# ibm Published articles for ibm. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## CVE-2026-8639 - IBM Verify Directory is affected by multiple vulnerabilities DevFeed: [CVE-2026-8639 - IBM Verify Directory is affected by multiple vulnerabilities]() Original publisher: [Read original article]() Author: CVEFEED.io; us.ibm.com Published: 2026-11-08T19:17:42Z Content type: news Language: en Sources: [Latest High/Critical Vulnerabilitiy Feed]() Topics: [Security]() Tags: [2026](), [console](), [cve](), [cve-2026-8639](), [directory](), [hijack](), [ibm](), [management](), [products](), [security](), [security-verify-directory](), [verify](), [verify-directory](), [vulnerabilities]() ### AI overview IBM Security Verify Directory 10.0.0 through 10.0.4 and IBM Verify Directory 11.0.0 are vulnerable to session fixation. An attacker can establish a session before a victim authenticates, then hijack that authenticated session to gain unauthorized access to the directory management console. The excerpt rates the vulnerability 8.8 (High); it provides no remediation details. ### Source excerpt CVE ID :CVE-2026-8639 Published : Nov. 8, 2026, 7:17 p.m. | 55 minutes ago Description :IBM Security Verify Directory 10.0.0 through 10.0.4 and IBM Verify Directory 11.0.0 is vulnerable to session fixation. An attacker can establish a session and, after the victim authenticates, hijack the authenticated session to gain unauthorized access to the directory management console. Severity: 8.8 | HIGH Visit the link for more details, such as CVSS details, affected products, timeline, and more... ## CVE-2026-19133 - TS4500 Tape Library addresses security vulnerability DevFeed: [CVE-2026-19133 - TS4500 Tape Library addresses security vulnerability]() Original publisher: [Read original article]() Author: CVEFEED.io; us.ibm.com Published: 2026-11-08T19:17:38Z Content type: release Language: en Sources: [Latest High/Critical Vulnerabilitiy Feed]() Topics: [account takeover](), [Access Control]() Tags: [2026](), [account](), [account-takeover](), [authorization](), [cve](), [cve-2026-19133](), [ibm](), [library](), [remote](), [security](), [storage](), [storage-ts4500-library](), [tape](), [vulnerability]() ### AI overview IBM Storage TS4500 Library versions 1.1.1.6 through 1.12.0.3 have an improper authorization vulnerability that could let a remote attacker who is already authenticated take over an account. The excerpt rates the issue HIGH with a severity score of 8.1. It provides no remediation steps or further details. ### Source excerpt CVE ID :CVE-2026-19133 Published : Nov. 8, 2026, 7:17 p.m. | 55 minutes ago Description :IBM Storage TS4500 Library 1.1.1.6 through 1.12.0.3 could allow a remote authenticated attacker to perform unauthorized account takeover due to improper authorization. Severity: 8.1 | HIGH Visit the link for more details, such as CVSS details, affected products, timeline, and more... ## CVE-2026-105467 - IBM QRadar SIEM is vulnerable to Open Redirect and Reflected XSS DevFeed: [CVE-2026-105467 - IBM QRadar SIEM is vulnerable to Open Redirect and Reflected XSS]() Original publisher: [Read original article]() Author: CVEFEED.io; us.ibm.com Published: 2026-11-08T17:16:35Z Content type: news Language: en Sources: [Latest High/Critical Vulnerabilitiy Feed]() Topics: [XSS](), [Security Information and Event Management (SIEM)]() Tags: [cve-2026-105467](), [ibm](), [qradar](), [qradar-security-information-and-event-manager](), [redirect](), [siem](), [xss]() ### AI overview IBM QRadar versions 7.5.0 through Update Pack 16 Interim Fix 02 and 7.6.0.0 through 7.6.0.5 are reported vulnerable to reflected cross-site scripting. Triggering it requires a victim who is authenticated to QRadar to open a specifically crafted URL. The source headline also mentions an open redirect, but the supplied description provides no details about that issue. ### Source excerpt CVE ID :CVE-2026-105467 Published : Nov. 8, 2026, 5:16 p.m. | 21 minutes ago Description :IBM QRadar 7.5.0 through 7.5.0 Update Pack 16 Interim Fix 02, and 7.6.0.0 through 7.6.0.5 is vulnerable to Reflected XSS. A victim must open a specifically crafted URL while authenticated to QRadar to trigger this. Severity: 8.1 | HIGH Visit the link for more details, such as CVSS details, affected products, timeline, and more... ## CVE-2026-17405 - Multiple Vulnerabilities in IBM API Connect DevFeed: [CVE-2026-17405 - Multiple Vulnerabilities in IBM API Connect]() Original publisher: [Read original article]() Author: CVEFEED.io; us.ibm.com Published: 2026-11-07T05:18:57Z Content type: article Language: en Sources: [Latest High/Critical Vulnerabilitiy Feed]() Topics: [Authorization](), [Access Control](), [ibm]() Tags: [2026](), [api](), [api-connect](), [authorization](), [cve](), [cve-2026-17405](), [ibm](), [information](), [keystore](), [multiple](), [object](), [remote](), [user](), [vulnerabilities]() ### AI overview IBM API Connect versions 10.0.8.0-10.0.8.12 and 12.1.0.0-12.1.1.2 have a broken object-level authorization flaw in keystore access checks. Because the checks do not verify that a requester may read a keystore object, a remote authenticated attacker could obtain sensitive keystore information. The supplied excerpt gives a severity score of 8.8 (high) but no remediation steps. ### Source excerpt CVE ID :CVE-2026-17405 Published : Nov. 7, 2026, 5:18 a.m. | 32 minutes ago Description :IBM API Connect 10.0.8.0 through 10.0.8.12 and 12.1.0.0 through 12.1.1.2 could allow a remote authenticated attacker to obtain sensitive information due to broken object-level authorization. Access checks on keystore objects do not confirm that the requesting user is permitted to read them, which could expose keystore information. Severity: 8.8 | HIGH Visit the link for more details, such as CVSS details, affected products, timeline, and more... ## IBM, AMD, And Oak Ridge Unveil Quantum Error Correction Platform DevFeed: [IBM, AMD, And Oak Ridge Unveil Quantum Error Correction Platform]() Original publisher: [Read original article]() Author: Jeff Burt Published: 2026-11-06T22:58:31Z Content type: news Language: en Sources: [The Next Platform: In-depth coverage of high end computing]() Topics: [Quantum Computing](), [High-Performance Computing](), [ibm](), [Distributed Systems & Parallel Computing]() Tags: [about](), [agentic](), [agentic-ai](), [ai](), [ai-models](), [algorithms](), [also](), [amd](), [api](), [compute](), [computing](), [error](), [error-correction](), [ibm](), [oak-ridge](), [over](), [platform](), [quantum](), [quantum-computing](), [quantum-error-correction](), [remarkable]() ### AI overview IBM, AMD and Oak Ridge are putting quantum and classical computing systems together at IBM's Poughkeepsie facility so researchers can test quantum error-correction methods and hybrid workflows with low latency. The platform pairs IBM's 120-qubit Nighthawk r2 with Oak Ridge's Umbra HPC system, using AMD Epyc Turin CPUs and Instinct MI355X GPUs for decoding, orchestration and AI workloads. Nighthawk's square-lattice