Go Flux Yourself: Navigating the Future of Work (No. 19)

TL;DR: July’s newsletter explores why servers are literally melting, teenagers are hacking from their bedrooms, and we’re wasting resources on an industrial scale. Better stewardship – of energy, talent, and attention – separates tomorrow’s winners from the rest.

Image created on Midjourney

The future

“We cannot build enough chips, we cannot build enough data centres, we cannot build enough power grids to meet the demand. But whether or not we need it and where do we draw the line, that’s a different topic.”

The servers are melting, literally.

On July 19, Google and Oracle revealed that their UK data centres had experienced cooling-related failures during Britain’s third heatwave of the summer. Google’s London-based region Europe-West2 went down, while Oracle’s local facilities faced similar technical difficulties. Our digital infrastructure buckled under the relentless heat.

These weren’t isolated incidents, either. OpenAI’s image-generation models regularly push graphics processing unit temperatures beyond sustainable limits, while cooling systems struggle to keep pace with computational demands. 

According to the Boston Consulting Group (BCG), generative AI is expected to account for 60% of data centre power growth through 2028. Every new AI iteration consumes exponentially more power, water, and computational capacity than the last.

What we’re witnessing is the collision of two unsustainable systems: our heating planet and our energy-hungry digital ambitions. It’s hard to ignore the brutal mathematics, which add up to a planet increasingly heating up and being sucked dry.

Notably, on May 20, the United Kingdom reached Earth Overshoot Day: the date when our demand for ecological resources exceeds what the Earth can regenerate in a year. Alarmingly, we hit this tipping point three days before supposed emissions-guzzler China, and far earlier than the global average. We’re over-consuming everything, not just digital resources.

Mehdi Paryavi, Chairman of the International Data Center Authority (IDCA), put it starkly when I spoke with him recently: the global race to build computing capacity is outpacing our ability to power it sustainably. “AI becoming the future of our world today does require massive investments,” he told me. “We cannot build enough chips, we cannot build enough data centres, we cannot build enough power grids to meet the demand,” he added, which is where this month’s opening quotation comes from.

The numbers tell the story. According to the International Energy Agency (IEA), data centres are on course to consume 3% of global electricity by 2030, roughly equivalent to the entire energy consumption of Japan. Elsewhere, the Center for Strategic and International Studies calculates that United States data centres alone will need 84 gigawatts by the end of the decade, a 2,100% increase from today’s 4 gigawatts.

Meanwhile, Morgan Stanley research published last September projects data centre carbon emissions will triple by 2030, reaching 2.5 billion tons of CO2. Forget about building a digital future, we’re constructing a carbon catastrophe.

The regulatory landscape is tightening, too. The upcoming Cyber Security and Resilience Act, expected to be introduced within weeks in the UK, will bring data centres into critical national infrastructure for the first time. Andy Green, a partner at cybersecurity consultancy Avella, explains the significance: “Where previously it was just best endeavours or they tried to comply with an ISO standard, now they’ll have actual regulations that they need to meet.”

This means mandatory incident reporting within specific timeframes, supply chain security assurance, and adherence to the National Cyber Security Centre’s Cyber Assessment Framework. As Green notes: “It’s a high bar that they’re going to have to hit in quite short order.”

The shift represents another constraint on an already strained system. Data centres now face not only physical limits from heat and power, but also regulatory compliance, which could further complicate their operations.

At South by Southwest London in early June, I enjoyed speaking with Will Alpine (he and his wife, fellow co-founder of Enabled Emissions Campaign Holly, chose their surname to reflect their climate activism, which I love).

Alpine is a former Microsoft sustainability engineer who helped coin the term “Green AI” back in 2020. He was among the first to warn about AI’s energy impacts before the generative AI boom made those concerns impossible to ignore. Alpine eventually left Microsoft after internal battles over what he calls “enabled emissions”: the carbon cost of AI accelerating high-pollution industries.

