Did you know training a single AI model could emit as much carbon as 108 cars in a year? and that's not even the worst part.
Every query, every generated response, every fraction of a second AI models operate—they continue drawing power.
A study from the University of Massachusetts Amherst found that training a single large-scale AI model can emit 284,000 kg of CO₂—equivalent to nearly 125 round-trip flights from New York to Beijing.
Once deployed, these models require even more energy for inference, as millions of users interact with them daily.
Big Tech companies, driven by profit and competitive advantage, continue scaling AI to unsustainable levels.
Instead of optimizing for efficiency, the race is to build bigger, more complex models—regardless of environmental cost.
The problem isn’t AI itself—it’s the centralized infrastructure that supports it.
AI, as it stands today, is an ecological time bomb fueled by unchecked expansion, and unless we rethink how intelligence is trained and deployed, the environmental burden will only grow.
I’ve been working on a piece that examines Nigeria’s tech and startup ecosystem.
With a focus on understanding the market, demographics, regulations, and products that sell.
Case spotlights on @ribhfinance, @Nectar_finance, and @airbillspay to ground the analysis in real-world examples of products that trailblaze.
And extract lessons to learn from Solana products and founders for success in this harsh market.
@eldivine The question is are you willing to pay them the talents their global worth? A startup offering 200k/m wants you to build openAi’s architecture from the ground