Google DeepMind Releases Agent Safety Benchmark For Enterprises

Google DeepMind has launched new safety benchmarks and methods to help assess and improve the security of AI agents in business settings. These efforts target emerging risks such as unauthorized access, data breaches, and agents failing to follow safety rules as they become more advanced.  

Key Developments in Agent Safety 

  • ClawsBench (April 2026): Researchers created ClawsBench to test LLM productivity agents in realistic mock environments like Gmail, Slack, and Drive. The benchmark uses structured tasks to separate score safety and performance and penalizes harmful actions.  
  • Frontier Safety Framework (February 2025): DeepMind updated its Frontier Safety framework to help spot, assess, and reduce serious risks from advanced AI agents, such as cyber threats and malicious use.  
  • Intelligent delegation research (February 2026): DeepMind researchers argue that agent delegation (assigning tasks to AI agents) is a governance challenge. Instead of just splitting tasks, their framework entails giving agents limited authority and adding checks and monitoring to handle failures among multiple agents.  
  • Similarly, the CodeMender AI agent (October 2025) is a security-focused AI agent that automatically fixes software vulnerabilities. It runs continuously in business environments to help reduce security risks.  

Enterprise Focus 

Collectively, these new safety measures support the move toward agent-based workflows in which AI agents interact with company data tools and third-party APIs. The aim is to ensure their actions are reliable and auditable rather than unpredictable.  

  • Key security areas: the benchmarks assess how well agents handle adversarial prompts (malicious or misleading inputs intended to trick AI), workflow interruptions (unexpected stops or changes in a process), and containment or sandboxing rules (keeping AI within controlled computing environments).  
  • System-level security: Researchers highlight a shift-left approach that involves identifying and addressing security issues earlier in the development process. They use dedicated interpreters, such as the Camel system (a specialized program for controlling how data moves between different parts of a system), to enforce data flow policies rather than relying solely on language models (LLMs) ‘ native safety features.  

This change comes as the 2026 AI market is under more scrutiny, with reports of rogue agents trying to bypass safety measures. As a result, uniform safety testing for businesses is now essential.  

Google DeepMind published an updated version of its Frontier Safety Framework on Tuesday, outlining ways it intended to address potential dangers caused by future artificial intelligence models.  

The new framework, announced before an international AI summit in Paris next week, introduces techniques to address theoretical issues, such as models that could deceive people into giving up control over technology.  

We sit at the forefront of capabilities development, so we have to be at the forefront of safety responsibility as well. Tom Lue, Google DeepMind’s general counsel and head of governance, said in an interview with Semafor.  

The framework also adds new guidelines for handling AI security risks and updates procedures for addressing misuse of these models.  

Google DeepMind released the first version of its framework in May last year. Since then, the AI landscape has changed.  

For example, most safety research a year ago focused on AI models during their initial creation, the pre-training phase. Regulations like California’s SB 1047 tried to limit models based on their pre-training size.  

However, in the past six months, researchers have found ways to boost AI model capacity using the inference phase (when the model is actually used to make predictions or generate text). Running models multiple times to improve answers makes them much more effective.  

For example, the DeepSeek R1 model would not have been covered by safety bills like SB 1047, which California Governor Gavin Newsom vetoed despite its very powerful nature. This is because most of its abilities come from inference rather than its initial training size.  

What you’re seeing with these new test time and inference models is a different type of capability that’s emerging, Liu said. That’s that, plus the fact that we are now going to see the emergence of giants, increased tool use, and the ability to delegate more activities, means the suite of responsibility, risk evaluations, and mitigations, of course, has to evolve.  

Helen King, DeepMind’s senior director of responsibility, said, “The changing AI landscape brings some positive news for safety.  

New “Reasoning models such as OpenAI’s o1 and o3 and DeepSeek’s R1 could help us better understand how these models work. “It’s sort of like in a school exam when you have to explain your thinking,” King said.  

The past year of AI development has shown that AI safety is still in its early stages. Any law passed now will likely become outdated soon.  

Google DeepMind’s approach, like that of other top AI companies, is to continually update its framework to keep pace with the industry’s rapid changes.   

Many “experts” predicted an AI disaster by now, but it hasn’t happened yet. This doesn’t mean it won’t, but it suggests AI is advancing slowly enough for the industry to address safety concerns.  

Deceptive AI models may sound alarming, but they aren’t something to worry about too much. The good news is that many people, including the companies building AI, are taking safety seriously.

Source: Google releases new AI safety framework 

Microsoft Expands Copilot Offering For Small Business Adoption

Microsoft has expanded its Copilot tools for small and medium-sized businesses, making AI more accessible, secure, and affordable.  

Here are the main highlights from Microsoft’s recent announcements about this expansion.  

  • Launch of Microsoft 365 Copilot Business: A new dedicated SKU for SMBs designed for organizations with fewer than 300 users.  
  • Firms can now purchase 1-299 seats of Microsoft Copilot Business for $21 per user per month, with promotional discounts available. Alternatively, for Microsoft 365 Copilot, pricing is $30 per user each month, or £24.7 per user with an annual commitment (excluding VAT).  
  • Copilot AI features integrate directly with Word, Excel, PowerPoint, Outlook, and Microsoft Teams.  
  • Copilot adheres to Microsoft 365’s security, privacy, and compliance standards, ensuring your data remains protected.  
  • A Copilot business subscription includes Copilot Studio, which lets you create custom agents to automate tasks.  
  • Microsoft is offering discounted prices until June 30, 2026. There are also bundles that combine Microsoft 365 Business Standard or Premium with Copilot.  

Targeted Benefits for SMBs: 

  • Save time by automating document, email, and report creation.  
  • Organizations can streamline the onboarding of new employees and manage recruitment processes more efficiently.  
  • Copilot generates responses tailored to each organization’s data, rather than generic answers.  

Partners and cloud solution providers can now help deliver these AI solutions, giving support with adoption and technical setup through special kits and training.  

