Author: Techtime

TSG to Unveil Fully Autonomous Counter-Drone System Next Month

Posted on by Techtime

By Yohai Schwiger

TSG is expected to unveil a new counter-drone system next month that, according to the company, will autonomously execute the entire engagement cycle—from detecting an aerial threat and fusing sensor data to making engagement decisions and activating the interceptor. If introduced as planned, the system will represent one of the first tangible demonstrations of the company’s expansion strategy, under which TSG has been broadening its portfolio beyond traditional command-and-control systems into complete Counter-UAS solutions.

“Today we already provide command-and-control integration with multiple sensor systems,” Golan Malka, TSG’s Vice President of Sales and Business Development, told TechTime. “Within the next month we will have a fully operational system capable of closing the entire loop—from detecting an attacking drone to deploying a net-based interceptor that neutralizes it. We are at an advanced stage of presenting a complete end-to-end solution.”

Malka joined TSG after serving as head of the dual-use technology directorate at Israel’s Directorate of Defense Research & Development (DDR&D), where he worked to bridge civilian technologies with defense requirements. He says that experience, together with lessons learned from both the Russia–Ukraine war and recent combat in Israel, reinforced his belief that today’s battlefield innovation increasingly originates in the civilian technology sector.

“In Ukraine, they realized that if you want to field solutions quickly, you have to adopt civilian technologies,” he says. “We’re seeing the same trend here, with growing adoption of robotics and autonomy. Warfare is becoming increasingly urban and asymmetric, and wherever a robotic platform can replace a human—whether in the air, on the ground, underground, or at sea—it will. That’s the direction modern warfare is taking.”

According to Malka, TSG’s new system combines data from a wide range of sensors, including radars, RF sensors, electro-optical systems, and acoustic sensors. It filters operational noise, builds a unified situational picture, prioritizes threats, and assigns the most appropriate interceptor from a range of available countermeasures.

“Everything has to happen extremely fast,” he explains. “Detection, localization, and classification all have to be completed within milliseconds.”

The new platform builds on TSG’s broader expansion strategy over the past several years. The company has strengthened its capabilities through the acquisition of 3D View (Mabat 3D), a developer of LiDAR and 3D mapping technologies; Production Floor, which specializes in manufacturing and integrating defense systems; and Puzzle, which provides testing and deployment capabilities. TSG has also invested in Robotican, an Israeli developer of autonomous drones and drone interceptors. Together, these moves are intended to transform TSG from a provider of command-and-control software into a supplier of fully integrated Counter-UAS solutions.

According to Malka, the company is already seeing early demand for the new platform across several markets. Beyond protecting borders, military forces, and critical infrastructure, he identifies data centers as an emerging target market.

“We’re also entering the data center protection market,” he says. “The range of scenarios is very broad—from safeguarding critical infrastructure to homeland security. We already have activities in Israel, Europe, and the United States.”

Europe, he adds, is one of the company’s primary growth markets. “European countries are significantly increasing their defense budgets in response to the changing nature of warfare, yet many still lack capabilities comparable to those developed in Israel. We therefore see Europe as a strategic market and are working closely with local partners there.”

ION FLUX Develops Breakthrough Hydrogen Generation Technology for Fuel Cells

Posted on by Techtime

photo above: ION FLUX founders Ofer Zohar (right) and Omer Hiram. photograph credit: Techtime

Haifa-based ION FLUX has developed a hydrogen generation technology that could transform the economics of using fuel cells for electricity generation. Fuel cells generate electrical energy through a chemical reaction between hydrogen and oxygen drawn from the atmosphere. This process produces water as the only byproduct (zero tailpipe emissions) while releasing electrons that generate an electrical curent. Although this technology has been known for decades, its adoption has remained limited because of the challenge of supplying hydrogen to fuel cells.

Today, hydrogen is typically produced at large centralized facilities and then transported to end users under high pressure. However, hydrogen is highly flammable, making transportation both hazardous and costly. ION FLUX’s technology, inspired by research conducted by Professor Alon Gany at the Technion – Israel Institute of Technology, is designed to overcome precisely this challenge. It produces hydrogen on demand using specially treated particles of recycled aluminum, a low-cost material that is safe to transport.