His frustration with corporate sustainability theatre was palpable: “It doesn’t matter how green your compute is if it’s being used to enable fossil fuel extraction.”

This insight points to the heart of our resource crisis: we’re optimising individual components while ignoring systemic waste.

Alpine warns that we must consider not just the operational energy costs, but “the physical materials that go into making the chips and precious materials, all the resource constraints associated with that, and then the waste at the end of the life cycle”. Often, hardware is recycled before its natural end to chase the latest specifications, and companies sleepwalk through needless software upgrades.

Here’s where the crisis becomes truly absurd. Tomás O’Leary, CEO of Origina, captured this waste perfectly when we spoke recently. “Every pointless software upgrade, every vendor-mandated migration, every compliance-driven refresh adds to infrastructure strain,” he said. “Companies waste enormous resources on changes that deliver zero business value.”

The infrastructure barons are exacerbating the situation. Hyperscalers now command around 60% of data centre capacity, with Amazon, Microsoft, and Meta leading the land grab. They’re securing entire campuses before they hit the market, creating what O’Leary calls “digital feudalism”, where smaller companies queue cap in hand for access to computational resources.

Yet amid this escalating consumption, some organisations prove that efficiency beats excess. At the start of the year, DeepSeek achieved generative AI performance comparable to ChatGPT at just 6% of typical costs: $6 million versus the industry standard of $100 million. The problem isn’t computational limits, though, but computational waste.

Chicago-based Roger Strukhoff, IDCA’s Chief Research Officer, sees this reflected in the organisation’s latest global digital readiness rankings. Despite hosting 44% of the world’s data centres, the US ranks just 38th globally because its approach prioritises scale over sustainability. He told me: “Are we environmentally sound? Are we socially, economically sound? There’s no point having a digital economy if your society is going in other directions.”

The Scandinavian countries top the IDCA rankings because they understand holistic resource management. Building the biggest data centres means nothing if you’re burning through energy, materials, and human potential at unsustainable rates.

This brings us to our most mismanaged resource: human talent. While we obsess over GPU shortages, we’re squandering cognitive capacity on an industrial scale. Arrests in mid-June, related to the Scattered Spider cybercriminal gang, highlight just how badly we’re failing young people with extraordinary capabilities.

On July 10, police arrested four people in connection with the M&S and Co-op cyberattacks from a couple of months ago: a 20-year-old woman from Staffordshire and three males aged 17-19 from London and the West Midlands, highlighting the poor support we provide to young people with extraordinary capabilities.

Indeed, these weren’t sophisticated criminal masterminds but teenagers operating from their bedrooms. A 2022 study by the University of East London found that 69% of European teenagers have committed some form of cybercrime. Additionally, there are 3.5 million vacant cybersecurity positions globally. We’re criminalising the people we need to protect our digital infrastructure.

For the New Statesman, I’ve been investigating the subject of teenage hackers: their motivations, the lack of career opportunities, and how initiatives like The Hacking Games, which uses gaming to identify skills to combat cybercrime, are addressing these issues. The pattern is depressingly predictable: bedroom gaming leads to online gaming, then gaming cheats, hacking forums, minor cybercrime, financial gain, and finally serious cybercrime. Each step feels logical and harmless, yet the pathway leads from curiosity to destruction. (I will share the essay once it is published. As a long-time NS subscriber, I am thrilled to feature its hallowed pages.)

The rise of AI should free humans to focus on tasks that are uniquely human, such as creativity, empathy, and complex problem-solving. Instead, we’re automating away entry-level opportunities while burning out experienced workers with increasingly demanding tasks. Two-thirds of HR professionals now believe that colleagues who do not use AI risk falling behind, potentially creating a two-tier workforce, according to new Culture Amp research. Yet 77% of those using AI are entirely self-taught, hinting at a shortfall in formal AI training across organisations.

The solution isn’t to abandon technology but to steward it more thoughtfully. As Paryavi explained, data centres “are not the drivers impacting our environment negatively. They are actually enablers of creating a cleaner environment if we do it right, if we think outside the box.”

