When Apple launched the iMac, in 1998, a TV ad for its colorful debut promised “three easy steps to the Internet”: “Step one: plug in. Step two: get connected. Step three…[incredulous chuckle]…there’s no step three. There’s no step three!”

That iMac intro made a compelling claim, in that the notion of getting Internet access simply by connecting to watts and bandwidth was a step-function change; we might even say that it started a ball rolling, toward today’s general expectation of abundant capability on demand. Everything digital in 1998, everything else via Amazon Prime in 2005? (Yes, it’s been that long.)

That expectation, though, now seems to be verging on challenge—and even threat—with climate change a major and growing factor.

If that Apple ad seems to make too much of nothing much, recall that prior to Windows 95 it was a science project to get on the Internet at all. TCP/IP support required technical fluency and involved a multi-step process – and we wouldn’t start using the term, “the cloud,” until roughly a decade later, so Apple wasn’t a laggard. People wanted to be told that “the future” was now just a matter of having the right outlets, close enough to your desk for a cable to reach them.

Today, the once-novel promises made by iMac & Amazon have become pervasive: much of our everyday life now relies on being confident that if you can connect to it, you can have it, 24×7 at a flat-rate price. That world view is now getting blurry.

What brings this to mind, just now, is an upcoming panel discussion at this Thursday’s “Phorum” conference in Philadelphia. I’ll be sharing (what I hope will be) useful data and recommendations on a growing complexity of getting stuff done.

We’ll be specifically exploring, according to the conference agenda,

AI, Data Centers, Chips & The Grid: The rapid growth of AI is raising big questions about energy demand, sustainability, and grid stability. From data centers to everyday enterprise use, how are AI systems impacting our environment and energy infrastructure? This session will explore the challenges of balancing innovation with responsibility.

I’ll be on the panel as “the chips guy,” which role I get from having spent years covering microchip and platform evolution during a time when the annual Microprocessor Forum [PDF] was a major tech-world event. There have been a few decades, since then, during which Apple (in particular) seemed to be proving that users could stop asking “what chip does it use?” – but in the current era of hyperscale AI, access to the right chips in huge quantities is becoming a geopolitical (not merely commercial or technical) concern.

Further, it’s critical to appreciate that this is not just about “the chips,” as if they emerge from a Star Trek replicator ready to plug in. It’s more complicated than that:

The supply of finished products such as chips, computers and electrical boards is dominated by Asian economies…[but] the minerals used in AI components span the periodic table… Minerals such as cobalt, lithium, aluminium, nickel, and copper (collectively, ‘the CLANCs’) are crucial for energy storage, battery materials and electrical wiring. China is the dominant buyer, purifier and processor of many of these minerals, but almost all of the actual mining happens in countries such as Chile (supplier of 20% of the world’s raw copper), the Democratic Republic of Congo (73% of raw cobalt) and Guinea (70% of aluminium ores). As a group, emerging and developing economies supply close to two-thirds (62%) of the world's raw rare earth elements and minerals.

As this reality sinks in, the current concern about AI’s real-world ingredients is echoing conversations that we were having six years ago – when pandemic chaos was exposing the complexity and brittleness of supply chains, then newly seen to be over-optimized for minimum cost when things were working as planned. “A humble cup of coffee requires 29 firms to collaborate across 18 countries,” observed The Economist in May 2020 – and not only cups of coffee, but also doses of vaccine, turned out to be produced by planet-encircling networks rather than point-source factories:

GlaxoSmithKline uses an extract from the soap-bark tree to produce a vaccine-enhancing ingredient called an adjuvant; for some of the vaccines now in development, the adjuvant may enhance their effectiveness or make a certain quantity stretch to more doses… The bark is harvested in Peru, Chile and Bolivia during the summer months of the southern hemisphere.

So, no, you couldn’t just order up a double batch on Amazon Prime to deal with pandemic demand – and today, those pandemic learnings about the value of supply-chain resilience had better be getting repurposed to deal with climate impacts. As noted in another report from The Economist, at the end of 2024:

A billion-dollar extreme weather event now occurs every three weeks. Four decades ago, one occurred every four months. These extreme weather events can take many forms, from floods to hurricanes to droughts. Yet, a consistent factor for all is that each makes global supply chains less reliable.

More disruptions mean longer delivery times, higher costs and lower output. The economic risks of climate change to global trade are predicted to sit around US$81bn. This rises to at least US$122bn when factoring in economic activity linked to industry output and consumption. Such losses are higher than the individual GDP of Slovakia, Ecuador or Kenya.

This matters, even to people who think that climate change is an abstract concern, as warned by the Trax Group in December 2025:

Production schedules will become increasingly unpredictable as manufacturers adjust output in response to component availability. Freight patterns will shift abruptly, with more small-batch and expedited shipments replacing steady, predictable flows. Warehousing operations will experience greater volatility as inventory levels swing between surges and sudden slowdowns. Distribution networks optimized for steady throughput may struggle with these irregular patterns. Procurement teams will face extended negotiations as suppliers manage allocation decisions across competing customers. Cost pressures will accelerate…

Whether we’re talking about chips that have been strategically developed into platforms, or about the power grids that are being challenged to make the resulting wheels go ’round, complexities are coming back and even emerging anew. As it turns out, “step three” was ahead of us all along – and it’s a big one.