The particles undergo pre-treatment process before being sealed inside sealed pods which are inserted into the power generation unit. The hydrogen is generated on-site and fed directly into the fuel cell—much like inserting a capsule into a household coffee machine.

The company was founded in 2024 by Professor Ofer Zohar, whose background is in neuroscience at Johns Hopkins University, and by chemical engineer and hydrogen expert Omer Hiram, who conceived the original idea.

RC vehicle equipped with ION FLUX aluminum fuel pod, an on-demand hydrogen generator, and a fuel cell that powers the vehicle's systems.
RC vehicle equipped with ION FLUX aluminum fuel pod, an on-demand hydrogen generator, and a fuel cell

The two first met just days after the October 7, 2023 Hamas attack. At the time, Hiram had established a neighborhood watch in Haifa’s Old Carmel neighborhood to help local residents cope with growing security concerns. Zohar joined the volunteer patrol, and the two got to know each other while serving together on neighborhood watch shifts.

Omer recalls: “I told him about the difficulties I had experienced transporting hydrogen to customer sites across Israel, and that we needed a completely different approach. Instead of transporting hydrogen, we should generate it at the point of use. Hydrogen is so volatile that, nearly a century after the Hindenburg disaster, it continues to suffer from a public perception problem.”

Those conversations gradually evolved into a business plan, and the two founded ION FLUX, developing new directions based on the technology originally demonstrated by Professor Gany.

Professor Ofer Zohar explains: “Aluminum is one of the most abundant metals on Earth, and it is easy to process and manufacture. It is generally regarded as a corrosion-resistant metal because it oxidizes almost instantly, forming an extremely thin surface layer. As a result, the aluminum becomes coated within seconds with a protective oxide layer that prevents further corrosion, making it appear as though it does not rust.”

A New Solution to an Old Idea

“Every water molecule contains two hydrogen atoms and one oxygen atom. More than a century ago, scientists proposed that aluminum’s strong tendency to oxidize could be used to strip oxygen atoms from water molecules, thereby releasing hydrogen. The challenge was finding a way to penetrate the protective oxide layer covering the aluminum surface.

Omer Hiram displays the RC vehicle's aluminum fuel pod, designed for safe and easy transportation
Omer Hiram displays the RC vehicle’s aluminum fuel pod, designed for safe and easy transportation

“Professor Gany demonstrated that coating aluminum with lithium hydride (LiH) weakens this protective layer, enabling the aluminum to react with water and extract hydrogen from water. The product we have developed consists of two complementary components: a proprietary aluminum fuel, branded ALF, and a hydrogen generator marketed under the name HyGen, which produces hydrogen on demand.”

The company’s ALF fuel consists of consists of aluminum particles coated with lithium hydride sealed inside closed pods. The material is solid and stable, making it easy to store, transport, and handle safely. These pods are inserted into the HyGen generator, where the particles are mixed with water to produce hydrogen gas, which is supplied directly to virtually any type of fuel cell. After hydrogen generation, the aluminum is converted into alumina (aluminum oxide), which can be recycled into new aluminum products.

Israel’s Directorate of Defense R&D Places Initial Order

The company’s concept has attracted considerable interest. ION FLUX raised its seed funding through the Road-2 incubator, a joint initiative of the Haifa Municipality and NVIDIA. It also secured funding from CreationsVC through the Creation-Space program and received its first product order from Israel’s Directorate of Defense Research & Development (DDR&D, MAFAT) for a field battery charging system.

Omer Hiram explains: “Our target market includes virtually every off-grid application: remote infrastructure, power tools, electric vehicles, backup power systems. and many others. The technology is fully scalable, making it suitable for both low-power and high-power applications. Recently, we demonstrated a small RC vehicle powered by our hydrogen generator. Within about two years, we expect to deliver a 5-kW generator capable of powering a forklift, charging an electric vehicle, or providing backup electricity for an average home.”

