FinOps for AI: How Companies Can Control AI Infrastructure Costs Artificial intelligence is becoming a strategic investment for businesses across industries. Companies are deploying generative AI applications, machine learning models, intelligent automation, and AI-powered analytics to improve productivity and accelerate innovation. However, these capabilities introduce new infrastructure expenses, including GPU computing, cloud services, data storage, networking, and model inference. As AI adoption expands, organizations need effective ways to manage these costs without compromising performance. This is where FinOps for AI becomes essential. FinOps for AI combines financial accountability, cloud cost management, and engineering optimization to help businesses understand, allocate, and control spending on AI infrastructure. It enables finance, technology, and business teams to make informed decisions about AI investments while connecting infrastructure expenses to measurable business outcomes. Why FinOps Matters for AI Infrastructure Traditional cloud cost management often focuses on virtual machines, storage, and predictable workloads. AI infrastructure introduces additional complexity because GPU resources can be expensive, workloads fluctuate, and model training may require substantial computing power. Inference costs can also grow rapidly as more employees and customers use AI applications. Organizations may pay for idle GPU instances, oversized models, duplicated experiments, or inefficient data pipelines without realizing how much these activities affect their budgets. FinOps provides a structured approach to identifying these expenses, improving resource utilization, and ensuring that AI investments deliver business value. Key Strategies for Controlling AI Infrastructure Costs 1. Establish Clear Visibility Into AI Spending Companies cannot control expenses they cannot accurately measure. Organizations should track costs across cloud providers, GPU clusters, AI platforms, development environments, and individual applications. Using cloud billing dashboards, infrastructure monitoring tools, and resource tagging helps teams understand which departments, projects, and workloads consume the most resources. Businesses should also establish key performance indicators such as cost per inference request, cost per training run, GPU utilization, and cost per business outcome. These measurements make it easier to identify waste and compare infrastructure alternatives. 2. Optimize GPU Utilization Graphics processing units are essential for many AI workloads, but they can become a significant source of unnecessary spending when poorly managed. Reserved GPU capacity that remains idle can increase operating costs without improving productivity. Companies can use workload scheduling, automatic scaling, and resource-sharing strategies to improve utilization. Batch processing can reduce the cost of suitable workloads, while shutting down unused development environments can prevent unnecessary charges. Organizations should also select GPU configurations based on actual workload requirements rather than automatically choosing the most powerful available hardware. 3. Choose the Right AI Models Larger AI models are not always necessary for every business application. A complex model may deliver excellent performance but introduce higher inference costs and latency. Companies can evaluate smaller language models, model quantization, caching, and optimized inference frameworks to reduce computing requirements. Routing simple tasks to smaller models and reserving larger models for complex requests can further improve cost efficiency. The goal is to balance model quality, response speed, reliability, and total operating expense. 4. Improve Cloud Resource Management Cloud infrastructure provides flexibility, but unmanaged consumption can lead to unpredictable bills. Organizations should review pricing options, including on-demand instances, reserved capacity, and spot resources where appropriate. Predictable workloads may benefit from committed-use discounts, while experimental projects may require flexible pricing. Hybrid infrastructure can also help companies balance owned hardware with rented cloud capacity. Regular cost reviews and automated budget alerts can identify unusual spending before it becomes a larger financial problem. 