connectivity offers more direct qubit connections than IBM's earlier heavy-hex design and supports standard surface-code error correction; IBM's API also enables mid-calculation measurements and real-time pulse control. Researchers can compare methods by qubit, time and classical-compute needs. One target is hierarchical error correction, which layers inner and outer codes to limit hardware and connectivity demands; IBM research suggests it could support a trillion logical operations at a physical error rate of one in 1,000. The excerpt frames the platform as support for research toward the Quantum Genesis Mission's 100-logical-qubit, 100,000-operation target by late 2028. ### Source excerpt Over the past few years, quantum computing companies have made remarkable strides in addressing the ... ## IBM, AMD Mix Quantum and AI at U.S. National Lab DevFeed: [IBM, AMD Mix Quantum and AI at U.S. National Lab]() Original publisher: [Read original article]() Author: Alan Patterson Published: 2026-11-06T14:10:57Z Content type: news Language: en Sources: [EE Times]() Topics: [Quantum Computing](), [ibm](), [High-Performance Computing](), [Artificial Intelligence]() Tags: [2](), [2026](), [absolute](), [accelerate](), [accelerators](), [ai](), [amd](), [analog](), [analysis](), [business](), [community](), [designline](), [digital](), [education](), [ee](), [ee-times](), [ee-times-university](), [eelife](), [eelive-electronic-design](), [eet](), [electrical-engineering](), [electronic-components](), [electronic-engineering-times](), [electronics](), [embedded-systems](), [error](), [error-correction](), [error-correction-system](), [fault-tolerant-quantum-computing](), [first](), [fundamentals-courses](), [hybrid](), [hybrid-quantum-computing](), [ibm](), [learning](), [mixed-signal](), [news](), [oak-ridge](), [opinion](), [part-search](), [product-search](), [quantum](), [quantum-computing](), [quantum-error-correction](), [research](), [semiconductors](), [software](), [system](), [tech-papers](), [technology](), [times](), [video]() ### AI overview IBM, AMD, and Oak Ridge National Laboratory are building a co-designed quantum and AI/HPC platform to study quantum error correction, a key step toward fault-tolerant quantum computing. It colocates IBM quantum processors, including the 120-qubit Nighthawk, with ORNL's Umbra system, which uses AMD Instinct MI355X GPUs and 5th Gen EPYC CPUs. Researchers will compare error-correction approaches and measure their needs in qubits, time, samples, and classical resources, exploring where GPU acceleration can help. IBM cites mid-calculation qubit measurement, real-time control, and access to raw measurement signals as capabilities for these experiments. The effort supports the DOE Quantum Genesis Mission target of 100 logical qubits and 100,000 hard operations by late 2028; IBM reports demonstrations of up to 70 logical qubits and 438 hard operations, leaving a stated roughly 230-fold increase in hard-operation count to reach the target. The text describes the work as research under development, not a demonstrated path to that goal. ### Source excerpt IBM, AMD, and Oak Ridge launch hybrid quantum-classical system to accelerate quantum-error correction research ahead of 2028. The post IBM, AMD Mix Quantum and AI at U.S. National Lab appeared first on EE Times. ## Chicago's bid to lead the country in quantum computing DevFeed: [Chicago's bid to lead the country in quantum computing]() Original publisher: [Read original article]() Author: Rafi Letzter Published: 2026-11-06T05:00:00Z Content type: news Language: en Sources: [IBM Research]() Topics: [Quantum Computing](), [High-Performance Computing]() Tags: [algorithm](), [algorithms](), [applications](), [chicago-quantum-exchange](), [collaboration](), [collective](), [company](), [computer](), [computing](), [convergence](), [discovery](), [drug-discovery](), [economic-development](), [ecosystem](), [ibm](), [industry](), [news](), [quantum](), [quantum-community](), [quantum-computing](), [quantum-network](), [quantum-systems]() ### AI overview Illinois is building a quantum computing hub around the National Quantum Algorithm Center and the future Illinois Quantum and Microelectronics Park, backed by a $500 million state investment and partners including IBM, the University of Chicago, and the University of Illinois. IBM has installed a Quantum System Two with a Nighthawk r2 processor at Hyde Park Labs; it will later move to the park, which is intended to house the center and computer. The excerpt describes regional research and application efforts, including IBM and UChicago's LASSQD method, which uses quantum hardware for molecular fragments and modeled a 26-electron, 24-orbital iron porphyrin fragment, and projects exploring drug discovery and power-grid optimization. Researchers and organizers present the facility as both a practical resource and a public symbol, while emphasizing decades of Midwest talent and institution-building as the basis for Illinois' ambitions. ### Source excerpt New IBM Quantum System Two supports Illinois' place at the center of the American quantum computing industry. ## Supermicro KV Cache Platform With IBM Storage Scale and NVIDIA Dynamo Claims 56x Faster TTFT at 130K Tokens DevFeed: [Supermicro KV Cache Platform With IBM Storage Scale and NVIDIA Dynamo Claims 56x Faster TTFT at 130K Tokens]() Original publisher: [Read original article]() Author: Brian Beeler Published: 2026-11-05T20:45:04Z Content type: news Language: en Sources: [StorageReview.com]() Topics: [Cache](), [Dynamo](), [NVIDIA Spectrum-X Ethernet]() Tags: [ai](), [amd](), [amd-epyc](), [cache](), [dynamo](), [enterprise](), [ethernet](), [flash](), [gpu](), [ibm](), [inference](), [kv-cache](), [long-context](), [nvidia](), [nvidia-dynamo](), [nvidia-spectrum-x-ethernet](), [scale](), [servers](), [storage](), [supermicro](), [tokens](), [ttft]() ### AI overview The tested Supermicro, IBM, and NVIDIA KV cache stack reused long-context state from shared flash, cutting time to first token by up to 56x at 130,000 tokens and increasing throughput 22x versus recomputing context. NVIDIA Dynamo manages cache placement and transfers blocks over NIXL; IBM Storage Scale ECE provides the shared tier on eight Supermicro Petascale servers with 983TB raw NVMe capacity and Spectrum-X Ethernet. In a cold-cache test of 200 requests, the tier delivered 4.26 requests per second versus 0.19; under 200GB/s of unrelated network traffic it sustained 3.6 requests per second, about 18x baseline. These results came from one RTX PRO 6000 Blackwell GPU, with host and local cache tiers disabled, and were not independently audited. The tests do not establish performance under a full multi-GPU cluster load; pricing and availability were not announced. ### Source excerpt Supermicro announced a KV cache storage solution with IBM and NVIDIA today that moves the key-value state behind long-context inference off the GPU and onto a shared all-flash tier. The package combines Supermicro Petascale storage