Sovereignty, AI and Matthew Broderick
"Shall we play a game?"
You might recognise this question from the 1983 sci-fi thriller WarGames. I used this exact quote to open a recent presentation on Sovereign AI I gave at Holyrood.
The cult classic follows a young hacker named David (played by Matthew Broderick) as he accidentally connects to a military AI named Joshua. Through no fault of its own, Joshua turns out to be quite the handful, leaving Broderick to navigate a crisis with catastrophic global stakes.
WarGames arrived in theatres the very same year I was born. Everything about it screamed 'emergence'. It is widely credited with popularising hacking, cybersecurity, and AI. A heartbeat before setting foot in that theatre, audiences firmly believed humanity was the world's undisputed intelligence. Talking computers didn't exist, and our emotional relationship with technology was limited to a few loving remarks directed at the family car.
But try imagine being a newly minted teenager 1983, no mobile phones, no internet as we know it and no home computing. For teenagers this movie had more in common with Lewis Carrolls through the looking-glass where Alice steps through a mirror and enters a world of compete absurdity where the normal rules of life no longer seem to apply.
Recently, with all of the seismic shifts in the technology landscape in the last three years, I can imagine a fair number of us feel a bit like that right now, I know I certainly do.
Navigating the Absurdity
It feels as though we’ve crossed an unseen event horizon. We are now trying to navigate a complex world of:
- Rapidly shifting landscapes: AI landscape is being transformed every 9 months.
- Macroeconomic anxiety: Bond and gilt yields continue to fuel stress, while private credit risk and eyebrow raising IPOs are creating massive short-term uncertainty for businesses.
- Supply chain fragility: Security, availability and vendor execution disruptions supported by staggering cost increases have pushed once-credible partners to the very top of board risk registers.
So, what do we do? In a world of absurdity where variables are shifting daily, having a firm grip on the few things you can control is the best, and frankly, only strategy.
Let’s Talk About AI
Attempting to build a business case for AI investment when the foundations of the technology, and market sentiment, are constantly shifting feels practically infeasible for most organisations.
While the UK economy is picking up, this may be less about stable(ish) government and more aligned to capital flight from US markets. We are simply not in a low-interest-rate environment that allows for frivolous, exploratory investments across several unproven technology stacks.
To date, most enterprise organisations have deployed a proof of concept for Microsoft Copilot. It provides value, but measuring that value remains a moving target. With reports suggesting low conversion rates for paid adoption and big tech adjusting its "AI all the things" rhetoric, something isn't quite clicking.
I'm paraphrasing here but I recall Jeff Sutherland (co-creator of Scrum) famously highlighting: spending vast amounts of time and resources trying to increase the performance of a single individual is a fool’s errand. You hit a ceiling of diminishing returns incredibly fast. The value of focusing on improving a team’s performance or a function is orders of magnitude higher.
Copilot is good. It’s helpful. But it is not transformational. You aren't revolutionising your business model just by handing out Copilot and Claude subscriptions.
A friend of mine (who I'm increasingly jealous of) and who works in private equity repeatedly emphasises that PE firms have zero interest in businesses using AI superficially. They want to see businesses using AI fundamentally: to transform operational processes or serve customers in entirely novel ways. Most businesses are missing this bigger picture, focusing instead on spoon-fed, individual performance focused, monthly subscriptions from sovereign US cloud providers.
The True Economics of Token Costs
This hesitancy stems from the fact that we are in the earliest days of adoption. And if you are concerned about the impending cost of AI, you have every right to be.
Remember when Uber launched? You could get an air-conditioned ride home at midnight for the cost of a coffee, thanks to early venture capitalists subsidising your journey. The AI market is now shifting from flat-fee subscriptions to transactional, token-based costs for workloads. Like early Uber, we are currently reaping the benefits of heavily subsidised economics. If you look into the projected true economics of those token costs post-subsidisation... well, you might need a glass of water.
If you are designing an AI solution based on transactional, recurring token costs, you might solve an interesting problem today. But you must be honest with your leadership: you have absolutely no idea what this will cost the business in three to five years. 2x? 5x? 10x? more?
Compounding this is the realised risk that the technologies your revenue relies upon can easily be restricted or altered by a foreign government. Recent trepidation from decision-makers isn't naysaying - it’s cold, reasonable business logic.
The Case for Sovereign Compute
Ensuring your business has access to sovereign compute infrastructure, like our new sovereign cloud built with Canonical and Cisco that cannot be switched off by a third party in your supply chain and provides access to the latest open-source AI models for private data processing has never been more critical.
Open-source models are rapidly approaching feature parity with proprietary ones. In fact, they already achieve everything most enterprises require, and much more.
Over the next 24 months, savvy businesses will focus on securing guaranteed, predictable monthly infrastructure costs, avoiding endless transactional token fees that have diminishing subsidies and no price ceiling.
At the end of WarGames, the AI Joshua concludes: "A strange game. The only winning move is not to play."
In the modern business landscape, there is no option to sit this out. We're all playing a high-risk, high-reward game. But you can ensure you stay in control of your future, your security, your availability, your performance, and your costs.
You just need to ensure you're playing it with the right partners.
Visit our website at www.brightsolid.com
This article is sponsored by Brightsolid
Holyrood Newsletters
Holyrood provides comprehensive coverage of Scottish politics, offering award-winning reporting and analysis: Subscribe