The Signal

OpenAI introduced GPT-6.1 Terra for enterprise document processing, visual interpretation, computer-use automation and software development. The model was released through the API, ChatGPT Work and Codex, subject to applicable access controls. API pricing is US$1 per million input tokens, US$0.05 per million cached input tokens and US$5 per million output tokens—50% below GPT-5.6 Terra's standard rates. On OpenAI's Agents' Last Exam benchmark, Terra at medium reasoning effort improved performance by 15 percentage points at approximately one-third the cost of its predecessor. These are vendor-reported benchmark results, not independently verified customer ROI.

The Government of Canada announced a phased commercial equity investment of up to C$1.6 billion in Cohere, alongside private investors including NVIDIA, Bell and Schwarz Group. The initiative supports Canadian AI capabilities for enterprises and governments with requirements for data sovereignty, security and domestic infrastructure. Cohere committed to investing at least C$1.7 billion in additional Canadian computing infrastructure and creating at least 1,000 technical jobs by 2030. These are forward-looking commitments rather than completed infrastructure deployments.

Sema4.ai announced version 3.0 of its Enterprise AI Agent Platform, introducing governed team workspaces, reusable agent skills, shared memory and verified logic. The platform aims to reduce repeated model reasoning for recurring enterprise workflows while supporting collaboration across agent teams. Sema4.ai reports up to 80% lower LLM token usage and up to 4× faster results through reusable verified queries and logic. The release also adds interoperability with Claude Cowork. The reported performance improvements are vendor claims, without a disclosed independent enterprise-wide benchmark.

OpenAI and ServiceNow announced an expanded partnership connecting ChatGPT, Codex and OpenAI agents with ServiceNow's enterprise workflow platform. The planned integration targets IT incident management, change assessment, security workflows and application development, allowing authorised users to interact with ServiceNow records from conversational interfaces. The integration is intended to connect customers' existing ServiceNow instances while retaining ServiceNow Action Fabric permissions and business rules. ServiceNow also plans to expand ChatGPT Work and Codex adoption internally. No quantified customer ROI or general-availability date for the complete integration was disclosed.

IBM announced additional watsonx Orchestrate capabilities at TechXchange 2026, focusing on managing and evaluating agents across enterprise platforms. The enhancements include simulated workflow testing, configurable LLM-as-a-Judge criteria and deterministic verification of tool calls and parameters. IBM is expanding support for cross-platform agent discovery, evaluation, observability, identity and policy management. The announcement did not disclose a quantified productivity result or independently verified production-scale improvement.

Salesforce announced Agent Evals, API Evals and expanded Agent Fabric cost controls to help enterprises monitor AI-agent spending, quality and operational risk. The release combines spending limits, optimisation policies, a FinOps dashboard and a manually invoked agent kill switch with forthcoming automated evaluations. Salesforce reports Agent Fabric processing more than 1.2 billion requests daily. The FinOps Dashboard, Model Wallet, Cost Optimization Policies and manual Agent Kill Switch were generally available at announcement. Agent Evals and API Evals were scheduled for beta in November 2026 and general availability in January 2027. Salesforce also reported up to 75% lower token usage in optimisation test scenarios, not independently audited customer savings.

Akamai introduced Akamai Inference and Akamai Agent Sandbox, expanding its cloud services into managed AI inference and isolated agent execution. The offerings are designed to support model serving, enterprise data integration and constrained execution environments for agents accessing tools and applications. The announcement adds two infrastructure services but does not disclose directly comparable customer production figures for inference latency, execution isolation overhead or cost per workflow. The offering supports broader deployment choices across open-weight and hosted models.

OpenAI and Oracle announced plans to integrate Oracle applications and enterprise data with ChatGPT and to use OpenAI technology in additional Oracle offerings. Oracle also disclosed internal adoption figures and a recruiting research use case demonstrating significant reductions in preparation time. Oracle reported 130,000 active ChatGPT users and more than 95,000 active Codex users; these groups may overlap. Its recruiting team reported reducing market-research preparation from two to four days to 15–20 minutes using ChatGPT Work. The time comparison is vendor-reported and does not disclose total labour savings, sample size or review overhead. Oracle AI Workspace remains under development.

OpenAI published details of CodeRabbit's use of GPT-6.1 Sol as the primary model in its security agent. The workflow analyses pull requests, repository dependencies and related code to identify security issues that may not be apparent from an isolated code change. CodeRabbit reports applying the model to security findings across thousands of repositories and organisations, without publishing a precise detection or false-positive rate. CodeRabbit Secure can attempt exploit verification in an isolated sandbox and draft proposed fixes for human approval.

OpenAI's Enterprise and Edu release notes describe a change to custom GPT creation controls as part of a transition toward newer reusable workflows. The update affects the ability to create new custom GPTs in specified Enterprise workspaces, with FedRAMP workspaces treated differently. The 26 October release-note entry identifies the creation restriction and its workspace exception. It does not quantify affected organisations, migration costs or productivity impact.

OpenAI began rolling out GPT-6 Rosalind Research to a limited set of approved Enterprise customers for life sciences research through ChatGPT Work and Codex. The model is not available in the standard Chat tab and requires organisational and user-level Research approval. The rollout begins with a two-week Admin Preview, is disabled by default and requires administrators to enable approved researchers. Existing GPT-5.5 Rosalind Research workflows can continue. No measured research-productivity or scientific-accuracy improvement was disclosed in the release notes.

Dubai approved new expansion targets for Dubai AI Campus, part of the Dubai International Financial Centre. The initiative combines technology infrastructure, investment access, AI laboratories, regulatory support and facilities for companies developing and testing AI solutions. Since its May 2024 launch, Dubai AI Campus has attracted more than 500 AI companies, which collectively secured over AED 1 billion in investment, and contributed to more than 2,000 specialised technology jobs. Its new targets are more than 1,000 AI companies and 5,000 skilled jobs by 2030. A planned 100,000-square-foot facility will include GPU-ready infrastructure, secure AI development zones and digital twin environments.

Samsung Electronics reported third-quarter 2026 results highlighting strong semiconductor performance and demand for AI-related memory. The company identified AI servers, high-bandwidth memory and server storage as important contributors to the memory business outlook. Consolidated quarterly revenue reached KRW 195.3 trillion, up 14% quarter on quarter, while operating profit reached KRW 107.4 trillion. The Device Solutions division reported KRW 147.1 trillion in revenue and KRW 106.5 trillion in operating profit. Samsung also reported significantly increased HBM4 sales. These are consolidated and divisional figures, not AI-only financial results.

Enterprise Singapore and partner agencies announced initiatives to support AI-led business transformation during Singapore Week of Innovation and Technology 2026. The measures focus on structured transformation planning, smaller-business adoption and research-led innovation. The initiatives include tailored AI transformation roadmaps for 24 companies, workshops supporting smaller businesses and S$150 million for advanced-materials research and innovation. The roadmap count represents planned support rather than confirmed production AI deployments; the research allocation should not be treated as direct enterprise AI implementation spending.

Singapore and Germany announced a renewed strategic partnership and a Joint Work Plan for 2027–2028. The digitalisation and AI workstream includes cooperation on AI safety, security, testing, governance and standards, alongside trusted data flows and cybersecurity. The bilateral programme covers five strategic pillars, including digitalisation and AI, research and innovation, and trade. The 2027–2028 work plan is a policy cooperation commitment rather than a newly binding AI regulation or completed technical standard.