The Dawn of AI-Native Science: Inside Inherent’s $50 Million Quest to Redefine Discovery

In the history of human progress, the most transformative breakthroughs—from the discovery of penicillin to the development of the semiconductor—did not begin with an answer, but with a question that no one else thought to ask. For four centuries, the scientific method has relied on human intuition, "taste," and curiosity to navigate the vast unknown. However, a new London-based AI lab, Inherent, argues that the human bottleneck in scientific discovery is about to be broken.

Emerging from stealth on Wednesday with a massive $50 million seed round, Inherent is positioning itself not just as another AI startup, but as the architect of a new paradigm: "AI-native science." Led by a team of veterans from DeepMind and former government policy architects, the company aims to move beyond AI that simply processes data, toward AI that possesses the "open-ended curiosity" necessary to pioneer the next century of scientific advancement.

Main Facts: A Landmark Launch in the European AI Landscape

Inherent’s debut marks one of the largest "stealth-to-launch" seed rounds in Europe for 2026, signaling a robust appetite among venture capitalists for frontier AI research that moves away from consumer chatbots and toward deep-tech applications.

The Funding Powerhouse

The $50 million round was co-led by two of the most prominent names in global venture capital: Index Ventures and Radical Ventures. The participation of NVentures, Nvidia’s venture capital arm, underscores the computational intensity and strategic importance of Inherent’s mission. Other participants include Ex/Ante, Metaplanet, Macroscopic Ventures, and Mythos Ventures.

The Mission: Faraday and "AI-Native Science"

At the heart of Inherent is Faraday, an AI-driven platform named after the legendary 19th-century scientist Michael Faraday. Unlike current Large Language Models (LLMs) that are optimized to provide immediate answers to user prompts, Faraday is designed to identify which questions are worth asking in the first place.

The company describes its approach as "AI-native science." This concept suggests that while traditional science is legible, linear, and human-centric, AI-native science will be iterative, complex, and potentially "messy." It involves pairing human researchers with self-improving AI agents that can explore "hypothesis spaces"—the infinite range of possible scientific theories—at speeds and scales that are physically impossible for the human brain.

Governance as a Strategic Pillar

In a move that distinguishes it from many of its peers in Silicon Valley, Inherent is structured as a Public Benefit Corporation (PBC). This legal framework mandates that the company balance its fiduciary duties to shareholders with a commitment to societal well-being. By embedding governance into its corporate DNA, Inherent aims to turn ethical oversight into a competitive advantage, particularly when dealing with high-stakes scientific discoveries that could have profound dual-use implications.

Chronology: From DeepMind to the White House to Inherent

The story of Inherent is a convergence of high-level academic research and top-tier geopolitical strategy. The founding team’s journey reflects the evolution of the AI industry from pure research to practical, mission-driven application.

The DeepMind Roots (2018–2023)

The core of Inherent’s leadership was forged at DeepMind, the Google-owned AI powerhouse. Founders Tantum Collins and Edward Hughes previously collaborated on "cooperative AI," a field focused on how AI systems can work together and with humans to solve complex problems. This research laid the groundwork for the collaborative "human-agent" model that Faraday now employs. Another co-founder, Louis Kirsch, also hails from DeepMind, bringing deep expertise in the technical architecture of self-improving systems.

The Policy Pivot (2021–2024)

While many AI founders spend their careers in laboratories, Tantum Collins took a detour into the halls of power. Before co-founding Inherent, he served as a key AI policy advisor in the Biden White House. This experience provided him with a unique perspective on the societal risks and geopolitical stakes of frontier AI. It also likely influenced the decision to adopt the PBC structure and to focus on "governance-first" development.

The Stealth Period and Launch (2025–2026)

Inherent operated in stealth for over a year, recruiting specialized talent like Kaloyan Aleksiev, who joined from Reka AI and Microsoft. During this period, the team focused on building the initial architecture for Faraday, moving away from general-purpose generative AI and toward specialized agents capable of "scientific reasoning."

The company also secured the mentorship of Matt Clifford, the co-founder of Entrepreneur First and former AI tsar for the UK government. Clifford’s involvement bridges the gap between the startup’s technical ambitions and the UK’s broader strategy to remain a global hub for AI safety and innovation.

Supporting Data: The Shrinking Gap Between Europe and Silicon Valley

Inherent’s $50 million seed round is more than just a win for the company; it is a data point in a larger trend of European AI resurgence. Historically, European startups struggled to match the "mega-rounds" seen in San Francisco. However, 2025 and 2026 have seen a dramatic shift.

