Morgan Stanley Unpicks Oil Market Riddle as Stockpiles Swell

Stay Ahead, Stay ONMINE

Morgan Stanley Unpicks Oil Market Riddle as Stockpiles Swell

Global oil inventories have swollen at a rapid clip in recent months, but given the bulk of the increase has been in the Asia-Pacific, prices have been able to hold their ground for now, according to Morgan Stanley. While total stockpiles surged by about 235 million barrels in the five months to the end of […]

While total stockpiles surged by about 235 million barrels in the five months to the end of June, just 10% of that has been in the OECD, the region that’s “critical for price formation,” analysts including Martijn Rats said in a July 15 note, which posed the question “Is the oil market actually tight? Or not?”

Global benchmark Brent has gained ground so far this month, after advances in May and June, despite the drag from the US-led trade war and a major unwind of OPEC+ supply curbs. Although there are widespread expectations for a global glut in the coming quarters, crude’s near-term structure — with prompt prices above those further out — suggests current market tightness.

“What bridges this apparent contradiction is the uneven regional distribution of global inventory builds,” the analysts said. “Most of the inventory builds have taken place in locations that have less impact on prices, whilst inventories in key pricing centers have remained unusually tight – the builds have been in the Pacific, but Brent is priced in the Atlantic.”

Morgan Stanley cautioned that once the peak summer-demand season ends, a sizable surplus would be on the horizon again, although the bank still expected that only a “modest share” would show up in OECD stockpiles. These were seen rising by no more than 165 million barrels over 12 months, returning holdings to 2017 levels, when Brent fluctuated around $65 a barrel, the analysts said.

Brent price forecasts were retained at $65 a barrel in the fourth quarter, and $60 for each of the four quarters in 2026. Futures were last at $68.96.

Of the stockpile builds tallied in recent months, countries outside the OECD added about 100 million barrels, with China alone accounting for 48 million barrels of that figure, Morgan Stanley said. Elsewhere, the volume of so-called oil-on-water also increased, rising by 106 million barrels.

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.

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.

Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safety

This is a linkpost for https://bit.ly/cot-monitorability-fragile

Broadcom scales up Ethernet with Tomahawk Ultra for low latency HPC and AI

Broadcom Support for minimum packet size allows streaming of those packets at full bandwidth. That capability is essential for efficient communication in scientific and computational workloads. It is particularly important for scale-up networks where GPU-to-switch-to-GPU communication happens in a single hop. Lossless Ethernet gets an ‘Ultra’ boost Another specific area

F5 streamlines application delivery, security with new AI assistant and code generation service

The AI Assistant can also reduce documentation searches and help accelerate the handling of routine tasks, such as developing traffic configurations and deploying APIs, Ford said. In addition, it can help customers optimize traffic management and app delivery flows by offering real-time AI contextual analytics that can be applied enterprise-wide,

Nvidia to restart H20 exports to China, unveils new export-compliant GPU

China re-entry impact Nvidia’s announcements mark a bid to re-enter the world’s second-largest AI market under tightened US export controls. But this return may not mean business as usual. “Despite Nvidia’s market re-entry, Chinese companies will likely continue diversifying suppliers to strengthen supply chain resilience,” said Prabhu Ram, VP of

Will All Oil and Gas Companies Have to Adopt GenAI?

Will all oil and gas companies have to adopt generative artificial intelligence (AI) at some point? That was the question Rigzone posed to Toni Fadnes, the Chief Transformation Officer (CTO) of Pions, which recently changed its name from eDrilling. Responding to the question, Fadnes told Rigzone, “those in an actual competitive environment, yes”. “Those, for example, national oil companies that will get ‘favorized’ by governments etc. do not need to get better, be better, reduce emissions, etc., because there are no real consequences,” he added. When he was asked when that ‘some point’ will be, Fadnes told Rigzone that “it will happen sooner than we think”. “Everything in this ‘AI revolution’ has gone faster than first assumed and it keeps on speeding up,” he added. “For example, we can already now shave 40-70 percent of a drilling engineer’s time, allowing them do what they are/were supposed to,” he continued. Rigzone has contacted the International Association of Oil and Gas Producers (IOGP) for comment on Fadnes’ statements. At the time of writing, the IOGP has not responded to Rigzone. The IOGP describes itself on its website as the global voice of its industry, “pioneering excellence in safe, efficient, and sustainable energy supply”. “Our members, integrated energy companies, national oil companies, independent upstream operators, service companies, and industry associations operate around the globe, supplying over 40 percent of the world’s oil and gas demand,” the IOGP states on its site. In a release sent to Rigzone back in May by Fadnes, eDrilling announced that it had changed its name to Pions “to reflect the company’s broader vision, evolving technology, and growing ambitions”. The website of eDrilling has been updated to reflect the company’s rebrand. On the new website’s ‘about’ section, Pions states, “we build AI-powered engineers. Agents that think, act, and collaborate with

