Startup doxx.net hands the network controls to AI

Stay Ahead, Stay ONMINE

Startup doxx.net hands the network controls to AI

“The whole thing is run completely by AI, so behind the scenes, I mean, it’s a network with 31 locations around the world, and internally there’s a mesh network, and I can’t, as a human being, manage all of this by myself,” Lyon said. Lyon said the team modeled every site, down to the wire […]

Lyon said the team modeled every site, down to the wire and the optic, in a virtual model before building. An infrastructure management system running the company’s own AI on its own hardware then ordered the installation. It orchestrated shipping and delivery through data center APIs. Human technicians performed the remote smart hands installations.

The company also built tools for agents to work with users and with the network. An agent gateway gives an AI agent an identity in the doxx.net chat app. The user pastes a credential into the agent. The agent obtains its certificate and appears in the user’s chat. Users can create group chats with several agents. In one example, Lyon said one agent runs BGP while others handle other tasks.

Stay Ahead

Explore More Insights

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.

Aker BP, Höegh Evi secure EU funding for Atlas CCS project

Aker BP and Höegh Evi have been awarded €223 million (about US$250 million) from the European Union Innovation Fund to support development of the Atlas carbon capture and storage (CCS) project in Norway. The Atlas project is designed as a large-scale CCS network linking industrial CO2 emitters in Europe with

Nvidia’s Berlin GTC: Faster AI factories, lower power use and more secure agents

Germany’s Agile Robots introduced Agile ONE SN, a humanoid reference platform built on Nvidia’s Isaac GR00T technology and powered by Jetson Thor. It has 128GB of unified memory and is scheduled to become available in early 2027. Telecom Operators Test AI-Powered Networks Nvidia and Nokia announced expanded work with telecom

Extreme wants to use idle network compute capacity for local AI

“We’ve spent decades building networks that reach every corner of the enterprise, and those networks hold a lot of idle compute. Exchange zones will put that compute to work, letting AI run locally on infrastructure customers already own. That delivers sovereignty, local inferencing, and better economics than sending everything to

IT/OT convergence exposes skills gap. Cisco responds with new OT security certification

OT security certification training and exam details The Cisco OT Security Practitioner certification focuses on industrial control systems (ICS) architectures, network segmentation, traffic monitoring, and incident response. Candidates also study security standards, including ISA/IEC 62443 and NIST 800-82, alongside Cisco Industrial Ethernet switch hardening and Cisco Cyber Vision. Cisco will

Energy Department to Use Defense Production Act Funding to Restart American Refinery and Expand Refining Capacity

WASHINGTON—President Donald J. Trump today announced plans to provide up to $150 million in Defense Production Act (DPA) Title III funding to expand domestic petroleum production, storage, and logistics at refineries across the United States. By leveraging DPA authorities, DOE will help increase domestic production of gasoline, diesel, jet fuel, and other petroleum products, strengthening America’s fuel supply and lowering prices for American families and businesses. Together, this funding is expected to increase U.S. refining output by millions of gallons per day. “After 17 years of Democrat leaders enacting energy subtraction policies that led to the closure of refineries and more expensive refined oil products, the Trump Administration is taking action to expand our nation’s refining capacity and strengthen our energy security,” said U.S. Secretary of Energy Chris Wright. “Thanks to President Trump, America is producing more oil and gas than ever before and thanks to President Trump, we will soon be producing more refined products as well. These Defense Production Act funds will be used strategically and in partnership with private industry to rapidly bolster America’s existing energy infrastructure. By restoring energy addition policies, we will continue to unleash American energy dominance and lower costs for the American people.” The investment supports President Trump’s commitment to ensuring all Americans have access to affordable, reliable, and secure energy while strengthening energy infrastructure critical to national defense. Thanks to the Working Families Tax Cuts, $1 billion in federal funding has been appropriated for energy-related projects deemed critical to national security. Presidential Determination No. 2026-11 identifies domestic petroleum production, refining, and logistics capacity as essential to national defense. The selection process is ongoing and will be announced in the near future.

