Gibran

Gibran is an AI research and product company building a new class of adaptive, scale-free artificial intelligence systems. Founded in London in 2025, the company draws on principles from evolutionary biology and complex adaptive systems to create AI that evolves through ongoing interaction with humans, rather than relying on static pretraining on massive datasets. Its models are designed to augment human creativity and autonomy rather than replace it.

The dominant paradigm in AI development — scaling up language models trained on ever-larger corpora — has delivered remarkable capabilities but carries a fundamental limitation: these models are static once trained, perform poorly in domains where large labelled datasets simply do not exist, and are designed to automate tasks rather than collaborate with people over time. Gibran was founded to address this gap, targeting data-sparse domains such as drug discovery, theoretical science, and personalised education, where the recombination of existing knowledge — rather than pattern interpolation — is what generates genuinely novel value.

Gibran's platform fuses large language models with nature-inspired, self-organising architectures derived from evolutionary and ecological principles. Rather than training a fixed model and deploying it, the system is designed to continuously learn from contextual human interaction, adapting its reasoning and outputs in response to each person's knowledge state, creative goals, and feedback. In drug discovery applications, Gibran's agents can propose novel molecular hypotheses by recombining known patterns — augmenting the intuition of researchers in domains where training data is scarce by design.

The company raised $2.6 million in seed funding from Together Fund and Mercuri in July 2025. Together Fund is led by Freshworks founder Girish Mathrubootham and Eka Software founder Manav Garg, who described Gibran's fusion of LLMs with nature-inspired systems as one of the most original ideas they had encountered in AI. The capital is being used to expand the research team, build the core platform, and develop early applications in scientific domains.