The Frankenstein Computer That's Actually Real

Threat IntelligenceTechnology Risks
Written By Noel Bradford

It's Monday morning. You're checking your Azure invoice, and once again it's climbed another 10%. Your SaaS subscriptions just increased their AI feature pricing. Microsoft Copilot wants another £30 per user per month. And somewhere in your inbox, there's a cheerful email from AWS explaining that inter-availability zone transfer fees have doubled.

You're being told this is progress. Innovation. Digital transformation.

What they're not telling you: the AI revolution is burning through electricity faster than we can generate it, and you're footing the bill for an energy crisis that's spiralling out of control.

But here's where it gets interesting. While tech giants scramble to restart nuclear power plants, two Swiss scientists in a laboratory in Vevey have built something that sounds like science fiction. They're powering computers with living human neurons. Actual brain tissue, grown in dishes, processing information. And these biological processors use one million times less energy than the silicon chips consuming your budget.

What You'll Hear in This Week's Episode

The Swiss scientists who've gone properly mental: Dr. Martin Kutter and Dr. Fred Jordan already built one successful company in anti-counterfeiting. They could have retired comfortably. Instead, they're growing brain tissue in laboratory dishes and hooking it up to electrodes. Their company, FinalSpark, launched the world's first commercially accessible biocomputing platform in May 2024. It's operational right now. You can watch neurons computing in real-time on their website.

The energy numbers that should terrify you: Training a single large AI model produces the same carbon emissions as five cars create during their entire lifetime. That statistic is from 2019. Modern models like GPT-4 produce 50 to 100 times more emissions than that. Data centres worldwide will consume 945 terawatt-hours by 2030, equivalent to Japan's entire electricity consumption. And every single watt gets passed down to businesses like yours through rising cloud costs.

The nuclear power pivot nobody saw coming: Microsoft committed $1.6 billion to restart Three Mile Island. Google partnered with Kairos Power for 500 megawatts of Small Modular Reactors. Amazon announced multiple nuclear investments totaling billions. Meta issued a request for proposals for 1 to 4 gigawatts of new nuclear capacity. These aren't small bets. These are existential moves because renewable energy simply cannot meet AI's accelerating demands.

What this means for UK small businesses: Your cloud costs are rising 10 to 14% annually. SaaS inflation is running five times higher than consumer inflation. Business confidence in the UK hit a historic low of -58 points in Q3 2025. Almost one in three small firms expect to downsize, sell, or shut down in the next 12 months. The productivity tools that could save you are increasingly unaffordable due to the very AI advancement meant to help.

Why This Episode Matters More Than You Think

Listen, I know what you're thinking. Living neurons computing in Swiss laboratories? That's not my problem. I've got payroll to run, customers to serve, and a business to protect.

But here's the thing: the energy problem driving FinalSpark's research is already hitting your bottom line. It's in every Azure invoice. Every AWS bill. Every SaaS subscription that just added "AI-powered features" and raised prices by 20%.

The technology industry created an energy crisis, and they're solving it by passing the costs downstream. To you. To every small business owner who believed the cloud would make technology more affordable and accessible.

And while they scramble to restart nuclear power plants that won't come online until 2028 to 2035, your costs keep climbing. Right now. This quarter. This year.

That's why this episode matters. Because understanding where technology is heading helps you make better decisions about where to invest your limited resources. Because the AI tools eating your budget might not be the future everyone claims. Because biological computing, quantum systems, and neuromorphic chips are showing us there are fundamentally different ways to process information without burning through nuclear power plants' worth of electricity.

What Makes FinalSpark's Work Actually Credible

Before you dismiss this as science fiction, let's be clear about what's verified:

The company is real. FinalSpark Sàrl, Swiss company registration CHE-256.971.603, founded January 27, 2014, operating in Vevey, Switzerland. The founders, Dr. Fred Jordan and Dr. Martin Kutter, both have PhDs from EPFL Signal Processing Institute. They previously co-founded AlpVision, an anti-counterfeiting technology company with over 80 patents.

The platform is operational. Launched May 15, 2024. Nine research institutions granted free access from an initial pool of 36 universities. Commercial access available at $500 to $1,000 per month. The system has run continuously for four years, testing over 1,000 organoids and collecting 18 terabytes of data.

The research is peer-reviewed. Published in Frontiers in Artificial Intelligence (DOI: 10.3389/frai.2024.1376042). Achieved top 1% most-read status by October 2024. Multiple independent research groups validating findings at ETH Zürich, University of Michigan, Free University of Berlin, and several UK universities.