5. Connect Infrastructure Costs to Business Outcomes Reducing spending alone does not guarantee success. Companies must determine whether their AI investments generate sufficient value. For example, an AI-powered customer support system may increase infrastructure expenses while reducing manual workloads and improving response times. Evaluating these outcomes alongside operating costs provides a more complete picture of return on investment. Finance, engineering, and business teams should collaborate to establish shared budgets, performance targets, and accountability for AI spending. Building a Sustainable FinOps Framework for AI Successful FinOps requires continuous monitoring rather than a one-time cost reduction exercise. Organizations should establish spending policies, review resource utilization regularly, forecast future demand, and assign clear ownership to AI projects. Automated alerts, centralized dashboards, and consistent reporting can help teams identify inefficiencies early. As workloads evolve, companies should reassess model choices, infrastructure configurations, and pricing strategies. Conclusion FinOps for AI helps organizations control infrastructure costs while supporting innovation and scalability. By improving spending visibility, optimizing GPU utilization, selecting efficient models, managing cloud resources, and measuring business outcomes, companies can reduce waste and strengthen financial accountability. As enterprise AI adoption grows, organizations that integrate FinOps into their AI strategy will be better positioned to manage unpredictable expenses, maximize infrastructure investments, and build financially sustainable AI operations. Read More: https://theinfotech.info/

GPU Infrastructure: The Hidden Cost of Enterprise AI Enterprise artificial intelligence is transforming how organizations automate operations, analyze data, and develop innovative products. From generative AI applications to advanced machine learning models, businesses are investing heavily in AI capabilities to remain competitive. However, behind these advancements lies a significant financial challenge: GPU infrastructure. While companies often focus on AI software, model performance, and business outcomes, the underlying infrastructure can introduce substantial costs that are difficult to predict and manage. Why GPU Infrastructure Matters for Enterprise AI Graphics processing units (GPUs) are specialized processors designed to handle large numbers of computations simultaneously. They are essential for training complex AI models, running inference workloads, and supporting applications such as large language models, computer vision, and predictive analytics. Unlike traditional computing infrastructure, AI workloads can require substantial processing power, high-speed memory, and rapid data movement. Enterprises may deploy GPUs in their own data centers, rent capacity from cloud providers, or adopt a hybrid infrastructure model. Each approach has different financial and operational implications. As AI adoption grows, organizations must look beyond the initial cost of acquiring or renting GPUs and evaluate the total cost of ownership. The Major Hidden Costs of GPU Infrastructure 1. High Hardware Acquisition Costs Enterprise-grade GPUs can require significant upfront investment, particularly when organizations need multiple accelerators for large-scale AI workloads. The total expense extends beyond the GPUs themselves to servers, networking equipment, storage systems, and compatible components. Businesses must also account for hardware replacement cycles and the possibility that newer GPU generations may deliver better performance or energy efficiency. Purchasing infrastructure without a clear workload strategy can result in expensive resources that deliver limited business value. 2. Energy Consumption and Cooling GPUs consume substantial electricity when operating at high utilization. Large AI clusters can significantly increase data center power requirements, while cooling systems add another layer of operational expense. Organizations running their own infrastructure must consider electricity prices, cooling efficiency, power availability, and facility capacity. These expenses can become particularly challenging as AI workloads expand. Improving workload scheduling, consolidating infrastructure, and monitoring power consumption can help businesses reduce unnecessary energy expenditure. 3. Cloud GPU Rental Expenses Cloud platforms allow companies to access GPU resources without purchasing physical hardware. This flexibility can accelerate experimentation and reduce initial capital requirements. However, hourly GPU charges, data transfer fees, storage costs, and managed service expenses can accumulate quickly. Unused instances, oversized configurations, and poorly optimized workloads may increase cloud bills without improving results. Enterprises should regularly review GPU utilization and select appropriate pricing models for predictable and variable workloads. 4. Underutilization and Resource Waste GPU infrastructure is expensive even when it is not being used effectively. Some organizations reserve more capacity than necessary, while others experience bottlenecks caused by insufficient memory, slow data pipelines, or inefficient model configurations. Low utilization increases the cost per completed training job or AI prediction. Workload profiling, resource monitoring, and intelligent scheduling can help organizations maximize the value of their existing infrastructure. 