Professor Ofer Zohar adds: “We have developed a closed-loop (fully circular) energy process. The process produces no hazardous emissions. It is producing only water as a byproduct (resulting in zero tailpipe emissions). We believe our solution can unlock the hydrogen economy, which has long been constrainedremained by the fundamental challenge – and safety concerns – associated with transporting hydrogen in its conventional form.”

For more information, visit ION FLUX.

DriveNets Connects Two Distant Data Centers into a Single AI Cluster

Posted on by Techtime

By Yohai Schwiger

Israeli networking company DriveNets has announced the first commercial deployment of its Scale-Across architecture, a technology that enables geographically separated data centers to operate as a single AI cluster. According to the company, this is the world’s first commercial implementation of the concept.

The deployment is part of Project Redwood, an initiative by U.S.-based AI infrastructure provider WhiteFiber to build distributed AI infrastructure. The project connects two data centers located approximately 84 kilometers (52 miles) apart via a high-speed Ethernet network, allowing AI workloads to access GPUs across both facilities as though they were installed in a single location.

The system delivers 111.2 Tbps of bandwidth while maintaining an end-to-end latency of just 0.9 milliseconds between the two sites. According to WhiteFiber, this latency is close to the theoretical physical limit imposed by the speed of light in optical fiber.

The deployment also integrates WEKA, which provides the storage and data layer, enabling the entire infrastructure to function as a single logical system rather than two independent facilities connected by a network.

The announcement follows DriveNets’ recent expansion of its AI Fabric portfolio with new systems based on Broadcom’s Tomahawk 6 switch silicon, which also support Scale-Across architectures. The WhiteFiber deployment represents the first commercial implementation of the technology disclosed by the company.

Solving AI’s Next Infrastructure Bottleneck

The architecture addresses one of the industry’s emerging challenges.

Building larger AI clusters is no longer limited primarily by GPU availability. Increasingly, hyperscale operators face constraints related to electrical power, land availability and cooling capacity at individual data center sites.

Rather than continuing to expand a single facility, Scale-Across allows computing resources to be distributed across multiple locations while still functioning as one unified AI infrastructure.

Making Two Facilities Behave Like One

Achieving this is far from trivial.

Training large AI models requires thousands of accelerators to exchange enormous volumes of data continuously. Even small increases in latency or packet loss can significantly reduce training efficiency.

According to WhiteFiber, Project Redwood is designed not simply to connect two data centers with a fast network, but to make them behave as a single GPU cluster from the perspective of AI software frameworks.

To accomplish this, the project relies on DriveNets’ Fabric Scheduled Ethernet (FSE) technology, which orchestrates network traffic to eliminate packet loss while maintaining deterministic performance under the heavy communication loads generated by large AI clusters.

WhiteFiber also claims it achieved 111.2 Tbps using only part of the available optical spectrum—performance the company says is roughly twice the capacity reported in comparable field deployments. The company plans to scale the architecture further and has filed patent applications covering key elements of the design.

A New Model for AI Factories

WhiteFiber provides GPU-as-a-Service infrastructure and develops AI computing platforms for enterprise customers. The Nasdaq-listed company (NASDAQ: WYFI) operates facilities in the United States, Iceland and Europe, and recently announced a multi-year contract worth more than $160 million to build AI infrastructure near Paris.

The company promotes a vision of distributed AI Factories, where computing resources located across multiple facilities can be aggregated into a single logical platform.

For DriveNets, the deployment represents more than another customer win. Only last month, the company raised $410 million to accelerate the expansion of its AI Fabric business, arguing that the next generation of AI infrastructure will increasingly rely on high-performance Ethernet as an alternative to InfiniBand.

The WhiteFiber project provides the first commercial proof point for one of the most ambitious capabilities in DriveNets’ portfolio. If Scale-Across gains broader adoption among AI infrastructure providers, it could fundamentally reshape how future AI superclusters are built—allowing them to expand beyond the physical limits of a single data center.