The specifications are documented: 16 brain organoids, approximately 10,000 neurons per organoid, 160,000 neurons system-wide, grown from induced pluripotent stem cells derived from adult cells, interfaced with 8 electrodes per organoid using 30 kHz sampling frequency.

This isn't vaporware. This isn't a press release with no substance. This is functioning technology, albeit early-stage, tackling a problem that's costing your business real money right now.

The Brutal Truth About AI's Energy Appetite

Let me give you the numbers that should worry every business owner:

2019 baseline: Training a large AI model emitted 284 metric tons of CO2 equivalent. That's the lifetime emissions of five cars, including manufacture. This research from University of Massachusetts researchers (DOI: 10.18653/v1/P19-1355) established AI's carbon problem. What happened next makes those numbers look modest.

GPT-3 (2020): Consumed 1,287 megawatt-hours of energy. Emitted 502 metric tons of CO2. That's the annual energy consumption of roughly 120 average US households.

GPT-4 (2023): Consumed 51,773 to 62,319 megawatt-hours. Emitted 1,035 to 15,000 metric tons of CO2 depending on data centre location. That's 40 to 48 times more electricity than GPT-3 for training, despite having only 10 times more parameters. The efficiency-to-scale relationship is deteriorating, not improving.

Current projections: The International Energy Agency's April 2025 report projects global data centre electricity consumption will double by 2030 to 945 terawatt-hours, equivalent to Japan's total electricity consumption. AI-optimized data centres will account for 35 to 50% of data centre energy by 2030, up from 5 to 15% currently.

And here's the kicker: training is just the beginning. Every query you run on ChatGPT, every Microsoft Copilot suggestion, every AI-powered feature in your SaaS applications burns energy. With 1 billion daily queries at 0.34 watt-hours per query, GPT-4's inference energy surpasses its training energy after just 150 to 200 days. Operational energy is the larger long-term concern.

How This Reaches Your Business

You might think, "That's Microsoft's problem, not mine." Wrong. Every cost rolls downhill.

Cloud pricing increases: Microsoft Azure increased cloud service licensing by 9% on April 1, 2023, then another 5% on monthly-billed annual subscriptions on April 1, 2025. Premium SSD rates jumped 10 to 11% in February 2025. Amazon Web Services doubled inter-availability zone transfer fees from $0.01 to $0.02 per gigabyte. Some customers reported AWS bills doubling between January and mid-2025.

SaaS subscription inflation: The Vertice SaaS Inflation Index shows SaaS now costs approximately $9,100 per employee in 2025, up from $8,700 in 2024 and $7,900 in 2023. That's a 15% increase over two years. SaaS inflation runs nearly five times higher than standard consumer inflation. Seventy-three percent of software vendors increased prices in 2023, with 60% deliberately masking rising prices.

Energy costs passed to consumers: The Jack Kemp Foundation reported in November 2024 that American consumers and small businesses could see electricity bills increase up to 70% by 2029. The primary driver explicitly cited: surging AI data centre energy demand. This isn't theoretical. Northern Virginia saw peak demand metrics increase from $29 to $444 per megawatt-day in 2025. Chicago customers paid an extra $11 per month in summer 2024 due to 45% electricity price jumps.

UK-specific pressures: UK data centres currently consume 2.5% of UK's electricity, expected to quadruple by 2030. Data centre electricity costs in the UK are four times more expensive than the United States. The Beaming SME Cloud Survey found 26% of SME leaders complained rising cloud costs harmed their businesses. By 2024, UK IT leaders reported average estimated cloud price increases of 14% over 12 months, exceeding UK inflation of approximately 10%.

You're not paying for innovation. You're paying for an energy crisis disguised as digital transformation.

Why Living Neurons Matter for Business Owners

FinalSpark's biological processors currently store 1 bit of information each. They perform simple stimulus-response tasks. They live approximately 100 days before cultures require replacement. They are not yet competitive with silicon for complex operations.

So why should you care about brain tissue computing experiments in Switzerland?

Because the energy problem is real, immediate, and escalating. Because your costs reflect this crisis in every invoice. Because the technology industry's solution is nuclear power plants that won't arrive until 2028 to 2035, forcing interim reliance on fossil fuels and continued cost increases.

Someone will solve the energy equation for computing. Maybe it's biological processors. Maybe it's quantum computing. Perhaps it's neuromorphic chips like IBM's NorthPole (4,000 times faster than TrueNorth, 25 times more energy efficient than comparable GPUs). Maybe it's some approach we haven't imagined yet.