5. Software, Networking, and Engineering Costs GPU infrastructure also depends on software frameworks, orchestration tools, high-speed networking, and specialized engineering expertise. Teams may need to manage distributed training, optimize inference workloads, maintain drivers, and troubleshoot performance issues. These indirect expenses are often overlooked during initial budgeting, even though they influence long-term operating costs and deployment timelines. Strategies to Control GPU Infrastructure Costs Enterprises can improve cost efficiency by matching GPU capacity to specific business requirements. Smaller models, quantization, caching, and optimized inference pipelines may reduce the computing power required for certain applications. Organizations should also establish FinOps practices for AI infrastructure, combining financial accountability with engineering performance data. Tracking cost per training run, cost per inference request, GPU utilization, and application-level business outcomes helps decision-makers evaluate spending more accurately. A hybrid infrastructure strategy may also be beneficial. Predictable workloads can run on reserved or owned capacity, while temporary experimentation and demand spikes can use cloud-based resources. Conclusion GPU infrastructure is a critical foundation for enterprise AI, but its true cost extends far beyond hardware or cloud rental fees. Energy consumption, cooling, underutilization, networking, software, and engineering requirements all influence the economics of AI deployment. Businesses that evaluate total cost of ownership and continuously optimize infrastructure can reduce waste while maintaining performance. As enterprise AI adoption accelerates, effective GPU cost management will become an essential part of building scalable, reliable, and financially sustainable AI operations. Read More: https://theinfotech.info/

Hybrid Cloud in the AI Era: Building a Flexible and Intelligent IT Infrastructure Artificial intelligence (AI) is transforming how businesses manage data, automate operations, and deliver digital services. As organizations adopt generative AI, machine learning, and AI-powered applications, their cloud infrastructure requirements are becoming more complex. Public cloud platforms offer scalability and advanced computing capabilities, while private cloud environments provide greater control over sensitive data and critical workloads. Hybrid cloud combines these approaches, helping businesses build flexible, secure, and cost-effective IT environments for the AI era. Why Hybrid Cloud Matters for AI AI applications require significant computing power, large datasets, and reliable infrastructure. However, not every workload is suitable for a single cloud environment. Businesses must balance performance, cost, security, and regulatory requirements when deploying AI technologies. Hybrid cloud enables organizations to distribute workloads across public cloud, private cloud, and on-premises infrastructure. For example, a company can use private infrastructure to store confidential customer information while leveraging public cloud resources to train AI models or process large datasets. This flexibility allows organizations to adopt AI without completely replacing their existing IT systems. It also supports gradual modernization, enabling businesses to introduce new technologies while maintaining established applications and operations. Key Benefits of Hybrid Cloud in the AI Era 1. Improved Data Security and Compliance Data security remains a major concern as businesses expand their AI capabilities. Hybrid cloud allows organizations to retain sensitive information within controlled environments while using public cloud services for less sensitive workloads. This approach can help businesses meet industry-specific regulations and internal governance requirements. Financial institutions, healthcare providers, and government agencies can use hybrid architectures to maintain greater control over confidential information while exploring AI-driven innovation. 2. Greater Scalability and Performance AI workloads can fluctuate significantly depending on training requirements, user demand, and application complexity. Hybrid cloud enables organizations to scale computing resources when needed without maintaining all infrastructure on-premises. Businesses can use public cloud infrastructure for resource-intensive AI training and deploy inference workloads closer to users or data sources when lower latency is important. This flexibility supports faster application delivery and more responsive digital services. 3. Cost Optimization Building and maintaining dedicated infrastructure for every AI workload can be expensive. Hybrid cloud gives businesses more options for managing computing costs by combining existing infrastructure with cloud-based resources. Organizations can run predictable workloads on private infrastructure and use public cloud capacity for temporary demand spikes. However, effective cost management requires monitoring storage, data transfer, computing consumption, and AI service expenses. 