servers on AMD EPYC 9005 processors, IBM Storage Scale, NVIDIA Dynamo, and NVIDIA Spectrum-X Ethernet, and the companies back it The post Supermicro KV Cache Platform With IBM Storage Scale and NVIDIA Dynamo Claims 56x Faster TTFT at 130K Tokens appeared first on StorageReview.com. ## IBM, Quantinuum and Intel signal a new phase in the quantum computing race DevFeed: [IBM, Quantinuum and Intel signal a new phase in the quantum computing race]() Original publisher: [Read original article]() Author: Maria Korolov Published: 2026-11-05T17:43:28Z Content type: news Language: en Sources: [Network World]() Topics: [Quantum Computing](), [High-Performance Computing](), [ibm]() Tags: [acquisition](), [cpus-and-processors](), [cpus-and-processors-data-center-high-performance-computing](), [data-center](), [edition-us](), [high-performance-computing](), [ibm](), [intel](), [language-en](), [quantum](), [quantum-computing](), [storytype-news]() ### AI overview The excerpt points to a shift toward integrating quantum hardware with classical computing and development software. IBM, AMD, and Oak Ridge paired IBM's 120 programmable-qubit Nighthawk with Oak Ridge's Umbra HPC cluster, which has 16 AMD GPUs and EPYC CPUs and connects to Nighthawk over Ethernet. The research platform lets researchers study real-time quantum error correction, which requires nearby classical compute to measure qubits, decode results, and apply corrections; it is not a commercial product. Quantinuum says its Nexus platform, compatible with current Helios and planned 2029 Apollo trapped-ion systems, aims to make fault-tolerance automatic for developers by 2029, including automatic code selection. Intel says acquiring Classiq would combine its spin-qubit manufacturing and software with Classiq's development platform to build a full-stack quantum business; the reported deal value is $1.9-$2.3 billion. ### Source excerpt Announcements from IBM and Quantinuum plus a recent acquisition by Intel show that the race to develop a quantum computer has shifted from focusing just on the qubits to integrating these computers with classical systems and software development platforms. To start with, IBM, AMD, and the DOE-funded Quantum Science Center at Oak Ridge National Laboratory are collaborating on a new research platform that pairs IBM's 120-qubit Nighthawk quantum computer with a classical computer in order to allow researchers to better explore quantum error correction, the organizations announced. This is a research platform rather than a commercial product, says IDC analyst Heather West. "Quantum error correction requires measuring qubits, decoding the results, and applying corrections in real time, which depends on substantial classical compute close to the quantum hardware," she says. "This platform gives researchers a place to test different error correction approaches on real hardware and understand what each requires in quantum and classical resources." It's part of a broader trend in the industry, West added. "Quantum computing is increasingly being developed as part of a larger hybrid environment alongside HPC and AI infrastructure," West told Network World. "And we are seeing more collaboration among national laboratories, quantum hardware providers, and classical computing vendors." "A quantum computer doesn't work on its own," says Jerry Chow, CTO of Quantum-Centric Supercomputing at IBM. "It needs to work together with other classical computing resources." You can think of a quantum computer as similar to a graphics accelerator, Chow told Network World, where the GPU steps in when needed to handle graphics-related tasks or matrix operations while the CPU continues to handle most other common tasks. "CPUs are really good at arithmetic and they'll keep doing that," he says. "GPUs are really good at tensors -- the new math that's behind all the AI stuff. And quantum computers a ## IBM, ORNL, and AMD Co-Locate the 120-Qubit Nighthawk With an MI355X Umbra Cluster for Quantum Error Correction DevFeed: [IBM, ORNL, and AMD Co-Locate the 120-Qubit Nighthawk With an MI355X Umbra Cluster for Quantum Error Correction]() Original publisher: [Read original article]() Author: Harold Fritts Published: 2026-11-05T16:44:22Z Content type: news Language: en Sources: [StorageReview.com]() Topics: [quantum mechanics](), [rocm]() Tags: [ai](), [amd](), [enterprise](), [error](), [error-correction](), [ibm](), [nighthawk](), [ornl](), [quantum](), [quantum-error-correction](), [qubit]() ### AI overview IBM, AMD, and ORNL have co-located IBM's 120-qubit Nighthawk processor with ORNL's Umbra system in Poughkeepsie to research quantum error correction. Nighthawk provides square-lattice connectivity, high circuit throughput, and real-time pulse control; Umbra's AMD MI355X GPUs support decoder training and evaluation, while 5th Gen EPYC CPUs handle orchestration. Researchers can test decoding and conditional execution close to the quantum hardware, including mid-circuit measurement and real-time responses, and compare methods by qubit, time, sample, and classical-resource costs. The work spans several code and decoder approaches, with hierarchical error correction among the targets. Connections to Oak Ridge's broader computing infrastructure are planned, but the announcement provides no deployment counts, performance benchmarks, completion date, or external access schedule. ### Source excerpt IBM, AMD, and the Department of Energy's Quantum Science Center at Oak Ridge National Laboratory unveiled a co-design platform for quantum error correction today, built around quantum and classical hardware sitting in the same facility. At IBM's quantum site in Poughkeepsie, New York, IBM quantum computers including the 120-qubit IBM Quantum Nighthawk processor are co-located The post IBM, ORNL, and AMD Co-Locate the 120-Qubit Nighthawk With an MI355X Umbra Cluster for Quantum Error Correction appeared first on StorageReview.com. ## Pitting a Modern Chess Engine on a Pentium PC Against 1996's Deep Blue DevFeed: [Pitting a Modern Chess Engine on a Pentium PC Against 1996's Deep Blue]() Original publisher: [Read original article]() Author: Maya Posch Published: 2026-11-05T16:30:51Z Content type: news Language: en Sources: [Hackaday]() Topics: [performance-optimization](), [Algorithms](), [Artificial Intelligence]() Tags: [algorithms](), [bot](), [chess](), [chess-computer](), [compilation](), [computer](), [computer-hacks](), [crafty](), [deep-blue](), [elo-rating](), [engine](), [games](), [ibm](), [llm](), [modern](), [optimization](), [pc](), [pentium-pro](), [ram](), [stockfish](), [system](), [time]() ### AI overview A 200 MHz Pentium Pro with 64 MB of RAM ran ported versions of Crafty and Stockfish, which were tuned with help from an LLM coding bot to test compilation flags and algorithms. The reported results suggest this era-appropriate PC could have beaten Deep Blue or even a grandmaster such as Kasparov in some games, despite Deep Blue's cited 2700 Elo and modern Stockfish's roughly 3700. The comparison raises