The Cohere Blog Engineering section rounds up its latest releases and engineering articles, including North 2, Command A+, and the open-source speech recognition model Cohere Transcribe. The section also features developer technical articles on hardware-aware dynamic speculative decoding, automating fork maintenance with AI agents, LLM serving fairness, and vLLM integration for W4A8.
Matched fieldsBody
The Cohere Blog Secure AI thematic collection brings together multiple articles on enterprise-grade AI security, covering topics such as AI supply chain security risks, secure deployment of critical systems, enterprise AI security challenges and solutions, and mitigation of LLM security risks. Cohere has also released the Secure AI Frontier Model Framework and explained why enterprises choose private deployment of AI.
Matched fieldsBody
The Cohere Blog Research section rounds up its recent releases: North 2 is positioned as "AI without compromise," and the open-source speech recognition model Cohere Transcribe is launched, claimed to achieve the state of the art in open-source speech recognition. Cohere Labs has also released Tiny Aya to make multilingual AI more accessible, and introduced the megakernel serving engine for North Mini Code along with the RCP-nDCG@10 retrieval relevance metric.
Matched fieldsBody
Cohere releases North 2, touting "AI without compromise." Its public sector blog also lists updates including deepening cooperation with the Canadian government (August 19), modernizing Freedom of Information (FOI) request systems with AI (August 8), and collaborating with the Canadian and UK governments on safe AI (June 15).
Matched fieldsBody
Cohere has launched North 2, its enterprise-grade AI product, officially positioned as "uncompromising enterprise AI." The page also lists product release records for Embed 5, North Small Translate, Parse, North Automations, North Mini Code, Command A+, Cohere Transcribe, Model Vault, Rerank 4, and others, among which Embed 5 was released on September 30, 2026.
Matched fieldsBody
The Open Science section of the Cohere Blog features real-world use cases of Tiny Aya, showcasing projects people have built with open multilingual AI, as well as how Co/plot supports research workflows through visualization. The related articles were published on July 14 and June 3, respectively.
Matched fieldsBody
Cohere and PwC have entered into a partnership to jointly accelerate secure enterprise AI adoption. Cohere previously partnered with OpenText to bring trusted agentic AI to government and regulated industries, and established collaborations with the University of Waterloo, the University of Toronto, and Mila. Cohere announced its Partner Program on October 7, 2025, to advance enterprise AI development.
Matched fieldsBody
The manufacturing section of the Cohere blog brings together two articles: "Five Ways to Modernize Manufacturing with AI" introduces AI use cases, benefits, and implementation considerations in manufacturing, while "The Future of Intelligent Operations: Manufacturing and Supply Chains" focuses on manufacturing and supply chain scenarios.
Matched fieldsBody
Cohere has launched North 2, positioned as an enterprise-grade, no-compromise AI product. Cohere also lists model and product lines including North, Compass, Command, Transcribe, North Small Translate, Parse, Embed, and Rerank, as well as the Model Vault dedicated model inference platform and private deployment options for on-premises or isolated VPC environments.
Matched fieldsBody
The Cohere Blog financial services feature brings together multiple enterprise AI articles, including the August 15 piece "Building AI Agents in Finance: Reshaping Financial Services," the July 3 piece "The AI Advantage: How Financial Institutions Can Win with AI," and the June 3 piece "AI in Finance: Addressing Compliance, Efficiency, and Scale." It also covers banking AI terminology, 6 reasons banks choose private AI deployment, financial AI use cases and benefits, and the topic of de-risking the path from pilot to profitability.