4. Support for Legacy Systems Many enterprises depend on legacy applications that cannot be easily migrated to the public cloud. Hybrid cloud allows these systems to operate alongside modern AI applications, reducing disruption during digital transformation. By integrating existing databases, business applications, and cloud-based AI services, organizations can introduce intelligent automation and analytics without undertaking a complete infrastructure overhaul. Challenges Businesses Must Address Despite its advantages, hybrid cloud introduces operational complexity. Managing multiple environments can create difficulties involving data integration, application compatibility, security policies, and performance monitoring. AI also increases the importance of data governance. Organizations need clear policies covering data access, model training, information retention, and the movement of sensitive information between environments. Another challenge is skills availability. IT teams must understand cloud architecture, cybersecurity, AI infrastructure, and automation to manage hybrid environments effectively. Consistent monitoring and centralized management tools can help address these challenges. Best Practices for Hybrid Cloud Adoption Businesses should begin by identifying which workloads require strict security controls, high computing capacity, or low latency. This assessment helps determine where applications and data should reside. Organizations should also establish consistent identity management, encryption, access controls, and monitoring across all environments. Automation can simplify workload deployment and reduce configuration errors. Finally, companies should evaluate infrastructure costs and performance regularly. A clear hybrid cloud strategy should align technology investments with measurable business objectives, such as faster processing, improved customer experiences, and reduced operational costs. Conclusion Hybrid cloud is becoming an important foundation for enterprise AI adoption. By combining public cloud scalability with private infrastructure control, businesses can deploy AI applications while balancing security, performance, and cost. As AI technologies continue to evolve, organizations that develop well-governed, integrated, and adaptable hybrid cloud environments will be better positioned to modernize operations and respond to changing market demands. Success will depend not simply on adopting multiple cloud platforms, but on managing them as a unified infrastructure that supports long-term business innovation. Read More: https://theinfotech.info/

Confidential Computing And The Future Of Data Protection As organizations move more workloads to cloud platforms and increasingly rely on artificial intelligence, protecting sensitive data throughout its entire lifecycle has become a major priority. Traditional security measures typically focus on protecting data when it is stored or transmitted, but data can remain exposed while it is actively being processed. Confidential computing addresses this challenge by protecting data while it is being used, creating a stronger security model for modern cloud and enterprise environments. What Is Confidential Computing? Confidential computing is a security approach that protects data during processing by placing workloads inside a trusted execution environment (TEE). A TEE is an isolated area of a processor where applications can process sensitive information while limiting access from the underlying operating system, hypervisor, cloud provider, or other unauthorized processes. This creates protection across three key stages of the data lifecycle: data at rest, data in transit, and data in use. Encryption has traditionally provided strong protection for stored and transmitted information, while confidential computing extends that protection to data being actively processed. Why Confidential Computing Matters The growing adoption of cloud computing has created new security challenges. Businesses often rely on third-party cloud infrastructure to store and process financial information, customer records, intellectual property, healthcare data, and other sensitive information. Even when cloud environments are properly secured, organizations may have concerns about who or what can access data during processing. Confidential computing can reduce these risks by isolating sensitive workloads from privileged infrastructure access. For enterprises, this can support stronger security controls without requiring them to keep every sensitive workload inside their own physical data centers. Confidential Computing And AI The rapid development of generative AI and enterprise AI is making confidential computing increasingly relevant. AI systems frequently process proprietary business information, customer data, source code, financial records, and other sensitive datasets. Organizations may want to use cloud-based AI infrastructure without exposing confidential information to unauthorized systems or infrastructure operators. Confidential computing can help create protected environments where sensitive AI