Microsoft has made Microsoft 365 Copilot available to businesses of all sizes by removing the previous minimum requirement of 300 seats. Now, there is no seat minimum, allowing even the smallest organizations to access Copilot’s advanced language models and leverage their organization’s data in Microsoft 365 apps to work smarter and boost efficiency.  

This change is a big step towards making productivity AI solutions more accessible to organizations of any size.  

Copilot is an AI tool that uses large language models and your organization’s data in Microsoft 365 apps to help you work smarter and more efficiently. It used to be available only to enterprise customers with at least 300 licenses, but now it’s open to businesses of all sizes.  

Key Announcements 

General availability of Copilot for businesses of all sizes.  

  • Copilot for Microsoft 365 is available for small and medium-sized businesses using Microsoft 365 Business Premium or Business Standard.  
  • Microsoft 365 Copilot business is available for $21 per user per month with promotional discounts. Microsoft 365 Copilot is also available at $30 per user per month or £24.70 per user per year (excluding VAT) for 1-299 seats.  

No minimum purchase requirement  

  • There is no longer a 300-seat minimum for commercial planes.  
  • Copilot is now available for Office 365 E3 and E5 customers, even if you don’t have a Microsoft 365 license.  

Inclusion in the Microsoft CSP program  

  • Commercial customers can now buy Copilot for Microsoft through Microsoft’s cloud solution provider partners.  

Expanded availability of Copilot for education  

  • Microsoft has also made Copilot for Microsoft 365 available to education faculty and staff. 

Source: Microsoft News 

TSMC Arizona Fab Reaches Production Efficiency Milestones

By December 23, 2025, global semiconductor manufacturing is at a major turning point. Taiwan Semiconductor Manufacturing Company (NYSE: TSMC), the world’s top contract chipmaker, has sped up its plans for the large Fab21 complex in Phoenix, Arizona. Phase one is already producing large volumes of 4 nm and 5 nm chips, and the company has started installing equipment and preparing clean rooms for phase two, which will make 3 nm chips. This progress is a big win for the US effort to bring key technology back home and strengthen the supply chain for future artificial intelligence.  

The acceleration at the Arizona site, which previously faced labor issues and construction delays, marks a turning point for the American “Silicon Desert”. It not only demonstrates renewed confidence but also sets the stage for the next technical leap by moving up the 3nm production timeline to 2027, a year earlier than expected. TSMC is meeting strong demand from US tech companies seeking to protect their AI hardware from risks in the Pacific region.  

Technical Milestones and the 92% Yield Breakthrough 

Fab 21’s achievements have silenced early doubters about US advanced manufacturing. TSMC reported that its Arizona Phase 1 facility achieved a 92% yield rate in 4nm production, about 4 points higher than similar sites in Taiwan. This success stems from digital twin technology enabling virtual process optimization before real-world implementations.  

Phase two’s shift to 3 NM technology advances both transistor density and energy efficiency. The 3 NM process offers up to 15% higher speeds at the same power or 30% lower power at the same speed compared to 5 NM chips. By December 2025, phase two’s building was complete, and interior installation for clean rooms and equipment was progressing rapidly. EUV lithography machines are set to arrive in early 2026 for 2027 production.  

A Windfall for AI Giants End-to-End Supply Chain 

Accelerated 3 nm output in Arizona benefits large AI companies. Apple, Nvidia, and AMD have reserved most of Fab 21’s capacity. For Nvidia, domestic production reduces shipping risks associated with the Taiwan Strait. Amkor Technology is constructing a $7 billion advanced packaging facility in Peoria, Arizona, contributing to the supply chain.  

TSMC and Amkor’s partnership will enable the US supply chain to produce AI chips from fabrication through advanced packaging. Previously, US-made chips still needed to be shipped to Taiwan for packaging, creating risks. With local packaging, firms like Nvidia and AMD can achieve faster, more secure North American supply chains for AI.  

The Geopolitical Significance of the Silicon Desert 

TSMC’s Arizona expansion is extremely important. It is the highlight of the US Chips and Science Act, which gave TSMC 6.6 billion dollars in grants and up to 5 billion dollars in loans. By late 2025, the US Department of Commerce had released several rounds of this funding, noting TSMC’s strong technical progress. This puts the US in a better position against global competitors such as Samsung (KRX:005930) and Intel (Nasdaq:INTC), which are also working to launch advanced chip technologies.  

This shift toward geographic decoupling is a direct answer to rising tensions in the South China Sea. By building a gigafab cluster in Arizona, expected to include six fabs and $165 billion in investment, TSMC is creating a secure backup to its Taiwan operations. This move has changed the global semiconductor industry, bringing high-end manufacturing closer to Silicon Valley’s software and design centers.  

Looking Ahead: The Road to 2nm and Beyond 

TSMC’s Arizona ambitions extend well into the future,d beyond current achievements. In April 2025, construction began on phase three,which will eventually produce advanced 2 nm and 1.6 nm chips needed for the next generation of AI models requiring greater power and efficiency. According to projections, by 2030, Arizona could match the capabilities of TSMC’s Fab 18 in Tainan, delivering the world’s most advanced chips.  

Challenges remain, mainly a shortage of specialized talent to run automated fabs. The 92% yield shows early staffing problems are mostly solved, but expansion from two to six fabs over five years will require more engineers and technicians. Adding advanced packaging on-site will require TSMC and partners to work closely together. Accelerated 3nm equipment installation and high yields have turned the Silicon Desert from ambition to reality. This is a vital safeguard for US AI and national security.  

As the next phase takes shape, anticipation continues to grow. In 2026, attention will focus on the arrival of EUV tools for phase two and progress on phase three. Supported by the Chips Act and major technology partners, TSMC Arizona sets a benchmark for domestic advanced chip manufacturing.

Source: TSMC Arizona’s 3nm Acceleration: Bringing Advanced Manufacturing to US Soil 

FCC Proposes Drone Innovation Zones To Boost US Production

On April 1, 2026, the FCC’s Wireless Telecommunications Bureau and Office of Engineering and Technology released a public notice calling for feedback on steps the Commission should take to strengthen US leadership in drone manufacturing technology and use.  