Skapion Raises $36 Million to Develop Counter-Swarm Drone Defense System

Posted on by Techtime

[Photo caption: Skapion’s founding team (from right): Gal Goren, Yaron Karp, Pini Yungman, Zafrir Yoeli and Ido Bar-On. Photo: Nicholas Pfosi]

By Yohai Schwiger

Israeli defense technology startup Skapion has raised $36 million in seed funding to develop a new defense system designed specifically to counter autonomous drone swarms. The round was led by UP.Partners and Khosla Ventures—an early investor in OpenAI—with participation from Fusion VC, Stratos Ventures, TBD VC, and q Fund.

Beyond the size of the investment, the announcement offers insight into the company’s strategy—from the operational problem it aims to solve and the backgrounds of its founders to an organizational structure that suggests the U.S. defense market is already its primary commercial target.

Founded only in late 2025, Skapion’s $36 million seed round stands out even within the rapidly expanding defense technology sector. The financing reflects growing investor confidence in technologies addressing one of the fastest-evolving challenges in modern warfare: defending against coordinated drone swarms.

The company said the proceeds will be used to expand its engineering teams, accelerate product development, conduct system integration and testing, and support engagements with defense and government organizations in Israel, the United States and additional international markets.

Built by Air Defense Veterans

Skapion’s founding team provides an early indication of the company’s ambitions.

Chief Architect Brig. Gen. (Ret.) Pini Yungman previously led Rafael’s Air and Missile Defense Systems Division and played a central role in the development of major Israeli air defense programs, including Iron Dome and David’s Sling.

CEO Ido Bar-On joined the company from Israeli drone developer XTEND, where he led business activities with defense and government customers worldwide.

The founding team also includes CTO Gal Goren, Zafrir Yoeli, a co-founder of Enlight Renewable Energy, and Yaron Karp. The combination of senior air defense experts, deep-tech entrepreneurs and experienced business executives suggests that Skapion is building a complete operational platform rather than a single enabling technology.

Designed for Swarms, Not Individual Drones

One phrase appears repeatedly throughout Skapion’s announcement and corporate messaging: “Native Counter-Swarm System.” It also represents the company’s primary point of differentiation.

While most existing Counter-UAS systems were originally developed to detect and defeat individual drones or relatively small numbers of aerial threats, Skapion says its platform is being designed from the ground up to handle coordinated attacks involving dozens—or even hundreds—of unmanned aircraft operating simultaneously.

Since the war in Ukraine and the growing use of one-way attack drones across the Middle East, countering drone swarms has become one of the most urgent challenges facing modern air defense forces. The problem is not only operational but economic: using expensive interceptors against large numbers of inexpensive drones can rapidly exhaust defensive resources.

According to the company, its system is built around three core objectives: engaging many threats simultaneously, significantly lowering the cost per interception, and maintaining autonomous operation even in communications-denied environments.

At this stage, however, Skapion has disclosed very little about the underlying technology. The company has not revealed whether the system relies on directed energy, electronic warfare, interceptor drones, kinetic interceptors or a combination of multiple approaches, nor has it published performance specifications or demonstration results.

Israeli Engineering, Washington Headquarters

Skapion’s corporate structure also provides clues about its long-term strategy.

While the company’s R&D center is based in Ramat Gan, Israel, its headquarters are located in Washington, D.C.

The decision suggests that Skapion is positioning itself from the outset to serve the U.S. defense establishment, widely regarded as the world’s largest market for advanced counter-drone technologies.

The company currently employs more than 20 people in Israel and is continuing to recruit engineers across electronics, software, robotics, autonomous systems and aerospace engineering as it accelerates development of its platform.

Altera’s Growth Signals a Recovery in the FPGA Market

Posted on by Techtime

[Photo caption: Altera’s new Agilex 9 Direct RF-Series FPGA family targets high-performance communications, aerospace, defense and advanced AI infrastructure]

By Yohai Schwiger

Programmable chipmaker Altera is growing at an annual rate of about 20%, while its operating income has more than doubled following several years of declining revenue and market share losses. The figures were disclosed by CEO Raghib Hussain in an interview with Reuters.