Google Signs PPAs for 3,000 MW of Hydro Power with Brookfield

Google has signed an agreement for up to 3,000 megawatts (MW) of hydroelectric power across the USA from Brookfield Asset Management and Brookfield Renewable. The hydro framework agreement is the “first of its kind” and the “world’s largest corporate clean power deal for hydroelectricity,” Brookfield said in a statement. The first contracts executed under the agreement are for Brookfield’s Holtwood and Safe Harbor hydroelectric facilities in Pennsylvania, representing more than $3 billion of power and 670 MW of capacity, the company said. Under the agreement, Google can procure carbon-free electricity from up to 3,000 MW of hydroelectric assets that will be relicensed, overhauled, or upgraded to extend the asset’s useful life and continue adding power to the grid, according to the statement. The first contracted assets consist of hydroelectric facilities in Pennsylvania that Brookfield is re-licensing. Brookfield and Google will initially focus on the mid-Atlantic (PJM) and mid-continent (MISO) electricity markets, with the flexibility to expand into other U.S. regions, the statement said. The 20-year power purchase agreements (PPAs) for Brookfield’s Holtwood and Safe Harbor hydroelectric facilities in Pennsylvania will support Google’s operations across PJM. The transaction structure allows Brookfield to maintain existing commitments to power consumers such as Amtrak from the Safe Harbor facility, the company said. Amanda Peterson Corio, Head of Data Center Energy at Google, said, “At Google, we’re dedicated to responsibly growing the digital infrastructure that powers daily life for people, communities and businesses. This collaboration with Brookfield is a significant step forward, ensuring clean energy supply in the PJM region where we operate. Hydropower is a proven, low-cost technology, offering dependable, homegrown, carbon-free electricity that creates jobs and builds a stronger grid for all”. Brookfield Asset Management President Connor Teskey said, “Our partnership with Google demonstrates the critical role that hydropower can play in helping

Eni Places 20-Year Order for 2 MMtpa from CP2 LNG

Eni SpA said Wednesday it had signed an agreement to buy two million metric tons per annum (MMtpa) of liquefied natural gas (LNG) for 20 years from Venture Global Inc.’s under-construction CP2 LNG in Cameron Parish, Louisiana. The Italian state-backed energy major expects deliveries, from phase 1 of the project, to begin by 2030. CP2 LNG is expected to start up 2027. “The agreement is Eni’s first long-term LNG supply from the United States and represents a significant milestone in Eni’s strategy to expand and diversify its global LNG footprint, enhancing portfolio flexibility. Part of these volumes will contribute to the diversification of Europe’s gas supplies”, Eni said in a statement online. “Venture Global’s proven success in project delivery will support Eni’s ambitions to grow its LNG portfolio to approximately 20 MTPA [million metric tons per annum] of contracted volumes by 2030, and to expand its trading business while meeting the evolving needs of customers across key markets worldwide”. Arlington, Virginia-based Venture Global has already been supplying Italy with LNG through its Calcasieu Pass and Plaquemines LNG facilities, according to the company. “Italy is an important ally and trading partner to the United States, and we are grateful for the trust of Eni as our newest customer. This deal marks a significant milestone for the company and is further recognition of our growing global energy leadership and strong record of execution”, Venture Global chief executive Mike Sabel said in a separate press release on Venture Global’s website. Last week another European company, Germany’s state-owned SEFE Securing Energy for Europe GmbH, announced the finalization of an agreement to increase its offtake from CP2 LNG to three MMtpa. The deal adds 750,000 metric tons a year to the 2023 agreement between Venture Global and SEFE. “Venture Global is expected to become Germany’s largest

