Supercomputing, AI and Quantum: What Actually Matters for Businesses

I attended a Brighton AI event at Sussex University Innovation Centre last week on supercomputing, neural networks and quantum.

While it’s a very technical topic, there were some very practical insights that businesses should be paying attention to right now.

Insight 1: The scale - and scarcity - of computing is a global issue

One explanation really brought it to life:

Petascale = a quadrillion calculations per second

Exascale = a quintillion calculations per second

That’s the level of computing power now sitting behind modern AI systems which is truly a vast, almost unimaginable scale.

But there isn’t enough of it.

High-performance computing is limited, highly competitive to access, and increasingly expensive in the cloud.

At the same time, infrastructure is heavily concentrated among a small number of players dominated by US-based companies like Nvidia, with growing efforts in Europe around sovereign compute and AI ‘gigafactories’.

→ Where your compute comes from, and who controls it, is becoming strategically important, not just technologically, but economically and geopolitically.

Insight 2: The real constraint is energy - and not every country has it

One of the most striking points was that, on current trajectories, training AI systems could consume a significant proportion of global energy within the next decade.

Around 50% of the energy in high-performance computing goes into processing… and 50% into cooling the systems.

That might be manageable for energy-rich countries, with capability concentrating in these regions of the world. It may not be sustainable at a global scale and it means AI is as much an energy and infrastructure challenge as it is a technology one.

And this is where things get more complex.

The UK, for example, does not currently have the same level of energy capacity or infrastructure as some other regions, which raises questions about future competitiveness. One of the directions being explored is small modular nuclear reactors to support AI and computing demand.

→ The future of AI infrastructure may be shaped as much by energy availability as by technological innovation.

Insight 3: A new generation of AI could be dramatically more efficient

The human brain performs complex intelligence at around 20 watts. It does this through spikes and pathways, meaning highly efficient communication between neurons.

By contrast, in today’s AI systems, a significant proportion of energy is used simply moving data around, rather than doing the computation itself.

This is driving research into neuromorphic computing - systems that mimic how the brain works.

Early studies suggest this approach could deliver up to 2000x energy savings.

If that holds, it could fundamentally change the cost, accessibility and scalability of AI.

→ It also raises an important global question. More efficient models could allow countries with less energy infrastructure to leapfrog, much like we saw with mobile adoption.

Insight 4: Not all quantum is ‘future’

Quantum computing gets a lot of the attention, but it’s still developing. What’s much closer to market is quantum sensing, using quantum effects to measure things with far greater precision than traditional technologies.

The quantum shift isn’t one moment in the future - it’s already starting, in specific areas such as:

• Brain imaging

• Battery quality assurance

• GPS-free navigation

• Blood flow monitoring.

One example that stood out: if GPS were lost in the UK, the economic impact could be around £1.4 billion. Quantum sensing offers potential alternatives.

→ This moves quantum from ‘interesting’ to strategically important, particularly for resilience, security and national infrastructure.

So what does this mean for SMEs?

AI is becoming a core driver of competitiveness - not just a tool, but part of business strategy.

Which means businesses should already be thinking about how it fits into their growth plans.

We’re likely to see a shift towards more modular, scalable computing and greater self-sufficiency, building capability and control rather than relying entirely on external platforms. Upskilling is no longer optional.

Not every SME will move immediately, but the direction is clear.

The future will be shaped by how we choose to engage with these technologies. This makes this a particularly interesting moment to be building and growing a business.

Thanks so much to all the speakers, Brighton AI, Silicon Brighton and Sussex University Innovation Teams

• Andy Forrester (Hyperaccelerator Solutions)

• Thomas Nowotny (University of Sussex / Sussex AI)

• Thomas Coussens (Spectomiq).