hardware speed as a factor in Stockfish's advantage, but the excerpt gives no definitive proof: the author says certainty would require going back in time. ### Source excerpt Although premature optimization can be a really bad thing, late optimization is sometimes unavoidable because we simply didn't have know about certain algorithms yet to massively improve the solving of ...read more ## IBM, AMD and Oak Ridge join forces to launch quantum computing error-correction platform DevFeed: [IBM, AMD and Oak Ridge join forces to launch quantum computing error-correction platform]() Original publisher: [Read original article]() Author: Kyt Dotson Published: 2026-11-05T13:00:59Z Content type: news Language: en Sources: [SiliconANGLE]() Topics: [Quantum Computing](), [High-Performance Computing](), [ibm]() Tags: [accelerators](), [advanced-micro-devices](), [ai](), [algorithm](), [algorithms](), [amd](), [architectures](), [computing](), [emerging-tech](), [error-correction](), [high-performance-computing](), [ibm](), [ibm-corp](), [jay-gambetta](), [launch](), [nasdaq-amd](), [news](), [noise](), [nyse-ibm](), [oak-ridge-national-laboratory](), [platform](), [quantum](), [quantum-computers](), [quantum-computing](), [quantum-error-correction](), [the-latest](), [travis-humble]() ### AI overview IBM, AMD and Oak Ridge National Laboratory's Quantum Science Center announced a co-designed platform to study quantum error correction using IBM quantum computers, including the 120-qubit Nighthawk, connected to Oak Ridge's Umbra supercomputer with AMD Instinct MI355X GPUs and 5th Gen EPYC CPUs. The project targets a scaling challenge: physical qubits can err every few hundred operations, while complex algorithms may require billions of error-free operations. Researchers can measure qubits during calculations, respond in real time, control the timing of protective signals and inspect raw measurement data. IBM says Nighthawk's square-lattice connectivity supports fast testing and real-time control; the platform is intended to evaluate error-correction architectures, software and workflows at scale. The excerpt describes research capabilities and goals, not demonstrated error-correction performance. ### Source excerpt IBM Corp., Advanced Micro Devices Inc. and the Quantum Science Center at Oak Ridge National Laboratory today announced a new quantum error-correction platform that integrates quantum and high-performance computing. Gaining high performance and correctness in quantum computing requires protecting quantum information from noise. In many cases, researchers have paired quantum computers with classical computers to [...] The post IBM, AMD and Oak Ridge join forces to launch quantum computing error-correction platform appeared first on SiliconANGLE. ## ORNL, IBM, AMD launch co-design platform for QEC DevFeed: [ORNL, IBM, AMD launch co-design platform for QEC]() Original publisher: [Read original article]() Author: Rafi Letzter Published: 2026-11-05T13:00:00Z Content type: article Language: en Sources: [IBM Research]() Topics: [High-Performance Computing](), [ibm]() Tags: [amd](), [co-design](), [hardware](), [hpc](), [ibm](), [launch](), [news](), [oak-ridge](), [ornl](), [platform](), [quantum](), [quantum-community](), [quantum-network](), [science]() ### AI overview ORNL, IBM, and AMD have launched a quantum error correction co-design platform at IBM's site in Poughkeepsie, pairing IBM Quantum Nighthawk r2 processors with ORNL's Umbra HPC system, powered by AMD MI355X GPUs and 5th Gen EPYC CPUs. Co-locating quantum and classical hardware is intended to reduce delays in decoding and measurement-driven workflows. Researchers can develop and validate algorithms and error correction methods on real quantum systems without building the infrastructure themselves; the platform also enables exploration of GPU-based low-latency decoding and hierarchical codes. The text describes a research environment, not a completed fault-tolerant computer, and says its work will inform IBM Quantum Starling, planned for 2029. ### Source excerpt The U.S. DOE's Quantum Science Center at Oak Ridge National Lab is leading the integration of IBM quantum hardware with ORNL HPC clusters, powered by AMD. ## Porting a "Crappy" Flash Game to the IBM 5150 and Beyond DevFeed: [Porting a "Crappy" Flash Game to the IBM 5150 and Beyond]() Original publisher: [Read original article]() Author: Tyler August Published: 2026-11-05T09:00:11Z Content type: article Language: en Sources: [Hackaday]() Topics: [Game Development](), [ibm](), [browser-game]() Tags: [5150](), [adobe](), [assembler](), [browser](), [browser-games](), [bugs](), [c](), [c-plus-plus](), [cga](), [chip](), [coding](), [controller](), [cpc](), [crt](), [display](), [dos-3-3](), [flash](), [game](), [ibm](), [ibm-5150](), [pc](), [porting](), [retrocomputing](), [software-development]() ### AI overview The port displays a Flash game on an IBM 5150 by exploiting the CGA card's MC6845 controller: carefully timed assembler changes character height on the fly, creating a roughly 160 x 86 pixel window in text mode. The project uses Open Watcom V2 and implements the game in C as a large state machine, with testing across a range of DOS machines to find bugs. The excerpt identifies the display hack as the build's most impressive part; it does not provide further technical detail about the game port. ### Source excerpt Flash is obsolete and the IBM 5150 is obsolete. Therefore, by the transitive property of algebra, you should be able to run flash games on IBM's original PC, right? Okay, ...read more ## Trinetra.Q: Statistical and Behavioral Detection of Attacks on the Quantum-Error-Correction Layer and Classical Control Stack of Quantum Computing Systems DevFeed: [Trinetra.Q: Statistical and Behavioral Detection of Attacks on the Quantum-Error-Correction Layer and Classical Control Stack of Quantum Computing Systems]() Original publisher: [Read original article]() Author: Garrv Sipani Published: 2026-11-05T05:00:00Z Content type: article Language: en Sources: [arXiv: Cryptography and Security]() Topics: [Quantum Computing](), [Detection engineering](), [Cybersecurity]() Tags: [attacks](), [behavioral-detection](), [compilation](), [computer-security](), [cs-cr](), [decoder](), [detection](), [error-correction](), [hardware](), [heron](), [ibm](), [quant-ph](), [quantum-computing](), [security]() ### AI overview Trinetra.Q combines statistical and behavioral detectors to identify attacks on quantum error correction and the classical control stack that may leave physical error rates normal. It compares per-stabilizer firing, syndrome defects and repetition, decoder confidence margins, victim outputs against matched controls, canary-circuit results across jobs, and compilation/execution behavior with envelopes calibrated from honest operation; indications require statistically significant deviations. The abstract reports evaluation in simulation and on three production IBM Heron processors, including tests against adaptive attackers. A