fbpx Skip to content

Why we need to make hardware less hard

Hardware semiconductor

With the news agenda consumed by war in the Middle East and last month’s boardroom drama at OpenAI, you could be forgiven for having missed the release of the government’s Advanced Manufacturing and Battery Strategies.

While the reports were published without fanfare, both are important policy issues for the UK’s future competitiveness and our ability to respond to challenges like reaching net zero, rebuilding an industrial base, and future pandemic preparedness. The commitment to spend £4.5bn by 2030 in areas like autonomy and life sciences is a welcome signal of intent.

Managing the energy transition, improving disease prevention, responding to rising geopolitical tension and developing advanced computing are rooted in the physical world.

They require the attention of policymakers, but also the ingenuity of scientists and entrepreneurs working on breakthroughs in fields like material science, computational biology, autonomy and quantum mechanics.

VCs flocked to software. But some startups are bucking the trend

The notion of venture capital was first formalised in the 1950s and 60s to creatively finance companies like Fairchild Semiconductor and Intel that were developing technologies just like these.

The communications and computing infrastructure that resulted, as well as more recent innovations like sensors and GPS, enabled a subsequent wave of software companies to flourish.

Without the need for factories, inventory and supply chain risk, it’s no surprise that venture capitalists flocked to high-margin and capital-light internet-based software startups like Airbnb, Uber and Spotify which are now household names, rather than the hardware companies that underpinned them.

But hardware startups like Anduril (defence), SpaceX (space), Northvolt (batteries) and Wayve (autonomous vehicles) are bucking the trend on both sides of the Atlantic. Eight out of the 10 largest public companies by market cap rely to some extent on hardware.

To solve today’s challenges we need to invest in more of these deep tech companies that commercialise cutting-edge research and new discoveries in science and engineering. It was therefore reassuring to see that the share of investment going to European deep tech companies hit a record-breaking 44% of total capital invested in 2023, up from just 15% a decade ago, according to Atomico’s State of European Tech report.

But alongside a focus on startups working on tackling real-world problems that involve atoms, we should not forget the impact that bits can have in the physical world too.

Software’s role in manufacturing

Early hardware innovations in semiconductors, networking and communications created the infrastructure on which a wave of software solutions could be built. We now need to leverage the modern software stack to unlock a new wave of industrial progress.

We need to make hardware less hard.

Early pioneers in digital manufacturing design and simulation like Ansys, Autodesk and Siemens Digital Industries have been built over decades and developed large defensive moats.

They have complex features, broad coverage of the end-to-end engineering process for different industries, and a strong familiarity dividend – engineers have been using these products for most or all of their careers.

They are also highly acquisitive – Ansys, Autodesk and Siemens Digital Industries alone have bought more than 100 companies between them since the early 1990s. This means that they are entrenched in the manufacturing supply chain. Airbus claims that 85% of the code used to develop their aircraft is built and tested using Ansys software.

Software is therefore already an indispensable part of modern high-performance manufacturing and has been accelerating engineering since the early days of computing. The question is whether it can do more.

AI can disrupt incumbents

It is good news that the government has extended the Made Smarter programme, which supports industrial digitisation. This will help more companies adopt existing tooling.

But we need innovation too. Advances in machine learning, statistical physics, optimisation techniques and compute infrastructure open a door for startups to compete in different parts of the engineering process, including collaboration tools, virtual prototyping and robust generative design.

Just as Figma disrupted the digital design market and had been set to be acquired for $20bn by Adobe (before scrapping the deal amid regulatory scrutiny), cloud-first and AI-native startups like AI Build and PhysicsX are aiming to compete with the dominant incumbents.

AI Build (an IQ Capital portfolio company) has recently raised $8.5m to develop software that automates and leverages AI for 3D printing, while PhysicsX is aiming to help accelerate engineering simulations, fresh on the back of $32m in funding. Both are critical ‘software for hardware’ – the catalysts that will help power an industrial renaissance.

There is no question that hardware matters. But so does improving the software that will enable it.

Jonno Evans OBE is an investor at IQ Capital