Above: Classiq CEO Nir Minerbi. Photo: Eyal Touag

Intel announced yesterday that it is acquiring Israeli quantum software company Classiq in a deal estimated to be worth up to $2.3 billion in cash and stock. At a press briefing this morning, Classiq co-founder and CEO Nir Minerbi outlined how the company is expected to integrate into Intel, shared details of its commercial activity and presented his vision for the future of quantum computing.

According to Minerbi, once the deal is completed, Classiq will become a business unit within Intel and lead the chipmaker’s quantum strategy and commercial operations from Israel. The unit will report to the organization headed by Intel Chief Technology Officer Pushkar Nadha.

“They see us as providing the strategic and business leadership for Intel’s quantum computing operations—software, hardware, business and strategy. This will happen from Israel,” Minerbi said. The precise roles of Classiq’s executives within Intel have yet to be determined, he added, but the intention is to preserve much of the company’s existing structure.

“Intel’s management has emphasized how important it is for us to continue operating like a startup, moving quickly and serving the entire industry. They have also stressed that they need every single employee.”

Minerbi said Intel currently employs around 70 people in quantum computing and that the acquisition of Classiq is expected to double its quantum workforce. Intel plans to expand the operation significantly, he added. Asked whether the Classiq team would also be responsible for Intel’s existing quantum employees, Minerbi said it was too early to say definitively. “The overall direction is for us to play a significant leadership role in quantum computing under the CTO.”

A Platform for the Entire Industry

Minerbi said Nadha had presented US journalists and analysts with a slide comparing Microsoft’s DOS operating system, Nvidia’s CUDA platform, and Intel and Classiq’s place in the world of quantum computing.

“Intel’s vision is to provide the software that enables the hardware,” he said. “That hardware is not limited to Intel’s own hardware. Our software will remain hardware-agnostic and continue to serve the entire industry.”

Intel is developing its own quantum computer, but according to Minerbi, it plans to continue supporting hardware from other companies through Classiq’s platform. Beyond its own quantum computer, he said, Intel has capabilities in chip manufacturing, packaging, control systems, simulations and high-performance computing that can serve the industry as a whole.

“There are players whose goal is clearly to serve only their own computers. Intel is different, and so is its vision. Combining its hardware capabilities with our software creates a major opportunity.”

Customers Prepare for the Quantum Era

Classiq develops a software platform for designing and developing algorithms for quantum computers. Developers define the problem and its constraints, and the system generates a quantum circuit that can be adapted to different hardware platforms.

Minerbi said the company was founded approximately six years ago to address the shortage of software tools for quantum computers. “We developed software for computers that didn’t really exist yet. Two or three years after the company was founded, the computers evolved and an industry emerged, and we were well positioned to serve it.”

According to Minerbi, Classiq works with major corporations, governments, universities and defense companies. Responding to a question from TechTime about the pace of the company’s commercial activity, he said Classiq records annual bookings—orders or contracts worth tens of millions of dollars. He did not provide an exact figure for recognized revenue or the company’s growth rate.

Minerbi explained that customers are not yet paying for direct commercial benefits achieved through quantum computation, but rather for the ability to prepare for the future. “Organizations pay us because they are doing something that helps them commercially: preparing for the quantum computing era, building their teams, developing applications and creating intellectual property in the field.”

Before the Intel deal, Classiq had also seriously considered an IPO or a merger with a special purpose acquisition company, or SPAC, Minerbi revealed. “It’s fair to say we were close to taking that route,” he said. The company also considered a large private funding round and continuing to operate independently.

Looking ahead, Minerbi expects quantum computers to be integrated alongside existing computing infrastructure. “For at least the next 15 years, quantum computing will be hybrid. A quantum processor will work alongside CPUs and GPUs in a data center or a supercomputing center. It won’t replace your laptop; it will work together with the existing computing world.”

By Yohai Schwiger

NVIDIA CEO Jensen Huang arrived in Kazakhstan last week, October 23, for his first visit to the country, describing it as an emerging hub of the global artificial intelligence industry. The visit comes amid plans for a massive data center campus, with investment expected to reach up to $10 billion, that could help Kazakhstan evolve from an exporter of raw materials and energy into a supplier of advanced computing services.

During the visit, Huang met with Kazakhstan’s Minister of Artificial Intelligence and Digital Development, Zhaslan Madiyev, at the Alem.AI innovation center in Astana. Huang emphasized the advantage provided by the country’s energy resources and proposed turning them into a new product: computing power for the global AI industry.

