Get ONMINE

Stay Ahead, Stay ONMINE

The Rise of AI Factories: Transforming Intelligence at Scale

AI Factories Redefine Infrastructure The architecture of AI factories reflects a paradigm shift that mirrors the evolution of the industrial age itself—from manual processes to automation, and now to autonomous intelligence. Nvidia’s framing of these systems as “factories” isn’t just branding; it’s a conceptual leap that positions AI infrastructure as the new production line. GPUs […]

AI Factories Redefine Infrastructure

The architecture of AI factories reflects a paradigm shift that mirrors the evolution of the industrial age itself—from manual processes to automation, and now to autonomous intelligence. Nvidia’s framing of these systems as “factories” isn’t just branding; it’s a conceptual leap that positions AI infrastructure as the new production line. GPUs are the engines, data is the raw material, and the output isn’t a physical product, but predictive power at unprecedented scale. In this vision, compute capacity becomes a strategic asset, and the ability to iterate faster on AI models becomes a competitive differentiator, not just a technical milestone.

This evolution also introduces a new calculus for data center investment. The cost-per-token of inference—how efficiently a system can produce usable AI output—emerges as a critical KPI, replacing traditional metrics like PUE or rack density as primary indicators of performance. That changes the game for developers, operators, and regulators alike. Just as cloud computing shifted the industry’s center of gravity over the past decade, the rise of AI factories is likely to redraw the map again—favoring locations with not only robust power and cooling, but with access to clean energy, proximity to data-rich ecosystems, and incentives that align with national digital strategies.

The Economics of AI: Scaling Laws and Compute Demand

At the heart of the AI factory model is a requirement for a deep understanding of the scaling laws that govern AI economics.

Initially, the emphasis in AI revolved around pretraining large models, requiring massive amounts of compute, expert labor, and curated data. Over five years, pretraining compute needs have increased by a factor of 50 million. However, once a foundational model is trained, the downstream potential multiplies exponentially, while the compute required to utilize a fully trained model for standard inference is significantly less than that required for training and fine-tuning models for use.

The challenge shifts to post-training scaling and test-time scaling. Fine-tuning models to suit specific applications demands 30x more compute than the original pretraining. Meanwhile, the latest advanced inference tasks like agentic AI, where models reason iteratively before responding, can require 100x more compute than standard inference. These compute-intensive needs simply exceed the capacity of general-purpose data centers.

AI factories are designed with this exponential growth in mind. From the ground up, they are built to support massive inference demands, iterative reasoning, and adaptive model deployment.

AI’s New Cost Curve

This shift in workload dynamics rewrites the economic blueprint for infrastructure investment. Where once the ROI of data center capacity was measured against steady-state cloud or enterprise workloads, AI factories demand a forward-looking calculus based on scaling behavior and future inference velocity.

The cost per token or decision point becomes a more meaningful financial metric than simple cost per kWh or per-core performance. Operators must not only provision for peak demand but architect systems flexible enough to evolve with model complexity—supporting seamless upgrades in compute density, interconnect bandwidth, and software orchestration.

Moreover, these economics aren’t confined to hyperscale players alone. Enterprises deploying vertical-specific models—whether for fraud detection, supply chain optimization, or autonomous control systems—are increasingly recognizing that the benefits of faster, smarter AI decisions justify the infrastructure premium.

This drives demand for regional and modular AI factories tailored to industry use cases, where latency, data locality, and compliance matter as much as raw compute. As with previous inflections in the digital economy, those who internalize and invest early in the new cost curves will be best positioned to lead in a world where intelligence itself is the product.

AI Factory Development Around the World

Nvidia is not alone in recognizing the strategic importance of AI factories. Governments and enterprises across the globe are racing to deploy them:

India: Through a high-profile partnership with NVIDIA, Yotta Data Services has launched the Shakti Cloud Platform—one of the country’s first AI supercomputing infrastructures. Positioned as a national resource, Shakti aims to democratize access to high-performance GPU resources for startups, research institutions, and public sector innovation, reflecting India’s broader ambition to become a global AI hub.

Japan: Cloud providers like GMO Internet and KDDI are rapidly scaling NVIDIA-powered AI infrastructure to accelerate advancements in robotics, precision medicine, and smart cities. These efforts align with Japan’s Society 5.0 vision, which emphasizes the fusion of cyber and physical systems to tackle demographic and economic challenges through AI and automation.

Europe: The European Union is taking a coordinated, multi-national approach to AI factory development, investing in seven advanced computing centers across 17 member states via the High Performance Computing Joint Undertaking (EuroHPC JU). These sites are being positioned not just as data centers but as digital sovereignty assets—powering AI research, public sector applications, and secure industrial innovation.

Norway: Telenor’s NVIDIA-powered AI factory exemplifies how Nordic countries are integrating sustainability into digital transformation. With a strong emphasis on green energy, regional talent development, and cross-border collaboration, the initiative is laying a foundation for climate-conscious AI infrastructure that aligns with European ESG priorities.

United States: AI factory development is taking a dual-track approach. Public-private initiatives like the Stargate project—focused on frontier-scale computing—and executive directives from the White House underscore Washington’s intent to lead in both commercial and governmental AI capabilities. The U.S. sees AI infrastructure not just as a competitive edge but as a strategic imperative for national resilience.

Saudi Arabia: Through its Vision 2030 strategy, the Kingdom is investing heavily in AI infrastructure, including a partnership between the Saudi Data and Artificial Intelligence Authority (SDAIA) and global hyperscalers. Recent announcements include the creation of sovereign AI compute clusters designed to support Arabic-language models and AI-driven public services.

Singapore: Known for its methodical approach to digital infrastructure, Singapore is building out AI factories as part of its National AI Strategy 2.0. With investments in sovereign compute capabilities and robust data governance, the city-state is positioning itself as Southeast Asia’s AI nerve center—balancing innovation with regulatory foresight.

