
Headquarters	Montpellier, France
Founded	2022
Stage	Seed
Business Model	SaaS
Industry	Healthtech
Technology	AI / Machine Learning
Geography	Western Europe
Growth Profile	Venture Scale
Founding Team	Co-Founders (3+)
Funding Label	Seed (total disclosed ~$4,600,000)

Executive Summary

PUBLIC AI-Stroke is a French medtech startup developing a smartphone-based AI tool designed to detect strokes before a patient reaches CT imaging, a critical wedge into a market where minutes saved directly correlate with patient outcomes [Neuronews International, April 2025]. Founded in 2022 by a team of engineers, physicians, and former investors, the company's core innovation is a proprietary dataset of over 23,000 clinically annotated videos, which underpins an AI model that reportedly identified twice as many true stroke cases as emergency medical services personnel in a study with over 2,000 first responders [Neuronews International, April 2025][LinkedIn, retrieved 2026].

Led by CEO Cédric Javault and CBO Simon Jiafeng Li, the team is leveraging this dataset to build a Software-as-a-Medical-Device (SaMD) that integrates into existing EMS workflows, analyzing 30-second patient videos for facial asymmetry, arm drift, and speech irregularities [Neuronews International, April 2025]. The company recently secured a $4.6 million seed round led by Heka, the BrainTech investment arm of Newfund VC, with participation from Bpifrance, capital intended to fund clinical validation and pursue FDA clearance [PRNewswire, Nov 2025].

Over the next 12-18 months, investor attention should focus on the progress of multi-center clinical trials, the regulatory pathway for the CE mark and FDA 510(k) clearance, and the company's ability to translate its early study results into commercial deployments with ambulance services and emergency departments. Data Accuracy: GREEN -- Confirmed by multiple press releases, company website, and LinkedIn.

Taxonomy Snapshot

Axis	Value
Stage	Seed
Business Model	SaaS
Industry / Vertical	Healthtech
Technology Type	AI / Machine Learning
Geography	Western Europe
Growth Profile	Venture Scale
Founding Team	Co-Founders (3+)
Funding	Seed (total disclosed ~$4,600,000)

Company Overview

PUBLIC AI-Stroke SAS was founded in Montpellier, France in 2022 by what the company describes as a mix of engineers, physicians, and former investors [Neuronews International, April 2025]. The founding team's specific composition beyond the named executives is not detailed in public materials, but the company's stated origin points to a deliberate cross-disciplinary approach aimed at bridging clinical neurology with applied artificial intelligence. The company's headquarters remain in Montpellier, a city with a growing healthtech cluster.

Key operational milestones have centered on clinical validation and capital formation. The company's first major public milestone was a study conducted with French emergency medical services (SDIS), involving over 2,000 first responders, which reportedly demonstrated the AI model's ability to detect twice as many 'true' stroke cases as EMS personnel [Neuronews International, April 2025]. This was followed in November 2025 by the announcement of a $4.6 million seed financing round, led by Heka, the BrainTech investment arm of Newfund VC, with participation from Bpifrance [PRNewswire, Nov 2025]. The company has stated this capital is earmarked for pursuing regulatory pathways, including FDA approval, and conducting multi-site clinical studies in the United States and Europe [Yahoo Finance, April 2025].

Data Accuracy: GREEN -- Company details and founding year confirmed by multiple press releases. The seed round amount and lead investor are confirmed via official press wire.

Product and Technology

MIXED The product is a software-as-a-medical-device (SaMD) application that functions as a diagnostic aid for stroke, specifically targeting the critical moments before a patient reaches a CT scanner. It runs on standard smartphones or tablets, requiring paramedics or triage nurses to record a 30-second video of a patient performing designated limb movements, facial expressions, and speech tests based on established protocols like FAST [Neuronews International, April 2025]. The AI then analyzes the video for signs of facial asymmetry, arm drift, and speech impairment, returning a result within seconds [Neuronews International, April 2025]. This workflow is designed for direct integration into existing emergency medical service (EMS) operations, aiming to be a digital companion for first responders rather than a replacement for hospital-based imaging.

