Xeal Launches Laitent to Turn Idle EV Charging Capacity Into Distributed AI Computer
Xeal has launched Laitent, a new distributed edge-computing network designed to transform unused electrical capacity at EV charging sites into infrastructure for artificial intelligence inference. The company says the platform can provide AI compute closer to end users while making use of electrical infrastructure that has already been permitted, installed and connected to the power grid.
The launch represents an emerging convergence between electric vehicle charging infrastructure, AI computing, distributed energy management and edge computing. Rather than building entirely new data centers and waiting years for additional grid interconnections, Xeal intends to use the excess electrical capacity already available at its EV charging locations to deploy GPU-based computing.
According to Xeal, its existing network represents one of the five largest Level 2 EV charging footprints in the United States. The company says its infrastructure spans more than 1,600 properties and includes more than 200 megawatts of permitted and installed electrical capacity. Laitent is designed to tap into the portion of that capacity that is not being used by EV charging operations.
Xeal plans to deploy more than 100,000 NVIDIA GPUs alongside EV charging infrastructure as the Laitent network expands. The company has also announced partnerships covering AI infrastructure orchestration, telecommunications connectivity, property management and inference services.
The first Laitent Pod is expected to become operational with JVM Realty by the end of 2026, providing an initial demonstration of the concept in a real-world property environment.
Addressing the Growing Demand for AI Inference
The rapid expansion of generative AI and other AI applications has created growing demand for computing infrastructure. While large centralized data centers remain an important part of the AI ecosystem, many applications require computing resources to be located closer to users and devices.
AI inference—the process of running a trained AI model to generate an output—can require substantial computing capacity. Applications such as autonomous systems, smart buildings, robotics, industrial automation, connected vehicles and real-time analytics can benefit from lower latency between the user and the computing infrastructure.
Traditional data-center development, however, can require significant amounts of time because developers must secure suitable sites, obtain permits, establish power connections, construct facilities and install cooling and networking systems.
Xeal argues that electrical grid interconnection is one of the most challenging parts of expanding computing capacity.
Laitent takes a different approach by using electrical infrastructure that is already available at EV charging properties.
“ If you need inference compute today and can’t wait for a new data center to come online, Laitent can accelerate your timeline from years to months,” said Nikhil Bharadwaj, co-founder and CEO of Xeal.
According to Bharadwaj, the platform can take advantage of existing grid-connected infrastructure and place computing resources in locations where they can provide low-latency services. The company also emphasizes that Laitent Pods do not require a conventional water connection for cooling.
Turning Unused EV Charging Capacity Into Compute Capacity
The central idea behind Laitent is based on the difference between the maximum electrical capacity for which an EV charging site is permitted and the amount of power the site typically consumes.
Xeal says EV charging sites are generally permitted for a maximum charging load, but actual charging activity can remain well below that maximum. The company estimates that its charging sites may operate at less than 10% of their permitted capacity during typical periods.
That leaves substantial electrical headroom.
Laitent is designed to use that unused capacity for computing without taking away the electrical resources required for EV charging. Xeal describes the available capacity as a largely untapped infrastructure resource that can be dynamically allocated between charging and computing.
The company’s power orchestration software manages this process. The system can reportedly allocate power at different levels, ranging from a portion of an individual GPU to distributed computing capacity spread across multiple sites within a metropolitan area.
This approach could allow computing workloads to respond dynamically to available electrical capacity.
Instead of treating EV charging infrastructure and computing infrastructure as completely separate assets, Laitent combines them at the same location.
NVIDIA GPUs at the Metro Edge
Xeal is working with NVIDIA to deploy accelerated computing hardware throughout the Laitent network. The company says it plans to deploy more than 100,000 NVIDIA GPUs over time.
Individual Laitent Pods can support up to 48 NVIDIA Hopper or Blackwell Ultra GPUs, providing substantial accelerated-computing capacity within a compact physical footprint.
NVIDIA has also highlighted the potential of combining accelerated computing with Xeal’s existing EV charging network.
“As AI adoption grows, infrastructure must become more distributed, flexible and energy efficient—placing the right workloads where power and capacity are available,” said Marc Spieler, Senior Managing Director for the Global Energy Industry at NVIDIA.
The concept reflects a broader shift toward distributed computing architectures. Instead of sending every AI workload to a distant hyperscale data center, some workloads can potentially be processed at locations closer to customers.
Xeal calls these locations the “Metro Edge.”
The company says Laitent Pods can deliver sub-20-millisecond latency, depending on the application and network conditions. Lower latency can be particularly valuable for applications that require rapid responses.
Compact Compute Pods Designed for Existing Properties
One of Laitent’s notable characteristics is its physical design.
Xeal says a Laitent Pod is approximately the size of a single parking space. Each unit can accommodate up to 48 GPUs and is designed to operate independently without requiring a dedicated water connection.