workloads and data can be processed with additional hardware-based security controls. This could become particularly important as businesses deploy AI agents capable of accessing enterprise applications and making decisions based on internal information. Key Enterprise Applications Confidential computing can support several use cases across industries: Financial services: Protecting transaction information, financial models, and customer data during processing. Healthcare: Supporting secure analysis of sensitive patient and clinical information. Artificial intelligence: Protecting proprietary datasets, models, and prompts used by AI applications. Cloud computing: Adding another security layer for workloads hosted on shared infrastructure. Government: Supporting secure processing of sensitive public-sector information. Collaborative analytics: Allowing organizations to analyze sensitive data while limiting exposure between participating parties. Challenges To Adoption Despite its potential, confidential computing is not a complete replacement for traditional cybersecurity. Organizations still need identity management, encryption, secure software development, network security, monitoring, and access controls. Compatibility and performance can also influence adoption. Businesses need infrastructure, applications, and development tools that support confidential computing technologies. Teams may also require new skills to design and manage workloads using trusted execution environments. Another consideration is the complexity of building an end-to-end trusted environment. Protecting the processing layer does not automatically protect applications from vulnerabilities, compromised credentials, malicious code, or poor security practices. The Future Of Data Protection As cloud infrastructure, AI, edge computing, and data-intensive applications continue to expand, protecting data only when it is stored or transmitted may no longer be sufficient. Confidential computing provides an additional security layer by addressing the protection of data while it is actively being processed. In the coming years, confidential computing is likely to become an important component of enterprise security architectures, particularly for organizations handling highly sensitive information. Combined with encryption, zero-trust security, identity controls, and continuous monitoring, it can help businesses build stronger defenses around critical workloads. The future of data protection will increasingly depend not only on where data is stored, but also on how securely it is processed. Confidential computing offers a pathway toward that more comprehensive security model. Read More:https://theinfotech.info/

AI Infrastructure: The New Enterprise Technology Backbone Artificial intelligence is rapidly becoming a core component of enterprise technology. From generative AI and intelligent automation to predictive analytics and AI agents, organizations are embedding AI into everyday operations. As adoption grows, traditional IT infrastructure is evolving into something more capable: an AI infrastructure backbone designed to support large-scale data processing, model development, deployment, and real-time AI workloads. What Is AI Infrastructure? AI infrastructure refers to the combination of hardware, software, cloud platforms, networking, data systems, and specialized computing resources required to develop and operate AI applications. Unlike conventional enterprise infrastructure, AI workloads often require significant computing power, high-speed data movement, and specialized processors. Modern AI infrastructure can include GPUs, AI accelerators, high-performance CPUs, cloud computing platforms, data storage systems, networking technologies, model-management platforms, and security tools. As enterprises move AI applications from experimentation into production, infrastructure is becoming a strategic technology investment rather than simply an IT requirement. Why AI Infrastructure Is Becoming Critical The growing use of large language models, machine learning applications, computer vision, and AI agents is increasing demand for scalable computing environments. Enterprises need infrastructure that can process large datasets while supporting increasingly complex AI workloads. Cloud platforms have helped organizations access AI computing without building every component themselves. At the same time, many companies are adopting hybrid and multi-cloud approaches to balance performance, cost, security, and data-residency requirements. The result is a new infrastructure model where AI workloads influence decisions about servers, storage, networking, data architecture, and cloud strategy. GPUs and AI Accelerators Drive the Transformation Traditional CPUs remain important for enterprise applications, but AI workloads often require specialized acceleration. GPUs and other AI accelerators can process large numbers of calculations simultaneously, making them suitable for model training and inference. This has created a