Chairman Carr called this public notice the next major step in efforts to promote US drone leadership by cutting red tape, modernizing obsolete regulations, and securing a domestic drone supply chain. The agencies are seeking input on freeing up spectrum for drones, updating the licensing framework, supporting drone and anti-drone system testing, and other areas to reduce bureaucracy and accelerate improvements in this field.   

Comments are due by May 1, 2026, with reply comments due by May 18, 2026.  

Key Takeaways 

  • WTB and OET framed the public notice as supporting the commission’s ongoing work aligned with the American drone dominance agenda  
  • The public notice aims to collect information on various drone-related topics that could inform future FCC actions without setting a fixed regulatory path.  
  • It examines whether existing spectrum is suitable for drones and explores drone access to other spectrum bands, including those for licensed mobile services.  
  • WTB and OET are also seeking feedback on how to make it easier to develop and deploy drones by changing the commission’s experimental licensing process, possibly by adding a special drone license category.  
  • The focus on cutting red tape in processes, allowing greater flexibility in spectrum use, and using market-based approaches aligns with the direction of other recent actions by the commission and its bureaus.  

Background 

WTB and OET state that the public notice supports the Commission’s efforts in line with recent executive orders to accelerate US drone commercialization, encourage domestic manufacturing, broaden drone access to spectrum for advanced operations, and restrict foreign drone use in sensitive areas.  

In December 2025, the FCC restricted the import, sale, and marketing of certain foreign-made drones and components identified as security risks. The commission later provided conditional approvals for four drone devices found to cause no unacceptable risk after further review.  

The public notice also emphasizes the need to work with other federal agencies, including the FAA, the National Telecommunications and Information Administration, and national security agencies. This coordination helps ensure that communications policies support the administration’s broader goal of safely integrating drones into US airspace. These efforts include joining a multi-agency group focused on developing and launching advanced air mobility (AAM) technologies, such as electric vertical takeoff and landing (eVTOL) aircraft, in the United States.  

Increasing Spectrum Access for UAS 

The public notice seeks broad comment on any and all non-federal frequency resources that commenters believe are necessary to further America’s UAS leadership role. Specific areas of inquiry include:  

  • Unlicensed spectrum: WTB and OET note that most drones use unlicensed spectrum (902–928 megahertz, 2400–2500 megahertz, 5000-5725 megahertz, and 5725-5875 megahertz) and ask if these bands remain suitable.  
  • Opening licensed spectrum for UAS operations: The public notice asks whether drones should operate in more of the spectrum used for licensed mobile broadband, focusing on bands such as CBRS and the 3.7 gigahertz service, where aeronautical mobile use is currently barred.  
  • Accelerating UAS development in the 5,030-5,091 MHz band: The public notice seeks input on speeding up the rollout of rules adopted in 2024  
  • Additional comment on open proceedings: WTB and OET ask commenters to update the records on unresolved issues from the 2023 UAS Notice of Proposed Rulemaking, including drone access to the 960-1164 MHz band. They also invite feedback on other pending requests to open additional drone bands.  
  • Supporting interagency efforts: WTB and OET seek feedback on how the commission can support interagency efforts on drones and anti-drone systems. This includes setting up the National Training Center for Counter UAS Systems and working with third parties on air traffic management and surveillance.  

Streamlining UAS Licensing 

The public notice also requests input on ways to update the commission’s experimental licensing process for drone development and testing. WTB and OET note that the current system may be slow and limited in scope, especially for technologies that use multiple bands, support mobile operations, or enable BVLOS communications. To fix this, they are considering a dedicated experimental license for drones with longer terms, broader coverage, and faster renewal terms. They also want feedback on tiered licensing, pre-cleared test corridors, blanket authorizations, and modular licensing based on approved spectrum bands and use cases to speed up testing while still preventing interference. Finally, the public notice asks whether the current Part 5 rules, which limit counter UAS to research and development rather than operational use, are holding back the commercial development of anti-drone systems.  

Establishing Test Beds and Innovation Zones for UAS Operations 

WTB and OET are also asking whether and how to expand the Commission’s Innovation Zone program to support large-scale drone testing. Innovation Zones allow qualified licensees to test new technologies in a controlled environment. The public notice specifically seeks feedback on whether the AERPAW testbed at North Carolina State University has provided sufficient capacity and flexibility for large-scale drone development. In addition, if additional input is needed, WTB and OET are open to ideas for new testbeds, including those for commercial or defense use, maritime areas, or low-population regions where interference is less likely.  

Other Areas of Inquiry 

The public notice also requests feedback on several related efforts to update UAS regulations and accelerate drone deployment. These include:  

  • Clarifying the permissible applications of counter UAS technologies, including any barriers to counter UAS deployment, including 47 USC 333’s statutory prohibition on willful or malicious interference  
  • Making spectrum coordination and notification requirements simpler since current rules may limit UAS and counter-UAS operations.  
  • Offering market-based incentives to make it easier for UAS operations or aerial testing to access the spectrum  
  • Ways the FCC can encourage state and local law enforcement to use US-made UAS, such as publishing a list of trusted drones, offering public safety guidance, or using its private sector connections to promote US-made drones  
  • Considering that the FCC should set up a central resource to help operators understand UAS regulatory requirements  
  • How the agency can help develop the workforce needed to grow the US drone industry  

Impact Concerning Next Steps 

By requesting input on these UAS and counter-UAS topics, WTB and OET aim to build a strong record that could help guide future commission goals or actions to accelerate drone deployment and support the US drone industry. However, the public notice does not mean the FCC has chosen any particular regulatory path. Any changes to FCC rules would still require more action from the commission.  