Smart organisations are already making this shift. Rather than constantly upgrading systems, they’re optimising existing ones. Instead of demanding the latest AI models, they’re solving specific problems with targeted solutions. And rather than competing for scarce computational resources, they’re building efficiency into their operations from the ground up.

Those who choose efficiency over excess, sustainability over waste, and genuine innovation over artificial obsolescence will thrive within physical constraints while others chase computational fantasies.

As we enter an era where every joule of energy and every moment of human attention becomes precious, the future belongs to the resource stewards, not the resource raiders.

The present

Image created on Midjourney

By the time this month’s Go Flux Yourself is published, on the last day of July, I’ll be in Greece on a family holiday, taking a proper break for the first time in months. No newsletters or keynotes to write, podcasts to organise, workshops to facilitate, and no interviews to conduct. Just sea, sun, and the kind of decompression that only comes when you physically remove yourself from your usual environment.

This break feels particularly timely, as I facilitated a couple of workshops on the power of taking breaks earlier in July. I’ve been working with a global pharmaceutical company on their LinkedIn summer challenge, helping technical professionals find their voices during the holiday season. The core message was simple: time away helps creativity flourish in ways that always-on hustle simply can’t match.

During an unrelated visit to this client’s UK headquarters, I interviewed six young professionals about their international assignments. The company sends young European graduates anywhere in the world for up to two years through their early-career development programme, and their role is often very different from what they studied, for good reason. For example, one employee shifted from project management to employee engagement. Another moved from pricing to global market access. A third transitioned from pharmacy to brand management. The pattern wasn’t about perfect qualifications but recognising capabilities that traditional recruitment processes miss entirely.

This approach reminds me of February’s newsletter, where Harvard’s Siri Chilazi warned about the distinction between “performative fairness” and structural change. This company has embedded skills-based career development into its culture, actively cultivating potential across international boundaries.

On the subject of expanding opportunities, I’m pleased to announce I’m now hosting a new podcast for Clarion Events called DTX Unplugged, which delves into innovations, trends, and challenges shaping business evolution. After years of moderating the company’s main stage panels and writing event takeaways, this feels like a natural evolution. Good working relationships develop organically over time, and when they mature into new opportunities, that’s genuine relationship capital at work.

The timing of the podcast – which I will share in future newsletters, once live – aligns with broader shifts I’ve been tracking since launching Go Flux Yourself in January 2024. Ultimately, in our resource-constrained world, the most successful collaborations are no longer transactional but built on a shared curiosity about solving real problems together.

Back to switching off, now. Remember that rest doesn’t equal unproductive time. It’s an investment in capabilities that can’t be automated or optimised away. The ideas that emerge from genuine downtime, the connections that develop through unhurried conversations, and the perspectives that shift when you step outside familiar patterns all remain irreplaceably human.

It will take me a couple of days to properly decompress in Greece. And, no doubt, in the final days before my flight home, I’ll start thinking about my next moves: a new speaker-focused website, for instance, more speaking opportunities, and exciting podcast developments. But for now, the most productive thing I can do is absolutely nothing at all.

The past

Image created on Midjourney

As a bushy-haired 12-year-old football addict with a love of Alessandro Del Piero, I scored a hat-trick for Tigers against Panthers to win our internal school football competition. It remains one of my most cherished memories: being carried off the pitch on my teammates’ shoulders, feeling that perfect combination of individual achievement and collective celebration. Sad, I know!

Over three decades later, watching my own children at their sports day a couple of weeks ago, I was reminded why these moments matter so much.

The temperature had hit nearly 30 degrees as parents gathered on the Astroturf. It was Darcey’s first sports day, and I watched her throw herself into each event with uninhibited enthusiasm, a little taken aback. There was space hopping (where she excelled), the egg-and-spoon race, football dribbling, and the climactic tug of war, among other more traditional events. The children competed in teams representing continents of the world, with both Darcey and elder brother Freddie proudly wearing Europe’s (blue) colours.

