Quenching AI's Thirst for Power, Cleanly
There's a growing narrative that AI is an environmental disaster. You've seen the headlines: a single AI query uses as much energy as a lightbulb left on for an hour, or a training run uses more power than a small city.
These articles aren't wrong, but they are reductive.
They're framing energy consumption as a moral failure, a sign that AI is inherently "bad." This is a simplistic take on a complex topic. It's the same logic as saying, "Airplanes are bad because they use jet fuel," or "Hospitals are bad because they have a massive power bill."
This line of thinking misses the entire point. Energy consumption, by itself, is not inherently bad. We consume energy all the time for things we've decided are worthwhile.
Getting on an airplane is a massive energy expenditure, but we do it because we've decided that seeing family, closing a business deal, or experiencing another culture is worth the tradeoff. Having fresh food delivered to our city on diesel trucks is a huge carbon cost, but we accept it because it's part of a system that feeds millions.
Energy use is only "bad" when it's wasted. The real question isn't "Does AI use energy?" The real questions are:
"What is the return on that energy investment?"
"Is the tradeoff worth the cost?"
If we're spending massive amounts of energy to get AI to write funnier limericks or generate weirder cat photos, then yes, that's a waste. I confess, I’ve indulged every now and then, but I resisted the temptation to generate a cool image for this post. It wasn’t worth the carbon cost, nor the comments I’d inevitably get.
Sometimes that waste is delightful, but it's a cost with a benefit that’s harder to defend.
But what if that same energy cost is being invested to:
Cure diseases by modeling protein folds in seconds instead of decades?
Help educate millions of children in their native language with a personal tutor?
Propel businesses and organizations that are lifting people out of poverty?
When you frame it that way, it's not a "cost" anymore. It's an investment.
The most exciting part of this conversation is that it's not a zero-sum game. The reductive "AI is bad" narrative misses the ultimate irony:
AI is one of our best hopes for solving our energy problems.
We are getting incredibly close to the point where AI can help us discover new materials, design hyper-efficient solar panels, or manage power grids with a precision we've never had. Some of the most ambitious plans, like those from Google and xAI, even involve launching future solar-powered AI data centers into space, a move that could pay back its own carbon-launch debt in just 5-10 years. AI can, and likely will, help us develop cleaner, cheaper, and more abundant energy.
The debate shouldn't be "AI vs. the Environment." The real debate is how we ensure we're investing this powerful new tool in a way that lifts all of humanity, creating a return that is well worth the price.
Ready to build your AI strategy?
AI must be trained to follow your organization's mission and values, including the complex tradeoffs you navigate every day.
At GapJump AI, we help make sure that your AI strategy and deployment is deeply grounded in your values, not the default values reflected in out-of-the-box AI models. If you're ready to start that conversation, let us know.
