Perenco completes first UK CO2 injection

Stay Ahead, Stay ONMINE

Perenco completes first UK CO2 injection

Perenco has completed the UK’s first carbon dioxide injection at its Poseidon carbon capture storage (CCS) project. The firm kicked off the project’s CO2 injection test in February and has now reported its successful conclusion. Perenco, along with Carbon Catalyst and Harbour Energy conducted the CO2 injection test from the ERDA rig, newly equipped with […]

Perenco has completed the UK’s first carbon dioxide injection at its Poseidon carbon capture storage (CCS) project.

The firm kicked off the project’s CO2 injection test in February and has now reported its successful conclusion.

Perenco, along with Carbon Catalyst and Harbour Energy conducted the CO2 injection test from the ERDA rig, newly equipped with the injection test facilities.

Petrodec’s ERDA, the first rig in the UK to have achieved an approved safety case for CO2 injection support, has now sailed away from the Leman 27H platform.

There were a total of 15 injection cycles performed into the depleted Leman gas field, which mobilised 11 CO2 offshore batch refills.

The firm said that the test was completed without injection issues and that it collected valuable data throughout.

Perenco argued that the test proved that carbon “can be a reality for depleted fields of the Southern North Sea” and that it is possible to widely reuse North Sea production infrastructure to “unlock cost-effective projects”.

Armel Simondin, chief executive of Perenco, commented: “This test has met our expectations, both in terms of technical execution and the quality of data gathered.

“These insights are instrumental as we move to the next phase of developing the Poseidon Project.

“The successful test highlights the role our industry can play in the UK’s decarbonisation efforts.

“The Perenco CCS team and our joint venture partners are now fully focused on interpreting the results and converting new insights into an evidence-based development plan for Project Poseidon.”

The Poseidon project will use the Leman field to store CO2, which will be delivered from the Bacton Gas Terminal on the East Anglian coast.

The project is expected to have a capacity of 935m tonnes of CO2, going from an injection rate of around 10 million tonnes per annum (mtpa) when pumping starts around 2030 before growing to 40mtpa.

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.

IBM X-Force: Stealthy attacks on the rise, toolkits targeting AI emerge

Last year, X-Force predicted that once AI technologies “establish market dominance—when a single technology approaches 50% market share or when the market consolidates to three or fewer technologies—attackers will be incentivized to invest in attack toolkits” that target AI models and solutions. “Are we there yet? Not quite, but adoption

VMware (quietly) brings back its free ESXi hypervisor

By many accounts, Broadcom’s handling of the VMware acquisition was clumsy and caused many enterprises to reevaluate their relationship with the vendor The move to subscription models was tilted in favor of larger customers and longer, three-year licenses. Because the string of bad publicity and VMware’s competitors pounced, offering migration

CoreWeave offers cloud-based Grace Blackwell GPUs for AI training

Cloud services provider CoreWeave has announced it is offering Nvidia’s GB200 NVL72 systems, otherwise known as “Grace Blackwell,” to customers looking to do intensive AI training. CoreWeave said its portfolio of cloud services are optimized for the GB200 NVL72, including CoreWeave’s Kubernetes Service, Slurm on Kubernetes (SUNK), Mission Control, and

Kyndryl launches private cloud services for enterprise AI deployments

Kyndryl’s AI Private Cloud environment includes services and capabilities around containerization, data science tools, and microservices to deploy and manage AI applications on the private cloud. The service supports AI data foundations and MLOps/LLMOps services, letting customers manage their AI data pipelines and machine learning operation, Kyndryl stated. These tools facilitate

Fracker Liberty’s Profit Falls to 3-Year Low as Oil Slumps

Shale fracker Liberty Energy Inc. posted its worst earnings in three years amid plunging oil prices and mounting concerns about energy demand. Adjusted first-quarter profit fell to 4 cents a share, according to a statement Wednesday, matching the average estimate among analysts. Sales and capital spending both came in better than expected, prompting the shares to rise more than 9% before the start of regular trading in New York on Thursday. Current levels of fracking activity suggest US oil output will hold steady, “mitigating the possibility of steep declines experienced by the service industry in past cycles,” the company said. Meanwhile, major shale driller Diamondback Energy Inc. said Wednesday that it’s “actively reviewing its operating plan” for the rest of the year given market volatility, according to a separate statement. “While the current tumult in commodity prices is not immediately driving changes in North American activity, we expect oil producers are evaluating a range of scenarios in anticipation of oil price pressure,” Liberty said. Liberty’s broad footprint across North American shale provides it a unique scope of vision for domestic oil-production trends. The Denver-based oilfield contractor has tumbled roughly 40% this year as US President Donald Trump’s trade war punished crude prices and tarnished the outlook for near-term fossil-fuel demand. Liberty is the first major US-based oil-service company to post quarterly results, with rival Halliburton Co. set to follow Tuesday morning. WHAT DO YOU THINK? Generated by readers, the comments included herein do not reflect the views and opinions of Rigzone. All comments are subject to editorial review. Off-topic, inappropriate or insulting comments will be removed. MORE FROM THIS AUTHOR Bloomberg