What struck me wasn’t just the competition but the collaboration. Freddie, despite being one of the tallest and strongest in his year, made sure to encourage a teammate with special educational needs, offering words of support and gentle reassurance through gentle touches during the more challenging events. In that moment, I saw everything that matters for the future of work: competitiveness balanced with compassion, individual capability channelled toward collective success.

I thought about this before appearing as a human-work evolution expert on St James’s Place Financial Academy’s The Switch Podcast, where I discussed preparing careers for an AI-driven future with host Gee Foottit. I was asked whether it’s possible to future-proof careers anymore, or if adaptability is the only strategy. My answer drew on exactly these sporting metaphors. “You can’t future-proof a career,” I said, “but you can future-ready yourself by developing what I call the ‘six Cs’: the uniquely human capabilities that become more valuable as AI advances.”

I listed communication, creativity, compassion, courage, curiosity, and collaboration. These so-called soft skills are the defining capabilities of the age. They allow us to build trust, forge connections, and work across differences in ways that no algorithm can replicate. (Again, I will share the episode once it has been aired, later in the year.)

The sports day reinforced this: children learning to compete fairly, support teammates, handle disappointment gracefully, and celebrate others’ achievements. These are precisely the skills that will matter in a world where routine tasks become automated.

The most successful societies have always been those that channel competitive instincts toward collaborative ends. Whether we’re talking about Olympic teams, scientific research groups, or the small cohort of reformed teenage hackers now working to protect rather than exploit digital infrastructure, the magic happens when individual capabilities serve shared purposes.

The pre-digital era naturally fostered this kind of collaboration. You couldn’t build a cathedral, win a war, or explore new continents without complex coordination between different specialists. Physical constraints meant that waste was visible and resources had to be allocated thoughtfully.

We’ve gained tremendous individual capabilities since then, but we’ve lost something essential: the productive friction that forced people to work together, to consider long-term consequences, to balance personal ambition with collective welfare.

As we enter a period where every joule of energy, every moment of human attention, and every ounce of raw material becomes precious, those childhood lessons from the playing field become more vital than ever. Competition drives excellence, but collaboration determines whether that excellence serves human flourishing or just individual advancement.

The teammates carrying me on their shoulders weren’t just celebrating my goals but what we’d achieved together. In our resource-constrained future, that distinction may well determine which teams, organisations, and societies thrive.

Statistics of the month

🏢 AI governance vacuum
Some 93% of UK organisations use AI, but only 7% have fully embedded governance frameworks. Alarmingly, 35% of companies have no clear owner for AI strategy, despite EU AI Act obligations. [🔗] [🔗]

⚙️ AI testing failures exposed
Only 28% of organisations apply bias detection during AI testing, while just 22% test for model interpretability. Most rely on legacy development processes that have not been updated to address AI-specific risks, such as bias and explainability gaps. [🔗]

🎯 AI skills paradox
One in five UK business leaders now depend on freelancers to deliver critical AI skills they lack in-house, while 46% of freelancers report increased earnings from AI work. [🔗] 

💼 Freelance revolution accelerates
Over a third of UK businesses now save more than £40,000 monthly using freelancers, with 87% planning to engage them up to 10 times in the next six months. Meanwhile, 70% of self-employed workers earn more than in full-time employment. [🔗]

🚫 AI resistance grows
Meanwhile, across the Atlantic, nearly two-thirds of US adults (64%) say they’re more likely to resist using AI-powered technologies as long as possible, compared with 35% who say they’re more likely to embrace using AI as soon as possible. [🔗]

Thank you for reading Go Flux Yourself. Subscribe for free to receive this monthly newsletter straight to your inbox.

All feedback is welcome, via oliver@pickup.media. If you enjoyed reading, please consider sharing it via social media or email. Thank you.

And if you are interested in my writing, speaking and strategising services, you can find me on LinkedIn or email me using oliver@pickup.media.