The Cohere Blog financial services feature brings together multiple enterprise AI articles, including the August 15 piece "Building AI Agents in Finance: Reshaping Financial Services," the July 3 piece "The AI Advantage: How Financial Institutions Can Win with AI," and the June 3 piece "AI in Finance: Addressing Compliance, Efficiency, and Scale." It also covers banking AI terminology, 6 reasons banks choose private AI deployment, financial AI use cases and benefits, and the topic of de-risking
Matched fieldsBody
The Enterprise AI section of the Cohere blog aggregates its latest product releases and insights, including North 2, the document parsing model Parse, the workflow orchestration North Automations, as well as the speech recognition model Transcribe and the machine translation model North Small Translate. The content also covers the impact of small AI models on enterprises, the five stages of enterprise AI maturity, AI governance and security, and cooperation with NVIDIA to advance sovereign AI capabilities.
The Enterprise AI section of the Cohere blog aggregates its latest product releases and insights, including North 2, the document parsing model Parse, the workflow orchestration North Automations, as well as the speech recognition model Transcribe and the machine translation model North Small Translate. The content also covers the impact of small AI models on enterprises, the five stages of enterprise AI maturity, AI governance and security, and cooperation with NVIDIA to advance sovereign AI ca
Matched fieldsBody
Matched fieldsBody
HPE has launched its first four ProLiant Gen13 servers based on AMD's sixth-generation Epyc 9006 SP7 (codenamed Venice), targeting enterprise AI inference, agentic AI and analytics, virtualization, and other traditional workloads. The flagship DL585a Gen13 supports up to 8 GPUs and two Venice processors, will launch in March 2027, and will be compatible with GPUs from Nvidia, AMD, and Intel; the DL525 is a 1U single-socket system with up to 256 cores, launching in November 2026. Gen13 also introduces iLO 8 automated security and post-quantum cryptography readiness.
HPE has launched its first four ProLiant Gen13 servers based on AMD's sixth-generation Epyc 9006 SP7 (codenamed Venice), targeting enterprise AI inference, agentic AI and analytics, virtualization, and other traditional workloads. The flagship DL585a Gen13 supports up to 8 GPUs and two Venice processors, will launch in March 2027, and will be compatible with GPUs from Nvidia, AMD, and Intel; the DL525 is a 1U single-socket system with up to 256 cores, launching in November 2026. Gen13 also intro
Matched fieldsBody
NVIDIA used the same fine-tuning recipe to turn Nemotron into a specialized model for competitive programming and math olympiad: at IOI 2026, Nemotron-3-Ultra-CC scored 535.4 points (out of 600), and the IMO system scored 30 points (out of 42), exceeding the official gold medal line. The recipe includes a strong base model, domain problems and reasoning trajectory data, SFT and RL post-training, and a generate-evaluate-improve inference loop. The related SFT and RL checkpoints, training datasets, the Nemotron-IMO-Bench benchmark, and the NeMo-Skills inference pipeline have been open-sourced on Hugging Face.
NVIDIA used the same fine-tuning recipe to turn Nemotron into a specialized model for competitive programming and math olympiad: at IOI 2026, Nemotron-3-Ultra-CC scored 535.4 points (out of 600), and the IMO system scored 30 points (out of 42), exceeding the official gold medal line. The recipe includes a strong base model, domain problems and reasoning trajectory data, SFT and RL post-training, and a generate-evaluate-improve inference loop. The related SFT and RL checkpoints, training datasets
Matched fieldsBody
ElevenLabs Blog has published a contact center automation guide, explaining how AI can handle customer service processes with minimal human intervention, and presents three ElevenAgents deployment cases: Klarna achieved up to 10x faster resolution for 35 million US users, Rohlik's Maia automatically handles 90% of customer communications, and Getmobil achieved a 75% first-contact resolution rate and a 60% cost reduction.
Matched fieldsBody
DC Byte's latest EMEA analysis tracks 93 GW of IT capacity, with nearly a quarter concentrated in Frankfurt, London, Amsterdam, Paris and Dublin (FLAP-D), but land and power constraints are pushing development to spread into other markets. DC Byte's Oscar Gomez Comellas said about 60% of the region's project pipeline is still in early stages, with power supply being the main obstacle to delivery. EUDCA expects the number of European markets with more than 1 GW of installed capacity to grow from 5 to 12 to 14, and forecasts roughly 15% compound annual growth over seven years.