systematic negative map found most physical-layer attacks statistically null, with an implemented within-job feed-forward covert-channel primitive as the exception. ### Source excerpt arXiv:2611.05722v1 Announce Type: cross Abstract: Most quantum-computer security work targets physical-layer attacks on qubits, while two surfaces remain inadequately defended: the quantum-error-correction (QEC) syndrome-and-decoder pipeline, and the classical control and software stack that runs it. Tampering on these surfaces can corrupt a computation while the physical error rate appears normal, and can stay within a single-circuit integrity tolerance. We present Trinetra.Q, a detection layer of cooperating detectors, each monitoring a statistical or behavioral property of quantum-domain data against an envelope calibrated from honest operation, and raising an indication only on a statistically significant deviation. The detectors cover (A) QEC integrity over per-stabilizer firing rates, syndrome defect-count, and repeated-syndrome rate; (B) decoder confidence over the margin between minimum-weight error explanations in distinct logical classes; (C) behavioral output divergence, flagging crosstalk-class and novel attacks by their effect on a victim's output relative to a matched control; (D) control-stack tampering via known-answer canary circuits, built to be indistinguishable to a resident implant, with a cross-job consistency statistic; and (E) compilation/execution integrity. Each gates on significance, so honest operation raises no alarm. We evaluate in faithful simulation and on three production IBM Heron processors, and bound the physical attack surface with a systematic negative map: most physical-layer attacks return statistically null, the exception an implemented, measured within-job feed-forward covert-channel primitive. Every detector is also evaluated against an adaptive attacker with full knowledge of it. To our knowledge this is the first unified statistical-and-behavioral detection layer for the QEC layer and classical control stack of a quantum computing system, demonstrated on production hardware. ## webAI Releases TwIL-8B, an 8.8B Reasoning Model Scoring 98.2 on MATH-500 DevFeed: [webAI Releases TwIL-8B, an 8.8B Reasoning Model Scoring 98.2 on MATH-500]() Original publisher: [Read original article]() Author: Asif Razzaq Published: 2026-11-04T22:25:05Z Content type: news Language: en Sources: [MarkTechPost]() Topics: [llm-reasoning](), [LLM evaluation / benchmarking](), [Large Language Model](), [post-training](), [Local AI](), [Inference](), [Fine-tuning](), [model architecture](), [Post-training optimization](), [llama.cpp]() Tags: [2](), [2025](), [2026](), [agentic-ai](), [ai-infrastructure](), [ai-shorts](), [applications](), [artificial-intelligence](), [benchmarks](), [but](), [editors-pick](), [first](), [for-devs](), [granite](), [ibm](), [knowledge](), [language-model](), [large-language-model](), [livecodebench](), [machine-learning](), [math](), [model](), [models](), [new-releases](), [post](), [posts](), [python](), [reasoning](), [reasoning-model](), [releases](), [size](), [staff](), [tech-news](), [technology](), [webai]() ### AI overview TwIL-8B is an 8.79B dense reasoning model post-trained from IBM Granite 4.2 8B, with balanced and verbose variants. In webAI's self-reported, high-effort pass@1 results, it scores 98.2 on MATH-500 and 81.6 on LiveCodeBench (verbose), while knowledge benchmarks are weaker: GPQA Diamond is 70.2, 11.5 points below Qwen3.5-9B. On AIME 2026 it uses about 38% fewer output tokens than Qwen3.5-9B but scores 4.2 points lower; the article cautions that differing harnesses and tokenizers make token comparisons directional. Weights are available under a non-commercial license. Local use is supported through Hugging Face or llama.cpp GGUF builds; the 4.98 GiB Q4_K_M build is sized for CPU or modest-VRAM GPU inference. Reported scores used bf16, and the model card recommends a large generation budget (example: 32,768 tokens) to avoid truncating reasoning. ### Source excerpt webAI released TwIL-8B, an 8.8B reasoning model post-trained from IBM Granite 4.2 8B. It posts 98.2 on MATH-500 and 81.6 on LiveCodeBench, but trails similar-size models on knowledge benchmarks. The post webAI Releases TwIL-8B, an 8.8B Reasoning Model Scoring 98.2 on MATH-500 appeared first on MarkTechPost. ## Reading Why: Born-Measured Attention with Exact Word Evidence on Light and Qubits DevFeed: [Reading Why: Born-Measured Attention with Exact Word Evidence on Light and Qubits]() Original publisher: [Read original article]() Author: Teerawat Tongloy, Jiraporn Pongsopa, Rangsan Jomtarak Published: 2026-11-04T05:00:00Z Content type: news Language: en Sources: [arXiv: Machine Learning]() Topics: [Transformer](), [Quantum Computing]() Tags: [attention](), [circuit](), [class](), [coherent](), [cs-cl](), [cs-lg](), [embedding](), [error](), [gradient](), [hardware](), [ibm](), [inference](), [input](), [language-model](), [layer](), [light](), [measurement](), [natural-language](), [optics](), [physics-optics](), [quant-ph](), [qubits](), [tests](), [word]() ### AI overview The paper presents a transformer built from Born measurements that provides exact word-level evidence under fixed attention denominators: integrated-gradient contributions sum to each class probability above its no-information share, with reported numerical error at or below 10⁻¹¹. On e-SNLI, its 76.98% test accuracy is comparable to a same-size softmax transformer. Deleting the top-ranked evidence word changes its prediction in 45% of pairs, versus 32% for the softmax model using the same explanation method; attention rankings change predictions in 32%, despite better alignment with annotator-highlighted words. The model's weights compile to optical and qubit implementations. On IBM's ibm_kingston, Hadamard tests reproduced predictions and deletion behavior on two short pairs, but the word-level evidence did not yet survive hardware measurement; the abstract says it analyzes the shots and calibration needed. ### Source excerpt arXiv:2611.05355v1 Announce Type: cross Abstract: When a language model answers, we can usually only estimate why. We build a transformer in which every step is a Born measurement - attention $|\langle q|k\rangle|^4$ normalised over each row, a measured product in place of the feed-forward nonlinearity, and a Born readout. With its denominators held at their values for the input, each class probability is a polynomial in the word amplitudes, so its integrated-gradient evidence is computed by a finite quadrature with no quadrature error: the words' contributions sum to the class probability above its no-information share, to $10^{-11}$ or better. On e-SNLI natural language inference, with about 98k parameters outside the embedding, the model reaches $76.98 \pm 0.17$% test accuracy, level with a softmax transformer of the same size ($76.62 \pm 0.36$%). Deleting the word its evidence ranks first changes the answer in 45% of test pairs, out of