Experience says settle down. The data says prepare

They've been hype cycles before. But never ones that can write their own press releases

The pace of technological and political change will only accelerate in 2026. I can only guess as to exactly how that will play out.
Me, 7/11/25

Everywhere you turn these days binary systems seem to be giving way to poly ones… We are moving from programmable computing — where a computer could only ever reflect the insight and intelligence of the human who programmed it — toward polymathic A.G.I. That is where you basically describe the outcome you want, and the A.I. melds insight, creativity and broad knowledge to figure out the rest… It is the mother of all computing phase changes — and a species-level turning point.
Thomas L. Friedman, NYT, 10/11/25

Many writerly conceits will perish before we get anywhere close to Artificial General Intelligence (AGI).

Even in this age of plain old antediluvian AI, it’s difficult to claim you don’t know something or, indeed, “can only guess” about something. After all, you can always just ask AI for the answer. That may even apply to predicting the future.

See below for what AI has forecast for 2026.

(I’ve graciously allowed AI do all the heavy lifting – that is, the melding of “insight, creativity and broad knowledge”- then chimed in at the end with a human take on its predictions.)

2026: What’s Coming
The coming year won’t bring a revolution or a collapse. It will make several long-running shifts visible: AI transitioning from novelty to infrastructure, electricity grids straining under new loads, and policy finally catching up with technology.

None of this is speculative; it’s the natural continuation of what’s already in motion.

1. AI agents become standard business infrastructure
By late 2026, Gartner expects roughly four in ten enterprise applications to include task-specific AI agents, up from fewer than one in twenty in 2025. These systems are being embedded into customer-relationship and resource-planning platforms to handle scheduling, reporting and basic client support.

According to McKinsey’s Global Institute, almost every large company is experimenting with generative AI, but only a small minority have the managerial or technical maturity to redeploy staff productively. The outcome is less a wave of innovation than a gradual reduction in middle-tier administrative labour.

2. Demand from AI begins to strain grids
The International Energy Agency projects that global data centre electricity consumption will more than double by 2030, with AI training and inference accounting for the majority of the increase. The effect is already measurable.

In Ireland, regulators have halted new data-centre approvals after the sector’s power use climbed to nearly one-fifth of national demand. In Virginia, utilities are seeking permission to build new gas-fired plants to meet the surge in server-farm consumption.

The pattern suggests that AI growth will soon intersect with decarbonisation targets, forcing governments to reassess both.

3. Global growth remains weak despite productivity gains
The World Bank’s Global Economic Prospects report expects global GDP to expand by only about 2.4 per cent through 2026. Higher interest rates, demographic slowdowns, and a retreat from open trade are offsetting any productivity gains from automation.

This combination produces a quietly corrosive outcome: technological progress without broad prosperity. Economies are becoming more efficient, while living standards remain stagnant.

4. The EU AI Act becomes the global template for regulation
On 2 August 2026, the EU AI Act will take full effect, requiring companies to classify AI systems by risk level, document training data, and ensure human oversight for higher-risk models.

Rather than maintain multiple compliance regimes, most major technology firms will align with Brussels’ standards. That will raise legal and administrative costs but effectively establish the EU framework as the global baseline for AI governance.

5. Renewables keep expanding, but mineral supply becomes the choke point
The IEA’s Renewables 2025 forecast indicates that renewable power capacity is on track to double by 2030, primarily driven by the deployment of solar energy. At the same time, the International Renewable Energy Agency notes rising price volatility for lithium, copper, and gallium – materials essential for batteries and photovoltaic cells.

This means the limiting factor for decarbonisation is shifting from technology to access to resources. The supply of critical minerals is concentrated in China and a few resource-rich states, making energy security another geopolitical contest.

6. Defence spending underwrites AI and robotics research
All 32 NATO members have now met or exceeded the alliance’s 2-per-cent-of-GDP spending target, according to Reuters.* Much of the new funding is flowing to autonomous systems, data-fusion platforms, and battlefield analytics.

Defence budgets will continue to underwrite AI and robotics research. The boundary between civilian and military innovation is narrowing, creating a durable dual-use market insulated from ordinary business cycles.

7. Supply chains regionalise into competing trade zones
BCG analysis indicates that Mexico overtook China as the United States’ largest trading partner in 2023. This trend is expected to continue through 2026 as manufacturers relocate to reduce their geopolitical exposure. Vietnam, India, and Eastern Europe are absorbing similar shifts.

The result is not deglobalisation but a division into regional blocs. Supply networks are being rebuilt around political reliability rather than cost, trading efficiency for predictability.

8. Shorter workweeks spread selectively among high-skill sectors
Results from the U.K. four-day workweek pilot show stable or improved revenue and a 57 per cent decline in staff turnover, prompting more knowledge-based firms to follow suit. The change is most viable where AI tools automate repetitive tasks.

For lower-paid sectors – retail, logistics, care – automation is increasing output expectations rather than reducing hours. The shorter week remains a professional-class privilege, not a universal reform.