These projects highlight how AI factories are quickly becoming essential national infrastructure, akin to telecommunications and energy grids. More than just data centers, they represent strategic bets on where intelligence will be created, who controls its production, and how nations will compete in an AI-first global economy.

Inside the AI Factory: A Full-Stack Approach to Intelligence Production

Nvidia’s AI factory model isn’t just a high-powered compute stack—it’s a vertically integrated platform purpose-built to accelerate every stage of the AI lifecycle. From training foundational models to deploying them at scale in real-time applications, the architecture spans compute, networking, software, data pipelines, and digital twin simulation. Each layer is engineered for high-efficiency throughput, reflecting Nvidia’s belief that intelligence production requires the same rigor and precision as modern manufacturing.

1. Compute Performance: The Engine Room of Intelligence

At the core of the AI factory is GPU horsepower. Nvidia’s Hopper, Blackwell, and the forthcoming Blackwell Ultra architectures offer exponential leaps in performance. The flagship GB200 NVL72 system—a rack-scale unit with dual Blackwell GPUs connected by NVLink Switch—delivers 50x more AI inference throughput compared to the A100 generation. Integrated into DGX SuperPOD clusters, these systems can scale to tens of thousands of nodes, forming the compute backbone for hyperscale AI development.

DGX Cloud extends these capabilities into a managed, consumption-based model, allowing enterprises to access AI factory-grade infrastructure through major cloud platforms like Microsoft Azure, Google Cloud, and Oracle. It’s an operating model built for rapid deployment and elastic growth.

2. High-Performance Networking: Compute Without Bottlenecks

Scaling AI requires more than raw compute—it demands precision networking. Nvidia’s NVLink, Quantum-2 InfiniBand, and Spectrum-X Ethernet fabrics are designed to minimize latency and ensure lossless, high-bandwidth data movement between tens of thousands of GPUs. ConnectX-8 SmartNICs and BlueField-3 DPUs enable secure, multi-tenant environments while offloading network and storage tasks to free up GPU cycles. The result is a tightly-coupled infrastructure where compute and data flow at AI-native speeds.

3. Orchestration and Operational Intelligence

Orchestrating AI workloads at scale is non-trivial. Tools like Nvidia Run:ai, Base Command, and Mission Control provide full-stack visibility and GPU-aware scheduling, ensuring optimal utilization across heterogeneous environments. These platforms support multi-tenant operations, dynamic scaling, and fine-grained workload isolation—critical in enterprise and sovereign AI environments where uptime and performance cannot be compromised.

4. Inference Stack: From Model to Real-Time Decisions

The Nvidia inference stack—including TensorRT for optimized execution, NVIDIA Inference Microservices (NIMs) for containerized deployment, and NVIDIA Triton for scalable serving—enables low-latency, high-throughput AI services. These tools are optimized for transformer architectures and multimodal models, addressing the growing demand for agentic inference, edge reasoning, and continuous learning in production.

5. Data Infrastructure: Feeding the Intelligence Pipeline

AI performance is bound by the quality and availability of data. The Nvidia AI Data Platform enables seamless integration with modern data lakes, object stores, and streaming platforms. It provides end-to-end support for preprocessing, labeling, and versioning at scale—turning chaotic data pipelines into repeatable, high-performance processes. Certified storage partners (like NetApp, Dell, and VAST Data) ensure that storage throughput can keep pace with real-time inference and training demands.

6. Omniverse Blueprint: Digital Twin-Driven Infrastructure Planning

Designing an AI factory involves massive complexity—up to 5 billion components, 210,000 miles of cabling, and megawatt-scale power demands. Nvidia’s Omniverse Blueprint introduces a systems-level digital twin to simulate, validate, and optimize AI factory builds before breaking ground. This includes everything from airflow and thermals to rack placement and interconnect design.

By enabling real-time collaboration across electrical, mechanical, and IT disciplines, Omniverse reduces time-to-deployment and mitigates critical risk. In environments where an hour of downtime can equate to tens of millions in lost inference capacity, this level of planning precision is no longer optional—it’s a necessity.

AI factories represent more than just technical innovation—they are a new class of infrastructure, purpose-built for the intelligence economy. Nvidia’s full-stack platform provides the modularity, scalability, and performance required to manufacture intelligence at scale, redefining how enterprises and nations deploy AI as a core strategic asset.

Deep Dive on Omniverse Developments: Advancing AI Factory Design and Simulation

As AI continues to drive unprecedented demand for specialized infrastructure, NVIDIA is taking bold steps to help design and optimize the next generation of AI factories with its new Omniverse Blueprint for AI factory design and operations. Unveiled during NVIDIA’s GTC keynote, this innovative blueprint is designed to help engineers simulate, plan, and optimize the development of gigawatt-scale AI factories, which require the seamless integration of billions of components and complex systems.

In collaboration with leading simulation and infrastructure partners, including Cadence, ETAP, Schneider Electric, and Vertiv, the Omniverse Blueprint enables digital twin technology to support the design, testing, and optimization of AI factory components such as power, cooling, and networking systems long before physical construction begins.

Engineering AI Factories: A Simulation-First Approach

Using OpenUSD libraries, NVIDIA’s Omniverse Blueprint aggregates 3D data from multiple sources, including building layouts, accelerated computing systems like NVIDIA DGX SuperPODs, and power/cooling units from partners such as Schneider Electric and Vertiv. This unified approach allows engineers to address key challenges in AI factory development, such as:

  • Component Integration and Space Optimization: Seamlessly integrating NVIDIA systems with billions of components for optimal layouts.

  • Cooling Efficiency: Using the Cadence Reality Digital Twin Platform to simulate and evaluate cooling solutions, from hybrid air to liquid cooling.

  • Power Distribution: Designing scalable, redundant systems to simulate and optimize power reliability using ETAP.