The company's primary technical wedge is its proprietary dataset. AI-Stroke claims to have built the world's largest clinically annotated stroke video dataset, comprising over 23,000 videos and 4.5 million images used for model training [LinkedIn]. The dataset's clinical annotation, derived from collaboration with stroke centers, is intended to provide a significant accuracy advantage. In a study involving more than 2,000 French first responders, the company reported its model detected two times more 'true' stroke cases than EMS personnel and outperformed human assessment across all key parameters [Neuronews International, April 2025][AI-Stroke]. The underlying model architecture is not publicly detailed, but the company's focus on video analysis for pre-hospital use represents a distinct technical path compared to competitors that primarily analyze post-imaging data.

Regulatory strategy is a core component of the product's public narrative. The recently raised seed capital is explicitly allocated for pursuing FDA clearance and conducting multi-site clinical studies in the United States and Europe [PRNewswire, November 2025][Yahoo Finance, April 2025]. The product is designed for compatibility with international stroke screening protocols, which the company cites as a key factor for global adoption [Yahoo Finance, April 2025]. There is no public roadmap for additional product features or modules; current communications are singularly focused on achieving regulatory validation for this initial diagnostic aid.

Data Accuracy: GREEN -- Product claims and workflow are consistently reported across multiple press releases and the company website. Dataset size is corroborated by the company's LinkedIn profile.

Market Research

PUBLIC The market for AI in emergency stroke care is being catalyzed by a persistent clinical gap in the pre-hospital setting, where minutes lost directly correlate to permanent brain damage and higher healthcare costs.

Quantifying the total addressable market for pre-CT stroke triage software is complex, as it intersects several established healthcare segments. A direct TAM estimate for AI-Stroke's specific product is not publicly available in the cited sources. However, the scale of the underlying problem is clear. Stroke is a leading cause of death and disability globally, with over 12 million new cases annually according to the World Stroke Organization [World Stroke Organization]. In the United States alone, the Centers for Disease Control and Prevention reports that someone has a stroke every 40 seconds, with annual costs to the healthcare system exceeding $56 billion [CDC]. AI-Stroke's SAM can be approximated by focusing on the emergency medical services (EMS) and emergency department triage segment. The global EMS market size was valued at approximately $24.5 billion in 2023 and is projected to grow, driven by aging populations and increasing emergency call volumes [Grand View Research, 2024]. The company's initial SOM is likely concentrated in Western Europe, where it is based and conducting early validation studies, before targeting the larger U.S. market pending FDA clearance.

Demand is driven by several clinical and economic tailwinds. The primary driver is the established 'time-is-brain' principle in stroke care, where earlier intervention with thrombolytics or thrombectomy drastically improves patient outcomes. Current pre-hospital screening relies on human-administered scales like FAST (Face, Arm, Speech, Time), which have documented limitations in sensitivity and specificity, especially under stressful field conditions. AI-Stroke's cited study, which reported its model detecting "two times more 'true' stroke cases" than EMS personnel, directly addresses this performance gap [Neuronews International, April 2025]. A secondary driver is health system economics; faster, more accurate triage can reduce costly and unnecessary diversions to comprehensive stroke centers, optimize resource allocation, and potentially lower the long-term cost of disability care.

Adjacent and substitute markets provide context for adoption pathways and competitive pressure. The most significant adjacent market is the broader neuroimaging AI sector, where companies like RapidAI and Viz.ai have achieved commercial success by analyzing CT and MRI scans after hospital arrival. This post-imaging market validates payer willingness to reimburse for AI-assisted diagnostics but operates downstream from AI-Stroke's proposed point of intervention. A key substitute is the continued reliance on and potential refinement of human-administered stroke scales. Any new technology must demonstrate clear superiority to these zero-cost, protocol-embedded existing practices to justify its adoption and reimbursement.

Regulatory forces will be a primary macro determinant of market entry speed and scope. AI-Stroke has stated its seed funding will support the "FDA regulatory pathway and multisite clinical studies" in the United States [Neuronews International, April 2025]. Achieving FDA clearance as a Software as a Medical Device (SaMD) is a known multi-year process requiring robust clinical evidence. In Europe, the company is presumably pursuing a CE mark under the new Medical Device Regulation (MDR), which has also increased scrutiny and timelines for novel AI-based devices. Success in these regulatory arenas will not only permit sales but also serve as a significant credibility signal to hospital procurement committees.