Kinder Morgan Sees Increase in Profit on Higher Volumes

Kinder Morgan Inc. on Wednesday reported $615 million in adjusted net income attributable to common shares for the second quarter, up 13 percent from the same three-month period last year thanks to higher volumes of transported natural gas and refined oil products. The adjusted earnings per share (EPS) of $0.28 was in line with the Zacks Consensus Estimate. Kinder Morgan closed lower at $27.91 on the New York Stock Exchange on Wednesday. Adjusted net profit attributable to Kinder Morgan was $619 million. This metric is different from the company’s adjusted net profit attributable to common stock, which is equal to adjusted net profit attributable to Kinder Morgan minus net profit allotted to participating securities and adjusted net profit in excess of distributions for participating securities. Before adjustment for nonrecurring items, net profit attributable to Kinder Morgan was $715 million, up from $575 million for Q2 2024. Adjusted earnings before interest, taxes, depreciation and amortization grew six percent from $1.86 billion to $1.97 billion. Kinder Morgan’s natural gas pipelines carried 44.59 trillion British thermal units per day (Btupd) during Q2 2025, up from 43.12 trillion Btupd in Q2 2024, while sales volumes also rose to 2.83 trillion Btupd. Its products pipelines transported 1.02 million barrels per day (MMbpd) of gasoline, 369,000 bpd of diesel fuel and 325,000 bpd of jet fuel – up from one MMbpd, 354,000 bpd and 313,000 bpd respectively. “The Natural Gas Pipelines business segment’s improved financial performance in the second quarter of 2025 relative to the second quarter of 2024 was due primarily to continued higher contributions from both our Texas Intrastate system and Tennessee Gas Pipeline (TGP)”, said Kinder Morgan president Tom Martin. “Natural gas transport volumes were up 3 percent compared to the second quarter of 2024 primarily due to LNG deliveries on TGP, as well as new

Touchstone Posts Production Numbers from Trinidad Asset

Calgary, Alberta-based Touchstone Exploration Inc. said that its Central Block asset in Trinidad and Tobago had gross production volumes averaging 2,969 barrels of oil equivalent per day (boepd), or 1,930 boepd net, in the first quarter, The production consisted of approximately 16.74 million cubic feet per day (MMcfpd) of natural gas and 179 barrels per day (bpd) of natural gas liquids (NGLs), Touchstone said in a news release. The entity, acquired from Shell Trinidad Central Block Limited approximately $28.4 million and now renamed Touchstone Trinidad Central Block Ltd., holds a 65 percent operating interest in the onshore Central Block exploration and production license onshore in the Republic of Trinidad and Tobago. Heritage Petroleum Company Limited holds the remaining 35 percent participating interest. The Central Block asset includes four producing natural gas wells and a gas processing facility, according to the release. Based on preliminary field estimates, second-quarter gross production averaged 3,023 boepd, or 1,965 boepd net, consisting of approximately 17.05 MMcfpd of natural gas and 181 bpd of NGLs, the company said. Touchstone said that natural gas from the Central Block is sold under two separate contracts: one linked to liquefied natural gas (LNG) export pricing and the other to domestic market pricing, primarily for supplying Trinidad’s petrochemical sector. LNG sales are subject to vessel availability, which the company referred to as liftings. From January through April, 11 LNG liftings, including the associated liquids, were completed, totalling around 2.2 million British Thermal Units (MMBtu). An additional 11,065 MMBtu was sold into the domestic market, the company stated. Further, Touchstone said it completed site surveys for two additional well pads, each capable of supporting up to four drilling locations. The company said it is currently awaiting government construction approvals. Touchstone President and CEO Paul Baay, said, “We are pleased to