  • Networking Topology: Fine-tuning high-bandwidth networking infrastructure with NVIDIA Spectrum-X and NVIDIA Air.

The blueprint empowers engineers to collaborate in real-time across disciplines, reducing inefficiencies and enabling parallel workflows. Real-time simulations allow for faster decision-making and optimization, with teams able to adjust configurations and immediately see the impact — drastically reducing design time and avoiding costly mistakes during construction.

Building Resilience Into the AI Frontier

As AI workloads continue to evolve, the blueprint offers advanced features such as workload-aware simulations and failure scenario testing to ensure AI factories can scale and adapt to future demands. With the growing importance of minimizing downtime (which can cost millions per day in gigawatt-scale AI factories), the Omniverse Blueprint reduces risk, improves efficiency, and helps AI factory operators stay ahead of infrastructure needs.

NVIDIA’s ongoing efforts with partners like Vertech and Phaidra will bring AI-enabled operations into the fold, including reinforcement-learning agents that optimize energy efficiency and system stability. These advancements ensure that AI factories can adapt to changing hardware and environmental conditions in real-time, contributing to ongoing operational resilience.

The integration of digital twin technology into AI factory design is not just a theoretical enhancement—it’s essential for the future of AI-driven data centers. With over $1 trillion projected for AI-related upgrades, NVIDIA’s Omniverse Blueprint stands poised to lead this transformation, helping AI factory operators navigate the complexities of AI workloads while minimizing risk and maximizing efficiency.

To explore these developments further, watch the GTC keynote, and discover how NVIDIA and its partners are shaping the future of AI factory infrastructure.

The Age of Reasoning and Agentic AI

Nvidia defines its Blackwell Ultra platform not just as another leap in GPU performance, but as the gateway to a new phase in AI development—what it calls the age of reasoning. As workloads transition from static inference to dynamic decision-making, AI systems must increasingly mimic human-like cognition: analyzing context, planning multistep actions, and adapting behavior in real time. This shift is giving rise to two transformative paradigms—agentic AI and physical AI—both of which are redefining the infrastructure requirements for scalable intelligence.

  • Agentic AI involves AI models that operate autonomously to solve complex, multistep problems. These models reason iteratively, self-correct, and manage workflows across multiple domains. They’re already emerging in tools like AutoGPT, Devin, and AI copilots that can write code, generate research plans, or manage enterprise workflows. Unlike traditional inference, agentic AI requires continual interaction with large-scale memory, context retrieval, and recursive reasoning—all of which drive up compute needs by orders of magnitude.

  • Physical AI focuses on embodied intelligence—where simulation, sensor fusion, and real-world control intersect. Applications include real-time photorealistic simulation for digital twins, robotics, autonomous vehicles, and industrial automation. These workloads demand ultra-low latency and tight coupling between simulation and inference engines.

Blackwell Ultra is engineered for this new class of demands. It enables AI factories to scale compute across the full lifecycle—from massive pretraining runs to highly variable post-training tasks, including fine-tuning, retraining, and real-time inference. Crucially, Nvidia’s Dynamo software stack coordinates these large-scale operations, orchestrating token generation and communication across thousands of GPUs with efficiency that keeps latency low and throughput high.

In this new era, compute isn’t just about speed—it’s about intelligence per watt, adaptability per dollar, and the ability to support inference that behaves less like static prediction and more like dynamic reasoning. Blackwell Ultra and its supporting ecosystem are designed to meet that challenge head-on, reshaping not only how AI runs, but what it can become.

Oracle and NVIDIA Team Up to Accelerate the AI Factory Model with Agentic AI Integration

At NVIDIA’s 2025 GTC conference, Oracle and NVIDIA unveiled a major step forward in the buildout of enterprise-scale AI infrastructure — a key component of the emerging “AI Factory” model. The companies announced a deep integration between Oracle Cloud Infrastructure (OCI) and the NVIDIA AI Enterprise software platform, aimed at accelerating the deployment of agentic AI — autonomous AI systems capable of reasoning, planning, and executing complex tasks.

This collaboration brings NVIDIA’s inference stack — including 160+ AI tools and more than 100 NIM™ (NVIDIA Inference Microservices) — natively into the OCI Console. Oracle customers can now tap into a fully integrated AI stack, available in Oracle’s cloud regions, sovereign clouds, on-premises via OCI Dedicated Region, and even at the edge.

“Oracle has become the platform of choice for both AI training and inferencing,” said Oracle CEO Safra Catz. “This partnership enhances our ability to help customers achieve greater innovation and business results.”

NVIDIA CEO Jensen Huang underscored the significance of the integration for enterprise AI: “Together, we help enterprises innovate with agentic AI to deliver amazing things for their customers and partners.”

No-Code Blueprints and Turnkey Inference

A key element of the Oracle-NVIDIA collaboration is the launch of no-code OCI AI Blueprints, which allow enterprises to deploy multimodal large language models, inference pipelines, and observability tools without managing infrastructure. These blueprints are optimized for NVIDIA GPUs and microservices, and can reduce the time-to-deployment from weeks to minutes.

NVIDIA is also contributing its own Blueprints to the OCI Marketplace, preloaded with workflows for enterprise use cases in customer service, simulation, and robotics. For example, Oracle plans to offer NVIDIA Omniverse and Isaac Sim tools on OCI, bundled with preconfigured NVIDIA L40S GPU instances for simulation and physical AI development.

Pipefy, a business process automation platform, is already deploying multimodal LLMs on OCI using these AI Blueprints. “Using these prepackaged and verified blueprints, deploying our AI models on OCI is now fully automated and significantly faster,” said Gabriel Custódio, principal software engineer at Pipefy.

Enabling Real-Time Inference and Vector Search

Oracle is also integrating NVIDIA NIM microservices into OCI Data Science, enabling real-time inference with a pay-as-you-go model. These microservices can be deployed within a customer’s OCI tenancy for AI use cases ranging from copilots to recommendation engines, delivering rapid time-to-value while maintaining data security and compliance.