To have your organization’s perspectives considered, review the public notice and submit your comments to the FCC by the designated deadline. Companies interested in expanding drone use, including those focused on spectrum access, interference, domestic manufacturing, safety, labor, surveillance, noise, wildlife protection, or agricultural technology, are encouraged to participate in this process. Licensees and operators in spectrum bands that might be used for future UAS deployment should also provide feedback, ensuring their views are part of the discussion.

Source: FCC Releases Public Notice Seeking Comment on “Unleashing American Drone Dominance” 

Figure Deploys Humanoid Robots in US Automotive Manufacturing

Figure, a robotics company specializing in general-purpose humanoid robots, has begun deploying its autonomous systems in US automotive factories. Shifting from laboratory experiments to real-world operations marks a significant transformation for American manufacturing. These robots are now operating in high-traffic plants, performing repetitive and physically strenuous work. Because their form mirrors human bodies, they integrate seamlessly into existing human-centric spaces, eliminating the need for facility redesigns. This introduction aims to address workforce shortages. The rollout enables robots and human employees to collaborate, streamlining vehicle assembly.  

Transitioning From Lab To The Assembly Line 

The first stage of this rollout targets material manipulation jobs that require precise handling and strong spatial awareness. These tasks were previously difficult to automate due to irregularly shaped parts and unpredictable movements. Tigas’s robots use advanced end effectors that emulate human hands, enabling them to grasp wire harnesses, specialized fasteners, and delicate trim pieces. Rather than following rigid instructions, the robots interpret and adapt to their surroundings as they operate.  

With Integrated Vision Force Feedback, the robots modulate grip strength in response to sensed resistance. This capability prevents them from crushing lightweight plastic components or mishandling heavy metal parts. Such precision is crucial in auto manufacturing, where minor errors can have significant impacts. Unlike stationary robotic arms, these humanoid robots can navigate around production equipment to access storage bins. Their mobility ensures efficient part flow even during shift transitions or inventory updates.  

Synchronizing Machines and Human Labor 

One main goal for 2026 is to make sure humans and robots can work safely together. The figure uses Dynamic Proximity Buffers, so the robots slow down and stop right away if a person gets too close. These safety features are built into the hardware to ensure reliable operation in fast-paced environments. Right now, the robots help with line-side logistics, moving parts from delivery areas to the main conveyor. This saves human workers from having to walk long distances carrying heavy items.  

The robots also advance ergonomic neutralization by assuming tasks that demand frequent bending or overhead reaching movements commonly linked with repetitive strain injuries among assembly staff. Offloading these high-stress assignments to robots helps companies extend worker longevity and reduce injury rates. The machines learn through behavioral mapping, observing human demonstrations, and replicating the actions. This approach facilitates rapid retraining when assembly processes change for new vehicle models.  

Solving the Infrastructure Interoperability Puzzle 

A key advantage of the Figure platform is its form factor compatibility with current industrial setups. While most automated systems require costly custom railings, cages, or specialized docking stations, humanoid robots are designed to fit into the same spaces as the people they support. They can climb stairs, pass through circular doorways, and work at bench heights identical to those of humans. This zero-refit approach lets manufacturers add automation step by step without stopping production for major facility changes.  

The robots also use universal power connectivity, which allows them to recharge from standard industrial outlets or modular docking bays. This sustains the fleet running through multiple shifts with little downtime. A central fleet management system tracks the health and battery levels of every robot in the facility. If one unit needs maintenance, the system sends a backup unit to replace it. This kind of systemic redundancy is necessary for keeping modern automotive factories running smoothly.  

Expanding the Horizon of General Purpose Utility 

As figure robots are increasingly used, the company aims for cross-functional versatility, enabling a single robot to handle multiple tasks in a single shift. For example, a robot might sort engine parts in the morning and then move to quality control for visual inspections in the afternoon. This pliability sets general-purpose humanoids apart from single-task industrial robots. It gives manufacturers a liquid workforce that can adjust quickly to changes in market demand or supply chain needs.  

The 2026 software update adds collaborative problem-solving to the robots. If a robot encounters an unexpected problem, it can alert a nearby human supervisor for help via a haptic signal. The supervisor gives a quick fix, which the robot remembers and shares with the rest of the fleet through the cloud. This collective learning means the entire robotic team improves each time a robot faces a new challenge. This ongoing improvement increases the plant’s overall productivity.  

The New Pulse of American Production 

As these new techniques are introduced in factories, we are seeing a steady transformation in the workplace. The factory is becoming more responsive with systems that work closely alongside human needs. We are moving toward a time when labor and logic work together, and every task is supported by technology that is always ready to help. Over time, the line between tool and worker may blur, forming a space where people and machines work in harmony. The factory is no longer simply a piece of heavy machinery, but a lively, efficient environment, always ready to support production.

Source:  F.02 Contributed to the Production of 30,000 Cars at BMW 

IBM Integrates Quantum Safe Encryption Into Enterprise Storage

IBM has added quantum-safe encryption to its enterprise storage lineup, especially the new IBM FlashSystem x600 series, to protect data from harvest-and-decrypt-later cyberattacks. These solutions use advanced cryptographic algorithms built to resist future quantum computing threats. They are expected to be available in March 2026.  

Key aspects of this integration include:  

  • Secure storage systems: The new FlashSystem X600 uses the latest Flash Core Module 5 (FCM5), a specialized storage module to deliver Quantum Safe encryption for stored data  
  • Data protection: This technology protects data against future threats by enabling quantum-safe TLS, a protocol for encrypting data in transit, and safeguarding data at rest. It is designed for critical infrastructure, government, and finance sectors.  
  • IBM Z mainframe security: IBM Z16 and LinuxONE systems use quantum-safe technologies and algorithms to secure data at rest, in transit, and in use  
  • DataPower Gateway X4 is a new physical appliance that secures and automates hybrid workloads with built-in quantum-safe cryptography for long-term data protection.  
  • Quantum-safe portfolio: IBM provides tools such as IBM Guardian Quantum Safe (for tracking cryptographic use), Quantum Safe Explorer (for exploring quantum algorithms), and Quantum Safe Remediator (for deploying new cryptography).  

