Welcome back!
In this episode, Andreas Munk Holm sits down with Simon Thomas, CEO of Paragraf, one of Europe’s rare hard-tech success stories, taking graphene from scientific breakthrough to industrial-scale electronics.
Graphene has been called the “wonder material” for two decades. The promise has always been clear: faster, better, and dramatically more energy-efficient electronics. The missing piece has been execution at scale. Simon and the Paragraf team are building that missing bridge, with the world’s first graphene electronics foundry in the UK, a growing portfolio of real commercial products, and a deep conviction that the next era of computing will require new materials, not just bigger data centers.
This is a conversation about what it truly takes to build venture-backed hardware in Europe.
How you fund capex-heavy deep tech.
How do you keep investors aligned when timelines are long.
How you keep teams motivated through delays and national security reviews.
And why AI may accelerate materials discovery, but won’t replace the brutal, necessary work of turning atoms into real manufacturing.
What’s covered:
01:27 What Paragraf is building and why graphene matters now
03:50 Graphene wafers and the world’s first graphene electronics foundry
04:23 What graphene changes for power consumption and device life
05:01 Why graphene isn’t already inside data centers
08:02 Foundry versus product company: why Paragraf does both
09:40 Graphene’s 20-year journey from papers to real-world scale
13:15 When venture investors first showed up and what they needed to see
16:58 Sovereignty, British Patient Capital, and why “national backing” matters
27:36 Capex, equity limits, and the painful mechanics of deep-tech financing
30:22 Surviving hard moments: people, pivots, and the NSI Act review
38:10 How to structure boards over time, from tactical to strategic
42:23 Keeping teams committed through uncertainty
46:10 Where Paragraf is today: headcount, geographies, and commercialization
49:16 AI in materials discovery and why manufacturing is still the bottleneck
🎧 Listen on Apple or Spotify, or queue it for later with chapters ready to go.
Show Notes
Paragraf: Turning Graphene from a Breakthrough into a Factory
Simon starts with the simplest possible description of what Paragraf does. Graphene was discovered in 2004 in Manchester and quickly gained a mythical reputation because its properties are genuinely unusual. Electronics made with graphene can be far more energy efficient, while also offering higher performance.
The hard part has never been “does graphene work.” The hard part has been making it reliably, at wafer scale, in a way that fits into how the semiconductor industry actually builds things. Simon describes Paragraf’s core breakthrough as graphene grown directly on standard semiconductor substrates, so it can go straight into a processing line. That is the hinge between lab discovery and industrial adoption.
Today, Paragraf runs what Simon describes as the world’s first graphene electronics foundry in the UK. Their ambition isn’t just a better chip. It’s a new materials platform for electronics, built for a world where energy efficiency is becoming the limiting constraint.
Why Graphene Isn’t Already in Every Data Center
Andreas asks the obvious question: if graphene can reduce power requirements dramatically, why isn’t it already inside AI infrastructure?
Simon’s answer is blunt and physical. Graphene is one layer of carbon atoms thick. Making a uniform single-atom layer across an entire wafer is incredibly difficult, and handling it safely and consistently is not like working with conventional bulk materials like silicon. A lab can demonstrate extraordinary properties; the fab has to produce repeatable outcomes at scale.
That is exactly the gap Paragraf is trying to close.
Graphene Won’t Be Alone: The “2D Electronics” Future
One of the most interesting parts of the episode is Simon’s view that graphene will not simply replace silicon across everything. Graphene is the first true 2D material to prove what’s possible, but it is opening the door to a broader category of two-dimensional semiconductors. In practice, the long-term future might be a stacked ecosystem of graphene plus other 2D materials depending on the use case.
For now, it remains a commercial question. Where does graphene create an undeniable business case today, and which customers are ready to adopt it?
Foundry First, Products First — Why Paragraf Had to Do Both
A key theme is the awkward reality of deep tech commercialization. Spinning out “a foundry” is not a common venture model. To get investors and customers to take you seriously, you have to prove you can manufacture something real, and you have to prove that the material can be used to build functioning devices.
Paragraf started by creating its own “vanguard” products as proof that graphene is not just a scientific concept. Now, Simon says the company is shifting toward customer-driven development, where external teams bring their own ideas into Paragraf’s foundry.
It’s a loop: you put something in a customer’s hands, they test it, they come back with what they actually want, and then you build the real version together.