Market Underestimating Chance of Russian Gas Flow Disruption

The market is underestimating the chance of disruption to Russian gas flows. That’s what analysts at Standard Chartered Bank, including the company’s commodities research head Paul Horsnell, think, a report sent to Rigzone by the Standard Chartered Bank team late Tuesday revealed. “The ‘Sanctioning Russia Act of 2025’, introduced by U.S. senators Lindsey Graham (Republican) and Richard Blumenthal (Democrat), has 85 co-sponsors in the Senate (out of 100 senators),” the analysts stated in the report. “In a joint statement on 14 July the two senators noted President Trump’s decision to implement 100 percent secondary tariffs on countries that buy Russian oil and gas if a peace agreement is not reached within 50 days but pledged that they will continue to work on ‘bipartisan Russia sanctions legislation that would implement up to 500 percent tariffs on countries that buy Russian oil and gas’,” they added. “Market reaction to President Trump’s announcement was muted, perhaps as part of general trader fatigue in relation to proposed actions against Russia that were subsequently delayed or significantly watered down,” they continued. “There remains skepticism that the U.S. will take actions that might drive oil prices higher; however, we think it is much easier for the U.S. to put pressure on Russian gas flows as this would likely result in higher U.S. LNG exports at higher prices,” they said. “With a strong head of bipartisan steam behind measures in the Senate, we think the market is underestimating the chance of disruption to Russian gas flows,” the Standard Chartered Bank analysts noted. In the report, the analysts went on to state that they estimate that the EU’s net imports of Russian pipeline gas averaged 79.8 million cubic meters per day (mcmpd) in the first 14 days of July, based on European Network of Transmission System Operators for

Cisco upgrades 400G optical receiver to boost AI infrastructure throughput

“In the data center, what’s really changed in the last year or so is that with AI buildouts, there’s much, much more optics that are part of 400G and 800G. It’s not so much using 10G and 25G optics, which we still sell a ton of, for campus applications. But for AI infrastructure, the 400G and 800G optics are really the dominant optics for that application,” Gartner said. Most of the AI infrastructure builds have been for training models, especially in hyperscaler environments, Gartner said. “I expect, towards the tail end of this year, we’ll start to see more enterprises deploying AI infrastructure for inference. And once they do that, because it has an Nvidia GPU attached to it, it’s going to be a 400G or 800G optic.” Core enterprise applications – such as real-time trading, high-frequency transactions, multi-cloud communications, cybersecurity analytics, network forensics, and industrial IoT – can also utilize the higher network throughput, Gartner said.

Supermicro bets big on 4-socket X14 servers to regain enterprise trust

In April, Dell announced its PowerEdge R470, R570, R670, and R770 servers with Intel Xeon 6 Processors with P-cores, but with single and double-socket servers. Similarly, Lenovo’s ThinkSystem V4 servers are also based on the Intel Xeon 6 processor but are limited to dual socket configurations. The launch of 4-socket servers by Supermicro reflects a growing enterprise need for localized compute that can support memory-bound AI and reduce the complexity of distributed architectures. “The modern 4-socket servers solve multiple pain points that have intensified with GenAI and memory-intensive analytics. Enterprises are increasingly challenged by latency, interconnect complexity, and power budgets in distributed environments. High-capacity, scale-up servers provide an architecture that is more aligned with low-latency, large-model processing, especially where data residency or compliance constraints limit cloud elasticity,” said Sanchit Vir Gogia, chief analyst and CEO at Greyhound Research. “Launching a 4-socket Xeon 6 platform and packaging it within their modular ‘building block’ strategy shows Supermicro is focusing on staying ahead in enterprise and AI data center compute,” said Devroop Dhar, co-founder and MD at Primus Partner. A critical launch after major setbacks Experts peg this to be Supermicro’s most significant product launch since it became mired in governance and regulatory controversies. In 2024, the company lost Ernst & Young, its second auditor in two years, following allegations by Hindenburg Research involving accounting irregularities and the alleged export of sensitive chips to sanctioned entities. Compounding its troubles, Elon Musk’s AI startup xAI redirected its AI server orders to Dell, a move that reportedly cost Supermicro billions in potential revenue and damaged its standing in the hyperscaler ecosystem. Earlier this year, HPE signed a $1 billion contract to provide AI servers for X, a deal Supermicro was also bidding for. “The X14 launch marks a strategic reinforcement for Supermicro, showcasing its commitment

Moving AI workloads off the cloud? A hefty data center retrofit awaits

“If you have a very specific use case, and you want to fold AI into some of your processes, and you need a GPU or two and a server to do that, then, that’s perfectly acceptable,” he says. “What we’re seeing, kind of universally, is that most of the enterprises want to migrate to these autonomous agents and agentic AI, where you do need a lot of compute capacity.” Racks of brand-new GPUs, even without new power and cooling infrastructure, can be costly, and Schneider Electric often advises cost-conscious clients to look at previous-generation GPUs to save money. GPU and other AI-related technology is advancing so rapidly, however, that it’s hard to know when to put down stakes. “We’re kind of in a situation where five years ago, we were talking about a data center lasting 30 years and going through three refreshes, maybe four,” Carlini says. “Now, because it is changing so much and requiring more and more power and cooling you can’t overbuild and then grow into it like you used to.”