Energy Department Announces Over $125 Million for 63 Projects Across 16 States to Lower Energy Costs, Expand Electricity Access, and Advance Economic Development for America’s Tribal Communities

WASHINGTON—The U.S. Department of Energy today announced over $125 million for 63 tribe-led projects selected for award negotiations. The selected projects span 16 states, including Alaska, and include 61 tribal entities. This investment will support projects aimed at cutting electricity costs, expanding electricity access, and advancing economic development for tribes across the United States. Consistent with President Trump’s Executive Order, Unleashing American Energy, these selections highlight the fundamental role of energy in strengthening tribal economies. “President Trump has prioritized delivering affordable, reliable, and secure energy to the American people since day one, and this administration is taking decisive action to deliver on that commitment,” said U.S. Deputy Secretary of Energy James P. Danly. “This historic investment will help address longstanding energy challenges facing tribal communities by lowering electricity costs, expanding access to reliable power, and creating new opportunities for economic development. We are ensuring that tribal communities have the energy resources they need to grow and prosper.” “With this announcement, the Office of Indian Energy is making the largest-ever single investment in tribal energy development in the office’s 20-year history,” said DOE’s Office of Indian Energy Director Eric Mahroum. “These projects represent a new and expanded direction for the Office, and for the first time, we’re supporting projects at any phase of development and with a broad set of technologies and approaches. Our goal is to support tribes to advance high-value energy projects, including large scale projects like generation, transmission, distribution, and critical mineral projects.” The 63 selected tribe-led projects will deliver for tribal communities in three key areas. Thirteen projects will focus on the deployment and installation of tribal community-scale energy projects to meet the needs of the community, bringing electric service to 150 tribal homes that currently lack electricity, converting 500 homes and businesses to natural gas heating, and upgrading

Kinder Morgan takes FID on Southeast Connector, Permian Link natural gas pipelines

Kinder Morgan Inc. (KMI) has taken final investment decision (FID) on its Southeast Connector and Permian Link natural gas pipeline projects. Southern Natural Gas (SNG), a KMI subsidiary, secured sufficient long-term, take-or-pay transportation agreements with shippers to proceed with Southeast Connector, designed to transport up to 2.6 bcfd through about 470 miles of new pipeline. Subject to receipt of required permits and approvals, service is expected to begin in phases starting fourth-quarter 2030. The $3.1-billion project could be expanded if demand warrants. “Large-load interconnection requests across the Southeast now exceed 200 Gw,” said KMI chief executive officer and director Kimberly Dang in the company’s earnings call, noting that this could represent roughly 8 bcfd of natural gas demand if it all came to fruition and if it was all served by gas-fired generation. Natural Gas Pipeline Co. of America (NGPL), another KMI subsidiary, reached FID on the 2.8-bcfd Permian Link. Estimated cost of the project is $4.2-billion, of which KMI’s share is expected to be about $1.6 billion. The 434-mile pipeline will serve new power generation for an A-rated hyperscaler and is supported by long-term, take-or-pay transportation agreements. NGPL expects to place the project in service third-quarter 2030, subject to required permits and approvals. KMI also sanctioned the second and third phases of its Creekside Lateral project in central Texas, following execution of additional long-term, take-or-pay agreements with data center and industrial customers. The company has now contracted 750 MMcfd of the system’s 1.3-bcfd design capacity, with potential for future expansion to more than 2 bcfd. KMI expects to place the project in service in phases, beginning fourth-quarter 2026 and continuing throught fourth-quarter 2027, subject to permits and approvals. SNG and Tennessee Gas Pipeline (TGP), meanwhile, took FID on the 10-bcf Bear Creek Storage Expansion project, to be used by

GeoPark awarded Corralera Noreste block in Vaca Muerta bid round

The block was awarded through Neuquén Province’s Ronda 1/2026 licensing round, conducted by provincial energy company Gas y Petróleo del Neuquén (GyP). GeoPark plans to advance the block through a phased appraisal program beginning with the drilling and completion of two horizontal wells, followed by extended testing to evaluate productivity and resource quality. Future appraisal and development spending will depend on technical results and project economics, the company said. Corralera NE lies near Rincón de los Sauces in the northern Vaca Muerta trend. The award adds about 26,900 gross acres (24,200 net acres) to GeoPark’s portfolio at a total committed cost, including the entry bonus, of about $95 million. GeoPark estimates the block contains about 200 MMboe of unrisked gross resource potential based on two landing zones, with additional upside from two further prospective horizons. The company said potential gross peak production from the initial two landing zones could reach 40,000-50,000 boe/d, subject to successful appraisal and development.