A day earlier, NVIDIA announced that national telecommunications company Kazakhtelecom had already deployed 400 H200 GPUs and 480 B300 GPUs in Astana and Almaty. More than 2,000 additional processors are scheduled for deployment during 2027, including over 1,000 next-generation Vera Rubin processors. These installations represent the first stage of an expansion of the country’s computing infrastructure, intended to serve businesses, developers, and research institutions.

A New Model: Kazakhstan Supplies Energy, Technology Companies Sell Compute

At the heart of the plan is Data Center Valley, a project to be built in the Ekibastuz industrial area in northeastern Kazakhstan. The planned campus could reach a power capacity of up to one gigawatt, with a target of approximately 100,000 advanced NVIDIA GPUs, including GB300 and Vera Rubin systems.

The business model divides responsibilities among the partners. Kazakhtelecom, through the KT-Telecom venture, will provide the physical infrastructure, electricity, cooling, and connectivity. U.S. company Firebird will finance and own the computing equipment and market the capacity to international customers. NVIDIA is the technology supplier and a strategic partner, but will neither own the data center nor directly finance the investment.

Kazakhstan is therefore building more than infrastructure for domestic use: it is developing a commercial hub for exporting cloud and AI services. The government estimates that the project could generate at least $3 billion in annual export revenue. According to Kazakhtelecom, Firebird has already secured service agreements with major international customers, although their identities and the scale of their commitments have not been disclosed.

Kazakhstan’s government presents the initiative as a structural transformation of the economy: instead of exporting coal or electricity, the country would use its energy to produce a digital service with much greater added value. These remain ambitious projections, however, rather than existing revenues.

Why Kazakhstan?

The choice of Kazakhstan is no coincidence. The world’s ninth-largest country has substantial reserves of coal, oil, gas, and uranium, along with vast stretches of land and existing industrial infrastructure. Ekibastuz, the project’s planned location, is a major center of coal-fired power generation. The continental climate, with its exceptionally cold winters, could also help reduce some data center cooling costs.

Another advantage is its location between Europe and China, positioning it to serve as a regional computing hub for Central Asian countries. The government is also advancing fiber-optic infrastructure, including a communications route across the Caspian Sea that would reduce reliance on connections passing through Russia.

The project fits into President Kassym-Jomart Tokayev’s broader strategy to transform Kazakhstan into a digital economy. Alongside data centers, the country is promoting locally developed AI models, innovation centers, engineering training, and incentives for technology companies. It is a local version of a trend also taking shape in the Gulf states: using energy resources to move into the global AI value chain rather than relying solely on raw material exports.

Between China, Russia, and the United States

The initiative also carries geopolitical significance. Kazakhstan, which borders both Russia and China, pursues a foreign policy built on balancing relations among major powers. Russia is a longstanding partner in security, energy, and transportation; China invests in infrastructure and trade; and the United States is seeking to expand its economic and technological presence in the region.

The European Union also views Kazakhstan as a strategic partner, particularly because of its energy resources and critical minerals, as well as its importance to the Trans-Caspian transport corridor, which connects Asia and Europe while bypassing Russia. In June, Tokayev met with EU leaders in Brussels, where the two sides agreed to deepen cooperation in energy, digitalization, and infrastructure.

For Kazakhstan, working with NVIDIA offers a way to attract U.S. technology and capital while maintaining its ties with China and Russia. For Washington, it represents an opportunity to expand its technological influence in Central Asia.

The undertaking also involves risks. Beyond the enormous investment required and the need to secure long-term customers, the project depends on reliable electricity supplies, international connectivity, and U.S. export restrictions governing advanced AI chips. Its reliance on coal-fired electricity could also make it harder to attract customers committed to emissions reduction targets. The $10 billion figure is an investment target for a phased project, rather than an amount already invested in full.

For Israel, Kazakhstan’s emergence as a regional computing hub could create opportunities for companies specializing in cooling, energy management, cybersecurity, and cloud infrastructure. Over time, it could also become a market for Israeli AI applications in energy, mining, agriculture, and industry. No direct Israeli involvement in the project is currently known, but the expansion of computing infrastructure in Central Asia could create a new arena for Israeli technology companies.

Image credit: Government of Kazakhstan

For decades, software development followed a fairly familiar workflow: a programmer writes code, tests it, submits it for peer review, and, once approved, merges the change into the main codebase. An entire industry of version control tools, automated testing systems, quality assurance platforms, and collaboration software grew around that process. At Anthropic, the company behind the Claude AI model, that workflow is now undergoing a fundamental transformation. The reason is simple: humans no longer write most of the code.