Oil Surges Amid Iran Sanctions

Oil rose for a second day after the Trump administration ratcheted up pressure on Iran’s energy exports, while talks between the US and a handful of key trade partners stirred optimism that agreements on trade can be reached. West Texas Intermediate surged 3.5% to settle near $65 a barrel, marking the largest two-day increase since early January. President Donald Trump said he is confident a trade deal with the European Union could be achieved, and negotiations between the US and Japan bolstered expectations that deals can be struck to avoid the worst effects of tariffs. Futures were also propelled higher by investors covering short positions and algorithmic traders turning marginally more bullish ahead of the long weekend. Oil futures won’t trade on Friday, a holiday in many countries, crimping volumes. On the Middle East front, Treasury Secretary Scott Bessent said the US would apply maximum pressure to disrupt Iran’s oil supply chain as his department sanctioned a second Chinese refinery accused of handling crude from the Islamic Republic. The so-called teapot oil processor sanctioned by the US — Shandong Shengxing Chemical Co. — had allegedly handled more than $1 billion of Iranian crude, the Treasury Department said. Tehran, meanwhile, warned that nuclear talks with Washington may fall apart if the Trump administration “moves the goalposts.” “While the macroeconomic backdrop remains mixed, it has the potential to either amplify market rallies or derail them entirely, depending on how these geopolitical tensions evolve,” said Rebecca Babin, a senior energy trader at CIBC Private Wealth Group. The pressure of Trump’s sweeping trade measures has put crude on the back foot this month, with prices at one point trading about 30% lower than their high for the year. This week’s rebound was aided by US government data that showed inventory levels at Cushing, Oklahoma

EPA granted appellate stay after judge rules to end GGRF funding freeze

The U.S. Court of Appeals for the District of Columbia Circuit on Wednesday issued a stay on a district court judge’s decision ordering the Environmental Protection Agency to unfreeze Greenhouse Gas Reduction Fund funding. “The district court’s order is stayed insofar as it enables or requires Citibank to release, disburse, transfer, otherwise move, or allow access to funds,” wrote Circuit Judges Neomi Rao, Gregory Katsas and Nina Pillard. “It is further ORDERED that no party take any action, directly or indirectly, with regard to the disputed contracts, grants, awards or funds.” The stay came after the U.S. District Court for the District of Columbia on Tuesday granted an injunction to the Climate United Fund, Coalition for Green Capital, and Power Forward Communities — nonprofit “green banks” which received $6.97 billion, $5 billion and $2 billion, respectively, from the Inflation Reduction Act’s GGRF. Those funds, held at Citibank, have been frozen since Feb. 16. The nonprofits sued EPA and Citibank last month, alleging serious financial harm to both their institutions and their subgrantees should the freeze continue, and asked for an injunction to end it. In a Wednesday memorandum opinion regarding her order to grant the injunction, Judge Tanya Chutkan wrote that EPA “acted arbitrarily and capriciously when it failed to explain its reasoning and acted contrary to its regulations in suspending and terminating Plaintiffs’ grants.” “Though repeatedly pressed on the issue, EPA offers no rational explanation for why it suspended the grants and then immediately terminated the entire [National Community Investment Fund] and [Clean Communities Investment Accelerator] grant programs overnight,” Chutkan said. She continued, “Nor has EPA offered any rational explanation for why it needed to cancel the grants to safeguard taxpayer resources, especially when it had begun examining the grant programs to add oversight mechanisms, or why it needed