Valero wraps St. Charles FCC project, reports record earnings

Valero Energy Corp. has completed the $230-million fluid catalytic cracking (FCC) unit optimization project on schedule and within budget at its 340,000-b/d St. Charles refinery in Norco, St. Charles Parish, La., about 25 miles west of New Orleans, the San Antonio-based operator said in reporting its third-quarter 2026 earnings on Oct. 22. The FCC upgrade enhances the refinery’s ability to produce higher-value products, particularly alkylate and gasoline, positioning the site to capture premium margins as refining fundamentals remain underpinned by limited global capacity additions, Valero said. Originally announced in 2021 as part of Valero’s long-term capital allocation strategy, the St. Charles FCC unit optimization is now operational and anticipated to strengthen the refinery’s competitive position in heavy crude processing and product slate optimization. In its April presentation to investors, Valero said the FCC project was to include installation of a new compressor and upgrading of existing equipment to achieve an incremental 15,000-b/d FCC throughput rate using residual feedstock. The project aims to improve product yields to increase the refinery’s incremental production of propylene by 3,000 b/d to fill the site’s downstream alkylation unit capacity to produce an incremental 6,000 b/d of high-value alkylate. The FCC upgrade is also anticipated to yield an 8,000-b/d Incremental production of gasoline, according to the April presentation. Completion of the FCC project marks a strategic milestone for Valero as the company delivered third-quarter 2026 results underscored by exceptional financial performance and operational execution. Valero reported net income of $6.8 billion for the third quarter, compared with $1.1 billion during the same period a year earlier. Adjusted net income for the quarter was $6.8 billion, which was six times higher than its $1.1-billion adjusted net income in third-quarter 2025. Refining segment powers windfall The refining segment delivered the bulk of earnings gains, with the company reporting

Tight product supplies lift US summer refinery utilization to highest level since 1998

US refinery utilization averaged more than 95% during May–September 2026, the highest for those months since 1998, as tight global product supplies, particularly for diesel, strengthened margins, and demand for US exports, according to the US Energy Information Administration (EIA). The strong operating rates coincided with a widening profitability gap between US refiners and those in Europe and Asia. In its October Oil Market Report, the International Energy Agency (IEA) said relatively inexpensive crude supplies helped US refiners achieve record margins in September, while rising feedstock and freight costs squeezed competitors overseas. Despite elevated utilization, total US refinery output reached only its highest level since 2019. US refining capacity has declined 4% since then, limiting the output gains available from higher operating rates, EIA said. Utilization eased during the first 3 weeks of October from its summer highs, but remained above the 2021–25 seasonal average. US margins outperform IEA’s US Gulf Coast light sweet cracking margin indicator averaged $49.84/bbl in September, up $5.97/bbl from August. Heavy sour coking margins rose $11.57/bbl to $61.36/bbl, reflecting the advantage of access to lower-cost crude. By comparison, light sweet cracking margins averaged $30.15/bbl in Northwest Europe and $19.26/bbl in Singapore, down $5.97/bbl and $7.34/bbl, respectively. Medium sour cracking margins briefly turned negative in both regions during September. Record tanker freight rates widened regional crude cost differences, helping US refiners outperform competitors in Europe and Asia. IEA warned that deteriorating profitability could prompt some European refiners to reduce runs despite tight product supplies and low inventories. US product inventories also remained tight despite high refinery activity. Combined gasoline, jet fuel, and gasoil/diesel stocks entered October at 352 million bbl, their lowest seasonally adjusted level since 2008, according to IEA. Robust exports and autumn maintenance added pressure on inventories, while unplanned Midwest refinery outages further supported