In two technical articles published in October, Anthropic offers a glimpse of software development when AI agents become the primary workforce. According to the company, Claude now writes roughly 90% of the code produced by the Claude Code team. The implications, however, extend far beyond saving programming time: Anthropic has had to rebuild parts of its code management infrastructure, develop multi-agent review systems, and redefine where human intervention is required.

When Code Outpaces the Infrastructure

To understand the scale of this shift, it helps to understand Git, the version control system underpinning much of modern software development. Git allows developers to work simultaneously on separate branches of a project, record changes as commits, compare versions, and merge changes once they have been reviewed and approved. One of its core principles is distributed operation: each developer can maintain a local copy of the repository, including its change history, and work independently.

That approach serves human teams well. At Anthropic, however, the pace of agent activity created an entirely different kind of workload. Between May and September 2026, daily commits to its main repository surged from roughly 2,500 to more than 30,000—a twelvefold increase in four months. That amounts to an average of more than 20 commits per minute, around the clock. Meanwhile, the Git history of the main branch grew to nearly 100 gigabytes, and testing systems needed to perform approximately 25 million code checkout operations a day.

Storage capacity was only part of the problem. Each development or testing task may require setting up a workspace, checking out a particular version, comparing files, and calculating relationships between commits. Anthropic ran hundreds of internal Git servers to handle the load, but the infrastructure became costly and cumbersome.

Its answer was Scion, a new version control system developed in-house. Unlike Git’s distributed model, in which many operations run locally, Scion operates as a centralized service. Agents and testing systems request the information they need from a server. Version calculations, comparisons, and merges can run on the server side, without requiring every task to store and process the repository’s entire history.

The system incorporates multiple caching layers, from the machine running the agent through cache servers to the central service. When thousands of tasks request the same files, those files do not have to be fetched from the source each time. According to Anthropic, the cache hit rate exceeds 99%, and the move to Scion reduced the version control infrastructure’s CPU and memory requirements by a factor of 200 compared with the previous mirror-server setup. Some checkout operations that once took minutes now take seconds.

Who Reviews the Machine’s Code?

Version control, however, was only part of the challenge. When AI agents produce code at this pace, code review can also become a bottleneck. Traditionally, a programmer submits a pull request, and another engineer examines the code for bugs, security issues, or unintended effects on existing components before deciding whether to approve it.

Anthropic developed a review system called Antique, built around several layers of AI agents. During development, the coding agent is already expected to run checks and verify that its work functions as intended. A subagent operating with a separate context can then examine the change before submission. This separation is intended to reduce the risk that the reviewer will adopt the same mistaken assumptions that caused the original error.

Once a change is submitted, it undergoes multi-agent review: several agents examine the code in parallel, looking for problems. But reviewers can make mistakes, too, so another group of agents evaluates their findings and filters out false positives. This is a particularly important safeguard: when an AI agent reports a nonexistent bug, another agent may attempt to fix it and introduce a real defect.

Alongside agent-based review, conventional software checks continue to run, including validation checks, type checking, automated tests, and security scans. Anthropic therefore combines the judgment of language models with deterministic verification mechanisms.

What remains for humans? This is where the human-in-the-loop principle comes into play: keeping people involved in oversight. Anthropic uses a risk classification mechanism to determine the level of supervision required. Low-risk changes can proceed without further human approval, medium-risk changes require approval from one engineer, and particularly sensitive changes require two human approvers. The aim is to focus human attention where it adds the most value.

According to the company, after the new review system was introduced, the share of pull requests receiving substantive review comments rose from 16% to 54%. A retrospective evaluation also found that the system could have identified roughly one-third of the bugs that had previously caused incidents on claude.ai.

Seven Engineers, Five Months, and a New Human Role

One of the most striking details is that, according to Anthropic, a team of just seven engineers developed Scion, Antique, and additional systems, completing the transition in approximately five months. The engineers themselves believe Claude already writes better code than they do for many tasks. That does not mean people have become redundant.

Their role is gradually shifting from writing lines of code to making architectural decisions, defining constraints, identifying flawed solutions, and managing risk. In one example described by the company, agents repeatedly tried to retrieve information about changes by running expensive searches through the code history, rather than using a more efficient dedicated service. The code they produced might have worked, but the engineering decision was wrong. Sometimes, as Anthropic describes it, the engineer’s most important contribution was simply telling Claude “no.”