BP Chairman Lund Re-Elected to Board Despite Protest Vote

BP Plc’s outgoing Chairman Helge Lund was re-elected to the company’s board amid a sizable protest vote. Just under 76% of shareholder votes were in favor of his re-election, compared with support in excess of 90% for other senior company executives, according to preliminary figures shown at the annual general meeting on Thursday. Prior to the meeting, investors including Legal & General Group Plc had said they would oppose Lund’s re-election in protest at BP’s recent strategy reset and the impact it will have on climate commitments. The protest was largely symbolic, as Lund already announced his intention to step down as chairman. Lund, a key backer of BP’s previous unsuccessful strategy, was widely seen to be in a vulnerable position after activist investor Elliott Investment Management was revealed to have built up a 5% stake in the company, with the intention of pushing for change after years of underperformance. The pressure only increased after Chief Executive Officer Murray Auchincloss’s strategy reset received a lukewarm response from investors and fell short of what Elliott had wanted. BP announced on April 4 that Lund would step down as chairman “in due course,” while remaining as a director during the transition to his successor. BP has pledged to channel more investment into oil and gas, something Elliott and other investors have been urging. The resulting shift away from clean energy has prompted criticism from other shareholders, with representatives from wealth manager Rathbones and the Australasian Centre for Corporate Responsibility raising the matter at Thursday’s meeting. “With no other climate-related item on the AGM ballot, it is safe to assume that this vote against the Chair is in part a reaction to the ongoing climate governance concerns,” Nick Mazan, UK company strategy lead at the ACCR, said in a statement. The process to select a new chairman is “well

Trump administration orders a stop to Empire Wind construction

The U.S. Department of the Interior has directed the Bureau of Ocean Energy Management to order the 810-MW Empire Wind 1 project to cease all construction until further review, Interior Secretary Doug Burgum announced Wednesday. In a letter to BOEM, Burgum alleged that the project was “rushed through by the prior administration without sufficient analysis or consultation among the relevant agencies as relates to the potential effects from the project.” The letter said that construction will remain halted until “further review is completed to address these serious deficiencies.” Empire Wind developer Equinor said in a Thursday release that the project “will safely halt” all offshore construction, but it added that “Empire is engaging with relevant authorities to clarify this matter and is considering its legal remedies, including appealing the order.” Empire “is in the process of ascertaining the impact on the project and project financing” of the construction halt, Equinor said. “Equinor US Holdings Inc has provided guarantees for the equity commitment in the project financing. In a full stop scenario, the USD 1.5 billion will be repaid from the equity commitment to the project finance lenders and Empire Offshore Wind LLC will be exposed to termination fees towards its suppliers.” Upon taking office, President Trump issued an executive order that withdrew all federal waters from offshore wind leasing — and paused permitting, approvals and loans for all onshore and offshore wind projects — until the completion of a six-month review of offshore wind from Burgum. “Pursuant to that review, staff of the Department of the Interior [have] obtained information that raises serious issues with respect to the project approvals for the Empire Wind Project,” Burgum wrote. “In light of these revelations and consistent with the President’s instructions, I am directing you to exercise your authority to order Empire Wind to cease

USA EIA Boosts Henry Hub Natural Gas Price Forecasts

The U.S. Energy Information Administration (EIA) increased its Henry Hub natural gas spot price forecast for 2025 and 2026 in its latest short term energy outlook (STEO), which was released on April 10. According to its April STEO, the EIA now sees the Henry Hub spot price averaging $4.27 per million British thermal units (MMBtu) in 2025 and $4.60 per MMBtu in 2026. In its previous STEO, which was released in March, the EIA saw the Henry Hub spot price averaging $4.19 per MMBtu in 2025 and $4.47 per MMBtu in 2026. The EIA projected in its April STEO that the Henry Hub spot price will come in at $3.93 per MMBtu in the second quarter of 2025, $4.34 per MMBtu in the third quarter, $4.68 per MMBtu in the fourth quarter, $4.93 per MMBtu in the first quarter of next year, $4.18 per MMBtu in the second quarter, $4.61 per MMBtu in the third quarter, and $4.66 per MMBtu in the fourth quarter. The EIA highlighted in its latest STEO that the Henry Hub spot price averaged $4.15 per MMBtu in the first quarter of 2025 and $2.19 per MMBtu overall in 2024. In its March STEO, the EIA projected that the Henry Hub spot price would average $3.88 per MMBtu in the second quarter of this year, $4.30 per MMBtu in the third quarter, $4.49 per MMBtu in the fourth quarter, $4.66 per MMBtu in the first quarter of 2026, $4.13 per MMBtu in the second quarter, $4.50 per MMBtu in the third quarter, and $4.60 per MMBtu in the fourth quarter of next year. The EIA’s March STEO projected that the Henry Hub spot price would average $4.11 per MMBtu in the first quarter of 2025. This STEO also highlighted that the commodity came in at $2.19

