DeepSeek’s Real Disruption. Dam Bust-ups. And Deepening Digital Isolation. Plus Spies in the Office. Plus more! #241

How this week rewrote power.

Adrian Monck

Jan 30, 2025

Grüezi!

This week?

A Chinese startup just punched trillion-dollar holes in Big Tech’s AI monopoly dream, suggesting the future might belong not to those with the biggest data centres, but to those who can roll-out accessible AI for real-world value.
Two nuclear powers are playing hydroelectric chicken in the Himalayas, with China and India building massive dams that could turn water into a geopolitical weapon affecting a billion lives.
And Gen-Z isn’t choosing screens over friendship – they’re adapting to an economy that’s priced human connection out of reach, with digital entertainment filling the void left by unaffordable social spaces.

1️⃣ DeepSeek’s Disruption Isn’t About Cost – It’s About Industry Structure

That draining sound is AI snake oil.

Markets took a $1 trillion bath this week on news that a Chinese startup built a state-of-the-art AI model with older chips and a fraction of the usual energy. If only they’d read this newsletter they might have seen it coming. If only I had money to invest I would be writing this from the Maldives. But enough what ifs! There’s a bigger story.

We’ve been assuming AI would follow the familiar winner-take-all pattern of other tech sectors – think Microsoft in PC software or Google in search. As the Financial Times notes, “It is only a moderate exaggeration to say all that has mattered for investors in the past few decades was being on the right side of one or more of these winner-take-all stories.”

DeepSeek suggests a different future. Their breakthrough blows three holes in the winner-take-all theory:

You don’t need the latest, most expensive chips
You don’t need massive capital to build competitive models
Applications might not need to “rent” from the biggest companies

The market reaction is telling. While Nvidia fell hard, the biggest damage wasn’t to the major tech platforms. Instead, it hit the emerging data centre economy – infrastructure providers, power companies, and component makers.

But here’s what many are missing: when technology becomes more accessible, demand often grows “multiple orders of magnitude” with “more use cases and more demand.”

This mirrors what happened in the early 2000s with fibre optics:

“what they didn’t bet on was that engineers would figure out how to make those cables more efficient.”

For business leaders, this suggests:

Value may shift from infrastructure to applications
Implementation and use cases could matter more than underlying models
The barrier to entry isn’t just about computing power anymore
Innovation will come from unexpected places

Corporate Western adoption faces significant hurdles due to security and privacy concerns about Chinese-developed AI.

But that’s not stopping Chinese AI developers. They describe their progress as following a “step function” pattern – pioneers require massive resources to break new ground, but fast followers can achieve similar results with roughly one-tenth the computing power.

Watch this space.

2️⃣ My Davos Takeaways From Last Week

Deep dives on DeepSeek

Alpine insights, written last week:

💥 DeepSeek’s Geopolitical Ambush – Now Market-Validated

Markets caught up to what was evident at Davos;
China’s open-source strategy is rewriting the rules.

💣 Sanctions’ Spectacularly Backfire:

China’s forced efficiency innovations threaten US infrastructure investments;
Market carnage hit Nvidia and data centre stocks hardest.

🏭 Infrastructure Empire Strikes Back:

Massive data centre investments look vulnerable;
Power utilities saw record stock drops;
DeepSeek’s efficiency threatens US energy-backed AI strategy;
Trump’s emergency power plant authorisation suddenly seems … questionable.

⚡️ The AI-Climate Calculus Shifts:

Utility stocks priced for massive data centre buildout crashed;
Jevons paradox suggests cheaper AI = more total usage;
The “green AI” race adds new complexity to US-China competition.

🐧 Open Source as Strategic Weapon:

Threatens US companies’ closed-source profit models;
Attracts Global South developers to Chinese ecosystem;
Could accelerate AI democratisation beyond US control.

🔧 The Implementation Power Shift:

Value may shift from infrastructure to applications;
US advantage in chips may matter less than assumed;
Innovation could spread globally more rapidly;

🥡 Market Forces Meet Geopolitics

The “Silicon Curtain” looks more permeable than Davos feared;
Global AI power dynamics more fluid than assumed;
The “intelligent age’ isn’t just a power grab – it’s a race to rewrite industrial fundamentals.

3️⃣ Jevons’ Paradox Just Hit AI

A Victorian-era economist from Liverpool is having a moment.

Microsoft’s Satya Nadella invoked “Jevons Paradox” to explain his reaction to DeepSeek’s breakthrough. In 1865, economist William Stanley Jevons noticed something counterintuitive: when steam engines became more efficient, coal consumption increased rather than decreased. Better efficiency made the technology cheaper and more accessible, driving widespread adoption.

Nadella sees the same pattern with AI. As models become more efficient (like DeepSeek’s), we won’t see less infrastructure spending – we’ll see explosive growth in AI deployment. That’s why Microsoft is transforming into what some analysts call “an AI investment bank:”

$80B in direct infrastructure investment;
$100B fund with Blackrock for global AI deployment;
Partnerships with sovereign wealth funds.

The historical parallel is deeper than it appears. Just as Victorian banks of Jevons’ era funded railway expansion after steam engine efficiency improved, Microsoft is positioning itself to finance the global rollout of more efficient AI.