DC Byte's latest EMEA analysis tracks 93 GW of IT capacity, with nearly a quarter concentrated in Frankfurt, London, Amsterdam, Paris and Dublin (FLAP-D), but land and power constraints are pushing development to spread into other markets. DC Byte's Oscar Gomez Comellas said about 60% of the region's project pipeline is still in early stages, with power supply being the main obstacle to delivery. EUDCA expects the number of European markets with more than 1 GW of installed capacity to grow from
Matched fieldsBody
McKinsey's "Global Energy Perspective 2026" forecasts that data center power demand will climb at a compound annual growth rate of 24% through 2030, and in its most likely scenario, power supply will fall short of IT computing demand in all cases. More than 60% of operators have already combined self-built generation with grid connections, with gas turbines, fuel cells and battery storage becoming options for rapidly securing power, though competition for pipeline infrastructure could cause delays.
McKinsey's "Global Energy Perspective 2026" forecasts that data center power demand will climb at a compound annual growth rate of 24% through 2030, and in its most likely scenario, power supply will fall short of IT computing demand in all cases. More than 60% of operators have already combined self-built generation with grid connections, with gas turbines, fuel cells and battery storage becoming options for rapidly securing power, though competition for pipeline infrastructure could cause dela
Matched fieldsBody
AI data center power can surge and plunge within seconds or even milliseconds, and a sudden drop of hundreds of megawatts could cause wide-area grid disturbances; the industry is calling for chipmakers, operators, utilities and infrastructure parties to respond collaboratively. Nvidia electrical design engineer Sai Somayajula said rack power has risen from about 40 kW in the H100 era to about 150 kW with Blackwell, with Vera Rubin targeting about 240 kW, and under future 800V DC power distribution it could reach 1 MW per rack; Nvidia has added rack-level energy storage and control and introduced the DSX architecture. Oracle's Stargate data center campus is already 60% to 70% operational, with plans for gigawatt-scale capacity.
AI data center power can surge and plunge within seconds or even milliseconds, and a sudden drop of hundreds of megawatts could cause wide-area grid disturbances; the industry is calling for chipmakers, operators, utilities and infrastructure parties to respond collaboratively. Nvidia electrical design engineer Sai Somayajula said rack power has risen from about 40 kW in the H100 era to about 150 kW with Blackwell, with Vera Rubin targeting about 240 kW, and under future 800V DC power distributi
Matched fieldsBody
Uptime Institute's 2026 annual outage analysis shows that over the past three years, power problems have caused more than one-fifth of IT service outages, with UPS failures accounting for 39%, ranking first. As AI loads bring second-level fluctuations on the scale of tens of megawatts, the role of UPS is shifting from protecting IT loads to protecting the grid, and new architectures such as BESS and solid-state transformers (SST) are reshaping its position in the data center power supply chain.
Matched fieldsBody
AI writing is changing the structure of written language itself. There is no single ironclad proof for identifying it; judgment can only be built up from patterns such as vocabulary and sentence structure: words like quietly, genuinely, delve, and meticulous appear repeatedly in LLM output, and the "not X, but Y" construction and the three-part structure also form a common style. A 2023 Stanford study found that seven popular AI detectors on average misjudged 61% of essays by non-native English speakers as AI-generated, with one detector reaching a misjudgment rate of 97%. After analyzing more than 880,000 Reddit posts, news articles, and academic articles, Nature Human Behaviour found that the spread of LLMs is associated with a decline in writing style diversity.
AI writing is changing the structure of written language itself. There is no single ironclad proof for identifying it; judgment can only be built up from patterns such as vocabulary and sentence structure: words like quietly, genuinely, delve, and meticulous appear repeatedly in LLM output, and the "not X, but Y" construction and the three-part structure also form a common style. A 2023 Stanford study found that seven popular AI detectors on average misjudged 61% of essays by non-native English
Matched fieldsBody