A New Kind of AI Lab: Defining the Shift
The latest wave of AI startup funding rounds marks a shift from general-purpose models toward deep tech AI ventures that combine advanced research with concrete, high-value applications. This shift is defined by top researchers leaving large, established labs to found smaller organisations that target specific scientific or infrastructure problems, often with substantial early-stage capital and clear product roadmaps. Inherent, a newly unveiled AI lab, is a clear example: the company emerged from stealth with a USD 50 million (approx. RM230 million) seed round co-led by Index Ventures and Radical Ventures, and a mission to “write the playbook for AI-native science”. By building an AI system called Faraday to pair humans with self-improving AI for scientific discovery, Inherent shows how the frontier of AI development is moving into specialised, problem-first labs rather than staying only inside the largest platforms.
Inherent: DeepMind Talent Exodus and AI-Native Science
Inherent’s founding team highlights a growing DeepMind talent exodus into entrepreneurial projects. Co-founders Tantum Collins, Edward Hughes and Louis Kirsch previously worked at DeepMind, while Kaloyan Aleksiev brings experience from Microsoft and Reka AI. Collins also spent time in the White House working on AI policy under President Biden, giving the lab a mix of technical and policy expertise that is rare for a seed-stage company. According to Tech.eu, Inherent is building Faraday, “a system designed to help humans and self-improving AI work together on genuine scientific discovery”. The lab says it aims to rethink the scientific method “from first principles”, not simply attach AI to existing workflows. This suggests a priority shift: instead of chasing broad chatbots, top researchers are pursuing AI-native scientific tools that could accelerate discovery in physics, chemistry, biology and beyond.
Why Investors Back Deep Tech AI Ventures Early
The size of Inherent’s seed round signals strong investor appetite for deep tech AI ventures long before they have consumer-facing products. Index Ventures and Radical Ventures appear to be betting that AI-native science systems can open new markets in drug discovery, materials design and complex optimisation. Matt Clifford, co-founder of Entrepreneur First and an advisor to Inherent, called the team some of the “most impressive, thoughtful founders” he had met, underscoring how much investor confidence rests on talent quality in this segment. For investors, these AI startup funding rounds offer exposure to proprietary systems and data loops that could compound over time. Instead of competing directly with large foundation model providers, funds are backing teams that build domain-specific AI engines and workflows that sit closer to scientific and industrial problems—and could later become essential infrastructure.
From General Models to Use-Case Labs: Identity Verification AI
In parallel with labs like Inherent, specialised startups such as Didit are raising focused seed rounds to build identity verification AI infrastructure. These companies do not aim to train the largest general-purpose models; instead, they adapt advanced techniques to narrow but critical tasks like fraud prevention, compliance and secure onboarding. This use-case focus reflects a broader shift in AI development priorities. Rather than relying on one-size-fits-all models, enterprises need reliable, regulatory-aware systems that integrate tightly with existing workflows. Dedicated identity verification AI startups can iterate faster on accuracy, latency and auditability than large platforms serving broad consumer markets. Together with scientific labs such as Inherent, they show that the next competitive frontier lies in specialised AI labs that deeply understand a specific domain, build custom tooling, and turn advanced research into targeted products.
What the New AI Lab Wave Means for the Industry
The emergence of Inherent and focused ventures like Didit signals a new phase for the AI ecosystem. Talent that once concentrated in a few flagship labs is spreading out into smaller, mission-driven organisations, bringing frontier skills and inside knowledge of large-scale AI systems to startup environments. For big tech companies, this raises competitive pressure: they still dominate general-purpose models, but risk losing ground in specialised, high-margin applications if they move slowly. For startups, the bar is rising; investors are now willing to back deep tech AI ventures early, but expect credible roadmaps, clear problem selection and strong founding teams. Over the next few years, more AI-native labs are likely to appear at the intersection of science, infrastructure and regulation—turning the current wave of AI startup funding rounds into a long-term reshaping of how AI innovation is organised.