In the AI data stack, Oracle Database 23ai now supports accelerated vector search powered by NVIDIA GPUs and the cuVS library — enabling fast creation of vector embeddings and indexes for massive datasets. Companies like DeweyVision, which provides AI-driven media cataloging and search tools, are using this integration to ingest, search, and manage high volumes of video content efficiently.

“Oracle Database 23ai with AI Vector Search can significantly increase Dewey’s search performance while increasing the scalability of the DeweyVision platform,” said CEO Majid Bemanian.

Blackwell-Powered Superclusters Signal the AI Factory Future

Perhaps most notably, Oracle is among the first cloud providers to roll out NVIDIA’s latest generation Blackwell Ultra GPUs across its AI Supercluster. The NVIDIA GB300 NVL72 and HGX B300 NVL16 platforms — successors to last year’s GB200 — promise up to 1.5x performance gains and are designed for large-scale AI factories spanning tens of thousands of GPUs. Oracle’s Supercluster deployments will soon support up to 131,072 GPUs, connected by NVIDIA’s Quantum-2 InfiniBand and NVLink fabrics.

Companies like Soley Therapeutics and SoundHound AI are already leveraging this full-stack Oracle-NVIDIA platform to train next-generation models for drug discovery and voice AI, respectively. “The combination of OCI and NVIDIA delivers a full-stack AI solution,” said Yerem Yeghiazarians, CEO of Soley Therapeutics. “It provides us the storage, compute, software tools, and support necessary to innovate faster with petabytes of data.”

As AI workloads continue to demand ever-larger compute clusters and sophisticated software integration, partnerships like Oracle and NVIDIA’s are laying the foundation for scalable, enterprise-ready AI factories — designed to push the limits of reasoning, automation, and insight.

Secure AI Factories: The Cisco-NVIDIA Collaboration

As AI infrastructure becomes a foundational layer of national and enterprise strategy, its security posture can no longer be an afterthought—it must be embedded from the silicon up. Cisco and NVIDIA have partnered to deliver exactly that with the Secure AI Factory: a full-stack architecture that merges scalable compute and high-performance networking with zero-trust security principles and AI-native threat protection.

The collaboration tightly integrates Cisco’s security and networking stack—including Hypershield, AI Defense, and hybrid mesh firewalls—with NVIDIA’s BlueField-3 DPUs and AI Enterprise platform. The result is a unified framework that provides policy enforcement, observability, and real-time threat detection across every layer of the AI stack.

  • Hypershield applies adaptive segmentation and micro-isolation, using AI to identify and quarantine threats across east-west traffic inside data centers.

  • AI Defense leverages behavior-based analysis to protect against AI-specific risks such as prompt injection, model hijacking, adversarial inputs, and data leakage during runtime.

  • BlueField-3 DPUs offload security and network processing from host CPUs, enabling wire-speed telemetry, access control, and cryptographic operations without impacting AI performance.

This joint platform supports on-premises deployments through Cisco UCS AI servers and Nexus switches, or cloud and hybrid deployments using validated reference architectures optimized for AI factories. Security scales automatically with workload changes—eliminating blind spots in dynamic, multi-tenant environments where AI models evolve in real time.

By embedding security into every node, packet, and process, Cisco and NVIDIA are enabling enterprises to move fast without sacrificing control. In an era where AI models make mission-critical decisions and process sensitive data, the Secure AI Factory ensures that trust is not just assumed—it’s architected.

Chuck Robbins, Chair and CEO, Cisco, said:

Shape
Shape
Stay Ahead

Explore More Insights

More Articles

Stay ahead with more perspectives on cutting-edge power, infrastructure, energy,  bitcoin and AI solutions. Explore these articles to uncover strategies and insights shaping the future of industries.

Shape

6 trends that will shape the future of the cloud: Gartner

For this reason, Gartner recommends identifying specific use cases and planning the applications and data distributed across the organization that could benefit from a cross-cloud deployment model. This allows workloads to operate collaboratively across different cloud platforms, as well as different on-premises and co-location facilities. 4. Industry solutions According to

Read More »

New England Patriots kick off network upgrade

The longer-term roadmap with NWN includes a refresh of the stadium’s 1,800 Extreme Networks Wi-Fi 6 access points to either Wi-Fi 6E or 7, a refresh of the network’s 80 Cisco physical and virtual firewalls, followed by a network consolidation project. On top of all that, the Kraft Group is

Read More »

CompTIA cert targets operational cybersecurity skills

The SecOT+ certification will provide OT professionals with the skills to manage, mitigate, and remediate security risks in manufacturing and critical infrastructure environments, according to CompTIA. The certification program will provide OT positions, such as floor technicians and industrial engineers, as well as cybersecurity engineers and network architects on the

Read More »

Power Moves: Wood’s new vice-president and more

Steph Crawford has been appointed as vice-president of marketing, communications and operations at Aberdeen’s Wood. Crawford has worked at Wood since 2023, having started as marketing and communications manager before moving up to her most recent position, senior marketing and communications manager in August 2024. Wood has been subject to a takeover effort from Sidara, with the Middle Eastern company making a 35p per share offer, valuing the company at £240 million. The group has extended the deadline to make a decision on the bid multiple times. With uncertainty looming about the future of the business, shares in the company cannot be traded due to delays in Wood publishing its financial results. © Supplied by EnerMechEnerMech chief operating officer Tony McAnulty. Tony McAnulty has been appointed as Aberdeen-based EnerMech’s chief operating officer, a new role within the business with a focus on driving market share, operational efficiency and customer satisfaction. McAnulty, who has most recently served as EnerMech’s regional director for Australia and New Zealand, is relocating to Aberdeen to take up the position, which begins immediately. He will oversee the daily operations of the company and is tasked with cultivating a culture of high performance and continuous improvement. EnerMech CEO Charles ‘Chuck’ Davison Jr said: “In order to maintain our industry-leading position, we must be agile and forward-thinking as an organisation. The appointment of Tony will be instrumental in strengthening our global profile and positioning EnerMech for continued success. “With Tony at the helm of operations and a strong leadership team supporting him, I have no doubt that EnerMech will thrive in an increasingly competitive landscape.” McAnulty joined EnerMech in 2022 when it acquired Stork Australia and New Zealand, having previously spent 27-years at the Dutch-headquartered firm, where he held various senior leadership positions across the globe. He added:

Read More »

Phillips 66 Sells Majority Stake in German, Austrian Retail Unit

Pillips 66 has entered into a definitive deal to divest a 65 percent interest in its Germany-Austria retail business to a consortium of investment firms Energy Equation Partners and Stonepeak Partners LP. JET Tankstellen Deutschland GmbH operates 970 sites including 843 that are JET-branded, according to Phillips 66. “Phillips 66 will retain a non-operated 35 percent interest in the business through a newly formed joint venture”, the Houston, Texas-based refiner said in an online statement. “This transaction advances our strategy to optimize our portfolio and enhances long-term shareholder value”, said Phillips 66 chair and chief executive Mark Lashier. “The newly formed joint venture allows us to monetize this non-core asset while retaining the ability to benefit from its future growth”. Phillips 66 expects about EUR 1.5 billion ($1.68 billion) on pre-tax proceeds after customary price adjustments. “The transaction values the Germany and Austria retail marketing business at an enterprise value of approximately EUR 2.5 billion (approximately $2.8 billion), representing an implied Enterprise Value/EBITDA multiple of 9.1x based on expected 2025 EBITDA”, the company said. “The proceeds will be used to support the company’s strategic priorities, including debt reduction and shareholder returns”, it added. Phillips 66 said it would enter into a multi-year agreement to continue supplying the retailer through the Mineraloelraffinerie Oberrhein refinery in Karlsruhe, Germany, in which it owns an 18.75 percent stake. In a separate statement, Stonepeak and Energy Equation noted, “JET is one of the largest fuel retailers in Germany and Austria, serving more than 700,000 customers daily with quality products at fair prices through a network of 970 service stations. Located primarily in urban and high-traffic areas, JET also operates convenience stores, car washes and a rapidly growing EV charging network”. “Together with the outstanding JET team and its dedicated service station operators, we aim to strengthen

Read More »

Oklo looks to Trump to hasten nuclear permitting

Dive Brief: Oklo faces “good uncertainty” as the Trump administration considers executive orders to speed up the Nuclear Regulatory Commission licensing process, expand U.S. military and Department of Energy roles in nuclear deployments and revitalize anemic U.S. nuclear fuel supply chains, CEO Jacob DeWitte said Tuesday. The advanced nuclear technology company is engaged in a “pre-application readiness assessment” with the NRC that could smooth the formal application process for its first commercial reactor later this year, DeWitte said on Oklo’s first-quarter 2025 earnings call. The departure of OpenAI CEO Sam Altman as Oklo board chair earlier this year removes a conflict of interest should OpenAI choose to purchase power in the future from Oklo, which has already announced about 14 GW of nonbinding agreements with data center operators and others, DeWitte said. Dive Insight: Oklo completed key early work this quarter at its first commercial power plant site at Idaho National Laboratory, a 890-square-mile site in Idaho Falls where much of the United States’ nuclear research and commercialization activities take place. There are currently only two or three operational small nuclear reactors in the world, none of them in North America, but many companies competing to design and bring more online. DeWitte said Oklo’s deployment timeline for its first plant has not changed since it’s last investor update in March. Oklo aims to begin producing power at INL in late 2027 or early 2028, it said Tuesday. That target will depend on the outcome of Oklo’s NRC license application, which DeWitte said the company plans to submit in the fourth quarter of 2025. NRC feedback from the pre-application readiness assessment underway now will help Oklo address potential safety issues, information gaps and other items that could slow approval, DeWitte said. The application covers a newly upsized 75-MW reactor design, which

Read More »

Natural Gas Retreats Amid Repeated Substantial Injections, EBW Analyst Says

In an EBW Analytics Group report sent to Rigzone by the EBW team today, Eli Rubin, an energy analyst at the company, outlined that natural gas “retreat[ed]…amid repeated substantial injections”. “Yesterday’s EIA (U.S. Energy Information Administration)-reported 110 billion cubic foot injection (3.9 billion cubic feet per day looser than five-year norms and the third consecutive triple-digit injection in only the first full week of May) is deflating initial enthusiasm following the huge post-tariff April deleveraging sell-off,” Rubin stated in the report. “In our view, the bullish impulses of the gas market may prove correct – but are just early,” he added. The EIA’s latest weekly natural gas storage report, which was released on Thursday and included data for the week ending May 9, stated that “working gas in storage was 2,255 billion cubic feet as of Friday, May 9, 2025, according to EIA estimates”. “This represents a net increase of 110 billion cubic feet from the previous week. Stocks were 375 billion cubic feet less than last year at this time and 57 billion cubic feet above the five-year average of 2,198 billion cubic feet,” the EIA added. The EIA noted in its report that, at 2,255 billion cubic feet, total working gas is within the five-year historical range. Softening Enthusiasm for a Hot Summer In the latest EBW Analytics Group report, Rubin stated that near-term cooling degree days are ticking lower, softening enthusiasm for a hot summer. Rubin added, however, that next week could see late season heating demand in the Upper Midwest support spot demand. “Supply remains subdued and an emerging East Region storage surplus to normal may pressure Marcellus supply,” Rubin said in the report. “This week’s 36.5¢ (minus nine percent) sell-off in the 2025 injection season strip is lowering the projected injection season storage trajectory,” Rubin added. “While

Read More »