AMD Helion Puts CPU Density at the Center of Agentic AI Infrastructure

AMD expects the growth of agentic applications to substantially increase demand for server CPUs alongside GPU infrastructure. During her October 12 OCP keynote, CEO Lisa Su said the company now projects the server CPU market will exceed $200 billion by 2030, a significant upward revision from its earlier $120 billion estimate. While that remains AMD’s market forecast rather than an independent measure of future demand, it underscores the company’s conviction that general-purpose computing will play a much larger role in AI infrastructure. Two Rack Designs for Different Markets Helion’s two principal configurations reflect different approaches to deploying and managing agentic workloads. For hyperscale operators, the architecture emphasizes high-density CPU infrastructure dedicated to agent execution, with model inference handled by separate GPU-accelerated systems. This separation allows operators to expand CPU and GPU capacity independently, potentially improving resource utilization as the balance between agent activity and inference demand changes. The approach also introduces infrastructure considerations. Agents may communicate repeatedly with models hosted elsewhere in a data center or across distributed environments, making network performance, latency and workload orchestration important factors in overall system efficiency. The ability to scale computing resources independently is an architectural advantage, but realizing that benefit will depend on how effectively those resources are connected and managed. For enterprise deployments, Helion takes a more integrated approach, combining EPYC CPU blades, Instinct MI350P GPU blades and general-purpose computing resources within the same rack. The reference design accommodates up to 16 MI350P accelerators in two independent eight-GPU servers, each occupying five rack units. These GPU nodes operate separately rather than through a common rack-level scale-up fabric, reflecting the architecture’s emphasis on enterprise inference rather than the tightly coupled accelerator environments used for frontier-model training. The enterprise configuration also accommodates dedicated resources for databases, storage and technical computing, supporting the broader range

AI’s Fragmented Regulatory Landscape Becomes a Data Center Design Challenge

Designing for Verifiable Operations For an operator serving tenants across jurisdictions, a claim about where a workload runs must be one that the responsible service provider can substantiate. That means identifying who can administer the workload, how data moves between regions, where backups are maintained, and which records document those arrangements. A tenant promising to run a workload within a particular jurisdiction needs assurance that its data, backups, and access arrangements satisfy the applicable requirements. Workload placement alone cannot establish where a backup resides or from where an administrator can access a system. Those questions must be addressed across compute, storage, network paths, security controls, and third-party access, with responsibilities established during procurement and verified before service begins. Not all of those controls necessarily reside with the data center operator. A colocation provider may be responsible for physical access, facility security, and designated infrastructure services, while a tenant or cloud provider controls the applications, storage configuration, backup locations, and logical access to workloads. The operating agreement must distinguish between what the facility provider can verify directly and what remains within the tenant’s systems. The EU AI Act makes that division of responsibility particularly important. Its requirements for covered high-risk AI systems include logging, documentation, human oversight, robustness, and cybersecurity, subject to the law’s applicable implementation dates. These obligations attach to the relevant providers and deployers of AI systems, rather than automatically to every company hosting them. A data center operator may nevertheless be required by contract to provide physical-access records, security evidence, or information about infrastructure location so that a tenant can substantiate its own compliance. A further distinction concerns data residency. The AI Act is not itself a general requirement to keep all AI workloads or associated data within the European Union. Geographic restrictions may instead arise from data-protection

Microsoft exits UK AI data center project over electricity delays

“We remain committed to both the Loughton project and the UK, and we are exploring alternative power sources for the site, as well as other expansion sites in the UK,” Nscale added. “This decision relates to a single site,” a Microsoft spokesperson said of its termination of an agreement for the data center in Loughton, Essex, where Nscale had faced delays, with electricity grid ⁠officials telling the AI company they did not expect to supply power to the site until the 2030s. “Our relationship with Microsoft remains strong and long-term, and this unfortunate London-wide grid issue has no impact on our other contracts,” a spokesperson for Nscale told Reuters.