These advancements are part of IBM’s broader plan to offer crypto agility, helping businesses adapt to changing cryptographic standards. Building on these innovations, IBM is also focusing on the needs of businesses facing evolving security requirements.  

Continuing its focus on advanced security, IBM has introduced the next-generation DataPower Gateway X for a high-performance security gateway designed for enterprises. It secures, integrates, and automates modern and hybrid workloads, featuring quantum-safe cryptography for long-term data protection against future threats.  

The X4 appliance serves as a unified gateway to protect, control, optimize, and connect applications across on-premises, cloud, and hybrid environments.  

Why Application Security Matters for Business Agility 

Enterprise IT environments are becoming increasingly complex, spanning multiple clouds, data centers, and diverse architectures. At the same time, cyber threats are growing and becoming more advanced, including new risks to current encryption from future quantum computing.  

To innovate efficiently and securely, organizations must share devices with strong visibility, governance, and control. Secure, scalable integration is essential for delivering digital experiences and enabling automation to enhance efficiency.  

Without a modern approach to application security, organizations can slow innovation, face increased costs, and risk exposing critical assets to threats.  

Enterprise-Grade Security With Quantum Safe Production 

Protecting client assets is IBM’s priority. DataPower Gateway X4 appliances deliver robust application security and integration services, combined with ease of use and a low cost of ownership features that have been established in DataPower solutions.  

The gateway supports a unified security framework for on-premises, cloud, and hybrid environments. Placed at the network edge and in the DMZ, it blocks unauthorized access, helps prevent denial-of-service attacks, and optimizes traffic routing. With DataPower Virtual Edition, organizations can extend this protection to cloud deployments, ensuring consistent security across the entire environment. The gateway secures both traditional web services and modern workloads, including API-based apps, event-driven and streaming services using Kafka, gRPC, and GraphQL. In computing and IT security, IBM researchers developed cryptographic schemes that NIST adopted as standards to strengthen public key cryptography. DataPower Gateway X4 appliance includes post-quantum cryptography (PQC) capabilities that can be configured for both inbound and outbound connections using TLS server and client profiles. Hybrid cryptographic algorithms are also provided, combining quantum-safe and classical methods to balance security strength with performance. With IBM DataPower, organizations can confidently protect the WAN and optimize service delivery while reducing development effort and mitigating business risk.  

Key Capabilities of DataPower Gateway X4 Appliance 

DataPower Gateway X4 is a plug-and-play appliance for rapid deployment. It is tailored for enterprise architects, security teams, and platform engineers, offering:  

  • Improved performance and scalability: Cologne offers more processing power, memory, and network bandwidth than previous generations (data power X2 and X3 gateway appliances).  
  • Secure storage for cryptographic keys: The hardware security module (HSM), a dedicated device for managing cryptographic keys, stores them in secure hardware, speeding cryptographic operations and simplifying management by centralizing key security in a hardened unit.  
  • Enterprise-grade secure design: Features a hardened tamper-resistant build that supports advanced cryptographic operations and quantum-safe protection  
  • Reliable, seamless, high-speed integration: The front panel offers networking options with built-in Ethernet model modules that support 1 GB, 10 GB, 40 GB, and 100 GB speeds.  
  • High-performance storage: comes with 1.6 TB NVMe SSDs for faster performance and greater efficiency.  
  • Improved usability: features an optimized web management interface to help developers work more efficiently.  

Availability and Next Steps 

The DataPower Gateway X4 appliance will be available starting March 26th, 2026. DataPower V11.0 will also be released on that date for use with the DataPower Virtual Edition entitlement.

Source: Secure and automate hybrid IT workloads with IBM’s new DataPower Gateway X4 appliance 

AWS Expands Sovereign Cloud Infrastructure For US Governments

Amazon is making a landmark investment of up to $50 billion to expand AI and supercomputing for US government customers using AWS. Starting in 2026, this investment will add about 1.3 gigawatts of AI and supercomputing capacity by building new data centers with advanced technology in AWS Top Secret, AWS Secret, and AWS GovCloud regions. Federal agencies will have greater access to AWS’s full range of AI services, including Amazon SageMaker for training and customizing models, Amazon Bedrock for deploying models and agents, Amazon Nova, Anthropic Claude, top open-weight base models, AWS Trainium AI chips, and NVIDIA AI infrastructure. These tools will help agencies create custom AI solutions, manage large datasets, and boost productivity. The new capabilities will be available to all current and future US government customers in these regions, supporting America’s AI leadership and providing secure, scalable infrastructure for future innovation.  

This investment will enable government agencies to dramatically accelerate discovery and policy-making outcomes. By combining simulation and modeling data with AI, agencies can now conduct analyses in hours rather than weeks or months. This means faster trend identification, quicker insight generation from large datasets, and more efficient decision-making. Research teams will be able to analyze decades of global security data and translate complex patterns into actionable recommendations, thereby improving responses to national security threats. Advanced computing will also integrate supply chain infrastructure and environmental data, providing a clearer operational picture. Defense and intelligence teams will gain the capability to automatically identify threats and create more effective response plans by processing vast amounts of satellite imagery, sensory data, and historical trends. By integrating AI with modeling and simulation, agencies will address challenges more quickly and with greater accuracy, leading to tangible improvements in mission outcomes.  

This investment will transform how the US government and related industries accomplish high-impact missions, such as strengthening national security, advancing scientific research, and driving innovation. By supporting research in areas such as autonomous systems, cybersecurity, energy, and healthcare, the initiative directly enhances America’s leadership in computational discovery and innovation. It also aligns with the administration’s AI action plan and other advanced computing projects, prioritizing secure US-based AI and cloud infrastructure. Consequently, government and industry partners will see more efficient mission execution, faster research breakthroughs, and improved national competitiveness.  

Matt Garman, CEO of AWS, stated that this investment in purpose-built government AI and cloud infrastructure will change how federal agencies utilize supercomputing. Agencies will have broader access to advanced AI capabilities, allowing them to accelerate critical missions such as cybersecurity and drug discovery. The investment aims to remove technology barriers for government agencies and enhance America’s leadership in AI.  

