2025 was not simply a year of growth for Videntifier – it was a year of clarification. Over the course of the year, it became clear what kind of company we are becoming, what role our technology plays in the wider digital ecosystem, and where visual intelligence is heading as an industry.
What started as a specialist focus on harmful content detection has evolved into something more foundational: a visual recognition infrastructure that supports trust and safety while also enabling brands and advertisers to better understand, protect, and optimize the use of their visual assets.
“Over the last few years, we have worked hard to make our software more generally and easily applicable to a wider range of applications. Last year was the culmination of that effort, allowing us to effortlessly branch into many different visual recognition applications.” says Videntifier’s CEO Ari Kristinn Jónsson. “This has allowed us to deliver great value in areas where existing solutions fall short, such as in tracking visual assets for international brands.”
Expanding into Branding and Advertising
One of the most tangible outcomes of this shift was Videntifier’s expansion into branding and advertising use cases. Throughout 2025, organizations increasingly turned to visual intelligence to answer questions that traditional analytics struggle to address: where visual assets actually appear, how they are reused or altered, and whether they align with brand, affiliate, and influencer strategies.
“Marketing teams are drowning in impressions and dashboards, but still can’t answer a simple question: where did our visuals actually go, and in what context?” says Videntifier’s Head of Business Development Einar Sigurdsson. “Visual intelligence turns distribution, alignment, and brand safety into something you can measure – not guess.”
In practice, visual recognition enables:
- Visual performance intelligence across campaigns and platforms, giving teams real-time clarity on where and how their brand assets appear worldwide, helping ensure consistency, maximize impact, and avoid costly misalignment
- Influencer visibility and validation, revealing which visuals are actually used, on which channels, and in what context — empowering brands to optimize partnerships, refine creative strategy, and invest with confidence
- Competitive and market insight powered by visuals, uncovering emerging trends, high-performing formats, and content that gains traction early, so brands can move faster and stay ahead
What emerged is a clear parallel between the challenges faced by platforms and those faced by brands: fragmented visibility, limited control, and increasing regulatory and reputational risk. Visual intelligence offers a shared solution space.
Nexus in Real-World Deployment
In 2025, Nexus transitioned from a mature product into active deployment at scale. Several major social media platforms, including Medialab’s products, now use Nexus as part of their trust and safety infrastructure, integrating it directly into their workflows to filter out illegal videos and imagery.
Also, Nexus was selected by the Internet Watch Foundation (IWF) as a core technology supporting UK online safety efforts. Through the IWF, it is made available to UK-based platforms – including smaller operators – to support detection, reporting, and compliance with the Online Safety Act.
Strengthening the Foundations
Behind the scenes, 2025 was also a year of consolidation and long-term planning. Videntifier secured €1.9M in non-equity funding, enabling accelerated development while maintaining strategic independence.
These resources were primarily invested in:
- Core platform robustness and scalability
- Reliability improvements for high-volume, real-world deployments
- Research into next-generation AI architectures
The emphasis throughout the year was sustainability – building systems designed to operate reliably under real-world conditions.
Looking Beyond Recognition: VectorTree
As part of this longer-term vision, we launched VectorTree in Kaunas, Lithuania, a new venture focused on how AI systems store, structure, and retrieve contextual information.
While still at an early stage, VectorTree is being developed as a specialized database designed for generative AI applications – enabling large models to store, organize, and retrieve visual and multimodal context with greater precision. The goal is to support GenAI systems that do not just recognize content, but can reliably ground generation, reasoning, and decision-making in structured visual and contextual data.
Collaboration as a Force Multiplier
Progress in 2025 was closely tied to collaboration. Working with partners across trust, safety, and policy allowed us to validate technology in real environments and address complex, multi-category challenges.
Key collaborations included:
- Co-developing one's likeness protection and misuse detection solutions with Umanitek, a Switzerland-based AI company, and the prevention of image- and video-based abuse
- Supporting implementation of the UK Online Safety Act through accessible safety tooling provided to the Internet Watch Foundation, the UK's biggest hotline for reporting suspected online child sexual abuse imagery
- Strengthening multi-category harm detection through the integration of large, verified TVEC (terrorist and violent extremist content) hash database from Tech Against Terrorism, UN-backed partnership supporting smaller platforms to counter online terrorist propaganda while upholding human rights, directly to NEXUS
Rather than isolated deployments, these partnerships reinforced the importance of shared infrastructure and common standards.
Engagement Across the Ecosystem
Throughout the year, Videntifier contributed to industry and policy discussions across Europe and beyond, engaging with platforms, regulators, NGOs, and technology leaders.
Across these forums, a consistent theme emerged: visual intelligence is no longer optional. It is becoming a prerequisite for operating responsibly and competitively in digital environments dominated by images and video.
We actively participated in global trust, safety, and technology events, including IBC Accelerator (Amsterdam), Trust & Safety Summit (London), INCYBER Forum (Lille), and others. We also hosted an event in Reykjavik with our partner Umanitek, further strengthening dialogue across the ecosystem.
Looking Ahead to 2026
As we enter 2026, the focus shifts from validation to scale. Our priorities are to deepen the impact of visual intelligence in branding and advertising, expand into emerging sectors, and continue strengthening the AI foundations that support trust and safety at scale.
If you’re a platform, brand, NGO, regulator, or partner exploring how visual intelligence can strengthen trust, safety, and performance, we’d be glad to connect.
Contact: info@videntifier.com