Anthropic’s experience may offer a glimpse of software development’s future, but it does not mean every organization should replace Git or immediately adopt multi-agent review systems. Git is a mature, reliable, and flexible system, offering significant advantages in local operation, independence, and compatibility with a vast ecosystem of development tools. Scion, by contrast, was built for the unusual requirements of an organization where enormous volumes of code are generated and tested automatically, and the infrastructure remains continuously connected to the network.

Even so, the direction is clear: as AI agents write more code, the bottleneck in software development may shift from writing it to managing changes, verifying them, and ensuring their reliability. At Anthropic, the model already extends beyond programmers using AI tools. It is becoming a development system in which machines write code, other machines review it, and humans design the system, set the limits of autonomy, and remain accountable for the outcome.

For more than two decades, online marketing followed a familiar path: users typed a question into Google, received a list of links, and chose a website to visit. Businesses invested in search engine optimization to climb those rankings. Now, AI interfaces are changing how customers encounter businesses. Users ask for a recommendation, a comparison, or a solution, and the system consults sources on their behalf and puts together an answer. Businesses now need to find their way into that answer, too.

“A new marketing channel has opened up,” says Danny Shaked, an online marketing specialist and founder of SiteIT and Evil Twin. Working in the field since 2008, Shaked advises businesses in Israel and startups targeting international markets. “SEO, which many feared was about to die, has become the foundation of one of the most important areas in online marketing.” He expects the range of options presented to users to narrow over time: “My prediction is that in five years, you’ll ask the system to recommend a professional, and it will give you a single name.”

This emerging field is known as GEO, short for Generative Engine Optimization: adapting a business’s digital presence for engines that generate answers. While SEO focuses on improving visibility in search results, GEO aims to increase the likelihood that a system will use a business’s content, mention it, or recommend it. These are distinct goals: linking to an article does not necessarily mean recommending the company that published it.

The shift is already affecting clicks. In a study published in February 2026, based on 300,000 keywords, Ahrefs found that the presence of a Google AI Overview was associated with a 58% lower click-through rate for the top organic result than the rate projected using a comparison group. The finding applies to desktop searches and the first organic result, but it illustrates why a high ranking alone no longer guarantees traffic.

For Shaked, preparation starts with the basics: a properly functioning website, content that systems can crawl and understand, and clearly identifiable expertise. “SEO is the first step, the entry requirement. But beyond that, there is much more to learn and do,” he says. His recommendations include placing a clear answer near the top of the page, writing focused paragraphs, and making information easy for engines to extract.

Still, there is no established formula that guarantees a citation. Different engines select sources in different ways, and the systems continue to evolve. Success also requires new ways of measuring results. A Semrush analysis published in June, covering 3,981 domain appearances across four systems, found that approximately 62% consisted of a source link without the brand being mentioned in the answer itself. For a business seeking customers, having its content used is only part of the journey.

The challenge is particularly acute for startups. A new company must convince both the market and the systems delivering information to it that it is a credible source of expertise. “The challenge in marketing a startup is taking a company that, until recently, didn’t exist and establishing it as an expert in its field within a short period,” says Shaked.

To achieve that, he recommends building a body of knowledge: technical explanations, guides, practical insights, and content that reflects the expertise of the company’s people. This should be supported by a presence beyond the company website, including LinkedIn and video. “You need to think about how to become a source that search engines and language models will want to cite as an authority in your field.”

As content becomes cheaper to produce, however, Shaked cautions against flooding websites with automatically generated articles. “There’s nothing wrong with using AI to write content. But if anyone can produce exactly the same material, and it doesn’t reflect your experience and expertise, it’s poor content.” Even content written largely with AI can be good, he says, provided it offers originality and added value. As for automated optimization tools: “These tools can carry out the tasks, but they still lack the judgment needed to deliver good results over time.”

The next stage could take users even further away from websites. An AI agent could search for a supplier, compare offers, and eventually place an order. In that scenario, a website must enable the system both to understand what the business offers and to complete a transaction.

“Websites need to be built so that AI agents can use them technically. This is also closely connected to website accessibility,” says Shaked. He says support for AI agents is already becoming part of the requirements for websites he builds. Businesses therefore need to prepare for an additional audience: a system that reads on a customer’s behalf—and may soon make choices and take action for them, too.

Above: A Merkava tank equipped with a laser warning system. Photo: Elbit Systems.