growing ecosystem around accelerated computing. Enterprises are evaluating infrastructure based not only on processor performance but also on memory capacity, interconnect speeds, energy consumption, and workload efficiency. For organizations operating AI at scale, infrastructure optimization can directly influence application performance and operational costs. Data Is at the Center of AI Infrastructure AI systems depend heavily on data. Enterprises need infrastructure capable of collecting, storing, processing, governing, and delivering data to AI applications efficiently. Modern AI infrastructure therefore increasingly connects data lakes, warehouses, vector databases, analytics platforms, and enterprise applications. Data quality and accessibility are equally important because AI models can produce unreliable results when they are trained or supplied with incomplete, outdated, or poorly governed information. Strong data governance also helps organizations address privacy, compliance, and security requirements. Security Becomes a Core Requirement As AI becomes integrated into enterprise systems, infrastructure security becomes increasingly important. Organizations must protect models, training data, APIs, workloads, and the underlying computing environment. AI infrastructure strategies increasingly include identity management, encryption, access controls, network security, monitoring, and workload isolation. Enterprises also need safeguards against threats such as data leakage, unauthorized model access, prompt-based attacks, and compromised AI agents. The Future of Enterprise AI Infrastructure AI infrastructure is moving toward greater automation and flexibility. Organizations are increasingly looking for platforms that can dynamically allocate computing resources, monitor workloads, optimize costs, and support AI applications across cloud, on-premises, and edge environments. The infrastructure layer will also become important as enterprises deploy autonomous AI agents capable of interacting with business systems and performing multi-step tasks. Ultimately, AI infrastructure is becoming more than a collection of servers and cloud services. It is emerging as a technology backbone for AI-driven enterprises, connecting computing, data, applications, security, and automation. Companies that build scalable and well-governed infrastructure can create a stronger foundation for moving AI from experimental projects into reliable, production-level business applications. Read More: https://theinfotech.info/

The Trade Desk Launches Kokai Zuma, the Latest Release of Kokai The Trade Desk, a leading global advertising technology company, introduced Kokai Zuma, the latest release of its Kokai platform. Introduced in 2023, Kokai is The Trade Desk's platform for planning, buying, and measuring advertising across the open internet. The Zuma release makes Kokai easier to navigate, learn, and use by bringing together new agentic AI capabilities and a simplified measurement framework. As advertising becomes increasingly fragmented across channels, formats, and devices, media buyers face greater complexity in managing campaigns and understanding what drives performance. Zuma is designed to reduce that friction by helping traders find what they need faster, make changes with greater confidence, and spend less time managing workflows so they can focus on optimizing campaigns and improving results. The update also builds on the AI forecasting engine that has been at the heart of Kokai. Upgraded AI and Forecasting Infrastructure While the redesigned interface is the most visible change, The Trade Desk has also invested heavily in upgrading Kokai's forecasting engine and underlying AI infrastructure. The technology is designed to: Predict available advertising inventory. Model campaign outcomes. Power Koa's agentic capabilities in real time. Improve campaign optimization and performance. Initial results from Zuma's improved modeling and forecasting capabilities have shown an average 32% improvement in CPA performance. These backend investments are designed to enable workflow and campaign improvements while establishing the platform's AI infrastructure as a foundation for future client experiences. “Kokai changed the way digital media is bought and sold programmatically, and Zuma builds on that foundation, making our platform faster, simpler and smarter,” said Jeff Green, founder and CEO of The Trade Desk. “AI is helping our customers move faster than ever, while making measurement more intuitive and actionable. We've listened closely to our users about what they need to navigate an increasingly fragmented digital media ecosystem, and Zuma brings those learnings into Kokai. We're continuing to build the platform that advertisers deserve – one that removes friction from the work, puts agentic AI to practical use and gives them a clearer view of the outcomes their media dollars are driving.” Read More: https://theabm.info/the-trade-desk-launches-kokai-zuma-the-latest-release-of-kokai