Intel sells off majority stake in its FPGA business

Altera will continue offering field-programmable gate array (FPGA) products across a wide range of use cases, including automotive, communications, data centers, embedded systems, industrial, and aerospace. “People were a bit surprised at Intel’s sale of the majority stake in Altera, but they shouldn’t have been. Lip-Bu indicated that shoring up Intel’s balance sheet was important,” said Jim McGregor, chief analyst with Tirias Research. The Altera has been in the works for a while and is a relic of past mistakes by Intel to try to acquire its way into AI, whether it was through FPGAs or other accelerators like Habana or Nervana, note Anshel Sag, principal analyst with Moor Insight and Research. “Ultimately, the 50% haircut on the valuation of Altera is unfortunate, but again is a demonstration of Intel’s past mistakes. I do believe that finishing the process of spinning it out does give Intel back some capital and narrows the company’s focus,” he said. So where did it go wrong? It wasn’t with FPGAs because AMD is making a good run of it with its Xilinx acquisition. The fault, analysts say, lies with Intel, which has a terrible track record when it comes to acquisitions. “Altera could have been a great asset to Intel, just as Xilinx has become a valuable asset to AMD. However, like most of its acquisitions, Intel did not manage Altera well,” said McGregor.

Intelligence at the edge opens up more risks: how unified SASE can solve it

In an increasingly mobile and modern workforce, smart technologies such as AI-driven edge solutions and the Internet of Things (IoT) can help enterprises improve productivity and efficiency—whether to address operational roadblocks or respond faster to market demands. However, new solutions also come with new challenges, mainly in cybersecurity. The decentralized nature of edge computing—where data is processed, transmitted, and secured closer to the source rather than in a data center—has presented new risks for businesses and their everyday operations. This shift to the edge increases the number of exposed endpoints and creates new vulnerabilities as the attack surface expands. Enterprises will need to ensure their security is watertight in today’s threat landscape if they want to reap the full benefits of smart technologies at the edge. Bypassing the limitations of traditional network security For the longest time, enterprises have relied on traditional network security approaches to protect their edge solutions. However, these methods are becoming increasingly insufficient as they typically rely on static rules and assumptions, making them inflexible and predictable for malicious actors to circumvent. While effective in centralized infrastructures like data centers, traditional network security models fall short when applied to the distributed nature of edge computing. Instead, organizations need to adopt more adaptive, decentralized, and intelligent security frameworks built with edge deployments in mind. Traditional network security typically focuses on keeping out external threats. But today’s threat landscape has evolved significantly, with threat actors leveraging AI to launch advanced attacks such as genAI-driven phishing, sophisticated social engineering attacks, and malicious GPTs. Combined with the lack of visibility with traditional network security, a cybersecurity breach could remain undetected until it’s too late, resulting in consequences extending far beyond IT infrastructures. Next generation of enterprise security with SASE As organizations look into implementing new technologies to spearhead their business, they

Keysight tools tackle data center deployment efficiency

Test and performance measurement vendor Keysight Technologies has developed Keysight Artificial Intelligence (KAI) to identify performance inhibitors affecting large GPU deployments. It emulates workload profiles, rather than using actual resources, to pinpoint performance bottlenecks. Scaling AI data centers requires testing throughout the design and build process – every chip, cable, interconnect, switch, server, and GPU needs to be validated, Keysight says. From the physical layer through the application layer, KAI is designed to identify weak links that degrade the performance of AI data centers, and it validates and optimizes system-level performance for optimal scaling and throughput. AI providers, semiconductor fabricators, and network equipment manufacturers can use KAI to accelerate design, development, deployment, and operations by pinpointing performance issues before deploying in production.