Amazon’s investment highlights the critical roles of AI and supercomputing in advancing technology, safeguarding critical infrastructure, and driving industrial innovation. Federal customers and their partners are working to combine AI and high-performance computing, with the primary outcomes being accelerated problem-solving, improved research workflows, and the ability for researchers and engineers to address complex challenges more efficiently. This approach marks a significant shift from traditional high-performance computing to AI-powered discovery, enabling scientists to describe their challenges and receive actionable, simulation-backed recommendations from AI systems.  

Building Resting on a Foundation of Government Innovation 

Today’s announcement shows AWS’s leadership in government cloud computing, serving over 11,000 government agencies. AWS has supported large-scale government innovation for more than a decade, achieving several industry firsts:  

  • 2011: launched AWS GovCloud (US-West), becoming the first cloud provider to build infrastructure specifically for government security and compliance requirements.  
  • 2014: Introduced AWS Top Secret East, the first air-gapped commercial cloud accredited to support classified workloads  
  • 2017: launched AWS Secret Region, becoming the first cloud provider accredited across all US government data classifications: unclassified, secret, and top secret.  
  • 2018-2025: Expanded government cloud infrastructure with AWS GovCloud (US-East), AWS Top Secret West, and AWS Secret West regions  

AWS’s experience building infrastructure of all sizes and offering strong security, compliance, and governance tools for both unclassified and classified data enables federal agencies to focus on their missions rather than managing complex on-site systems. 

Source:  Amazon to invest up to $50 billion to expand AI and supercomputing infrastructure for US government agencies 

CISA Issues Guidance on Securing AI-Driven Infrastructure Systems

On Wednesday, the US Cybersecurity and Infrastructure Security Agency (CISA) and the Australian Cyber Security Center (ASD’s ACSC), along with other partners, released joint cybersecurity guidance for critical infrastructure owners and operators using AI in their operational technology (OT) systems. The document presents the four main principles to help organizations benefit from AI in OT while managing risks. It highlights machine learning, large language models, and AI agents because of their complex security challenges. The guidance also covers systems that use traditional statistical models and logic-based automation.  

The document, “Principles for the Secure Embedding of Artificial Intelligence in Operational Technology,” outlines key steps for safely integrating AI into OT systems. It highlights staff AI risk training, secure development, and careful consideration of business needs. The guidance uses organizations to address short and long-term data security, implement strong governance to comply with regulations, and regularly test AI models. It also emphasizes ongoing oversight, transparency, and the inclusion of AI in incident response plans to protect safety and security.  

The Purdue model is still a common way to organize OT and IT devices and networks. The guidance gives examples of current and possible AI uses in critical infrastructure based on this model. Predictive machine learning models are usually in operational layers (0-3). Large language models are more often in business layers (4-5) and often work with OT data.  

Level zero covers field devices such as sensors, actuators, and other components that interact directly with physical processes. These devices generate OT data that can be used to train AI models, particularly predictive machine learning models, or to flag marked deviations that may signal anomalies or emerging issues.  

Level one includes local controllers, which are systems designed to provide automated regulation for a process cell or production line. This category includes devices such as programmable logic controllers and remote terminal units. Some modern PLCs and edge control controllers can run lightweight, pre-trained predictive systems that support tasks like anomaly detection, load balancing, and maintaining a known safe state.  

Level two covers local supervisory systems that manage a specific process line or cell. These include SCADA systems, distributed control systems, and human-machine interfaces. AI models, mostly predictive machine learning, analyze data from these systems to spot early equipment anomalies and notify operators when corrective action is needed.  

Level three involves site-wide supervisory systems that oversee an entire facility or major sections of it. These include manufacturing execution systems and historians. Predictive machine learning models analyze aggregated historian data to predict maintenance needs and plan repairs. These models can also be used in local supervisory tools to offer recommendations for operator decision-making on performance and measurements.  

Levels 4 and 5 refer to enterprise and business networks, which include IT systems that manage corporate processes and support decision-making in critical infrastructure settings. This can involve OT data analysis and autonomous security capabilities that span both OT and IT environments. AI systems, including agents and large language models, can be applied to improve business workflows, especially where engineering needs intersect with wider business objectives. AI can also analyze OT data alongside IT data to measure operations, detect anomalies and threats, identify hardening opportunities, and generate insights that help enterprises prioritize resiliency decisions.  

Transitioning to the first principle, it focuses on understanding AI’s impact on operational technology. It describes the distinctive risks posed by integrating AI into OT systems and outlines potential impacts. Key risks for critical infrastructure owners and operators are presented, though organizations are advised that the list is not exhaustive and should supplement it with their own assessments. Later sections of the guidance explain how to address these risks, providing cross-references and mitigation strategies.  

Principle two urges organizations to assess how AI fits in OT. Before adding AI to OT, owners and operators should check if AI fits their needs and offers any advantages over other technologies. They should also consider whether AI’s existing capabilities meet their needs before using more complex AI solutions.  

AI delivers unique benefits, but as principle two reminds us, it is still developing and needs continuous risk assessment. Organizations should consider factors such as security, performance, complexity, cost, and impact on OT safety before. For each use case, they should consider the pros and cons of using AI against the application’s needs.  

Owners and operators must assess their ability to manage AI in their OT environment. They should understand how AI could introduce risks, such as the need for additional hardware, software, or security measures. If AI is used, they must follow secure development practices and a risk management framework, such as the NIST AI Risk Management Framework, to keep the system safe.  

The guidance notes on how OT vendors influence the entry of AI into OT. Some devices now have built-in AI features that sometimes require an internet connection. Vendors mainly add AI tools, such as models that predict grid frequency, and develop smart devices for engineering and control tasks.  

Critical infrastructure owners should ask vendors for transparency about AI in their products. Vendors must commit to strong security. Contracts should clearly state AI features and operation. Vendors should explain their AI use, share a software bill of materials, and provide insight into their supply chain. If a vendor finds that an AI feature could cause errors, they should notify operators.  

