Testimony on AI, energy, and distributed compute
I’m Greg Osuri, Founder and CEO of Overclock Labs, the team behind the Akash Network. I’m an open-source developer and cloud infrastructure engineer.
Akash is a decentralized cloud platform used by hundreds of thousands of developers. By tapping into idle servers worldwide, it delivers compute power at a fraction of the cost of traditional cloud providers.
Our open network—supported by hundreds of global contributors—empowers developers to build and deploy AI and other applications without relying on Big Tech.
Artificial Intelligence isn’t just a breakthrough—it’s becoming the substrate of civilization.
Like electricity in the 20th century and the internet in the 21st, AI is the next foundational layer. It’s not a feature—it’s the foundation. From curing cancer to defending the homeland, AI will enhance everything we do.
There is no domain that AI won’t transform.
Yet we’re only at the starting line. Superintelligence is coming—an invention that could define the next 500 years. We will either master it—or be mastered by it.
America is leading—for now. But we’re in a sprint, and China is gaining ground. Their DeepSeek R1 model, trained for just $5 million, delivered results that rival OpenAI’s. They’re scaling faster, investing more, and moving with purpose.
We cannot afford to fall behind. The future of AI must be built on freedom, transparency, and trust—not surveillance and control. That is America’s edge. And we must defend it.
To lead in AI, we must confront the energy crisis it’s creating.
According to the Department of Energy, U.S. data centers consumed 176 terawatt-hours of electricity in 2023—4.4% of the nation’s total. By 2028, that number could triple. At the high end, AI infrastructure could consume 12% of all U.S. electricity—an enormous burden on our grid.
Here’s the problem: our national power generation isn’t doubling every two years, but the energy needed to train and run AI is.
AI demand is on an exponential curve. Our response can’t be linear.
Gartner warns that by 2026, energy demands may exceed utility capacity, leaving 40% of AI data centers without sufficient power.
Meanwhile, China is racing ahead—outbuilding us, outpowering us, and scaling fast. DeepSeek R1 is a glimpse of how quickly the lead can slip.
Without bold action, this energy bottleneck won’t just slow AI innovation—it could cost us global leadership.
To maintain momentum, we must use every tool available—including fossil fuels as a temporary bridge. But that’s just a stopgap, not a long-term plan.
Fortunately, we don’t need to wait for massive new data centers or power plants. Leading AI labs have shown that training doesn’t need to happen in hyperscale facilities. It can be distributed across smaller compute clusters.
That opens the door to a decentralized grid, powered by solar, renewables, and local energy sources from homes, campuses, and communities.
This model isn’t just more efficient—it’s more secure. Decentralization builds resilience. No single point of failure. No single target. That matters, especially in a world of cyber threats and geopolitical risk.
We need distributed infrastructure to meet AI’s energy demands today, while we scale sustainable energy for tomorrow.
Yes, we must ease barriers for advanced nuclear to ensure long-term energy stability. But decentralization is how we scale now.
That’s exactly what Akash Network enables: a decentralized AI cloud—open, efficient, and built for the future.
AI will shape the future of every industry, every economy, and every nation. The question is: who will shape the infrastructure behind it?
We cannot rely solely on Big Tech or centralized power. We must build a distributed, resilient, and inclusive AI infrastructure—rooted in American ingenuity, powered by clean energy, and open to all.
Congress has a chance to lead:
-
Incentivize decentralized cloud infrastructure
-
Fund federal R&D for distributed computing
-
Open federal lands to power edge innovation
-
Accelerate clean, flexible energy solutions
Let’s ensure America doesn’t just participate in the AI revolution—we lead it, shape it, and share it.
Let’s unleash AI’s full promise—not just through code, but through the infrastructure that defines our future.
Thank you. I look forward to your questions.