My take on the Gartner Magic Quadrant for LAN infrastructure? Highly inaccurate

Fortinet being in the leader quadrant may surprise some given they are best known as a security vendor, but the company has quietly built a broad and deep networking portfolio. I have no issue with them being considered a leader and believe for security conscious companies, Fortinet is a great option. Challenger Cisco is the only company listed as a challenger, and its movement out of the leader quadrant highlights just how inaccurate this document is. There is no vendor that sells more networking equipment in more places than Cisco, and it has led enterprise networking for decades. Several years ago, when it was a leader, I could argue the division of engineering between Meraki and Catalyst could have pushed them out, but it didn’t. So why now? At its June Cisco Live event, the company launched a salvo of innovation including AI Canvas, Cisco AI Assistant, and much more. It’s also continually improved the interoperability between Meraki and Catalyst and announced several new products. AI Canvas is a completely new take, was well received by customers at Cisco Live, and reinvents the concept of AIOps. As I stated above, because of the December cutoff time for information gathering, none of this was included, but that makes Cisco’s representation false. Also, I find this MQ very vague in its “Cautions” segment. As an example, it states: “Cisco’s product strategy isn’t well-aligned with key enterprise needs.” Some details here would be helpful. In my conversations with Cisco, which includes with Chief Product Officer and President Jeetu Patel, the company has reiterated that its strategy is to help customers be AI-ready with products that are easier to deploy and manage, more automated, and with a lower cost to run. That seems well-aligned with customer needs. If Gartner is hearing customers want networks

Equinix, AWS embrace liquid cooling to power AI implementations

With AWS, it deployed In-Row Heat Exchangers (IRHX), a custom-built liquid cooling system designed specifically for servers using Nvidia’s Blackwell GPUs, it’s most powerful but also its hottest running processors used for AI training and inference. The IRHX unit has three components: a water‑distribution cabinet, an integrated pumping unit, and in‑row fan‑coil modules. It uses direct to chip liquid cooling just like the equinox servers, where cold‑plates attached to the chip draw heat from the chips and is cooled by the liquid. The warmed coolant then flows through the coils of heat exchangers, where high‑speed fans Blow on the pipes to cool them, like a car radiator. This type of cooling is nothing new, and there are a few direct to chip liquid cooling solutions on the market from Vertiv, CoolIT, Motivair, and Delta Electronics all sell liquid cooling options. But AWS separates the pumping unit from the fan-coil modules, letting a single pumping system to support large number of fan units. These modular fans can be added or removed as cooling requirements evolve, giving AWS the flexibility to adjust the system per row and site. This led to some concern that Amazon would disrupt the market for liquid cooling, but as a Dell’Oro Group analyst put it, Amazon develops custom technologies for itself and does not go into competition or business with other data center infrastructure companies.

Intel CEO: We are not in the top 10 semiconductor companies

The Q&A session came on the heels of layoffs across the company. Tan was hired in March, and almost immediately he began to promise to divest and reduce non-core assets. Gelsinger had also begun divesting the company of losers, but they were nibbles around the edge. Tan is promising to take an axe to the place. In addition to discontinuing products, the company has outsourced marketing and media relations — for the first time in more than 25 years of covering this company, I have no internal contacts at Intel. Many more workers are going to lose their jobs in coming weeks. So far about 500 have been cut in Oregon and California but many more is expected — as much as 20% of the overall company staff may go, and Intel has over 100,000 employees, according to published reports. Tan believes the company is bloated and too bogged down with layers of management to be reactive and responsive in the same way that AMD and Nvidia are. “The whole process of that (deciding) is so slow and eventually nobody makes a decision,” he is quoted as saying. Something he has decided on is AI, and he seems to have decided to give up. “On training, I think it is too late for us,” Tan said, adding that Nvidia’s position in that market is simply “too strong.” So there goes what sales Gaudi3 could muster. Instead, Tan said Intel will focus on “edge” artificial intelligence, where AI capabilities Are brought to PCs and other remote devices rather than big AI processors in data centers like Nvidia and AMD are doing. “That’s an area that I think is emerging, coming up very big and we want to make sure that we capture,” Tan said.

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