London-based artificial intelligence infrastructure startup Potential Sciences has secured a multi-million-dollar seed investment backed by the United Kingdom government’s Sovereign AI fund and Stripe Chief Executive Officer Patrick Collison. The capital injection, valued in the tens of millions of dollars, is intended to accelerate the development of the company’s physics-driven computational models designed to fundamentally alter how researchers simulate molecular behavior and discover novel therapeutics.
The funding round highlights an increasing alignment between national industrial strategy and private venture capital in the United Kingdom, particularly within high-stakes sectors such as biotechnology, quantum chemistry, and advanced computing. Alongside the Sovereign AI fund and Collison, the financing syndicate includes prominent venture capital firms such as Twin Path Ventures, Odyssey, Portfolio Ventures, Blue Wire Capital, rule30, Nauta Capital, and Lightfield Equity, supplemented by a cohort of strategic angel investors. While the company has opted not to disclose the exact valuation or the precise financial breakdown of the round, the commitment from both public and private stakeholders underscores the anticipated commercial and scientific impact of the startup’s core technology.
Pioneering Physics-First AI for Molecular Simulation
Founded in November 2024 and officially launched into the broader market ecosystem in early 2025, Potential Sciences is tackling one of the most persistent bottlenecks in pharmaceutical research: the prediction of molecular interactions. Traditional computational drug discovery heavily relies on vast repositories of historical experimental data to train machine learning models. While effective in certain contexts, these data-driven models frequently struggle when extrapolating beyond known chemical spaces, leading to costly clinical failures and extended development timelines.
Potential Sciences is pursuing a fundamentally different architectural paradigm. Rather than training models exclusively on legacy laboratory datasets, the startup is building foundational artificial intelligence models rooted in the fundamental laws of quantum mechanics and physics. By simulating the actual physical forces governing atomic and molecular behavior from first principles, the company aims to bypass the limitations inherent in historical training data.
According to preliminary disclosures from the startup, this physics-centric approach could slash the computational costs associated with chemical design by a factor of one hundred. By making advanced molecular calculations radically cheaper and more accessible, Potential Sciences aims to democratize high-throughput drug discovery, enabling smaller biotechnology firms and academic institutions to conduct simulations that were previously economically unfeasible.
Strategic Backing from the UK Sovereign AI Fund
The investment marks a significant milestone for the UK government’s Sovereign AI fund, an initiative designed to cultivate, retain, and scale indigenous artificial intelligence champions within the United Kingdom. As global competition for technological sovereignty intensifies among major economies, governments are increasingly deploying public capital to secure domestic supply chains and foundational research capabilities in critical technology domains.
Beyond financial capital, the Sovereign AI fund is providing Potential Sciences with critical computational resources, committing dedicated compute capacity to train and scale its heavy physics-based models. Access to scalable compute infrastructure remains one of the primary operational hurdles for early-stage AI startups, making the provision of state-backed processing power an invaluable asset for the young firm.
The participation of the Sovereign AI fund in the Potential Sciences financing round aligns with its broader mandate of fortifying the UK’s technological ecosystem. Previous investments by the fund include semiconductor startup Olix and specialized artificial intelligence ventures such as Ineffable Intelligence and Callosum. By casting a wide net across chip design, foundational models, and domain-specific applications, the UK government is attempting to construct a vertically integrated domestic AI sector capable of competing on the global stage.
Joséphine Kant, an investment partner at the Sovereign AI fund, reflected on the leadership and potential of the startup following the investment announcement. "I first met Hatem in Cambridge, and he immediately struck me as a formidable founder," Kant noted. "He combines technical excellence with low ego and unapologetic ambition. He knows the science in extraordinary detail, but he’s also deliberate about bringing in people who are exceptional in areas different from his own."
Kant added that the overarching vision for the company extends far beyond incremental improvements to existing software suites. "Potential Sciences has the ambition to become a true frontier lab for chemistry, with the potential to fundamentally change how the field works."
Leadership and Origins
Potential Sciences was founded by Dr. Hatem Helal, a prominent technologist and former executive at Graphcore, the pioneering British AI semiconductor company that was ultimately acquired by Japanese tech conglomerate SoftBank. Dr. Helal’s background at the intersection of high-performance computing, hardware acceleration, and machine learning heavily influenced the architectural design of Potential Sciences. Recognizing that classical computing architectures often buckle under the immense computational weight of quantum mechanical simulations, Dr. Helal assembled a multidisciplinary team capable of optimizing software algorithms to run efficiently on modern processing hardware.
The company’s establishment in late 2024 and rapid scaling through 2025 coincided with a broader wave of consolidation and specialization within the European artificial intelligence sector. While early waves of generative AI focused predominantly on large language models and natural language processing, the current wave of venture investment is increasingly targeting deep science applications—often referred to as AI for Science. These ventures require founders who possess advanced doctoral-level understanding of complex physical sciences alongside commercial execution capabilities, a profile that investors identified in Dr. Helal.
The inclusion of high-profile angel investors, most notably Stripe CEO Patrick Collison, further validates the startup’s technical thesis. Collison has long maintained an active interest in scientific progress, technological acceleration, and the institutional mechanisms required to foster scientific breakthroughs, frequently backing ambitious companies operating at the frontier of computation and biology.
Market Context and Industry Implications
The broader drug discovery sector is undergoing a structural transformation as pharmaceutical executives face the "patent cliff"—a period where numerous blockbuster drugs lose exclusivity, forcing companies to refill their pipelines rapidly. Traditional drug discovery is notoriously slow, taking an average of ten to fifteen years and costing billions of dollars per approved therapeutic, with failure rates exceeding ninety percent during clinical trials.
Computational approaches have long promised to alleviate these inefficiencies, but early generations of computational chemistry software were often hampered by slow execution speeds and limited predictive accuracy. The convergence of advanced machine learning techniques, specialized AI hardware, and quantum chemical simulations represents a new frontier in the quest for rational drug design.
By anchoring its models in the unalterable laws of physics rather than historical observations, Potential Sciences aims to solve the generalization problem that plagues data-driven drug discovery. If the startup can successfully deliver on its promise of reducing computational design costs by a factor of one hundred, it could fundamentally alter the economics of early-stage pharmaceutical research. Pharmaceutical companies could evaluate vastly larger libraries of chemical compounds in silico, identifying promising candidates with higher binding affinities and fewer off-target toxicities before committing physical resources to synthesis and wet-lab testing.
Chronology and Future Outlook
The trajectory of Potential Sciences highlights the rapid pace at which deep-tech startups can secure institutional validation when operating in high-priority technological domains:
- November 2024: Potential Sciences is officially founded by Dr. Hatem Helal, drawing upon his expertise from Graphcore and advanced computational research.
Early 2025: The company establishes its operational headquarters in London and begins developing its physics-first foundational AI architecture for molecular simulation.
- Mid-2025: The startup engages with venture capital syndicates and public sector investment vehicles to secure early-stage funding.
- Current Announcement: The UK Sovereign AI fund, alongside Stripe CEO Patrick Collison and a diverse syndicate of venture capital funds, confirms a seed financing round valued in the tens of millions of dollars, coupled with a strategic allocation of sovereign compute capacity.
Looking ahead, Potential Sciences plans to deploy the newly acquired capital to expand its engineering and scientific research teams, scale its proprietary computational infrastructure, and initiate pilot partnerships with pharmaceutical and biotechnology enterprises. While the company has chosen to maintain a measured public profile regarding its precise commercial timelines, the convergence of state-backed infrastructure support, private sector capital, and rigorous scientific leadership positions Potential Sciences as a notable entity to watch within the global computational chemistry landscape.