According to Hussain, the company grew by more than 20% last year and expects revenue to increase by approximately 25% this year. As a privately held company, however, Altera no longer publishes detailed financial results.

Hussain attributed part of the turnaround to bringing engineering teams closer to customers. “I believe in engineer-to-engineer conversations,” he said, adding that the company has reorganized its engineering organization to increase direct customer engagement and is already seeing stronger collaboration.

Over the past year, Altera also produced working prototypes for six new chips while reducing the number of service agreements inherited from Intel from 125 to just 15, underscoring the speed of its transition back into an independent company.

The improvement follows a difficult period. Altera’s revenue declined from approximately $2.9 billion in 2023 to about $1.5 billion in 2024. Reuters attributed the downturn to customers shifting spending toward AI GPUs, market share losses to rival Xilinx—now part of AMD—and an extended inventory correction across the FPGA industry following the semiconductor shortages of previous years.

Today, however, Altera argues that the AI revolution that initially hurt its business could become its next major growth engine.

Reinventing Itself—Twice

Founded in Silicon Valley in 1983, Altera became one of the pioneers of the FPGA industry after introducing programmable logic devices that could be erased and reprogrammed. For decades, it competed with Xilinx across communications, industrial automation, aerospace, defense and data center markets.

Intel acquired Altera in 2015 for $16.7 billion, hoping to combine Xeon processors with FPGA accelerators to improve data center performance and workload acceleration. However, the rapid rise of GPUs as the dominant platform for AI computing, together with Intel’s broader technological and business challenges, prevented that vision from fully materializing.

In October 2023, Intel announced plans to separate its programmable logic business. The company restored the Altera brand in early 2024 and relaunched it as an independent business. In September 2025, investment firm Silver Lake acquired a 51% stake in Altera for $4.46 billion, valuing the company at $8.75 billion. Intel retained the remaining 49%, while Altera began preparing for a potential future IPO.

The company also maintains a major R&D center in Jerusalem, Israel, where engineers develop FPGA hardware and software technologies for communications, AI and data center applications.

Not Replacing GPUs—Working Alongside Them

Hussain argues that FPGAs should no longer be viewed as competitors to GPUs for AI training or large-scale inference. Instead, they add value in the surrounding infrastructure—handling connectivity, data preprocessing, protocol conversion, scheduling and sensor fusion. “If the GPU is the brain, the FPGA is the nervous system,” Hussain said.

That distinction is becoming increasingly relevant in robotics and physical AI applications. While GPUs excel at massively parallel computation, FPGAs offer deterministic, ultra-low-latency processing and can be reconfigured even after deployment.

Modern robots must simultaneously process information from cameras, radar, LiDAR, motion sensors and control systems before forwarding synchronized data to AI processors in real time. FPGAs can perform these tasks efficiently while remaining flexible enough to accommodate new sensors, interfaces or system architectures without requiring new silicon.

From Data Centers to the Edge

Altera continues to target AI infrastructure as well. An expanded partnership with Arm is designed to integrate FPGAs alongside Arm-based processors in AI systems, where programmable logic can accelerate networking, SmartNIC functionality, specialized data processing and communication between processors and AI accelerators.

At the same time, the company’s commercial focus is shifting increasingly toward edge AI applications—including robotics, autonomous systems, industrial automation and other real-time environments where deterministic performance is essential.

The recovery may extend beyond Altera. AMD’s Embedded segment, which includes the former Xilinx business, reported 6% year-over-year growth in the first quarter of 2026 after a prolonged downturn.

While the numbers do not yet confirm a broad FPGA market recovery, they suggest demand is stabilizing.

The more important question, however, is not whether FPGAs are returning to the role once envisioned for them as alternatives to AI accelerators. They probably are not.

If the FPGA market is indeed entering a new growth cycle, it is doing so in a different role—not as the primary compute engine, but as the flexible hardware layer that connects sensors, networks, processors and AI accelerators. As AI increasingly moves beyond data centers into robots, industrial machines, vehicles and autonomous systems, the ability to reconfigure hardware after deployment may once again become one of the FPGA’s most valuable strategic advantages.

