Some Thoughts Prompted by News of Super El Niño

---

I was reading the news this morning about the massive 2026 "Super El Niño" currently brewing in the Pacific, and it kinda trigger a few interesting thoughts.

The "cloud" that we have shifted over >90% of enterprise workload to is made of concrete, copper, water, and an absolutely staggering amount of electricity. And as we barrel into the 2026 that is more and more defined by AI, the physical constraints of planet Earth are crashing headfirst into our algorithmic ambitions.

Here is the reality check few people are talking about.

The Super El Niño Stress Test

We are currently watching the formation of a massive Super El Niño. This isn’t just a weather alert on the evening news; it is a global infrastructure stress test.

El Niño acts as a massive thermal engine. For the world's rising AI ambition, it means two things: extreme heatwaves that spike consumer grid demand and severe droughts that dry up the hydroelectric dams powering our "green" data centers. The drought will also mean bad news to chip fabs which consumes crazy amount of water daily.

The Geopolitics of Megawatts

The country that can deliver the most power at scale wins the AI race. Period. But how the two biggest global players are solving this gridlock couldn't be more different.

China faces a geographic mismatch: the tech demand is in the East, but the abundant, cheap renewable energy is in the West. Their solution is a masterclass in brute-force, strategic engineering: the "Eastern Data, Western Computing" initiative.

Instead of trying to shove more electricity into the crowded grids of Beijing or Shenzhen, they are building massive data centers in the western deserts. Thanks to their dominance in Ultra-High Voltage transmission, they simply route the latency-insensitive AI training workloads out to where the cheap power lives. They’ve turned their entire geography into one optimized motherboard.

The U.S. grid is famously fragmented, and waiting ten years for a public transmission line to be approved isn't an option for Big Tech. So, American companies are going "behind the meter."

They are literally buying data centers physically attached to nuclear power plants to siphon the atomic energy before it ever touches the public grid. Microsoft is funding the resurrection of Three Mile Island. Tech giants are throwing billions at Small Modular Reactors (SMRs) and even building private natural gas plants just to keep the GPUs humming.

If the public grid can't handle it, Silicon Valley is simply buying its own.

The Great Tech Backlash of 2026

But increasingly, I am thinking more about a harsh socioeconomic reality. AI is creating a massive "Energy-Labor Disconnect."

Hyperscale data centers are essentially giant, power-hungry dark boxes. They consume the electricity of a medium-sized manufacturing town, but they employ a fraction of the people. While a traditional factory might employ thousands and build a robust local supply chain, an AI data center employs maybe 50 security guards and IT technicians.

Worse, they are pricing traditional manufacturing out of existence. AI companies have virtually unlimited capital and an inelastic demand for power. They will outbid a local steel mill or chemical plant for electricity every single time.

The takeaway? The macroeconomic benefits of AI are global, the profits go to the selective few tech oligarchs, but the physical costs—strained grids, soaring local utility bills, and hollowed-out manufacturing jobs—are hyper-local.

Right now, the average person is getting squeezed from every possible direction. The Super El Niño is wrecking crop yields in the global south, sending the price of a basic grocery run through the roof. The ongoing geopolitical powder keg in the Middle East is doing the exact same thing to fuel prices.

And at the exact moment when people are feeling the financial walls close in, they open their phones to a polarized, culture-war-torn internet that feels increasingly alien. Recently, news of job displacement by AI has spread from tech native firms to broader industry. Technology is no longer viewed as the great equalizer. It is increasingly being treated as an occupying force.

I recently also read that anti-AI groups have realized that our neural networks are only as good as the information they consume, so they are poisoning the well. It is digital guerrilla warfare.  

• Text Poisoning: Communities on platforms like Reddit are intentionally flooding forums with contradictory garbage, hidden "jailbreak" text, and fake consensus specifically designed to break AI scrapers.

• Visual Sabotage: Artists are applying digital cloaks like Nightshade and Glaze to their portfolios—invisible pixel alterations that teach AI models that a picture of a dog is actually a handbag, quietly corrupting the model's logic from the inside out.  

The takeaway? Technology is not the fabric that society builds itself, society is built by people, family, community united by trust.