IOWN Global Forum Addresses Sovereign AI Infrastructure Challenges in Amsterdam The Innovative Optical and Wireless Network Global Forum (IOWN Global Forum) will host FUTURES Amsterdam 2026 – Building Europe’s AI Infrastructure, bringing together policymakers, technology leaders and infrastructure experts to address the complex infrastructure decisions shaping global competitiveness in the AI era. As AI reshapes computing demands worldwide, regions face distinct infrastructure challenges, including balancing rapid AI deployment with digital sovereignty objectives, managing energy and sustainability requirements, and ensuring access to reliable, high-capacity connectivity. Attracting delegates from across the IOWN Global Forum's network of member organizations, including data center operators, telecom providers, cloud infrastructure specialists, equipment manufacturers and European government officials, FUTURES Amsterdam will facilitate practical dialogue on infrastructure investment, policy and technology strategies. The event will explore Europe's unique economic and strategic priorities while contributing to broader solutions for the global AI infrastructure landscape. Infotech Insights: F5 Accelerates AI-Powered WAF and Runtime Security for Emerging Cyber Risks Event Highlights Attendees will have the opportunity to: Hear opening remarks from Dr. Katsuhiko Kawazoe, President and Chairperson, IOWN Global Forum & Chief Executive Fellow, NTT, Inc. Listen to a keynote from Ian Redpath, Research Director, Omdia, titled The AI Infrastructure Market: The Forces Shaping What Comes Next, examining the forces shaping AI and data center infrastructure, including growing demand for computing capacity and the increasing importance of connectivity, energy, sustainability, latency and resilience. See a presentation from Stijn Grove, Managing Director, Dutch Data Center Association, on the strategic role of data center infrastructure in the Netherlands and what Europe needs to translate its AI ambitions into practical, scalable infrastructure. Attend a panel discussion titled Building Sovereign AI Infrastructure: From Ambition to Implementation. Experts will discuss what is required to build secure, scalable and sovereign AI infrastructure, examine practical barriers and explore how stronger collaboration across the infrastructure ecosystem can accelerate implementation. The panel will be moderated by Ronald van Loon, Founder & CEO of Intelligent World. Infotech Insights: Fortanix and CrowdStrike Partner to Secure Enterprise AI and Protect Models and Data Attend a presentation from Dr. Gonzalo Camarillo, Marketing Steering Committee Chair, IOWN Global Forum, and Head of Implementation Components, Ericsson, on how the IOWN Global Forum's work supports the evolving needs of AI and data center infrastructure. Connect with peers driving the next generation of network infrastructure and adoption pathways. Explore exhibits from IOWN Global Forum members showcasing their technologies and solutions. Read More: https://theinfotech.info/iown-global-forum-addresses-sovereign-ai-infrastructure-challenges-in-amsterdam

HPE and Oracle Expand Networking Collaboration for Gigawatt-Scale AI Infrastructure HPE announced an expanded collaboration with Oracle to help scale Oracle’s global AI infrastructure by deploying HPE Juniper Networking across Oracle’s AI data centers. The expanded collaboration builds on more than a decade of engineering work between Oracle and Juniper Networks and includes networking support services and financing capabilities. Today, HPE Juniper Networking routing and switching platforms support key elements of Oracle Cloud Infrastructure (OCI) data center and edge networks, providing the scale, performance, and availability needed for Oracle’s global cloud operations. That foundation is becoming increasingly important as OCI expands its AI superclusters. Large GPU clusters place significant demands on bandwidth, latency, congestion management, and fault recovery, while Oracle’s regional and edge networks must scale to connect clusters with services, customers, and external networks worldwide. “Oracle is building AI infrastructure at extraordinary scale, and HPE is uniquely positioned to support OCI with one of the industry’s most comprehensive networking portfolios for AI data centers,” said Rami Rahim, Executive Vice President, President and General Manager, Networking, HPE. “Together, we are advancing a high-performance networking architecture that helps Oracle address the traffic management, power, space, and operational requirements of increasingly large AI environments.” Read More: https://theinfotech.info/hpe-and-oracle-expand-networking-collaboration-for-gigawatt-scale-ai-infrastructure