What Innoviz Reveals About Next-Generation Counter-UAS

Posted on by Techtime

By Yohai Schwiger

Innoviz this week announced another partnership in the counter-unmanned aerial systems (Counter-UAS) market, this time with Israeli company AeroNous, developer of an open command-and-control (C2) platform designed to protect critical infrastructure, strategic facilities and sensitive sites. Under the agreement, Innoviz’s LiDAR technology will be integrated into AeroNous’ C2 platform to enhance real-time localization by providing highly accurate 3D positioning of aerial targets.

At first glance, this appears to be just another integration announcement. Viewed alongside the company’s recent string of partnerships, however, it points to a much broader shift in how modern Counter-UAS systems are being designed.

The AeroNous announcement follows several similar collaborations unveiled in recent weeks. Regulus is integrating Innoviz LiDAR to improve drone tracking in complex environments; Givon Defense is using it to enhance precision localization; Cogniteam has incorporated it into a new AI-powered aerial target classification module; and AeroNous is now embedding it within an open C2 platform. While these are different companies addressing different applications, the announcements repeatedly emphasize the same concepts: Localization, Perception, 3D Spatial Awareness, Classification and Situational Awareness.

Taken together, these partnerships suggest something more significant than Innoviz’s expansion into defense. They highlight the emergence of a new generation of Counter-UAS architectures.

From Single Sensors to Multi-Sensor Perception

Until only a few years ago, most counter-drone systems were built around a single primary sensor. In some cases it was radar for initial detection; in others, electro-optical cameras or RF sensors capable of identifying communication links between a drone and its operator. Once the target was detected, the system activated an appropriate jammer or interceptor.

Today’s battlefield has changed dramatically.

Small FPV drones, autonomous drones operating without RF links, drone swarms and low-altitude flight through dense urban environments have exposed the limitations of that approach. The challenge is no longer simply detecting that an object exists in the air. Modern systems must rapidly determine what the object is, precisely where it is located, where it is heading and how its trajectory is likely to evolve over the next few seconds.

As a result, Counter-UAS platforms are increasingly adopting an architecture that closely resembles autonomous driving perception systems.

Rather than relying on a single sensor, they build a sensor fusion layer that combines multiple information sources. Radar provides long-range detection and tracking; RF sensors identify communication signals when available; EO/IR cameras contribute visual identification and classification; AI algorithms correlate the incoming data; and the command-and-control system produces a unified operational picture from which engagement decisions are made.

Within this architecture, LiDAR plays a distinctly different role from the other sensors.

Across nearly every Innoviz announcement, the technology is described not primarily as a detection sensor but as a localization and perception sensor. In other words, its purpose is not simply to indicate that an aerial threat exists, but to provide the system’s spatial perception layer.

LiDAR contributes information unavailable from the other sensors: a detailed 3D point cloud of the environment, precise spatial coordinates, distance, altitude, shape and motion. This enables highly accurate localization, continuous target tracking and real-time positional updates for the C2 platform.

In Cogniteam’s implementation, for example, the point cloud also serves as the foundation for an AI module capable of distinguishing drones from birds and other airborne objects, significantly reducing false alarms.

In this sense, LiDAR is becoming the perception layer of modern Counter-UAS systems. It does not replace radar, cameras or RF sensors. Instead, it complements them by providing the spatial information upon which sensor fusion and tracking algorithms depend.

It is also noteworthy that most of Innoviz’s recent partners are not building interceptors. AeroNous, Cogniteam and Regulus focus on software layers, perception engines and command-and-control platforms. This reflects a broader shift in the industry, where competitive advantage is increasingly moving away from the interceptor itself and toward the software architecture that fuses multiple sensors into a single operational model of the airspace.

Why LiDAR? Lessons from Autonomous Driving

The comparison with autonomous vehicles is difficult to ignore.