Operators might not want vendors to train AI on operational data as it could contain intellectual property or sensitive information. A data usage policy should state where data is stored, how it is sent, and how it is encrypted. Buyers should check if the product can run on-site or without the vendor’s cloud. Operators should decide when and how to enable or disable AI features. These actions help organizations control and manage AI risks in OT systems.  

The third principle stresses the need for strong guidance to safely integrate AI into OT. This includes clear policies, procedures, and accountability for AI decisions. The governance structure must involve key stakeholders and AI vendors. This ensures oversight across buying, development, design, deployment, and operations.  

Each key stakeholder helps build effective AI governance. Senior leaders such as the CEO and CISO must support the effort. Their backing is essential to strong governance and to addressing AI security risks. In terms of functionality, experts in OT, IT, and AI should join in, as their knowledge reveals dangers and obstacles that others might miss.  

Cybersecurity teams add protection by making policies to keep OT data used by AI models safe. They find vulnerabilities and suggest ways to reduce risks. This helps secure systems and information.  

Principle four urges strong oversight and reliable backup practices for AI in OT systems. People remain responsible for safety. AI tools should support oversight and safe operation. This principle calls for AI systems that can be monitored, checked, and fixed when needed. The guidance explains that organizations should set up monitoring and oversight for AI in OT. This ensures operators always have control as systems change.  

Critical infrastructure owners should track all AI components and dependencies. They should log and monitor their inputs and outputs. It is important to set and maintain clear standards for safe OT operations so they know when maintenance or backup is needed.  

The document sets key performance indicators (KPIs) to track AI results. Owners and operators should meet regularly with stakeholders, such as vendors and boards. These meetings help review results, discuss issues, and identify opportunities for improvement.  

Commenting on the guidance, Hugh Carroll, Vice President of Corporate and Government Affairs at Fortinet, wrote in a written statement, “Leading global cybersecurity agencies, including the US’s CISA and the UK’s NCSC and Canada’s CCCS, have released much-needed guidance outlining principles for the secure deployment of artificial intelligence in operations technologies. Fortinet is honored to have the privilege to contribute to this important effort as we collectively work to best safeguard OT environments from today and tomorrow’s threats.”  

These new principles deliver timely and practical guidance to safeguard resilience and security as AI becomes central to OT. Marcus Fowler, CEO of Darktrace Federal, said, “It’s encouraging to see a strong focus on behavioral analytics, anomaly detection, and safe operating limits. These can identify AI drift, model changes, or emerging security risks before they influence operations.” This move from static thresholds to behavior-based oversight is vital. It helps defend cyber-physical systems, even when small deviations carry great risk.  

Fowler highlighted that the guidance also urges caution with LLM-first approaches to safety decision-making in OT environments. These approaches are unreliable and hard to explain. They create unacceptable risk when human safety and process continuity are at stake. It is important to use the right AI for the right job.  

Taken together, these principles reflect a maturing understanding that AI in OT must be paired with uninterrupted monitoring and transparent and separate identity controls. According to Fowler, we welcome this guidance and remain committed to helping operators implement these safeguards to strengthen resilience across critical infrastructure. We continue to see growing recognition of AI’s operational value in cybersecurity, as evidenced by recent NDAA provisions from bipartisan members of the House Armed Services Committee that emphasize AI-driven anomaly detection, securing operational technology, and incorporating AI into cybersecurity training. That’s an active step toward strengthening US cyber readiness.  

Floris Dankaart, Lead Product Manager at the cybersecurity consulting firm NCC Group, said this worldwide coordination is noteworthy. CISA, Australia’s ACSC, NSA, and other partners are coming together to address a shared challenge. This kind of coordination is rare and signals the importance of this issue. Equally important, most AI guidance addresses IT, not OT. It’s refreshing and necessary to see regulators acknowledge OT-specific risks and provide actionable principles for safely integrating AI in these environments.  

A major challenge will be addressing skill gaps in audit teams, especially those related to AI. OT environments are typically much more structured and deterministic than IT environments, which might be at odds with many modern LLM-based AI applications, according to Dankaart. At the same time, anomaly detection based on machine learning models has been commonplace in OT threat identification and monitoring for some time and continues as a key component of the defender’s arsenal.  

He added that balancing these factors and getting to the heart of what we really mean by AI will be key for critical infrastructure owners. Luckily, some of the best practices in OT and AI use overlap. The idea that you must always have a manual fallback procedure, the ability to operate in island mode, and human-in-the-loop controls, to name a few.  

In conclusion, the guidance identified that adopting AI in OT presents both opportunities and risks for critical infrastructure owners and operators. While AI can increase efficiency, productivity, and decision processes, it also introduces new challenges that require diligent management to support the safety, security, and dependability of OT systems.  

To successfully manage the risks of adding AI to OT systems, critical infrastructure owners and operators must follow the guidance’s principles, understand AI, consider its use in OT, set up governance and assurance frameworks, and build safety and security into AI and AI-enabled OT systems. By adhering to these steps and frequently monitoring, testing, and improving AI models, organizations can achieve a balanced, secure integration of AI into OT systems that support vital public services. 

Source: Global security agencies issue joint guidance to help critical infrastructure integrate AI into OT systems 

Qualcomm Launches New Chip To Expand Affordable AI PC Market

Highlights 

  • Qualcomm and Snap extended their decade-long collaboration with a multi-year strategic agreement.  
  • The long-term agreement will integrate Snapdragon XR solutions into future spec devices.  
  • This collaboration provides developers and customers with a scalable platform to create advanced eyewear experiences.  

Qualcomm Technologies, Inc., and Specs, Inc., a Snap subsidiary, have announced a multi-user agreement to outfit future Specs devices with Qualcomm Technologies’ Snapdragon system-on-a-chip (SOC)  

Powering the Next Generation of Eyewear 

This marks Specs Inc’s first flagship project, launching Specs Advanced Eyewear that integrates digital experiences into the physical world for consumers later this year. Specs are standalone, see-through glasses that enable users to see, hear, and interact with digital content within their physical environment.  