Rotterdam: A living case study for a successful hydrogen future

Across continents, governments and industries are looking to hydrogen not just as a future fuel, but as an actionable part of the net-zero equation. The upcoming World Hydrogen Summit 2025, returning to Rotterdam next week, offers a timely and grounded opportunity to examine how hydrogen is moving from concept to implementation and what challenges remain. This year’s event aims to capture the global momentum and highlights the common challenges faced across borders. Rotterdam isn’t just the host city, it exemplifies what a large-scale hydrogen transition looks like in practice. Becoming a focal point for infrastructure development, cross-border hydrogen trade and industrial decarbonisation, the city currently stands as a case study in progress – not without its hurdles, but rich in lessons on systems integration, permitting, investment planning and cross-sector collaboration. Rotterdam is a real-world checkpoint for the international hydrogen movement. From pipelines and storage hubs to import terminals and hydrogen-powered transport, the city is demonstrating how theory translates into industrial action. Projects like Shell’s Holland Hydrogen 1, Europe’s first major renewable hydrogen plant at 200 MW, has reached advanced construction, and is now connected to the high-voltage grid and linked to a growing hydrogen pipeline network. Meanwhile, Air Liquide’s ELYgator is set to add another 200 MW of electrolyser capacity by 2027, powered by offshore wind and serving the refining sector. The 2025 theme, “Navigating Challenges with Solid FIDs”, reflects an increasingly important milestone for the hydrogen sector: reaching final investment decision (FID). As macroeconomic pressures, regulatory uncertainty, and infrastructure constraints persist, the summit will explore what it takes for nations to move beyond feasibility studies to fully committed projects. Sessions led by industry experts and energy ministers will unpack case studies from around the globe, with a focus on how developers have overcome bottlenecks, particularly around demand aggregation, policy

Read More »

Philippines Charts Ambitious Plan to Export Green Power

The Philippines is looking at exporting renewable energy to Taiwan and other countries, Energy Secretary Raphael Lotilla said, a lofty strategy that would take advantage of surplus capacity. There are ongoing discussions between officials in Manila and Taipei, he said. “It’s not only Taiwan that has approached us but other countries in anticipation of excess capacity in renewable energy,” Lotilla told a media briefing on Thursday. Still, the strategy is among the most ambitious efforts to export clean energy in Asia. The region has struggled and largely failed to create cross-border transmission networks despite decades of proposals due to political and technical hurdles. Talks with those in Southeast Asia are being made in the context of the Association of Southeast Asian Nations’ power grid, an initiative to interconnect electricity grids in the region, according to Lotilla. “In the case of Taiwan and others which are not members of Asean, then we will have to look further into the regional framework for those.” The Philippines is among Southeast Asian countries with aggressive plans for clean-power projects with the government easing investment restrictions and policies to attract domestic and foreign cash. The nation is aiming to increase the share of renewable energy in its power generation mix to 35 percent by 2030 and 50 percent by 2040. But it had to pause from accepting and processing renewable contracts for five months last year following a flood of applications. Taiwan has been considering building renewable power plants in nearby countries like the Philippines and importing the generated electricity to feed the needs of its manufacturers.  Taiwan’s Economic Minister Kuo Jyh-huei said earlier this month that green energy imports from the Philippines could help reduce the carbon footprint of Taiwanese exporters as costs could remain under NT$5 (17 US cents) per kilowatt-hour.  The Philippines’ Department of

Read More »

US companies are helping Saudi Arabia to build an AI powerhouse

AMD announced a five-year, $10 billion collaboration with Humain to deploy up to 500 megawatts of AI compute in Saudi Arabia and the US, aiming to deploy “multi-exaflop capacity by early 2026.” AWS, too, is expanding its data centers in Saudi Arabia to bolster Humain’s cloud infrastructure. Saudi Arabia has abundant oil and gas to power those data centers, and is growing its renewable energy resources with the goal of supplying 50% of the country’s power by 2030. “Commercial electricity rates, nearly 50% lower than in the US, offer potential cost savings for AI model training, though high local hosting costs due to land, talent, and infrastructure limit total savings,” said Eric Samuel, Associate Director at IDC. Located near Middle Eastern population centers and fiber optic cables to Asia, these data centers will offer enterprises low-latency cloud computing for real-time AI applications. Late is great There’s an advantage to being a relative latecomer to the technology industry, said Eric Samuel, associate director, research at IDC. “Saudi Arabia’s greenfield tech landscape offers a unique opportunity for rapid, ground-up AI integration, unburdened by legacy systems,” he said.

Read More »

AMD, Nvidia partner with Saudi startup to build multi-billion dollar AI service centers

Humain will deploy the Nvidia Omniverse platform as a multi-tenant system to drive acceleration of the new era of physical AI and robotics through simulation, optimization and operation of physical environments by new human-AI-led solutions. The AMD deal did not discuss the number of chips involved in the deal, but it is valued at $10 billion. AMD and Humain plan to develop a comprehensive AI infrastructure through a network of AMD-based AI data centers that will extend from Saudi Arabia to the US and support a wide range of AI workloads across corporate, start-up, and government markets. Think of it as AWS but only offering AI as a service. AMD will provide its AI compute portfolio – Epyc, Instinct, and FPGA networking — and the AMD ROCm open software ecosystem, while Humain will manage the delivery of the hyperscale data center, sustainable power systems, and global fiber interconnects. The partners expect to activate a multi-exaflop network by early 2026, supported by next-generation AI silicon, modular data center zones, and a software platform stack focused on developer enablement, open standards, and interoperability. Amazon Web Services also got a piece of the action, announcing a more than $5 billion investment to build an “AI zone” in the Kingdom. The zone is the first of its kind and will bring together multiple capabilities, including dedicated AWS AI infrastructure and servers, UltraCluster networks for faster AI training and inference, AWS services like SageMaker and Bedrock, and AI application services such as Amazon Q. Like the AMD project, the zone will be available in 2026. Humain only emerged this month, so little is known about it. But given that it is backed by Crown Prince Salman and has the full weight of the Kingdom’s Public Investment Fund (PIF), which ranks among the world’s largest and