Graphene’s Timeline: From Scotch Tape to Investable Company
Simon walks through graphene’s arc. The first papers emerged in 2004. The manufacturing question began around 2007 and 2008, with companies like Samsung exploring whether graphene could be produced at scale. From 2007 to 2013, the challenge was quality. Researchers could build amazing devices in labs, but production-scale graphene clean enough for real electronics remained elusive.
By 2015, Simon and Professor Sir Colin Humphreys at Cambridge began tackling the “lab to fab” problem directly. That work ultimately became Paragraf’s founding spark.
When Venture Investors Finally Leaned In
Investor interest became real around 2016 and 2017. Simon explains that the turning point was learning what venture capital actually needed. Graphene alone wasn’t investable. Investors needed proof that graphene could become a product with commercial relevance. Once Paragraf had tangible proof-of-concepts, the company became legible to capital.
It’s a recurring deep-tech pattern. Novel science becomes fundable once it becomes touchable.
Sovereignty and Signaling: Why British Patient Capital Matters
Andreas links Paragraf’s funding journey to the broader sovereignty shift sweeping Europe. Simon agrees that government-aligned capital plays a catalytic role. When a national institution backs a company like Paragraf, it becomes a credibility anchor that unlocks other investors.
He also acknowledges the reality that novel materials can carry national security implications. As geopolitical tension increases, founders building foundational industrial capabilities have to think carefully about where they locate operations, which partnerships they accept, and how shifting politics changes the business environment overnight.
Deep Tech Capital: Why This Journey Isn’t SaaS
Simon is direct about the brutal financing math. Paragraf is capital intensive. It needs machines, facilities, and specialized production lines. Early on, equity is often the only option because there are no assets or revenue streams to borrow against. That means founders must model capital needs obsessively and understand the path from spending money to reaching margins.
He highlights the moment when companies like Paragraf can begin to access debt financing, which can be cheaper than equity but adds its own complexity and risk.
The deep-tech venture job is not just raising money. It is managing the capital stack while trying to build something the world has never manufactured before.
Hard Times: The NSI Review and the Importance of People
When asked about the hardest moments, Simon points to a National Security and Investment Act review that put the company in limbo for seven months. It was a forced pause that could have become existential. Instead, the team used the time to explore adjacent opportunities and emerged with a new product line.
His core point is simple. When uncertainty hits, the difference is people. A team that believes in the vision will find another path forward rather than waiting for permission to continue.
How to Run Boards Over Time
One of the most practical parts of the episode is Simon’s explanation of how board engagement evolves.
Early on, Paragraf held monthly board meetings. That cadence wasn’t bureaucracy. It was operational leverage. The board helped with tactical execution: suppliers, services, legal support, and operational decision-making. As the company matured, meetings dropped to every two months and eventually quarterly, with the focus shifting from tactical reporting to strategic projection.
The pattern is clear. Early boards help build the machine. Later boards help steer it.
Keeping Teams Aligned When the Road Is Long
Simon’s approach to team retention is unusually transparent. He emphasizes purpose, vision, and values as anchors, and he prioritizes regular internal communication about risk and fundraising. The goal is not to pretend everything is fine, but to avoid the morale damage that comes from people feeling excluded from reality.
His point is that teams don’t leave because things are hard. They leave when they don’t understand what is happening and stop believing the goal is real.
Where Paragraf Is Today
Paragraf now employs around 110 people, operates two sites in Cambridgeshire, and has a U.S. subsidiary in San Diego after acquiring a complementary company. The company is scaling commercialization, expanding globally, and preparing new offices in Shanghai and the UAE.
Simon reflects on the surreal nature of the journey. A single experiment on a wafer becomes a global company, a global supply chain, and potentially a new industrial layer in electronics.
AI and Materials: Acceleration, Not Replacement
The conversation closes on AI. Simon sees AI as a major accelerator for materials discovery and modeling, but he makes a critical distinction: modeling doesn’t equal manufacturability. There is still a missing bridge between simulating a new material’s properties and reliably producing it in the real world.
AI may compress discovery cycles. It won’t eliminate the hard work of the lab-to-fab transition.
One-Line Takeaway
Graphene has been promised for twenty years, but Paragraf is building the missing piece: industrial-scale graphene electronics, funded with patience, manufactured with discipline, and aimed squarely at the energy limits of modern computing.