Specs use Snapdragon XR platforms, which combine edge AI with high-performance, low-power computing. This enables intelligent context-aware experiences to run on the device, supporting faster, more private interactions. This initiative demonstrates both companies’ commitment to advancing human-centric integrated computing.  

Building on a Decade-Long Relationship 

Snap and Qualcomm Technologies have a history of collaboration in immersive technology. Snapdragon platforms have powered several previous generations of Snap’s spectacles, and this agreement extends that partnership.  

The companies will align their strategic roadmaps and collaborate on technical development to deliver industry-leading capabilities for the specs platform, including on-device AI, advanced graphics, and multi-user digital experiences.  

This program creates a scalable foundation for developers and partners supporting reliable product development and more sophisticated digital experiences over time.  

We believe the future of computing will be more human and grounded in the real world, said Evan Spiegel, co-founder and CEO, Snap Inc. Our work with Qualcomm Technologies provides a firm foundation for the future of specs, bringing developers and consumers advanced technology and performance that pushes the limits of what’s possible.  

“StarCore, the next era of computing, will be defined by devices that understand what you see, hear, ask, and say, as well as context, and respond instantly to the world around you,” said Cristiano Amon, president and chief executive officer, Qualcomm Incorporated. “Our work on future generations of specs will enable power-efficient interactive AI devices that deliver agentic experiences that feel natural, intuitive, and integrate seamlessly within daily life.”  

About Qualcomm 

Qualcomm relentlessly innovates to deliver intelligent computing everywhere, helping the world tackle some of its most important challenges. Building on our 40 years of technology leadership in creating era-defining breakthroughs, we deliver a broad portfolio of solutions built with our leading-edge AI, high-performance low-power computing, and unrivaled connectivity. Our stack-driven platforms deliver extraordinary engineering experiences and underpin all our revenue products, strengthening businesses and industries to scale to greater heights. Together with our ecosystem partners, we enable next-generation digital transformation to individuals, businesses, and advanced societies. At Qualcomm, we are enabling human progress.  

Qualcomm Incorporated includes our licensing business, QTL, and the vast majority of our patent portfolio. Qualcomm Technologies Inc. is a subsidiary of Qualcomm Incorporated that, together with its subsidiaries, operates substantially all of our engineering and research and development functions and substantially all of our products and services in businesses including our QCT semiconductor business, Snapdragon, and Qualcomm-branded products. Qualcomm patents are licensed by Qualcomm Incorporated.  

About Specs Inc. 

Specs were available only in April, and subsequently, Snap Inc. developed advanced eyewear that integrates digital experiences into the physical world. See-through lenses that place digital objects directly into programming. Snap Space, powered by Snap OS, is a proprietary contact-covered operating system designed for natural interaction between hands and eyes.  

Specs Inc. also offers Lens Studio, a set of advanced developer tools for building immersive augmented reality across Specs, Snapchat, and other platforms.

Source: Qualcomm and Snap Expand Strategic Collaboration to Advance Intelligent Computing Experiences on Specs 

Microsoft Enables Direct NPU Access For Windows AI Developers

Microsoft now gives developers direct access to neural processing units (NPUs) on Windows 11, with a focus on Copilot+ PCs. With updates to the DirectML API and the Windows AI platform, developers can build and run AI models efficiently on NPU hardware from partners such as Qualcomm (Snapdragon X Elite), Intel (Core Ultra), and AMD.  

Here are the main highlights of this update:  

  • DirectML NPU support: DirectML, which is part of the DirectX family, now lets apps use NPUs directly. This means AI workloads can shift from the GPU or CPU to the NPU, improving performance and saving battery life.  
  • Targeting Copilot plus PCs: These improvements are designed for Copilot plus PCs, which have high-performance NPUs (40+ TOPS) needed to run local AI models like Phi Silica.  
  • Windows AI APIs and Studio Effects: Developers can build apps that use Windows 11’s built-in AI features like background effects, voice focus, and real-time transcription through standard Windows APIs  
  • ONNX Runtime integration: The ONNX Runtime now supports NPU acceleration, making it easier to migrate existing models from GPU to NPU with only minor changes.  
  • Microsoft Foundry on Windows: Previously called Windows AI Foundry, this updated platform now supports the full AI lifecycle from selection to optimization and deployment on GPU, CPU, and NPU.  

With these improvements, on-device generative AI is now faster, more private, and uses less energy. For example, you can run local language models like Phi 3.5 directly on Snapdragon X-powered devices.  

We’re excited to work with Intel, one of our main partners, to launch the first neural processing unit (NPU) powered by DirectML on Windows. AI is changing the world, powering innovation and creating value in many industries. NPUs are vital for delivering great AI experiences to both developers and consumers.  

An NPU is a processor designed for machine learning (ML) tasks that require substantial computing power but don’t require graphics. NPUs also use power efficiently. These new devices will change how AI improves our daily lives. Early next year, we’ll release DirectML for support for Intel Core Ultra processors with Intel AI Boost, the new built-in NPU.  

DirectML is a basic API that provides direct access to hardware features of modern devices, such as GPUs, for machine learning tasks. It is part of the DirectX family, the Windows graphics and gaming platform, and works with other DirectX components, such as DirectX 12. DirectML also connects with popular machine learning tools such as the ONNX Runtime and Olive, making it easier to develop and deploy AI across Windows.  

Adding NPU support to DirectML opens up new possibilities for AI on Windows. DirectML with NPU support will be available as a developer preview in early 2024, along with the latest ONNX Runtime release. We’ll share additional updates shortly about new features, partners, and how to use DirectML for NPUs.  

We can’t wait to see the amazing AI experiences you will create on Windows using direct tunnels and Intel Core Ultra processors.

Source: DirectML: Accelerating AI on Windows, now with NPUs