U.S. Advances AI Data Center Push with RFI for Infrastructure on DOE Lands

ORNL is also the home of the Center for Artificial Intelligence Security Research (CAISER), which Edmon Begoli, CAISER founding director, described as being in place to build the security necessary by defining a new field of AI research targeted at fighting future AI security risks. Also, at the end of 2024, Google partner Kairos Power started construction of their Hermes demonstration SMR in Oak Ridge. Hermes is a high-temperature gas-cooled reactor (HTGR) that uses triso-fueled pebbles and a molten fluoride salt coolant (specifically Flibe, a mix of lithium fluoride and beryllium fluoride). This demonstration reactor is expected to be online by 2027, with a production level system becoming available in the 2030 timeframe. Also located in a remote area of Oak Ridge is the Tennessee Valley Clinch River project, where the TVA announced a signed agreement with GE-Hitachi to plan and license a BWRX-300 small modular reactor (SMR). On Integrating AI and Energy Production The foregoing are just examples of ongoing projects at the sites named by the DOE’s RFI. Presuming that additional industry power, utility, and data center providers get on board with these locations, any of the 16 could be the future home of AI data centers and on-site power generation. The RFI marks a pivotal step in the U.S. government’s strategy to solidify its global dominance in AI development and energy innovation. By leveraging the vast resources and infrastructure of its national labs and research sites, the DOE is positioning the country to meet the enormous power and security demands of next-generation AI technologies. The selected locations, already home to critical energy research and cutting-edge supercomputing, present a compelling opportunity for industry stakeholders to collaborate on building integrated, sustainable AI data centers with dedicated energy production capabilities. With projects like Oak Ridge’s pioneering SMRs and advanced AI security

Generac Sharpens Focus on Data Center Power with Scalable Diesel and Natural Gas Generators

In a digital economy defined by constant uptime and explosive compute demand, power reliability is more than a design criterion—it’s a strategic imperative. In response to such demand, Generac Power Systems, a company long associated with residential backup and industrial emergency power, is making an assertive move into the heart of the digital infrastructure sector with a new portfolio of high-capacity generators engineered for the data center market. Unveiled this week, Generac’s new lineup includes five generators ranging from 2.25 MW to 3.25 MW. These units are available in both diesel and natural gas configurations, and form part of a broader suite of multi-asset energy systems tailored to hyperscale, colocation, enterprise, and edge environments. The product introductions expand Generac’s commercial and industrial capabilities, building on decades of experience with mission-critical power in hospitals, telecom, and manufacturing, now optimized for the scale and complexity of modern data centers. “Coupled with our expertise in designing generators specific to a wide variety of industries and uses, this new line of generators is designed to meet the most rigorous standards for performance, packaging, and after-treatment specific to the data center market,” said Ricardo Navarro, SVP & GM, Global Telecom and Data Centers, Generac. Engineering for the Demands of Digital Infrastructure Each of the five new generators is designed for seamless integration into complex energy ecosystems. Generac is emphasizing modularity, emissions compliance, and high-ambient operability as central to the offering, reflecting a deep understanding of the real-world challenges facing data center operators today. The systems are built around the Baudouin M55 engine platform, which is engineered for fast transient response and high operating temperatures—key for data center loads that swing sharply under AI and cloud workloads. The M55’s high-pressure common rail fuel system supports low NOx emissions and Tier 4 readiness, aligning with the most

CoolIT and Accelsius Push Data Center Liquid Cooling Limits Amid Soaring Rack Densities

The CHx1500’s construction reflects CoolIT’s 24 years of DLC experience, using stainless-steel piping and high-grade wetted materials to meet the rigors of enterprise and hyperscale data centers. It’s also designed to scale: not just for today’s most power-hungry processors, but for future platforms expected to surpass today’s limits. Now available for global orders, CoolIT is offering full lifecycle support in over 75 countries, including system design, installation, CDU-to-server certification, and maintenance services—critical ingredients as liquid cooling shifts from high-performance niche to a requirement for AI infrastructure at scale. Capex Follows Thermals: Dell’Oro Forecast Signals Surge In Cooling and Rack Power Infrastructure Between Accelsius and CoolIT, the message is clear: direct liquid cooling is stepping into its maturity phase, with products engineered not just for performance, but for mass deployment. Still, technology alone doesn’t determine the pace of adoption. The surge in thermal innovation from Accelsius and CoolIT isn’t happening in a vacuum. As the capital demands of AI infrastructure rise, the industry is turning a sharper eye toward how data center operators account for, prioritize, and report their AI-driven investments. To wit: According to new market data from Dell’Oro Group, the transition toward high-power, high-density AI racks is now translating into long-term investment shifts across the data center physical layer. Dell’Oro has raised its forecast for the Data Center Physical Infrastructure (DCPI) market, predicting a 14% CAGR through 2029, with total revenue reaching $61 billion. That revision stems from stronger-than-expected 2024 results, particularly in the adoption of accelerated computing by both Tier 1 and Tier 2 cloud service providers. The research firm cited three catalysts for the upward adjustment: Accelerated server shipments outpaced expectations. Demand for high-power infrastructure is spreading to smaller hyperscalers and regional clouds. Governments and Tier 1 telecoms are joining the buildout effort, reinforcing AI as a

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