The automotive industry reached a similar conclusion years ago: no single sensor can safely perceive the surrounding environment. Radar excels at measuring range and velocity but provides limited shape information. Cameras deliver rich visual context but depend heavily on lighting conditions. LiDAR adds precise three-dimensional mapping. Sensor fusion combines all of these inputs into a unified representation of the world, enabling autonomous driving decisions.

Counter-UAS systems are now following a remarkably similar path.

The difference is that, instead of detecting pedestrians and vehicles, they must identify small, fast and often autonomous drones navigating among trees, buildings and power lines. Here too, success depends on maintaining an accurate, continuously updated 3D representation of the environment.

This may also explain why Innoviz’s technology appears particularly well suited to these applications.

For more than a decade, the company developed its LiDAR platform for one of the world’s most demanding perception challenges: autonomous driving. Beyond the optical hardware itself, Innoviz built an extensive software stack that includes firmware, signal processing, point cloud processing, perception algorithms and object-level data generation.

Rather than outputting millions of raw laser points, the system can produce structured object data—including position, velocity and direction—ready for integration into higher-level command-and-control software.

The hardware itself offers additional advantages. Automotive-grade LiDAR must operate continuously under harsh environmental conditions while maintaining high reliability, low power consumption and resistance to vibration and temperature extremes—qualities equally valuable in military applications.

Ultimately, Innoviz’s recent partnerships reveal less about the LiDAR market than about the future direction of Counter-UAS technology.

The next generation of systems is no longer being built around a single sensor or a single interceptor. Instead, it relies on a multi-sensor perception architecture in which each component contributes a different layer of information to a unified operational picture.

If the previous objective was simply to detect a drone, the new objective is to construct an accurate, continuously updated three-dimensional model of the airspace—and LiDAR is rapidly becoming one of the key building blocks of that architecture.

CEVA Enters a New Market: Software Companies Designing Their Own AI Chips

Posted on by Techtime

Photo above: CEVA CEO Amir Panush. Photo: Netti Levy.

Israeli semiconductor IP company CEVA has signed a strategic AI licensing agreement with a major U.S. software and AI platform company that is developing its own custom AI processor. Under the agreement, CEVA’s NeuPro-M neural processing technology will be integrated into a custom AI chip designed for intelligent computing devices.

CEVA described the deal as strategically significant because it expands the company’s addressable market beyond traditional semiconductor vendors and OEMs to include software platform companies.

The agreement reflects a broader industry trend in which software and AI platform providers—including Google, Meta, Microsoft, OpenAI, and Anthropic—are increasingly designing proprietary silicon to optimize AI performance, improve energy efficiency, and better integrate hardware with their software stacks.

“This is one of the world’s leading software and AI platform companies,” said CEVA CEO Amir Panush. “Its decision to build an AI processor based on NeuPro-M reflects the industry’s shift toward AI-centric computing architectures. As AI workloads become increasingly distributed between the cloud and edge devices, platform companies need to optimize the entire technology stack—from silicon and software to operating systems and user experience. This represents one of the most strategically important AI licensing agreements in CEVA’s history.”

CEVA’s NeuPro family of neural processing units (NPUs) is designed to accelerate AI inference across a wide range of applications, from ultra-low-power embedded systems to high-performance computing platforms. According to the company, the architecture can deliver up to 3,500 tokens per second per watt when running large language models such as Llama 2 and Llama 3.2 with more than one billion parameters.

The customer will use the NeuPro-M architecture to accelerate AI inference directly on edge devices. The jointly developed processor is intended to support generative AI, multimodal AI, and agentic AI applications, enabling advanced AI capabilities to run efficiently on custom silicon. As part of the collaboration, CEVA worked closely with the customer to optimize neural networks specifically for the company’s AI workloads.

The deal also highlights a structural shift taking place across the semiconductor industry. Increasingly, companies whose core business has traditionally been software are investing in custom chip development, seeking tighter integration between hardware and software while reducing dependence on off-the-shelf processors. For IP suppliers such as CEVA, this emerging customer base represents a new growth opportunity beyond the company’s traditional markets.