the 53% any single deletion can change, against 32% for the softmax model explained the same way and 40% by its attention x gradient. Attention can look more like the words annotators highlighted, yet accounts for the answer less well: deleting the word it ranks first changes the Born Transformer's answer in 32% of pairs. The same weights compile exactly to optoelectronic optics and to qubits, and a coherent attention layer becomes one circuit equal to the model (852 CNOTs on 23 qubits). On IBM's ibm_kingston, reading only the attention overlaps by Hadamard tests, the device gave the model's answers on two short test pairs and kept or changed them under deletion as the model does; the evidence itself did not yet survive the device, and we work out the shots and calibration that measuring it on hardware would need. ## IBM puts agentic workloads at the heart of its AI operating model DevFeed: [IBM puts agentic workloads at the heart of its AI operating model]() Original publisher: [Read original article]() Author: Victoria Gayton Published: 2026-11-03T16:45:43Z Content type: news Language: en Sources: [SiliconANGLE]() Topics: [Agentic AI](), [ibm](), [AI Integration Strategies](), [AI Strategy](), [ai evaluation](), [ai observability](), [Artificial Intelligence](), [AI Infrastructure]() Tags: [actions](), [agentic](), [agentic-workloads](), [ai](), [ai-agents](), [ai-operating-model](), [applications](), [artificial-intelligence](), [autonomous](), [autonomous-systems](), [cube-event-coverage](), [data](), [data-governance](), [dave-vellante](), [enterprise-ai](), [enterprise-operations](), [enterprises](), [hybrid-cloud](), [ibm](), [ibm-corp](), [ibm-s-ai-operating-model-2026](), [ibmaioperatingmodel](), [infra](), [infrastructure](), [john-furrier](), [model](), [news](), [nyse-ibm](), [organizations](), [production](), [ritika-gunnar](), [salesforce](), [sap](), [scaling](), [scaling-ai](), [siliconangle](), [thecube](), [video-exclusive]() ### AI overview IBM's proposed AI operating model combines four components for agentic workloads: enterprise data, applications and workflows, automation, and hybrid cloud. The article says they must work together and integrate with existing systems such as SAP and Salesforce; automation provides observability that can inform model and agent selection. IBM's Ritika Gunnar emphasizes governing data access while preserving context and provenance, and keeping people involved in essential systems where tacit employee knowledge or AI maturity calls for human judgment. Logs and prompts can feed evaluation loops that compare decisions for accuracy, timeliness, and cost against intended business outcomes. The excerpt presents these as IBM's approach, without detailing implementation or independently evaluating results. ### Source excerpt Agentic workloads present a new operating challenge for enterprises putting artificial intelligence into production. As autonomous systems take on ongoing responsibilities, organizations must support decisions and actions across their data, applications and infrastructure. IBM Corp.'s approach draws on the experiences of clients scaling AI across their organizations. That work helped identify the requirements underpinning its [...] The post IBM puts agentic workloads at the heart of its AI operating model appeared first on SiliconANGLE. ## Seeing Why: Exact Evidence Maps for Quantum and Optical Image Classifiers DevFeed: [Seeing Why: Exact Evidence Maps for Quantum and Optical Image Classifiers]() Original publisher: [Read original article]() Author: Teerawat Tongloy, Rangsan Jomtarak Published: 2026-11-03T05:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Quantum Computing](), [Image](), [fourier transform](), [cifar-10]() Tags: [cifar-10](), [circuit](), [classifiers](), [cs-lg](), [equals](), [hardware](), [headroom](), [ibm](), [image](), [kingston](), [lens](), [loading](), [maps](), [mask](), [measurement](), [optical](), [physics-optics](), [quant-ph](), [quantum](), [qubits](), [simulation](), [test]() ### AI overview For quantum and all-optical image classifiers that process amplitude-encoded images linearly up to measurement, Born evidence maps (BEM) exactly decompose each class's probability above its no-information share into pixel contributions. For the quadratic class-probability form, this equals zero-baseline Integrated Gradients in closed form and can be measured on the classifying hardware. Removing the 5% of pixels with the highest BEM reduced accuracy from 73% to 9%, matching occlusion's effect without its 1,024 extra runs per image; gradient saliency removed less. Simulations with four phase plates reached 72.3% on Caltech-10, while IBM's ibm_kingston classified 300 Caltech-10 test images at 37.7% versus 10% chance (56.3% in exact simulation). Chip-estimated 2x2 maps for ten confident correct classifications correlated with exact maps (r=0.77). The maps also exposed a dataset shortcut; the excerpt does not specify it. ### Source excerpt arXiv:2611.01993v1 Announce Type: cross Abstract: When an image classifier decides, we can usually only estimate why. For quantum and all-optical classifiers that act linearly on amplitude-encoded pictures up to the measurement, with no trained classical network before or after them, the decomposition needs no approximation: each class's probability above its no-information share splits exactly into contributions of the encoded amplitudes, shown as Born evidence maps (BEM) over the picture and measurable on the hardware that classifies. For the quadratic form behind each class probability it equals Integrated Gradients from a zero baseline, in closed form. Deleting the 5 % of pixels with the highest evidence drops accuracy from 73 % to 9 %, as occlusion does with 1,024 extra runs per picture, whereas gradient saliency removes much less. One architecture serves both: a Fourier-plane mask followed by a brickwall circuit on qubits or by phase plates on a lens bench, read by a Born headroom readout with no trained parameter; centring each picture on its own mean lets dark features count as much as bright ones. In simulation, four phase plates at a single wavelength reach 72.3 % on Caltech-10 colour photographs (96.8 % on MNIST, 86.7 % on Fashion-MNIST, 47.9 % on colour CIFAR-10). On IBM's ibm_kingston, with the mask applied before approximate loading, 300 Caltech-10 test pictures were classified at 37.7 % against 10 % by chance (56.3 % in exact simulation), and 2x2 maps of ten correctly and confidently classified pictures, estimated on the chip, agreed with the exact ones (r = 0.77). The maps also revealed a shortcut in the data. ## IBM aims to simplify mainframe modernization with AI-powered insight DevFeed: [IBM aims to simplify mainframe modernization with AI-powered insight]() Original publisher: [Read original article]() Author: Michael Cooney Published: 2026-11-02T21:12:10Z Content type: news Language: en Sources: [Network World]() Topics: [Mainframe](), [Artificial Intelligence](), [Legacy Modernization](), [knowledge-management]() Tags: [ai](), [ai-powered](), [artificial-intelligence](), [artificial-intelligence-data-center-high-performance-computing-mainframes-networking](), [atlas](), [data-center](), [dependencies](), [discovery](), [edition-us](), [high-performance-computing](), [ibm](), [language-en](), [mainframe](), [mainframe-modernization](), [mainframes](), [map](), [mapping](), [middleware](), [modeling](), [modernize](), [natural-language](), [networking](), [os](), [outages](), [patch-management](), [storytype-news]() ### AI overview IBM's proposed Z Atlas would map dependencies across a client's z/OS infrastructure, middleware, and applications to help teams assess the effects of changes before production. IBM says automated discovery, topology modeling, dependency analysis, natural-language interactions, and AI workflows would support impact assessments and change evidence, easing work such as patch management, regulatory response, testing, and cross-team coordination. For example, teams could trace how a middleware upgrade may affect dependent applications and identify systems and processes to validate. The aim is to reduce outages caused by changes not proven safe and preserve shared system knowledge as experienced staff retire. Z Atlas is only at the statement-of-direction stage; the text describes intended capabilities, not a released product or demonstrated results. ### Source excerpt IBM is going to use AI to help and hasten mainframe customers looking to modernize seriously complex Big Iron operations. Understanding how the entire IBM Z environment is interconnected has become increasingly important as many mainframe teams still rely on manual analysis, fragmented tooling, and institutional knowledge to understand dependencies and evaluate risk across teams, creating a manual process that is slow and ineffective, wrote Skyla Loomis, general manager, IBM Z Software in a blog about the news. "As experienced professionals retire and modernization initiatives accelerate, the siloed knowledge problem is magnified with mainframe teams having trouble measuring the full impact of their changes before they reach production," Loomis stated. Enter Z Atlas, a native AI-based package Loomis says will map a client's entire z/OS stack across infrastructure, middleware, and applications. The package will then use that system intelligence to reduce the manual effort of tasks such as patch management, regulatory response, and knowledge sharing across every team, from systems programmers to application developers. The idea behind Z Atlas, which is only in the "statement of direction" phase, is to eliminate the leading cause of unplanned z/OS outages -- changes that aren't proven to be safe -- by mapping how a change impacts the z/OS stack by testing and validating it, Loomis stated. "For example, z/OS system programmers can use this solution to evaluate how a middleware upgrade might affect dependent applications, identify related infrastructure and operational processes that may require validation, determine which systems should be included in a test plan, or coordinate validation activities across application, infrastructure, and operations teams," Loomis wrote. "IBM Z Atlas' system intelligence is built on a foundation of automated discovery, topology modeling, dependency analysis, natural-language interactions, and AI-powered workflows to create impact assessme ## Critical Langflow Flaws Could Let Unauthenticated Attackers Run Code on Servers DevFeed: [Critical Langflow Flaws Could Let Unauthenticated Attackers Run Code on Servers]() Original publisher: [Read original article]() Author: The Hacker News Published: 2026-11-02T11:50:00Z Content type: news Language: en Sources: [Cyber Newsroom]() Topics: [Security research](), [code security](), [Local AI]() Tags: [agents](), [ai](), [ai-agents](), [api-security](), [application-security](), [artificial-intelligence](), [authentication](), [authentication-bypass](), [automatic](), [bypass](), [code-execution](), [control](), [critical](), [cve](), [cve-2026-93680](), [database](), [docker](), [enumeration](), [ibm](), [langflow](), [open-source](), [remote](), [vulnerabilities](), [vulnerability]() ### AI overview IBM reports that Langflow releases 1.0.0 through 1.12.8 contain two critical flaws that may allow unauthenticated authentication bypass or remote code execution; version 1.13.0 is the stated fix, and the excerpt says neither flaw was in CISA's Known Exploited Vulnerabilities catalog as of October 30. A Metasploit module for the authentication bypass can connect to the MCP endpoint and enumerate tools, but does not run server commands; its author lists a narrower affected range than IBM. Langflow 1.12.8 enables automatic login by default, while 1.13.0 disables it by default. To upgrade, back up the database, secret key, and uploaded files; prepare any automatic-login account for password sign-in; then upgrade the Python package or Docker image to 1.13.0 while preserving persistent data. Use the linked recovery guidance if sign-in fails; the excerpt warns that deleting or reinstalling Langflow can permanently remove flows. IBM gives no workaround for servers that cannot upgrade. ### Source excerpt Critical Langflow Flaws Could Let Unauthenticated Attackers Run Code on Servers Two critical flaws in Langflow could let a remote attacker with no account bypass authentication or run arbitrary code, IBM said last week. Langflow is an open-source tool for building AI agents and w ... Read more Published Date: Nov 02, 2026 (13 hours, 59 minutes ago) Vulnerabilities has been mentioned in this article. ## Chip Industry Week In Review DevFeed: [Chip Industry Week In Review]() Original publisher: [Read original article]() Author: The SE Staff Published: 2026-10-30T07:01:27Z Content type: news Language: en Sources: [Semiconductor Engineering]() Topics: [Chip design](), [tsmc](), [intel](), [AI Factory](), [High-Performance Computing](), [post-training](), [Nemotron]() Tags: [3d](), [3d-integration](), [advanced](), [advanced-packaging](), [advantest](), [agentic](), [agentic-ai](), [agnisys](), [ai](), [ai-models](), [alluxa](), [altera](), [amd](), [amkor](), [anthropic](), [applied-materials](), [arm](), [ase](), [axiomise](), [bank-of-america](), [benchmarking](), [blackstone](), [bosch](), [broadcom](), [brown-university](), [business](), [business-news](), [cadence](), [chinese-academy-of-sciences](), [chip](), [chipagents](), [citi](), [classiq](), [computing-news](), [cpu](), [cxmt](), [domain](), [earnings](), [eda-news](), [engineering](), [enpro](), [ferric](), [fraunhofer-iis](), [furiosaai](), [fz-julich](), [google](), [hardware-security-news](), [hbm](), [ibm](), [imec](), [industry](), [news](), [review](), [week-in-review]() ### AI overview Memory and advanced-packaging capacity are emerging constraints on AI expansion: SK hynix expects shortages through 2030 and says DRAM and HBM are effectively sold out through 2029; Samsung expects tight supply through 2028 and says its 2027 HBM is sold