Accelerating AI development through cost effective green data center solutions.
We’re a new type of compute provider supporting AI model training.
.webp)
.webp)
We’re helping companies train AI at the absolutely lowest cost—enabling you to save more.
We also do it in a way that supports the planet by bringing compute to where energy happens.
Models can pause at checkpoints & latency isn’t a major factor, so why pay to train your AI model in an ultra luxury, low latency, 99.99999% uptime hyperscale data center?
- Minimal redundancy
- Low cost energy
- Operate servers dynamically based on energy prices
We’re a team of specialists that wants to make AI training more accessible and environmental
We’re combining stranded renewable energy with high-performance computing to fundamentally reduce training costs.
Unlike other cloud providers, we only focus on AI model training—not inference. This means we can take advantage of the unique properties of AI training and optimize accordingly.
Our systems are able to flex on speed, cost, and model size, whereas cloud solutions are only expensive/fast. We know one of the biggest barriers to model training is cost, so that’s what we’re optimizing.
Cloud offerings
BuildAI
FAQs
On the buildout cost we lower cost by using less expensive chips, rapidly deploying our modular systems, and requiring less redundancy because workloads can be paused. For lowering ongoing cost, we use remote low cost renewable energy, shut down or raise prices to offset spiking power costs, and run our modular data centers remotely & autonomously.
We have strong relationships with GPU providers (NVIDIA, AMD), with system integrators (ICC, CDW), and with leasing firms. We also plan to co-locate with data center providers to meet demand that may outpace our short-term supply.
We only focus on optimizing our infrastructure for training and fine-tuning machine learning models, not inference. Competitors like CoreWeave, Lambda, and Crusoe require millions of dollars worth of backup power generation infrastructure to locate their data centers with remote oil fields and wind farms. We're able to shut down and place our data centers closer to the existing power infrastructure to optimize energy consumption and price.
To get a much better price! By making a 5–10% sacrifice on training speed, massive energy savings can be unlocked by not operating during high energy priced periods. Our customers are happy to make this tradeoff to conserve their runway, train larger/more models, and to ultimately do it in a way that’s better for the planet
We shut down during daily set time blocks (e.g. 5-8pm “shoulder period” when solar supply goes offline and demand jumps as people come home from work). We work with customers on checkpointing / saving their models ahead of these periods, so we can resume training, without replicating any work, when energy prices fall. Over time, we will progress to a more dynamic AI-enabled model where we can take advantage of turning off our servers throughout high-cost periods within a day. Customers who would prefer speed > cost can pay a premium to keep their training workloads running during these periods. At the platform level we will have orchestration software to manage the systems and the workloads to be able to shut the systems on and off.
We locate our modular centers near commercial distribution substations for aggregating and stepping down voltages for industrial and commercial use. In the long-term, we’re planning to cut energy deals with renewable developers to lock in even more favorable pricing.