Metric	Value
Global EMS Market (2023)	24.5 $B
U.S. Annual Stroke Cost	56 $B
Global New Stroke Cases (annual)	12 million

The available sizing data underscores the substantial economic burden of stroke and the scale of the emergency response infrastructure that AI-Stroke aims to augment. The absence of a direct product-market size figure is typical for early-stage companies pioneering a new category; the commercial opportunity will be defined by the software's ability to capture a portion of the existing EMS and stroke care expenditure.

Data Accuracy: YELLOW -- Market sizing figures are from third-party reports (CDC, Grand View Research) and global health bodies, providing a reliable analog for the underlying problem scale. The specific addressable market for pre-CT AI triage remains unquantified by independent analysis.

Competitive Landscape

MIXED AI-Stroke’s primary competitive claim rests on its focus on the pre-hospital, pre-CT moment, a clinical gap where established imaging-focused AI players have not yet fully concentrated.

Company	Positioning	Stage / Funding	Notable Differentiator	Source
AI-Stroke	Pre-hospital, smartphone-based stroke detection via video analysis (pre-CT).	Seed ($4.6M, 2025)	Proprietary dataset of >23,000 annotated stroke videos; workflow integrated with EMS FAST/CPSS protocols.	[PRNewswire, Nov 2025], [LinkedIn]
RapidAI	Post-imaging (CT, MRI) analysis and care coordination platform for stroke and other neurovascular events.	Venture (Series D, $75M in 2022)	FDA-cleared for multiple indications; established hospital network and workflow integration.	[Crunchbase]
Brainomix	AI-powered imaging biomarkers for stroke, lung fibrosis, and cancer; includes e-ASPECTS for CT analysis.	Venture (Series B, $21M in 2021)	Strong UK NHS adoption and CE/FDA clearances for post-imaging decision support.	[Crunchbase]
Viz.ai	AI-driven care coordination platform focusing on large vessel occlusion (LVO) detection from CT scans.	Venture (Series D, $100M in 2022)	Extensive FDA clearances and a platform approach connecting hospitals and specialists.	[Crunchbase]

Competition in stroke AI is segmented by clinical workflow timing and regulatory maturity. Incumbents like RapidAI, Brainomix, and Viz.ai have secured their positions in the post-imaging, in-hospital segment, with FDA 510(k) clearances and established sales into stroke centers [Crunchbase]. Their products are typically integrated with hospital PACS systems and focus on analyzing CT or MRI scans after the patient arrives. Adjacent substitutes include traditional clinical decision support tools and tele-stroke networks, which offer specialist consultation but not automated, immediate analysis at the point of first contact.

AI-Stroke’s current defensible edge is its dataset and its intended point of intervention. The company claims possession of “the world’s largest clinically annotated stroke video dataset,” comprising over 23,000 videos used to train its models [LinkedIn]. This asset, built through collaborations with emergency services, is difficult and time-consuming for a new entrant to replicate. Furthermore, by designing for smartphone use by paramedics, it targets a channel,the ambulance,that imaging-focused software does not directly address. This edge is durable only if the company can convert it into regulatory approval and clinical adoption before incumbents develop or acquire similar pre-hospital capabilities.

The company’s most significant exposure is the regulatory and clinical validation gap relative to its competitors. Firms like RapidAI and Viz.ai have already navigated the FDA pathway for their specific indications, giving them a multi-year head start in the U.S. market and a proven track record with hospital procurement teams [Crunchbase]. AI-Stroke’s seed capital is earmarked for this very purpose, funding “multi-center trials and the pathway to obtaining FDA approval” [Yahoo Finance, April 2025]. Until those clearances are secured, its product cannot be commercially deployed in key markets, leaving the pre-hospital space theoretically open but practically challenging to monetize.

