Has Renewable Energy Bit the Dust? By James Reed

Jo Nova seems to think so: https://joannenova.com.au/2025/03/tech-giants-quietly-drop-renewables-and-sign-pledge-to-triple-nuclear-power/?utm_source=feedly&utm_medium=rss&utm_campaign=tech-giants-quietly-drop-renewables-and-sign-pledge-to-triple-nuclear-power.The article examines a significant pivot by major tech companies—Google, Amazon, Meta, and others—away from renewable energy toward a commitment to triple global nuclear power by 2050. Here's the breakdown of Jo's article:

The piece opens with a claim that after two decades of championing wind and solar, tech giants have abruptly shifted to nuclear power, signing a pledge facilitated by the World Nuclear Association. This pledge, backed by 14 global banks and firms like Occidental and Japan's IHI Corporation, marks a public endorsement of nuclear energy by non-nuclear industries—a first, per the article.

A key observation is the language shift: terms like "carbon," "low emissions," and "CO₂" are absent from the tech giants' rhetoric. Instead, they tout nuclear as "safe, clean, and firm," emphasizing "round-the-clock" energy and "resilience." Nova interprets this as a retreat from climate activism, possibly to avoid clashing with U.S. President Donald Trump's administration (inaugurated 2025), which has slashed renewable subsidies.

The article highlights recent moves: Microsoft's deal to revive Three Mile Island's nuclear plant, Google's purchase of seven small modular reactors (SMRs), and Amazon's $500 million stake in a nuclear energy firm. This contrasts with Microsoft's "biggest ever" renewable energy agreement less than a year ago, suggesting a rapid U-turn driven by practicality over ideology.

Nova critiques the renewables era as a "bubble" that's burst, arguing tech firms now prioritise raw energy over green dogma. She ties this to the AI boom—data centres powering AI need constant, high-density power, which renewables allegedly can't deliver. The piece ends with a jab at governments lagging behind, urging policy shifts to match this "pragmatic" pivot.

Renewable energy—wind and solar—has been sold as the future, but it's failing, and it can't power the AI revolution.

First, renewables are inherently unreliable. Wind blows when it wants; the sun shines on its schedule. Nova's point about tech giants craving "round-the-clock" energy nails it—AI data centres, like those Google's building, run 24/7, processing exabytes for models like Gemini. A 2024 Nature study pegs AI training energy at 100-1000 megawatt-hours per run—multiply that by thousands of runs globally. South Australia's 2016 blackout, despite heavy wind investment, showed the grid's fragility—hours offline from storm-downed turbines. Nuclear's "firm" output (e.g., SMRs at 300 MW steady) trumps solar's 20 percent capacity factor or wind's 35 percent. AI can't wait for clouds to clear.

Second, energy density is a killer flaw. Solar farms and wind arrays sprawl—think the 650 MW Broken Hill solar plant, which Nova's earlier posts note failed to prevent blackouts in 2024. AI needs gigawatts packed tight—data centres can't sprawl across hectares like farms. Nuclear delivers: a single reactor (e.g., Three Mile Island, 800 MW) outpowers dozens of wind turbines (3 MW each, intermittent).

Third, cost and scale expose renewables' limits. Nova cites Microsoft's renewable flip—billions spent, yet they're resurrecting nuclear. Australia's renewable push—$20 billion on wind/solar since 2010—still sees prices soar ($1,300 more per household since 2022, per opposition stats). AI's appetite is insatiable—Forbes (2025) says data centre demand will hit 35 GW by 2030 globally. South Australia's 56 percent renewable grid struggles at 2 GW peak—tripling that for AI is a pipe dream with windmills. Nuclear's upfront cost (e.g., $5 billion for an SMR cluster) pays off in decades of steady juice; solar's cheap panels fade fast.

Fourth, the AI revolution demands resilience, not fragility. Nova's "energy resilience" buzzword reflects tech's panic—outages crash servers, lose data, kill profits. SA's 2023 storms left renewable-heavy regions dark—battery backups (e.g., Hornsdale's 193 MWh) lasted hours, not days. AI can't gamble on weather; nuclear's 90 percent uptime (vs. wind's 30-40 percent) is the lifeline.

Finally, renewables' green promise is a sham for AI's scale. Tech giants dropping "CO₂" talk, as Nova notes, signals they know it—emissions cuts don't trump survival. Building wind farms emits concrete and steel carbon; solar's rare-earth mining scars landscapes. Nuclear's footprint is tiny—0.4 grams CO2/kWh vs. solar's 20-50 (per IPCC 2014). AI's not sipping power—it's guzzling. Renewables can't fuel that beast without drowning us in turbines.

In short, renewable energy's a busted flush—too flaky, too weak, too costly to drive AI's relentless hunger. Nuclear's the answer tech's chasing, as Nova's giants prove. And let us not forget fossil fuels such as coal as well, only side-lined because of climate change alarmism.