Read More »

Check Point CISO: Network segregation can prevent blackouts, disruptions

Fischbein agrees 100% with his colleague’s analysis and adds that education and training can help prevent such incidents from occurring. “Simulating such a blackout is impossible, it has never been done,” he acknowledges, but he is committed to strengthening personal and team training and risk awareness. Increased defense and cybersecurity budgets In 2025, industry watchers expect there will be an increase in the public budget allocated to defense. In Spain, one-third of the budget will be allocated to increasing cybersecurity. But for Fischbein, training teams is much more important than the budget. “The challenge is to distribute the budget in a way that can be managed,” he notes, and to leverage intuitive and easy-to-use platforms, so that organizations don’t have to invest all the money in training. “When you have information, management, users, devices, mobiles, data centers, clouds, cameras, printers… the security challenge is very complex. You have to look for a security platform that makes things easier, faster, and simpler,” he says. ” Today there are excellent tools that can stop all kinds of attacks.” “Since 2010, there have been cybersecurity systems, also from Check Point, that help prevent this type of incident from happening, but I’m not sure that [Spain’s electricity blackout] was a cyberattack.” Leading the way in email security According to Gartner’s Magic Quadrant, Check Point is the leader in email security platforms. Today email is still responsible for 88% of all malicious file distributions. Attacks that, as Fischbein explains, enter through phishing, spam, SMS, or QR codes. “There are two challenges: to stop the threats and not to disturb, because if the security tool is a nuisance it causes more harm than good. It is very important that the solution does not annoy [users],” he stresses. “As almost all attacks enter via e-mail, it is

Read More »

HPE ‘morphs’ private cloud portfolio with improved virtualization, storage and data protection

What do you get when combining Morpheus with Aruba? As part of the extensible platform message that HPE is promoting with Morpheus, it’s also working in some capabilities from the broader HPE portfolio. One integration is with HPE Aruba for networking microsegmentation. Bhardwaj noted that a lot of HPE Morpheus users are looking for microsegmentation in order to make sure that the traffic between two virtual machines on a server is secure. “The traditional approach of doing that is on the hypervisor, but that costs cycles on the hypervisor,” Bhardwaj said. “Frankly, the way that’s being delivered today, customers have to pay extra cost on the server.” With the HPE Aruba plugin that now works with HPE Morpheus, the microsegmentation capability can be enabled at the switch level. Bhardwaj said that by doing the microsegmentation in the switch and not the hypervisor, costs can be lowered and performance can be increased. The integration brings additional capabilities, including the ability to support VPN and network address translation (NAT) in an integrated way between the switch and the hypervisor. VMware isn’t the only hypervisor supported by HPE  The HPE Morpheus VM Essentials Hypervisor is another new element in the HPE cloud portfolio. The hypervisor is now being integrated into HPE’s private cloud offerings for both data center as well as edge deployments.

Read More »

AMD targets hosting providers with affordable EPYC 4005 processors

According to Pinkesh Kotecha, chairman and MD of Ishan Technologies, AMD’s 4th Gen EPYC processors stood out because they offer the right combination of high performance, energy efficiency, and security. “Their high core density and ability to optimize performance per watt made them ideal for managing data-intensive operations like real-time analytics and high-frequency transactions. Additionally, AMD’s strong AI roadmap and growing portfolio of AI-optimised solutions position them as a forward-looking partner, ready to support our customers’ evolving AI and data needs. This alignment made AMD a clear choice over alternatives,” Kotecha said. By integrating AMD EPYC processors, Ishan Technologies’ Ishan Cloud plans to empower enterprises across BFSI, ITeS, and manufacturing industries, as well as global capability centers and government organizations, to meet India’s data localization requirements and drive AI-led digital transformation. “The AMD EPYC 4005 series’ price-to-performance ratio makes it an attractive option for cloud hosting and web services, where cost-efficient, always-on performance is essential,” said Manish Rawat, analyst, TechInsights. Prabhu Ram, VP for the industry research group at CMR, said EPYC 4005 processors deliver a compelling mix of performance-per-watt, higher core counts, and modern I/O support, positioning it as a strong alternative to Intel’s Xeon E-2400 and 6300P, particularly for edge deployments. Shah of Counterpoint added, “While ARM-based Ampere Altra promises higher power efficiencies and is ideally adopted in more cloud and hyperscale data centers, though performance is something where x86-based Zen 5 architecture excels and nicely balances the efficiencies with lower TDPs, better software compatibilities supported by a more mature ecosystem.”

Read More »

Shell’s immersive cooling liquids the first to receive official certification from Intel

Along with the certification, Intel is offering a Xeon processor single-phase immersion warranty rider. This indicates Intel’s confidence in the durability and effectiveness of Shell’s fluids. Yates explained that the rider augments Intel’s standard warranty terms and is available to data center operators deploying 4th and 5th generation Xeon processors in Shell immersion fluids. The rider is intended to provide data center operators confidence that their investment is guaranteed when deployed correctly. Shell’s fluids are available globally and can be employed in retrofitted existing infrastructure or used in new builds. Cuts resource use, increases performance Data centers consume anywhere from 10 to 50 times more energy per square foot than traditional office buildings, and they are projected to drive more than 20% of the growth in electricity demand between now and 2030. Largely due to the explosion of AI, data center energy consumption is expected to double from 415 terawatt-hours in 2024 to around 945 TWh by 2030. There are several other technologies used for data center cooling, including air cooling, cold plate (direct-to-chip), and precision cooling (targeted to specific areas), but the use of immersion cooling has been growing, and is expected to account for 36% of data center thermal management revenue by 2028. With this method, servers and networking equipment are placed in cooling fluids that absorb and dissipate heat generated by the electronic equipment. These specialized fluids are thermally conductive but not electrically conductive (dielectric) thus making them safe for submerging electrical equipment.