The pods are also designed for outdoor deployment. Xeal says they use a NEMA 4 enclosure, providing protection suitable for outdoor environments.
Noise is another consideration for deploying computing infrastructure at commercial and residential properties. Xeal says Laitent Pods operate at less than 65 decibels, which the company compares with the approximate noise level of a washing machine.
The compact form factor could allow the infrastructure to be deployed at locations where constructing a conventional data center would not be practical.
According to Xeal, installation can take place within hours, potentially allowing properties to begin generating computing-related revenue without undergoing a large-scale construction project.
Partnerships Build the Laitent Infrastructure Ecosystem
Xeal is developing Laitent through a network of technology and infrastructure partners.
One of the company’s key partnerships is with Rafay Systems, which provides infrastructure orchestration technology. Rafay’s platform is intended to help manage distributed GPU resources across multiple Laitent locations.
Haseeb Budhani, CEO of Rafay Systems, said distributed EV charging locations could be transformed into a production-oriented computing network through the combination of orchestration and GPU infrastructure.
The company is also working with Spectrum Business to provide dedicated enterprise-grade fiber connectivity for inference workloads. Reliable connectivity is critical for distributed computing because applications depend not only on local GPU capacity but also on high-performance communications between users, applications and infrastructure.
Xeal has additionally secured partnerships with dozens of real estate companies and property managers.
A Tier 1 inference provider is expected to use up to 5 MW of Laitent compute capacity, providing an early commercial application for the network.
Opportunity for Property Owners
Laitent is also being positioned as a new revenue opportunity for property owners.
Property owners hosting Laitent Pods can potentially receive fixed rental payments or ancillary revenue from the computing infrastructure. Xeal says the model can require little or no upfront investment from participating property owners.
The company estimates that a Laitent installation could add as much as $1 million in property value, depending on the property and deployment.
Steve Boyack, Chief Operating Officer of JVM Realty, said the arrangement could create an additional revenue stream from infrastructure already located within communities.
JVM Realty has worked with Xeal since 2023 and is expected to host the first Laitent Pod.
For property owners, the model creates the possibility of monetizing underutilized electrical capacity without replacing the primary function of the site.
Potential Benefits for EV Charging Customers
Xeal also suggests that the model could potentially produce benefits for EV charging customers.
Because Laitent uses capacity that may otherwise remain unused, the company says the additional economic activity could potentially contribute to savings associated with EV charging.
The system is designed to prioritize charging requirements while allocating available electrical capacity to computing workloads.
This dynamic approach is important because EV charging demand can fluctuate substantially. Charging sites may experience higher utilization at certain times and much lower utilization at others.
A flexible power-management system could therefore shift resources based on real-time requirements.
Expanding Beyond 200 Megawatts
The initial Laitent network is based on more than 200 MW of permitted and installed electrical capacity across Xeal’s existing footprint.
However, the company sees considerably more potential.
Xeal says it intends to unlock more than 1 gigawatt of existing electrical headroom across real estate and EV charging deployments.
That expansion would significantly increase the amount of computing capacity that could potentially be deployed without relying exclusively on newly constructed grid connections.
Xeal’s experience in EV infrastructure is central to this strategy. The company says it has built more than 1,600 EV charging sites during the past seven years and has experience deploying energy infrastructure across more than 160 metropolitan areas in the United States.
That existing footprint gives the company access to locations, electrical infrastructure and relationships that could support the expansion of Laitent.
A New Intersection Between EVs and AI Infrastructure
The Laitent launch illustrates how EV infrastructure may increasingly serve purposes beyond vehicle charging.
EV charging networks already involve significant investments in electrical distribution, grid connections, communications systems, property access and power-management technology. When charging demand does not consume the full permitted capacity, some of that infrastructure can potentially support additional applications.
AI computing provides one such application.
The combination could become particularly relevant as both EV adoption and AI workloads continue to expand. EV charging networks require increasingly sophisticated energy management, while AI applications require additional distributed computing capacity.
Laitent attempts to address both trends through a single infrastructure model.
The first commercial deployment with JVM Realty will provide an early test of whether distributed GPU infrastructure can operate effectively alongside EV charging equipment at property-level locations.
If the model scales as planned, Xeal’s existing charging footprint could become part of a broader distributed computing network rather than functioning solely as an EV energy service.
For AI companies and enterprises, the proposition is access to computing resources closer to customers without waiting for conventional data-center construction. For property owners, it creates another potential revenue stream. For Xeal, it creates a new use for electrical capacity that is already connected to the grid.
With plans to expand from more than 200 MW of existing infrastructure toward more than 1 GW of potential headroom, Xeal is positioning Laitent as a new category of energy-aware edge AI infrastructure, connecting EV charging, distributed power management, GPU computing and metropolitan connectivity in a single network.
Source Link:https://www.businesswire.com/