CyberCube Creates New Framework for Evaluating AI-Driven Losses CyberCube, the leading provider of cyber risk analytics, unveils six new AI Event Families and a new I2T2 framework, providing carriers, brokers, and risk managers with a practical way to think about and evaluate AI-driven loss. The I2T2 framework addresses AI risk in a way that remains applicable to cyber insurance, while also reflecting the wider array of insurance risks now possible. The framework sets out four dimensions through which AI affects insurance risk: Information, Intelligence, Tactics, and Technology. CyberCube’s framework is set out in “Machine State of Mind: A Framework for Quantifying AI-Driven Insurance Risk,” examining how artificial intelligence is reshaping risk across the insurance industry. It notes that AI is no longer a niche technical risk confined to a single line of insurance. As AI systems take on greater autonomy in decision-making, content generation, and operational processes, they introduce new forms of loss that cut across established coverage lines, including, but not limited to, Errors & Omissions (E&O), Technology E&O, Cyber, General Liability, and Directors & Officers, as well as introducing novel impacts that insurance has never covered before. Infotech Insights: Castrol ON PG25 Cooling Fluids Validated for NVIDIA AI Infrastructure As this new intelligence evolves, CyberCube argues that the industry's ways of thinking about insurance risk must evolve with it. Rather than treating AI risk as just an extension of Cyber risk, the framework examines where AI and Cyber overlap, where they diverge, and what this means for how claims are likely to be triggered and allocated across policies. Key Findings AI risk shares several characteristics with Cyber risk, including single points of failure (SPoFs), human error, data loss, and the potential to amplify existing threats. However, it also differs in important ways. For example, AI can cause harm through hallucination or flawed output even when functioning exactly as designed, without the involvement of any malicious actor. AI is best understood as a technology and information revolution, rather than merely an evolution of Cyber risk. As the use of AI extends into many corners of the global economy, new frameworks are required for understanding how AI could affect property-casualty insurance coverages. (Re)insurers grappling with AI risk will follow a path similar to that previously taken with non-affirmative (“silent”) Cyber risk. Lessons from the last decade suggest that the exact application of coverages will vary by carrier. Given the breadth of transformation AI is creating, however, (re)insurers must reckon with AI risk as a multi-line challenge and opportunity. Read More: https://theinfotech.info/cybercube-creates-new-framework-for-evaluating-ai-driven-losses

The Trade Desk Launches Kokai Zuma, the Latest Release of Kokai The Trade Desk, a leading global advertising technology company, introduced Kokai Zuma, the latest release of its Kokai platform. Introduced in 2023, Kokai is The Trade Desk's platform for planning, buying, and measuring advertising across the open internet. The Zuma release makes Kokai easier to navigate, learn, and use by bringing together new agentic AI capabilities and a simplified measurement framework. As advertising becomes increasingly fragmented across channels, formats, and devices, media buyers face greater complexity in managing campaigns and understanding what drives performance. Zuma is designed to reduce that friction by helping traders find what they need faster, make changes with greater confidence, and spend less time managing workflows so they can focus on optimizing campaigns and improving results. The update also builds on the AI forecasting engine that has been at the heart of Kokai. Upgraded AI and Forecasting Infrastructure While the redesigned interface is the most visible change, The Trade Desk has also invested heavily in upgrading Kokai's forecasting engine and underlying AI infrastructure. The technology is designed to: Predict available advertising inventory. Model campaign outcomes. Power Koa's agentic capabilities in real time. Improve campaign optimization and performance. Initial results from Zuma's improved modeling and forecasting capabilities have shown an average 32% improvement in CPA performance. These backend investments are designed to enable workflow and campaign improvements while establishing the platform's AI infrastructure as a foundation for future client experiences. “Kokai changed the way digital media is bought and sold programmatically, and Zuma builds on that foundation, making our platform faster, simpler and smarter,” said Jeff Green, founder and CEO of The Trade Desk. “AI is helping our customers move faster than ever, while making measurement more intuitive and actionable. We've listened closely to our users about what they need to navigate an increasingly fragmented digital media ecosystem, and Zuma brings those learnings into Kokai. We're continuing to build the platform that advertisers deserve – one that removes friction from the work, puts agentic AI to practical use and gives them a clearer view of the outcomes their media dollars are driving.” Read More: https://theabm.info/the-trade-desk-launches-kokai-zuma-the-latest-release-of-kokai