Read More »

Microsoft will invest $80B in AI data centers in fiscal 2025

And Microsoft isn’t the only one that is ramping up its investments into AI-enabled data centers. Rival cloud service providers are all investing in either upgrading or opening new data centers to capture a larger chunk of business from developers and users of large language models (LLMs).  In a report published in October 2024, Bloomberg Intelligence estimated that demand for generative AI would push Microsoft, AWS, Google, Oracle, Meta, and Apple would between them devote $200 billion to capex in 2025, up from $110 billion in 2023. Microsoft is one of the biggest spenders, followed closely by Google and AWS, Bloomberg Intelligence said. Its estimate of Microsoft’s capital spending on AI, at $62.4 billion for calendar 2025, is lower than Smith’s claim that the company will invest $80 billion in the fiscal year to June 30, 2025. Both figures, though, are way higher than Microsoft’s 2020 capital expenditure of “just” $17.6 billion. The majority of the increased spending is tied to cloud services and the expansion of AI infrastructure needed to provide compute capacity for OpenAI workloads. Separately, last October Amazon CEO Andy Jassy said his company planned total capex spend of $75 billion in 2024 and even more in 2025, with much of it going to AWS, its cloud computing division.

Read More »

John Deere unveils more autonomous farm machines to address skill labor shortage

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Self-driving tractors might be the path to self-driving cars. John Deere has revealed a new line of autonomous machines and tech across agriculture, construction and commercial landscaping. The Moline, Illinois-based John Deere has been in business for 187 years, yet it’s been a regular as a non-tech company showing off technology at the big tech trade show in Las Vegas and is back at CES 2025 with more autonomous tractors and other vehicles. This is not something we usually cover, but John Deere has a lot of data that is interesting in the big picture of tech. The message from the company is that there aren’t enough skilled farm laborers to do the work that its customers need. It’s been a challenge for most of the last two decades, said Jahmy Hindman, CTO at John Deere, in a briefing. Much of the tech will come this fall and after that. He noted that the average farmer in the U.S. is over 58 and works 12 to 18 hours a day to grow food for us. And he said the American Farm Bureau Federation estimates there are roughly 2.4 million farm jobs that need to be filled annually; and the agricultural work force continues to shrink. (This is my hint to the anti-immigration crowd). John Deere’s autonomous 9RX Tractor. Farmers can oversee it using an app. While each of these industries experiences their own set of challenges, a commonality across all is skilled labor availability. In construction, about 80% percent of contractors struggle to find skilled labor. And in commercial landscaping, 86% of landscaping business owners can’t find labor to fill open positions, he said. “They have to figure out how to do

Read More »

2025 playbook for enterprise AI success, from agents to evals

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More 2025 is poised to be a pivotal year for enterprise AI. The past year has seen rapid innovation, and this year will see the same. This has made it more critical than ever to revisit your AI strategy to stay competitive and create value for your customers. From scaling AI agents to optimizing costs, here are the five critical areas enterprises should prioritize for their AI strategy this year. 1. Agents: the next generation of automation AI agents are no longer theoretical. In 2025, they’re indispensable tools for enterprises looking to streamline operations and enhance customer interactions. Unlike traditional software, agents powered by large language models (LLMs) can make nuanced decisions, navigate complex multi-step tasks, and integrate seamlessly with tools and APIs. At the start of 2024, agents were not ready for prime time, making frustrating mistakes like hallucinating URLs. They started getting better as frontier large language models themselves improved. “Let me put it this way,” said Sam Witteveen, cofounder of Red Dragon, a company that develops agents for companies, and that recently reviewed the 48 agents it built last year. “Interestingly, the ones that we built at the start of the year, a lot of those worked way better at the end of the year just because the models got better.” Witteveen shared this in the video podcast we filmed to discuss these five big trends in detail. Models are getting better and hallucinating less, and they’re also being trained to do agentic tasks. Another feature that the model providers are researching is a way to use the LLM as a judge, and as models get cheaper (something we’ll cover below), companies can use three or more models to

Read More »

OpenAI’s red teaming innovations define new essentials for security leaders in the AI era

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More OpenAI has taken a more aggressive approach to red teaming than its AI competitors, demonstrating its security teams’ advanced capabilities in two areas: multi-step reinforcement and external red teaming. OpenAI recently released two papers that set a new competitive standard for improving the quality, reliability and safety of AI models in these two techniques and more. The first paper, “OpenAI’s Approach to External Red Teaming for AI Models and Systems,” reports that specialized teams outside the company have proven effective in uncovering vulnerabilities that might otherwise have made it into a released model because in-house testing techniques may have missed them. In the second paper, “Diverse and Effective Red Teaming with Auto-Generated Rewards and Multi-Step Reinforcement Learning,” OpenAI introduces an automated framework that relies on iterative reinforcement learning to generate a broad spectrum of novel, wide-ranging attacks. Going all-in on red teaming pays practical, competitive dividends It’s encouraging to see competitive intensity in red teaming growing among AI companies. When Anthropic released its AI red team guidelines in June of last year, it joined AI providers including Google, Microsoft, Nvidia, OpenAI, and even the U.S.’s National Institute of Standards and Technology (NIST), which all had released red teaming frameworks. Investing heavily in red teaming yields tangible benefits for security leaders in any organization. OpenAI’s paper on external red teaming provides a detailed analysis of how the company strives to create specialized external teams that include cybersecurity and subject matter experts. The goal is to see if knowledgeable external teams can defeat models’ security perimeters and find gaps in their security, biases and controls that prompt-based testing couldn’t find. What makes OpenAI’s recent papers noteworthy is how well they define using human-in-the-middle

Read More »
More Articles