But someone will solve it, because the current trajectory is economically and environmentally unsustainable. Understanding where technology is heading helps you make better decisions about where to invest your limited resources.

What Small Business Owners Need to Do Now

Stop paying for AI features you don't need. Seriously. Every SaaS vendor is slapping "AI-powered" on features and charging 20 to 30% premiums. Microsoft Copilot wants £30 per user per month as an add-on. Salesforce CRM charges $500 per month per seat for AI features. Unless you can demonstrate concrete productivity gains that exceed these costs, you're subsidising the AI energy crisis for zero return.

Audit your cloud spending monthly, not quarterly. Use cloud cost management tools. CloudHealth, CloudZero, native provider dashboards. One UK telecommunications company saved £40,000 monthly through rightsizing alone, totalling £400,000 saved in one year. That's real money that could fund actual business growth instead of vanishing into energy costs.

Challenge every price increase. When vendors cite "infrastructure costs," "energy expenses," or "AI capabilities," demand specifics. What are you getting for the extra money? Can you opt out of AI features? Are there alternative pricing tiers? Vendors are testing how much they can push costs downstream. Push back.

Watch for energy-efficient alternatives. Neuromorphic computing is already delivering orders of magnitude efficiency gains for edge workloads. Cloud providers will eventually need to compete on energy efficiency as costs become untenable. Early movers who identify and adopt more efficient platforms will have cost advantages.

Stay informed about technology shifts. Fifteen minutes a month staying informed about emerging developments pays dividends when those developments suddenly become relevant to your business. That's why this podcast exists. That's why understanding FinalSpark's work matters even though you won't be buying biological computers next year.

The Takeaway: Mad Ideas Sometimes Win

After 40 years in this industry, I've learned that the maddest ideas sometimes win. Especially the really mad ones.

Nobody believed cloud computing would replace on-premises servers. Everyone said Software-as-a-Service was a fad. The experts dismissed smartphones as toys. The technology that slashes fundamental operating costs always migrates downstream. What starts in research laboratories eventually reaches small businesses, usually faster than anyone expects.

FinalSpark's living neurons computing in Swiss laboratories might be the future. Or they might be a fascinating dead-end. But the energy problem driving their research? That's real and immediate. That's in your Azure bills. That's forcing technology giants to restart nuclear power plants.

And that's why, however barking mad it sounds, this episode matters.

Listen to the full 21-minute episode now. Graham nearly falls out of his chair when I explain how brain organoids actually work. Mauven provides the NCSC perspective on what happens when computing paradigms shift and security frameworks haven't caught up. And we dig into the specific numbers that explain why your cloud costs keep climbing.

Then come back tomorrow. We're spending the entire week unpacking this story, because it's too important to leave as just a podcast episode.

Stay curious. Stay sceptical. Stay secure.

And maybe keep one eye on the Swiss scientists growing computers in dishes. Because mad ideas sometimes win.

Noel Bradford

Noel Bradford – Head of Technology at Equate Group, Professional Bullshit Detector, and Full-Time IT Cynic

As Head of Technology at Equate Group, my job description is technically “keeping the lights on,” but in reality, it’s more like “stopping people from setting their own house on fire.” With over 40 years in tech, I’ve seen every IT horror story imaginable—most of them self-inflicted by people who think cybersecurity is just installing antivirus and praying to Saint Norton.

I specialise in cybersecurity for UK businesses, which usually means explaining the difference between ‘MFA’ and ‘WTF’ to directors who still write their passwords on Post-it notes. On Tuesdays, I also help further education colleges navigate Cyber Essentials certification, a process so unnecessarily painful it makes root canal surgery look fun.

My natural habitat? Server rooms held together with zip ties and misplaced optimism, where every cable run is a “temporary fix” from 2012. My mortal enemies? Unmanaged switches, backups that only exist in someone’s imagination, and users who think clicking “Enable Macros” is just fine because it makes the spreadsheet work.

I’m blunt, sarcastic, and genuinely allergic to bullshit. If you want gentle hand-holding and reassuring corporate waffle, you’re in the wrong place. If you want someone who’ll fix your IT, tell you exactly why it broke, and throw in some unsolicited life advice, I’m your man.

Technology isn’t hard. People make it hard. And they make me drink.

https://noelbradford.com
Previous

No MFA? No Certification. The Cyber Essentials Rule That Changes Everything
Next

Weekend Reflection - Efficiency Theatre and the Tyranny of the Measurable