Elbit Systems Sweden plans to establish Elbit’s first European manufacturing and maintenance center for laser warning systems (LWS), expanding its defense industrial footprint in the region.

Laser warning systems are passive detection systems that provide rapid, precise warning of laser-based threats. These include lasers associated with guided bombs and missiles, airborne and ground-based target designators, laser rangefinders and laser beam-riding missiles.

The systems use optical sensors sensitive to the wavelengths employed by military lasers. When laser radiation reaches a sensor, the system analyzes several characteristics to identify and classify the source. Wavelength helps identify the type of laser, while pulse repetition frequency helps distinguish between different systems. The direction of the incoming radiation indicates the threat’s bearing, and other signal characteristics help determine whether the source is a rangefinder, target designator or another guidance system.

Elbit’s technology can also detect laser beam-riding guidance systems, which present a particular challenge. In this guidance method, the launcher projects a laser beam toward the target, and the missile uses that beam as a reference to correct its flight path. Unlike a missile that homes in on laser light reflected from the target, a beam-riding missile typically receives its guidance signal through a rear-facing sensor.

Serving European and NATO Customers

Elbit said the new center forms part of its effort to strengthen its European defense industrial infrastructure. Scheduled to begin operating in 2027, it will provide integration, assembly, maintenance and lifecycle support for European laser warning system customers.

The systems produced at the facility are designed to detect laser threats immediately and determine their direction of origin. Elbit’s latest generation uses a distributed architecture designed for integration with existing command-and-control systems. This approach also allows multiple sensors to be installed on a platform, with configurations tailored to operational requirements.

Established in 2020, Elbit Systems Sweden now employs more than 100 people and works closely with the Swedish defense industry. It participates in Swedish military programs, including advanced digitization initiatives.

In 2023, the subsidiary secured a contract worth approximately $170 million to serve as the integration partner for the Swedish Army’s LSS Mark digitization program. It also won a roughly $48 million contract to supply the Swedish Army with Technical High Mobility Shelters, mobile units designed to house technical equipment and operate in demanding terrain.

Tobias Wennberg, CEO of Elbit Systems Sweden, said the center would provide customers in Sweden, across Europe and in NATO countries with locally supported self-protection solutions backed by secure supply chains under sovereign control.

Israel-based High Lander, a developer of drone traffic management software, has completed its $30 million Series A funding round after securing an additional $16 million. Participants included UAE defense technology group EDGE, which increased its investment in the company, alongside Protego, STEP and Kinetica. The funding will support international expansion, with an emphasis on the U.S. market and defense and homeland security applications.

As part of that expansion, High Lander is establishing a U.S. hub to serve as the base for its North American operations. The facility will include control rooms and testing and demonstration facilities, where the company can integrate its software with third-party sensors, detection systems and response technologies. It is intended to support projects and deals already under development in the United States and allow defense customers to see integrated solutions in operational settings.

The move reflects High Lander’s expansion from civilian drone management into defense through its Andromeda platform. Using AI capabilities, the system combines and cross-references information from multiple detection and monitoring systems to create a unified operational picture of low-altitude airspace.

Andromeda operates as a hardware-agnostic software layer, allowing customers to connect radars, sensors and response systems from different manufacturers under a single command-and-control platform. It supports aircraft detection, classification and tracking. A key capability is distinguishing authorized drones from unidentified or unauthorized aircraft, helping operators assess threats and coordinate a response.

Andromeda complements High Lander’s two existing platforms. Orion enables autonomous operation and management of drone fleets and their missions, while Vega provides unmanned traffic management, or UTM, including coordination of aircraft operating in shared airspace. Andromeda adds a security and operational intelligence layer, linking information about authorized flights with detection and defense systems.

The expansion reflects a broader shift in the drone market. As civilian and commercial applications grow, so does the need to protect airports, strategic sites and critical infrastructure from unauthorized drones. Governments are also developing regulatory frameworks to monitor and manage increasingly dense air traffic at low altitudes.

Israel is preparing significant regulatory changes in this area, including expanded drone monitoring and identification capabilities and plans to establish a national civilian authority for low-altitude airspace. The proposed authority would develop policy and regulate operations in airspace expected to accommodate growing numbers of autonomous aircraft.

Founded in 2018, High Lander aims to build on its autonomous aviation management infrastructure to create a software layer connecting aviation, autonomy and security, according to co-founder and CEO Alon Abelson. The company sees that broader role as a foundation for further international expansion.

Photo: The High Lander team. Courtesy of High Lander.