9. Autonomous vehicles expand, legal frameworks lag
Companies such as Waymo and Baidu are scaling their robotaxi services across major US and Chinese cities. Early data indicate lower accident rates than those of human drivers, but the legal responsibility remains unresolved.

When collisions occur, courts must decide whether liability rests with the passenger, the manufacturer, or the software provider. The technology is progressing faster than tort law, creating a gap between capability and accountability.

10. Cyber-conflict becomes a routine feature of geopolitics
Both Google Cloud’s Cybersecurity Forecast 2026 and NATO Review analysis predict further growth in AI-enabled hacking, ransomware, and disinformation. These attacks are now embedded in state competition, rather than a criminal subculture.

A significant infrastructure outage causing civilian casualties would likely trigger overt military retaliation, formalising ‘cyber-deterrence’ as doctrine. Cyberwarfare is no longer a secondary theatre; it has become the default background condition of geopolitics.

Conclusion
2026 will not upend the world; it will reveal how unevenly it has already changed. AI will become an operational assumption rather than a novelty. Electricity grids and regulatory systems will begin to reveal their limitations.

The economic divide will widen between those who can use these tools to multiply their output and those whose output becomes the data set.

The novelty phase is over; administration of the future begins.

My take

1. AI agents become standard business infrastructure
I fear so. Take it from one who knows, middle-tier administrative labour – refusing to acknowledge your looming obsolescence makes the inevitable hard landing harder than it might otherwise have been if you’d been willing to bear just a little more reality.

2. Electricity demand from AI begins to strain grids
Seems plausible, but what if an anti-AI movement – or, at least, an anti-AI-hogging-all-the-power movement – springs up? I suspect it might, especially when the job losses start coming thick and fast.

3. Global growth remains weak despite productivity gains
Here’s what came back when I asked for further information from my digital overlord assistant:

Global GDP will grow by about 2.4 per cent in 2026, a level consistent with stagnation rather than recovery. Automation is improving efficiency inside firms, but higher interest rates, ageing populations and geopolitical friction are cancelling out the gains. Output per worker is rising, while consumption per household remains flat, and most of the benefits accrue to companies that own the algorithms. The result is a world that looks technologically advanced but economically tired.

I fear “technologically advanced but economically tired” will increasingly describe many developed nations.

4. The EU AI Act becomes the global template for regulation
Yeah, nah. (See deglobalisation comment below.)

5. Renewables keep expanding, but mineral supply becomes the choke point
Not my field of expertise, but that sounds plausible, as well as good news for Australia. However, it’s possible the AIs I consulted have been trained on a vast amount of renewable-energy-championing academic studies and media reports. As a result, they may be discounting the possibility of Net Zero targets being abandoned once the importance of power-gobbling AI is made brutally apparent.

6. Defence spending turbocharges AI and robotics research
Take it to the bank.

7. Supply chains regionalise into competing trade zones
Yes, deglobalisation is only going to accelerate from here on in.

8. Shorter workweeks spread selectively among high-skill sectors
Maybe. It’s a more comforting scenario than the alternative, which is lots more people having non-existent work weeks.

9. Autonomous vehicles expand, legal frameworks lag
Driverless vehicles have operated legally in the US since 2020 and in China since 2022. Take it from a former journalist – if there had been any Blood Flows Red on the Highway incidents involving Waymos or Apollo Gos in the last half-decade, you would have heard about it.

Societies now face a difficult choice. Do they give their legal blessing to robotaxis and robotrucks, thereby throwing vast numbers of blue-collar and downwardly mobile middle-class men on the scrapheap? Or do they continue to bear the significant costs of human-piloted vehicles?

10. Cyber-conflict becomes a routine feature of geopolitics
Hasn’t that long been the case? Well, yes, it has. However, my trusty info-manservant elucidated before we parted company for the night.

AI is making cyber operations faster, cheaper and harder to defend against — generating floods of synthetic phishing, autonomous malware and deepfake propaganda. At the same time, improved attribution tools, many of which are also AI-based, are giving states greater confidence in identifying and punishing attackers. This feedback loop — automation raising the volume, detection raising the stakes — pushes cyber conflict from deniable espionage into open confrontation. The technology removes both restraint and plausible deniability.

If this assessment is even half-true, 2026 could make 2020 look like a placid idyll.

What if this time it really is different?
As Freddie de Boer periodically notes, humans are prone to assuming the technological breakthroughs of their era will have civilisation-upending effects.

Freddie, who is courageously correct about most things, makes a fair point.

For instance, many Boomers and some Gen Xers grew up believing there was a decidedly non-zero chance of nuclear war. A war to end all wars that would cause a mass extinction event for humanity. Sting even wrote a cringe song about it.

But a boy falsely crying wolf many times in the past doesn’t rule out the possibility that a canid will eventually turn up to unleash a “species-level turning point”.

By the end of next year, it should be clear whether or not the wolf cometh.

*Apparently, it’s only 23 NATO members out of 32 who’ve met the spending target, as of mid-2025. (Said the AI I got to check the other AI’s copy.)