Comparative Funding Scales

Inherent’s launch sits alongside several other European success stories that suggest the continent is no longer just a talent exporter:

• Mistral AI: The French champion continues to see massive valuations and has reached an estimated $300 million in Annual Recurring Revenue (ARR).
• Lovable: Recently reported a staggering $100 million in revenue within a single month of a major product launch.
• Peec AI: Based in Berlin, it reached $10 million ARR in just six months.

The Efficiency of AI Discovery

The logic behind Inherent’s "Faraday" platform is supported by emerging benchmarks in the industry. For instance, Anthropic’s Glasswing project recently demonstrated that frontier AI models could identify software vulnerabilities at a rate that significantly outpaces human developers. Inherent is betting that this same delta applies to the natural sciences.

If an AI agent can test 10,000 hypotheses in the time it takes a human to test one, the primary role of the human shifts from "executor" to "curator." Inherent’s data suggests that the most valuable use of AI is not in automating existing labor, but in expanding the boundaries of what is "knowable."

Official Responses: Vision from the Boardroom

The announcement of Inherent’s funding was accompanied by statements that highlight the philosophical shift the company is attempting to lead.

Danny Rimer, a partner at Index Ventures, emphasized that current AI models are being used incorrectly by the scientific community. "Most AI is built to answer questions," Rimer noted. "What it can’t do yet is figure out which questions are worth asking—the open-ended curiosity that produced penicillin, the microwave, and the GPU. That’s the gap Inherent is building into."

Index Ventures further elaborated on the investment in a blog post, acknowledging that the future of science might look unrecognizable to traditionalists. "AI-native science will be messier, less legible, but capable of exceptional outcomes," the firm wrote. This suggests a willingness to move away from the "explainability" requirement that often slows down scientific progress, provided the outcomes are verifiable and beneficial.

Matt Clifford, acting as an adviser, pointed to the importance of the team’s pedigree. He noted that the combination of DeepMind’s technical rigor and the founders’ policy experience makes Inherent uniquely suited to handle the "dual-use" nature of scientific AI, where a breakthrough in drug discovery could potentially be misused in bioweaponry.

Implications: A 400-Year-Old Method Under Pressure

The emergence of Inherent and its Faraday platform carries profound implications for the future of research, education, and the global economy.

1. The End of the "Lone Genius" Era

For centuries, science has been defined by individual "geniuses" like Newton, Curie, or Einstein. Inherent’s model replaces the lone researcher with a "human-AI collective." In this model, the AI handles the brute-force exploration of data and hypothesis, while the human provides the "taste"—the ethical judgment and the intuition for which discoveries will actually benefit humanity. This could democratize high-level research, allowing smaller teams to achieve results that previously required the resources of a Tier-1 university.

2. The PBC Model as a Standard

As AI labs face increasing scrutiny over safety, Inherent’s choice to be a Public Benefit Corporation could set a new standard for the industry. If Inherent can prove that a PBC can still deliver venture-scale returns, it may pressure other major labs (like OpenAI or Anthropic) to formalize their societal commitments in more legally binding ways.

3. Geopolitical Positioning

The UK government has long sought to position London as the "AI safety capital of the world." With Inherent, the UK now has a flagship company that combines safety-conscious leadership (Clifford and Collins) with the technical firepower to compete with Silicon Valley. This strengthens Europe’s hand in international AI treaty negotiations, as they are no longer just regulators, but also the home of the most advanced scientific AI.

4. The Transformation of Basic Research

If Faraday works as intended, the timeline for solving "unsolvable" problems—such as room-temperature superconductivity, carbon capture, or Alzheimer’s—could shrink from decades to years. However, this "messy" and "less legible" science will require a complete overhaul of how we validate discoveries. Peer review, as it exists today, may not be able to keep pace with an AI that generates thousands of verifiable breakthroughs per year.

Conclusion: The $50 Million Bet on Curiosity

Inherent is not building a tool to replace scientists; it is building a tool to replace the limitations of science. By focusing on curiosity rather than just computation, and by prioritizing governance as a core feature rather than a hurdle, the team is attempting to build the "operating system" for the next scientific revolution.

Whether Faraday can truly replicate the "happy accidents" that led to the world’s greatest inventions remains to be seen. However, with $50 million in the bank and the backing of the world’s most influential AI investors, Inherent has the runway to try. The next four hundred years of the scientific method may look nothing like the last four hundred.