Patent covers technology designed to reduce wasted GPU cycles in large AI clusters and creates potential future licensing and product opportunities

SOAR AI, Inc. (NYSE American: SOAR) (“SOAR AI” or the “Company”) today announced the issuance of U.S. Patent No. 12,775,572, covering technology underlying NeuralSync™, the Company’s proprietary AI infrastructure technology platform.

The patent covers systems and methods for intelligently coordinating accelerator resources across computing systems using programmable data processing units (“DPUs”). The patented architecture includes technology for forecasting the availability of GPUs and other accelerators, sharing that information across systems, and coordinating the movement of data between accelerator memory resources.

The patent represents an important milestone in SOAR AI’s strategy to build proprietary technology alongside its AI data centers and the accelerated cloud infrastructure it is developing and intends to operate. SOAR AI combines AI data center development, NeoCloud infrastructure and deep systems engineering within a single platform—giving the Company the ability not only to deploy accelerated computing infrastructure, but to engineer technology within the infrastructure stack itself.

“We believe the next generation of AI infrastructure companies will be defined by more than access to power and GPUs,” said Chris Ensey, Chief Executive Officer of SOAR AI. “The differentiation will increasingly come from how intelligently that infrastructure is engineered and operated. NeuralSync reflects what makes SOAR AI different: we build the infrastructure, we will operate the compute and we engineer technology within the stack.”

Making AI Infrastructure More Intelligent

The rapid expansion of AI is transforming data centers into increasingly complex computing systems containing thousands of interconnected accelerators. As these environments grow, performance increasingly depends not only on the GPUs themselves, but on how efficiently compute, memory and networking resources work together. When GPUs sit idle waiting on data, those wasted cycles represent expensive compute that produces no useful work.

NeuralSync is designed to reduce those wasted cycles.

Rather than treating the network solely as a mechanism for moving data, NeuralSync is designed to introduce greater intelligence into the infrastructure connecting accelerated compute. The patented technology enables DPUs associated with GPUs and other accelerators to generate forward-looking information about accelerator availability, share that information across computing systems, and use it to help coordinate accelerator resources and data movement.

The architecture is designed around a simple but increasingly important idea: AI infrastructure should be able to understand the resources available across a computing environment and intelligently coordinate how those resources interact.

“The economics of an AI factory ultimately depend on turning expensive accelerated infrastructure into useful compute,” Ensey continued. “As clusters scale, simply adding more GPUs is not enough. How effectively those GPUs communicate, access data and coordinate with other resources becomes increasingly important. NeuralSync is being engineered around that problem.”

Building the NeuralSync™ Technology Portfolio

The newly issued patent covers innovations within NeuralSync related to DPU-based accelerator coordination, availability forecasting and accelerator-memory data movement.

The patent contains 20 claims, including independent system and method claims covering key elements of the architecture.

SOAR AI intends to continue expanding the intellectual property portfolio surrounding NeuralSync through continuing applications and additional filings as the technology evolves.

The Company views the patent as an early component of a broader technology strategy focused on developing proprietary infrastructure capabilities alongside its planned AI data center and NeoCloud operations.

The Company views NeuralSync as a potential future licensing and product opportunity that complements its AI infrastructure business. SOAR AI intends to make NeuralSync available to other data center operators with custom network fabric and optimization needs.

About NeuralSync™

NeuralSync™ is SOAR AI’s proprietary AI infrastructure technology platform focused on the intelligent coordination of compute, memory and networking resources across accelerated computing environments.

NeuralSync is being developed to address infrastructure challenges associated with increasingly large and distributed accelerator deployments, including resource coordination, data movement and communication among accelerated computing resources.

About SOAR AI, Inc.

SOAR AI, Inc. (NYSE American: SOAR) is an AI infrastructure and software company focused on the development and operation of AI infrastructure and AI-powered software solutions. The Company is constructing an AI infrastructure campus in Ohio that combines data center infrastructure, high-density GPU compute, high-performance networking and related technology to support AI, machine learning and high-performance computing workloads. The name SOAR, which stands for Sovereign AI Resources, reflects the Company’s strategy of developing AI infrastructure that provides customers greater control over their data, models and computing environments.

The Company also develops AI-powered operational systems for aviation businesses through its Aviation AI platform and operates Vaunt, a technology-enabled private aviation membership platform.

For more information, visit soarai.com.

Forward-Looking Statements

This press release contains forward-looking statements within the meaning of the Private Securities Litigation Reform Act of 1995, including statements regarding the development, capabilities and potential benefits of NeuralSync; potential licensing and product opportunities; the Company’s intellectual property strategy, including its intellectual property enforcement strategy, continuing applications and additional patent filings, and the enforceability and validity of its patents; the Company’s ability to prevent third parties from designing around its patents or developing competing technologies; the Company’s competitive position; the development and operation of the Company’s AI infrastructure; and the Company’s ability to execute its growth strategy.

Forward-looking statements can often be identified by words such as “expects,” “anticipates,” “intends,” “plans,” “believes,” “seeks,” “estimates,” “projects,” “targets,” “would,” “will,” “should,” “could,” “may,” “potential,” “opportunity” and the negative of these terms or other similar expressions, but the absence of these words does not mean that a statement is not forward-looking.

Forward-looking statements are based on current expectations, assumptions, estimates and projections and are not guarantees of future performance or events. Actual results may differ materially from those expressed or implied by these forward-looking statements as a result of various risks and uncertainties, including risks related to the Company’s ability to successfully develop, license or commercialize NeuralSync; the possibility that the Company’s patents may be challenged, invalidated, narrowed or circumvented, may not provide meaningful competitive advantage, or may not provide sufficient scope or strength to prevent competitors from designing around them; the possibility that pending or future patent applications may not result in issued patents; claims of infringement by third parties; the possibility that the Company may need to resort to litigation to protect its intellectual property rights, which may be costly and time-consuming and may not be successful; rapid technological change in AI infrastructure, including the possibility that such changes could reduce the relevance or competitive value of the Company’s patented technology; construction and commissioning of the Company’s Ohio campus; obtaining financing on acceptable terms; customer concentration; integrating Alignment Engine’s business; continued listing on NYSE American; volatility in the Company’s common stock; and other factors described in the Company’s filings with the U.S. Securities and Exchange Commission, including its most recent Annual Report on Form 10-K and subsequent Quarterly Reports on Form 10-Q and Current Reports on Form 8-K.

Readers should not place undue reliance on forward-looking statements, which speak only as of the date made. The Company undertakes no obligation to update or revise any forward-looking statements, whether as a result of new information, future events or otherwise, except as required by law.

Trademarks

NeuralSync™ is a trademark of Sovereign AI Chillicothe, LLC (f/k/a Alignment Engine Inc.), a wholly owned subsidiary of SOAR AI, Inc. All other trademarks, service marks and trade names referenced in this press release are the property of their respective owners. Solely for convenience, certain trademarks and trade names referred to in this press release may appear without the ^® or ™ symbols, but such references are not intended to indicate, in any way, that the applicable owners will not assert, to the fullest extent under applicable law, their rights to these trademarks and trade names.

View source version on businesswire.com: https://www.businesswire.com/news/home/20261027230321/en/