Neuramancer, a Munich-based AI startup, has raised €1.7 million ($2 million) in pre-seed funding led by Vanagon Ventures. The platform performs multi-stage forensic analysis on images and videos to detect deepfakes and partial manipulations via statistical noise irregularities. The funds will scale the technology platform, expand to audio and text detection, and target insurance fraud prevention.
Deepfake Funding Heats Up in Europe
This raise follows a wave of investments in deepfake defenses: Doppel raised $70M Series C in November 2025, while Resemble AI secured $13M in December 2025. Other players include Reality Defender with $33M and Hive Moderation exceeding $120M. Neuramancer differentiates with its academic-rooted, transparent forensics optimized for compressed media from unknown sources.
Insurance Fraud Losses Fuel Demand
Deepfake fraud caused $1.56B in losses in 2025 alone. The German Insurance Association highlights billions in annual damages from manipulated media claims. Current black-box models struggle with evolving generative AI and partial edits, leaving fact-checkers and insurers vulnerable.
Noise Analysis Enables Transparency
Neuramancer's suite generates heatmaps and contextual reports explaining detections in 20 seconds. Unlike real-time enterprise tools, it excels in post-upload forensics for challenging cases like compressed videos. Products include a Fakechecker app, analysis API for integration, and on-premise options with European hosting for GDPR compliance.
Forensic Reports Aid Investigations
The content-agnostic approach analyzes noise artifacts resilient to generator advancements. This positions it for media houses verifying election content and insurers probing claims. Expansion to audio and text fakes builds on visual strengths from FAU Erlangen-Nürnberg research.
As Anika Gruner, co-founder, noted:
"While many providers rely on intransparent black-box models, we pursue a scientifically grounded, fully transparent approach. For us, it is clear: European, explainable AI will become a strategic competitive advantage."
Regional Backers Signal EU Focus
Vanagon Ventures led with participation from Bayern Kapital, ZOHO.VC, Lightfield Equity, and finance/tech angels. This mix underscores conviction in explainable AI amid EU AI Act mandates for deepfake disclosure starting 2026. Bavarian support via prior SPRIND €375k grant validates the startup's trajectory.
Market Scales Amid Regulations
The deepfake detection market stands at $209M in 2024, projected to reach $1.5B by 2030 at 38% CAGR. EU rules require labeling and technical solutions, driving demand for compliant tools. Platforms like YouTube now flag deepfakes of politicians, amplifying need for backend forensics.
Academic Roots Bolster Credibility
Founders hail from FAU Erlangen-Nürnberg with IT security and media forensics expertise. New CCO Martin Sondenheimer arrives from Munich Re and Allianz, targeting insurance as a priority. Distinguished Engineer Aron Homberg brings 20+ years in scalable software.
Hiring Drives Platform Expansion
Post-funding, Neuramancer seeks AI engineers for ML pipelines and MLOps. A public beta demo launched in December 2025, with closed beta earlier. Political visits from Bavarian leaders and TV features on ProSieben signal growing validation ahead of 2026 commercialization.