Bitcoin miners chase AI demand as Nvidia confirms Rubin production

TLDR

• Nvidia announced that its next-generation Rubin AI platform is now in full production.
• The Rubin system includes 72 GPUs and 36 CPUs and supports scaling into larger computing pods.
• Jensen Huang said Rubin improves AI token generation efficiency by around ten times.
• The performance increase comes with only a 1.6 times rise in transistor count.
• Nvidia emphasized the growing speed of AI development and the need for faster hardware.

Nvidia has confirmed that its next-generation AI platform, Vera Rubin, is now in full production and set for release this year, as CEO Jensen Huang unveiled new hardware details at CES 2026 in Las Vegas, highlighting its capabilities for accelerating AI workloads and drawing fresh attention from Bitcoin miners seeking new revenue opportunities.

Nvidia launches Rubin platform with AI-focused design

Nvidia CEO Jensen Huang announced that Rubin is now in full production and aims to increase AI computing capacity by five times. He stated, “Rubin systems improve token generation efficiency by 10 times using our proprietary data format,” which he urged the industry to adopt. This claim was made while emphasizing that the transistor count increased only 1.6 times.

Rubin’s main server will include 72 GPUs and 36 CPUs and can scale to pods with over 1,000 chips. Huang said this setup boosts performance while optimizing energy use, crucial for large-scale AI tasks. The design supports faster model inference rather than training, targeting the expanding deployment phase of AI.

Huang described AI development as a competitive race, requiring rapid investments in hardware, networking, and infrastructure. He said the fastest systems will shape future innovation timelines. The launch reinforces Nvidia’s leadership in AI hardware while preparing for increased market demand.

Bitcoin miners repurpose capacity to meet AI demand

Bitcoin miners have started offering their energy and data-center assets to AI firms, focusing less on mining and more on infrastructure. These firms market rackspace, power contracts, and cooling to AI startups needing scalable hosting. That shift allows miners to earn more stable cash flows when Bitcoin prices fall.

However, increased AI demand is also making prime data-center sites more expensive to secure and operate. Hyperscalers, cloud providers, and AI startups are competing for the same resources. As costs rise, miners with efficient infrastructure may thrive, while smaller players relying only on mining struggle.

Large-scale miners with existing facilities are better positioned to benefit from hosting Rubin and other AI workloads. Their energy contracts and space allow for rapid deployment without major upgrades. This positions them to gain new revenue streams as demand grows.

CoreWeave and hyperscalers to receive Rubin systems

Nvidia stated that cloud service provider CoreWeave will be among the first to deploy Rubin systems in its operations. The company also expects Microsoft, Amazon, Oracle, and Alphabet to adopt the new platform. These companies are likely to use Rubin to improve language model inference and system scaling.

Rubin will use new networking switches featuring co-packaged optics, which link thousands of machines with faster data transfer rates. This feature allows larger clusters to operate with better bandwidth and less power loss. The technology is key to building large, connected systems needed for AI inference.

Nvidia’s Rubin systems will enter data centers during 2026 as production continues and deliveries expand globally. Huang said the launch represents a critical step for the AI industry. Meanwhile, miners shifting to AI infrastructure remain focused on maximizing existing assets as deployment begins.

The post Bitcoin miners chase AI demand as Nvidia confirms Rubin production appeared first on CoinCentral.