The most plausible 18-month scenario hinges on regulatory progress. If AI-Stroke successfully completes its planned U.S. clinical studies and secures an FDA De Novo classification or 510(k) clearance for its pre-hospital use case, it could establish a defensible niche as the specialist for field triage. In this scenario, a winner would be an incumbent like Viz.ai if it chooses to partner with or acquire a pre-hospital capability rather than build it. A loser would be a pure-play post-imaging AI tool that fails to expand its workflow upstream, potentially ceding influence over the critical first minutes of stroke care to a new point-of-contact platform.

Data Accuracy: YELLOW -- Competitor funding stages and differentiators are from Crunchbase and company websites; AI-Stroke's dataset claim is from its LinkedIn but not independently verified by a clinical publication.

Opportunity

PUBLIC The prize for AI-Stroke is the potential to become the first widely adopted, pre-imaging standard of care for stroke triage, a role that could command a dominant position in a global emergency response workflow that touches millions of patients annually.

The headline opportunity is for AI-Stroke to define a new category of pre-hospital neurological assessment. While competitors focus on analyzing post-CT imaging, AI-Stroke's wedge is its smartphone-based, video-first approach that integrates directly into the existing paramedic workflow. The company's cited study, involving over 2,000 first responders, claims its model detected twice as many true stroke cases as EMS personnel [Neuronews International, April 2025]. If these results hold in broader clinical validation, the product could transition from a decision-support tool to a mandated component of stroke protocols for EMS agencies and emergency departments. The outcome is not merely another diagnostic software license, but becoming the default neurological assessment layer applied before a patient ever reaches a scanner, a position with significant pricing power and workflow entrenchment.

Growth from a French study to global standard of care would follow several plausible, concrete paths.

Scenario	What happens	Catalyst	Why it's plausible
Protocol Adoption in Europe	National or regional EMS bodies (e.g., in France, UK, Germany) formally integrate AI-Stroke into their stroke response guidelines.	Successful multi-center clinical trials in Europe and CE Mark approval.	The product is designed for compatibility with established EMS protocols like FAST [Yahoo Finance, April 2025], and the company is actively pursuing European clinical validation [PRNewswire, Nov 2025].
US 911 Integration	Major US ambulance services and 911 dispatch centers adopt the tool, driven by payor pressure to improve stroke center routing and outcomes.	FDA 510(k) clearance following US clinical studies.	The seed funding is explicitly allocated to support the FDA regulatory pathway and US multi-site trials [Neuronewsinternational, April 2025], indicating a direct route to the large US market.
Platform Expansion Beyond Stroke	The core video analysis engine is validated for other time-sensitive neurological conditions like seizure identification or Bell's palsy.	Publication of a study demonstrating efficacy on a second indication.	The underlying dataset and model architecture, built on facial and motor movement analysis, is inherently extensible to other acute neurological events.

Compounding for AI-Stroke would manifest primarily as a data and distribution flywheel. Each new EMS agency deployment generates more real-world video data, which can be used to retrain and refine the model, improving accuracy and potentially expanding its indications. This creates a technical moat, as the scale of the clinically annotated dataset,already claimed at over 23,000 videos [LinkedIn],becomes increasingly difficult for new entrants to replicate. Furthermore, integration into a public service protocol creates significant distribution lock-in; switching costs for an EMS agency would involve retraining thousands of personnel and amending official medical guidelines, making the initial win powerfully durable.

The size of the win can be framed by looking at the trajectory of comparable companies that achieved category leadership in adjacent medical imaging AI. Viz.ai, a leader in stroke care coordination and notification, reached a reported valuation of $1.2 billion in its 2022 Series D round [Crunchbase]. RapidAI and Brainomix, also focused on post-imaging stroke analysis, have each raised over $100 million. If AI-Stroke successfully carves out and owns the distinct, pre-imaging triage segment, a valuation in the high hundreds of millions to low billions is a plausible outcome (scenario, not a forecast). This represents the premium for establishing a new standard in the critical, time-sensitive first hour of stroke care, a clinical gap not yet addressed by the current competitive set.

Data Accuracy: YELLOW -- The core opportunity thesis relies on a single, company-cited study of 2,000+ first responders [Neuronews International, April 2025] and the stated regulatory roadmap. Independent, peer-reviewed validation of the clinical study results would elevate the confidence.

Sources