out. TSMC is reportedly considering a 20% CoWoS price increase. The roundup also reports fab and technology investments by Infineon, UMC, Intel Foundry, Lam Research and imec, alongside new domain-specialized chip-design AI models from Cadence and an Arm framework for benchmarking CPUs on agentic workloads. Security items include a critical EDK II UEFI flaw that could enable code execution during network boot before OS protections load, and nine Supermicro BMC vulnerabilities, including a critical unauthenticated command-execution flaw. These are selected developments from the supplied roundup excerpt. ### Source excerpt HBM shortages threaten AI expansion; TSMC weighs price hikes; advanced packaging ramps; domain-specialized AI models; Infineon and UMC boost fab investments; Intel's quantum move; Lam and imec advance 3D integration; Google expands TPUs; AI-driven earnings; CPU benchmarking methodology for agentic AI. The post Chip Industry Week In Review appeared first on Semiconductor Engineering. ## Latent Style-based Quantum Wasserstein GAN for Drug Design DevFeed: [Latent Style-based Quantum Wasserstein GAN for Drug Design]() Original publisher: [Read original article]() Author: Julien Baglio, Yacine Haddad, Richard Polifka Published: 2026-10-30T04:00:00Z Content type: article Language: en Sources: [arXiv: Artificial Intelligence]() Topics: [Quantum Computing](), [Artificial Intelligence](), [VAE]() Tags: [adversarial](), [artificial-intelligence](), [autoencoder](), [capacity](), [circuits](), [cs-ai](), [cs-lg](), [design](), [drug-discovery](), [drugs](), [gans](), [heron](), [ibm](), [inference](), [metrics](), [noise](), [performance](), [pipeline](), [q-bio-bm](), [quant-ph](), [quantum](), [qubits](), [sampling](), [training]() ### AI overview The proposed latent style-based quantum Wasserstein GAN combines a variational autoencoder for mapping molecules into latent space with a hybrid quantum GAN. It reuploads style-based noise at each rotational gate and uses a Wasserstein gradient penalty to improve trainability and mitigate mode collapse. On noiseless simulators, the authors trained and sampled with circuits up to 15 qubits; an inference-only IBM Heron run used 5 qubits to sample 2,500 molecules and examine quantum noise. The 110-parameter generator achieved statistically comparable overall performance and significant gains on selected drug-relevant metrics versus both the roughly 700,000-parameter MOSES latent GAN and a classical latent GAN matched for generator capacity. The excerpt reports these as design-model results, not evidence of clinical efficacy. ### Source excerpt arXiv:2603.22399v2 Announce Type: replace-cross Abstract: The development of new drugs is a time-consuming and expensive process, with average costs estimated to be up to around $2.5 billion. The first step in this process is the design of the new drug, for which de novo drug design, assisted by artificial intelligence, has blossomed in recent years and has revolutionized the field. In particular, generative artificial intelligence has delivered promising results in drug discovery and development, reducing costs and time to solution. However, classical generative models, such as generative adversarial networks (GANs), can be difficult to train and are prone to mode collapse. Quantum computing may be an avenue to overcome some of these issues, providing compact models with far fewer trainable parameters, thereby enhancing the generalizability of GANs. We propose a new latent style-based quantum Wasserstein GAN (QGAN) for drug design, combining a variational autoencoder that maps molecular structures into a latent space with a hybrid quantum GAN. Our pipeline is benchmarked against classical models of the industry-relevant MOSES suite. Style-based noise reuploading at every rotational gate and a Wasserstein gradient penalty are used to improve trainability and to mitigate mode collapse. Training and large-scale sampling with circuits of up to 15 qubits are performed on noiseless quantum simulators to validate our architecture. We present an inference-only demonstration on an IBM Heron quantum computer using 5 qubits, sampling 2,500 new molecules, to investigate quantum noise effects. Our default QGAN, using 110 trainable generator parameters, reaches statistically comparable overall performance and statistically significant gains on selected drug-relevant metrics relative to both the MOSES classical latent GAN baseline, which contains about 700,000 trainable parameters for the generator, and a classical latent GAN with matched generator capacity. ## One Array for AI: Inside the IBM FlashSystem and Broadcom Fibre Channel Architecture Behind Oracle AI Database 26ai DevFeed: [One Array for AI: Inside the IBM FlashSystem and Broadcom Fibre Channel Architecture Behind Oracle AI Database 26ai]() Original publisher: [Read original article]() Author: Brian Beeler Published: 2026-10-29T18:46:40Z Content type: article Language: en Sources: [StorageReview.com]() Topics: [Oracle Database](), [Database](), [Transactions]() Tags: [400gbe](), [ai](), [ai-infrastructure](), [architecture](), [array](), [broadcom](), [database](), [enterprise](), [enterprise-storage](), [fibre-channel](), [gpu](), [ibm](), [index](), [infrastructure](), [oracle](), [oracle-ai-database](), [oracle-ai-database-26ai](), [search](), [transactions](), [vector](), [vector-index](), [vector-search]() ### AI overview IBM and Broadcom describe validating Oracle AI Database 26ai on IBM FlashSystem over Broadcom Fibre Channel, with vector retrieval sharing storage queues with transactional and backup workloads. In the reported tests, capping vector volumes at 45,000 IOPS per volume improved OLTP throughput under contention; a separate BioASQ benchmark reached 85 queries per second on two 64G FC ports versus 43.7 on 32G, while search quality stayed within 0.1%. The text also reports ransomware alerts on an RMAN backup target in under 60 seconds and AutoZoning that creates host zones after SAN staff enable Target Driven Zoning on eligible ports. Capacity can grow sharply: the tested million-vector dataset occupied 16.24GB including its index, about 18 times its source text, before temporary build and transaction space. In the healthcare demonstration, images remained in object storage while embeddings and metadata stayed in Oracle. The findings are specific to IBM's lab setup; the report says clinical-scale GPU index-build performance still needs measurement and is sponsored by Broadcom. ### Source excerpt The biggest AI infrastructure decision many enterprises make in the next year won't be a GPU investment; it will be a database upgrade. Oracle AI Database 26ai is the long-term support release Oracle is steering its 19c installed base toward. 26ai puts vector search in the same engine that holds transactions. When the vector index The post One Array for AI: Inside the IBM FlashSystem and Broadcom Fibre Channel Architecture Behind Oracle AI Database 26ai appeared first on StorageReview.com. [Next page]()