A Swedish wave energy company is partnering with NVIDIA to develop artificial intelligence tools that could help renewable energy systems keep pace with data centers' explosive power demands. Eco Wave Power, which converts ocean waves into electricity using nearshore technology, announced it has joined the NVIDIA Inception program to integrate AI into its wave energy operations and explore new applications for powering data center infrastructure.

Why Are Data Centers Suddenly Interested in Wave Energy?

The connection between AI and renewable energy has become urgent. Data centers are projected to consume up to 12 percent of total U.S. electricity by 2028, according to the Lawrence Berkeley National Laboratory. This explosive growth is driven largely by artificial intelligence systems that require enormous amounts of computing power. Wave energy, which generates electricity directly from ocean waves near coastal cities and ports, offers a location-based renewable solution that could be deployed close to where data centers actually operate.

Eco Wave Power operates Israel's first grid-connected wave energy station and recently launched the first onshore wave energy pilot at the Port of Los Angeles in collaboration with Shell Marine Renewable Energy. The company's global project pipeline includes 404.7 megawatts of capacity, positioning it as a potential contributor to meeting rising electricity demand from AI infrastructure.

What Specific AI Tools Will Eco Wave Power Develop?

Through the NVIDIA Inception program, Eco Wave Power U.S. will gain access to developer tools, technical resources, training, and ecosystem support to accelerate AI-driven applications. The company is evaluating integration of artificial intelligence technologies across multiple aspects of its operations, including real-time optimization of wave energy generation, predictive maintenance and infrastructure monitoring, digital twin modeling of wave energy power stations, AI-driven analysis of ocean and weather data, and intelligent energy management for coastal and port infrastructure.

The U.S. subsidiary will serve as the central hub for AI-related initiatives across Eco Wave Power's global portfolio. The company is advancing research and development collaborations with Florida-based universities and other technology and infrastructure stakeholders, focusing on integrating AI layers into wave energy infrastructure, including digital twin technologies, predictive maintenance systems, and operational optimization.

"Infrastructure supporting artificial intelligence requires massive amounts of electricity, and we believe renewable energy generation must evolve alongside it," said Inna Braverman, Founder and Chief Executive Officer of Eco Wave Power. "By joining the NVIDIA Inception program through our U.S. subsidiary, we aim to accelerate the development of intelligent energy management capabilities and explore opportunities at the intersection of renewable energy infrastructure and AI-driven energy demand."

Inna Braverman, Founder and Chief Executive Officer of Eco Wave Power

How Can Data Center Operators Reduce Energy Waste More Efficiently?

Beyond renewable energy generation, researchers are developing faster methods to predict and manage power consumption. Scientists from MIT and the MIT-IBM Watson AI Lab created a rapid prediction tool called "EnergAIzer" that estimates how much power will be consumed by running a particular AI workload on a specific processor or graphics processing unit (GPU) accelerator chip. This tool produces reliable power estimates in seconds, unlike traditional modeling techniques that can take hours or even days to yield results.

• Speed Advantage: The EnergAIzer method generates accurate power estimates in seconds rather than hours or days, enabling data center operators to make rapid resource allocation decisions.
• Broad Applicability: The prediction tool can be applied to a wide range of hardware configurations, including emerging GPU designs that haven't been deployed yet, helping operators plan for future infrastructure needs.
• Practical Decision-Making: Data center operators can use these estimates to effectively allocate limited resources across multiple AI models and processors, improving overall energy efficiency and reducing wasted power.

"The AI sustainability challenge is a pressing question we have to answer. Because our estimation method is fast, convenient, and provides direct feedback, we hope it makes algorithm developers and data center operators more likely to think about reducing energy consumption," explained Kyungmi Lee, an MIT postdoc and lead author of the research.

Kyungmi Lee, MIT Postdoc, MIT-IBM Watson AI Lab

The MIT researchers developed this faster approach by using less-detailed information that could be estimated more quickly, rather than breaking down workloads into individual steps and emulating how each module inside a GPU is utilized one step at a time. Traditional emulation methods are impractical for operators who need to compare different algorithms or configurations to find the most energy-efficient approach, since a single emulation could take days to complete.

What Are Governments Doing to Optimize Data Center Power Use?

International efforts are also underway to reduce data center energy consumption. Japan's New Energy and Industrial Technology Development Organization (NEDO) selected Rakuten Mobile and KDDI Corporation for a research and development project aimed at reducing power consumption in mobile data centers. The project, titled "Development of Simultaneous Optimization Technology for Virtualized Base Stations and Computing Infrastructure," aims to reduce data center and radio access network (RAN) power consumption by approximately 40 percent by 2030.

The collaborative initiative spans five key research domains designed to realize communication infrastructure that efficiently processes the immense computational load of AI and virtualized RAN processing while minimizing power consumption. These include virtualized network optimization to transition network equipment between GPU clusters to optical networking, AI-driven RAN optimization to achieve energy savings while maintaining mobile network quality, energy-efficient and compact cooling for computational infrastructure, enhanced security measures, and cooling solutions for ultra-high-performance GPUs in the 3000-watt thermal design power class.

The convergence of renewable energy innovation, faster power prediction tools, and government-backed infrastructure optimization projects reflects a growing recognition that AI's energy demands require coordinated solutions across multiple fronts. As data center electricity consumption continues to accelerate, the combination of nearshore wave energy systems enhanced by AI, rapid power estimation methods, and optimized cooling infrastructure could help ensure that artificial intelligence development remains sustainable.