From Victorian coal mines to modern data centres – Nadella’s betting that Jevons was right: efficiency gains don’t reduce infrastructure needs, they explode them. He’s hoping that Microsoft finances that explosion.

#AI #Technology #Innovation #Economics #Leadership

4️⃣ Dam, Dam and Double Dam

Geopolitics gets some hydro hysteria.

The world’s next major conflict might not start with missiles or tanks, but with a dam switch.

Deep in the Himalayas, the world’s largest hydroelectric dam is being built. At 60 gigawatts, it’s three times bigger than China’s Three Gorges Dam. Its location? Right before Tibet’s largest river flows into India, effectively giving China a tap on India’s water supply.

India’s response? It wasn’t diplomatic notes or UN resolutions. They’re building their own massive dam downstream – explicitly as a defensive measure against potential “water bombs” from upstream releases.

This isn’t just another gigantic infrastructure story. It’s about what happens when two nuclear powers compete for control over water that sustains a billion people. Both dams are being built in one of the Earth’s most earthquake-prone regions, where climate change is already melting the glaciers that feed these rivers.

The stakes extend far beyond India and China. Bangladesh, which gets 65% of its water from this river system, watches helplessly as two giants upstream reshape its lifeline. Local communities on both sides face displacement from ancestral lands.

We often talk about future water wars. But this isn’t the future – it’s happening now. These mega-projects represent a new kind of geopolitical competition where infrastructure becomes a weapon and water becomes power.

As we race toward a climate-changed world where water security becomes paramount, this Himalayan water race offers a preview of challenges we’ll all face. The question isn’t just whether these dams will generate clean energy – it’s whether they’ll remake the power balance of Asia.

#Geopolitics #WaterSecurity #Infrastructure #ClimateChange #GlobalRisk

5️⃣ Is Technology Replacing Our Desire for Human Connection?

Did Home Alone predict the future?

Social scientist Alice Evans wonders if we are gaming and podcasting ourselves into digital isolation. Solitude is the new Friday night out. Sounds compelling.

But what if we’ve got this back to front?

Young people aren’t choosing screens over friendship. They’re responding to the rising costs of social connection:

Can’t afford city apartments where friends live;
Work irregular hours in the gig economy;
Living with parents longer;
Priced out of traditional social spaces.

The result? A 50% increase in time spent alone since 2010.

Evans brilliantly identifies how digital entertainment (podcasts, gaming, social media) is replacing “non-market demand” for human interaction.

But here’s the deeper insight. Digital isolation isn’t the disruptor – it’s an adaptive response to economic precarity.

When a drink with friends means an expensive Uber, when rent consumes 50% of income, when working hours are unpredictable... digital connection becomes the only affordable form of social capital.

We’re not seeing technology replace social connection. We’re watching the market price it out of reach. That’s a much scarier disruption than AI taking our jobs.

#Economics #Society #FutureOfWork

6️⃣ Spies: Professional Alienation Down The Ages

Slow Horses for dead-end career courses.

The spy thriller has always been our best mirror for workplace anxiety.

In The Spy Who Came in from the Cold (1965), Richard Burton is the ultimate alienated worker – skilled, dedicated, and destroyed by his profession. His expertise renders him unemployable elsewhere. Sound familiar?

That same year, Michael Caine’s Harry Palmer in The Ipcress File brought class consciousness to the office. A working-class professional in upper-class institutions, this is about merit vs privilege – workplace politics with life-or-death stakes.

Palmer’s rebellion through competence – mastering filing systems, speaking German, cooking fancy meals – shows expertise as a weapon against class prejudice. His boss calls him “insubordinate” for being too good at his job.

By 1979, Alec Guinness’s George Smiley in Tinker, Tailor, Soldier, Spy represents a different archetype: the career civil servant as hero. His almost priestly power comes from institutional memory and patient expertise – precisely what many organisations are losing today.

Smiley wins through deep knowledge of systems, files, and human weakness – even his own. But the series hints that this type of career professional is endangered – even as it celebrates him.

Today, Gary Oldman’s Jackson Lamb presides over skilled professionals doing meaningless work while maintaining the appearance of purpose. Slough House reflects our current anxieties:

Promised careers vs diminished reality
Technology vs human expertise
Generational divides
“Family” rhetoric masking dysfunction
Work stripped of meaning but keeping up appearances

The evolution:

1960s: Profession destroys your humanity (Burton)
1970s: The twilight of expertise as power (Smiley)
2020s: Protection through workplace relationships not professional mission

What changed?

Class conflict to generational divide
From principled opposition to office politics
From craft knowledge to data metrics
From institutional loyalty to workplace relationships
From professional identity to performative work

Slow Horses resonates because it captures our moment: underemployed skilled workers, tension between tradition and tech, protective rather than purposeful management, eroding work-life boundaries.

The genius of spy fiction? Using extraordinary circumstances to illuminate ordinary truths. From Burton’s alienation to Lamb’s cynical protection racket, these stories track how our relationship with work evolved – and not always for the better.

Each era gets the spy story it deserves. From Burton’s alienated worker to Caine’s class-conscious professional, through Smiley’s masterful civil servant, to Lamb’s cynical defence of workplace rejects...

Our era is about skilled professionals maintaining appearances while doing work stripped of meaning. What does that say about modern office life?