AI networks drive SONiC boom

Aviz Networks: Aviz, which received early funding help from Cisco, offers a pre-tested, multi-ASIC, hardware agnostic, enterprise-ready SONiC distro with added features such as bug fixes, telemetry and support. Hedgehog Networks: Founded by former Cisco exec Marc Austin in 2022 to bring SONiC-based disaggregated networking to the enterprise, Hedgehog’s mission is to simplify the deployment of private cloud AI networks with automation and orchestration based on Kubernetes. Upscale AI: This startup has raised $500 million with the goal of building a new networking platform from ground up, starting with its own silicon. The company is targeting AI data centers as its sole market and is using SONiC as the NOS. Most recently, Upscale introduced Token Fabric architecture, which combines the company’s own SkyFabriX scale-up silicon, scale-out systems built on Nvidia Spectrum-X Ethernet, and a common software layer for operations. General availability is planned for early 2027. The announcement builds on two earlier milestones. Upscale, based in Santa Clara, Calif., emerged from stealth in September 2025 with a $100 million seed round. In June, it raised an additional $190 million 4. SONiC expands beyond the data center Community support is the lifeblood of any open-source project. Under the Linux Foundation, the SONiC community has grown to more than 4,300 contributors. Heavyweight members include Alibaba, Arista, Broadcom, Cisco, Dell, Google, Microsoft, Nokia, and Nvidia. The SONiC community is currently working to extend the technology from the data center to the network edge. The PoE Edge Networks with SONiC (PENS) workgroup, led by Aviz, Wistron, Cisco and Celestica, is working to adapt SONiC for edge LANs. Specifically, the group is integrating protocols consistent with Wi-Fi networks and IoT devices, such as Power Over Ethernet (PoE) and 802.1x.

SONiC momentum continues at OCP Global Summit

This will complete SONiC integration across the Cisco multi-silicon Nexus switching portfolio, according to a blog post by Surbhi Paul, director, product management, for data center AI networking at Cisco. “For organizations deploying Cisco 8000 Series across WAN backbones and core transport, extending Cisco-validated SONiC to the Nexus switches creates a unified, end-to-end, open network fabric,” Cisco stated. “Customers eliminate vendor silos, leveraging a single supply chain, shared FRR routing tools, and one Technical Assistance Center (TAC) contract from the WAN edge deep into the data center core.” “Built for AI infrastructure providers scaling 8000 to 32,000+ GPU clusters. High-density 400G/800G Cisco Nexus (N9100) switches (NVIDIA Spectrum-X) and Cisco Nexus (Cloud Scale) switches allow architects to tune RoCEv2 congestion algorithms (PFC/ECN) within SONiC, maximizing GPU training efficiency without risking multi-vendor support fragmentation,” Cisco stated.

AI’s GPU Refresh Cycle Is Becoming a Stranded-Capital Problem

The AI infrastructure buildout is measured in gigawatts and successive generations of accelerators. But as operators bring new compute capacity online, a less visible question is gaining financial urgency: What happens to GPU systems when their first production assignment ends? For hardware that remains functional but no longer meets the performance or efficiency requirements of a leading-edge AI cluster, retirement need not mean obsolescence. Equipment may be redeployed, resold, harvested for components or processed to recover materials. The outcome depends on decisions made long before a system leaves its rack. That was the central argument advanced by Linda Li, chief strategy officer of Re-Teck and its parent company, Li Tong Group, in the latest onsite installment of the DCF Show podcast, hosted by Data Center Frontier Contributing Editor Doug Black at the DCF Trends Summit 2026 in Northern Virginia. For Li, the acceleration of GPU refresh cycles is making lifecycle management a question of capital preservation, facilities design and executive governance—not merely IT asset disposition. Plan for the Exit Before the Equipment Arrives Li draws on more than two decades of her company’s work in reverse supply chains for consumer electronics and telecommunications equipment. Both sectors have managed rapid generational shifts without assuming that every displaced device had exhausted its useful life. “We are dealing with a totally different organism,” she said, describing the combination of rapid accelerator iterations, rising rack power density and increasingly sophisticated thermal requirements. The response, in her view, has two levels. The first is physical infrastructure: facilities designed to accommodate successive generations of servers without extensive reconstruction. Scalable power distribution, adaptable raceways, modular rack systems and high-capacity air and liquid cooling can create room for hardware transitions that may arrive sooner than originally expected. The second is the network of partners and contracts around the

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.

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

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

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

Stay Ahead, Stay ONMINE