No tags match.
In one month, the enterprise software industry bought its way toward agents that act rather than advise. The controls to govern them are running behind.
The Model Context Protocol has shifted from a technical curiosity to a default question in enterprise software evaluation. Following CircleCI’s June 2026 MCP server release and Databricks’ move to govern MCP services in Unity Catalog, this article explains what MCP changes for integration cost, vendor lock-in, and AI governance, and sets out the practical procurement questions technology leaders should write into their evaluations.
At WWDC 2026, Apple unveiled a rebuilt Siri powered by a custom Google Gemini model reported to carry 1.2 trillion parameters, at a reported cost of roughly $1 billion a year. This article argues the deal is a template rather than a surrender: Apple rented the frontier model but kept the customer relationship, the data path, the context layer, and the interface. It sets out the four-part playbook enterprises should copy – task-level model bake-offs, gateway architecture that keeps providers swappable, owned small models for routing economics, and treating context as the proprietary asset – alongside the dependency and concentration risks of renting without an exit.
More than $40 billion has gone into enterprise AI, and 56% of CEOs say they have nothing to show for it – the gap has surprisingly little to do with the models
After every major outage the same advice circulates: go multi-cloud. For most organisations, the resilience that matters is an architecture decision, not a procurement one.
On 27 May 2026 the European Commission presented its Tech Sovereignty Package, including the Cloud and AI Development Act, restricting sensitive public-sector data from US hyperscalers. The restrictions stop at the public sector, but the underlying issue, namely US CLOUD Act jurisdiction over American-incorporated providers regardless of where data is stored, applies to every regulated business. This article explains the legal mechanism, the Dutch precedent blocking the Kyndryl and Solvinity acquisition, the trap of overcorrecting toward immature European providers, and the practical steps CIOs should take around data classification, portability, hybrid and multi-cloud design, and encryption key custody.
A CloudBees survey of more than 200 enterprise technology leaders found that 81% reported an increase in production issues linked to AI-generated code, even as 92% remained confident their code was production-ready before shipping. This article examines the verification gap created when AI generates code faster than teams can validate it, the rising and largely untracked costs that follow, the absence of clear governance ownership, and the practical steps engineering leaders can take to close the gap.
Google now processes 3.2 quadrillion tokens a month, up from 480 trillion a year ago. While vendors compete on per-token prices, the organisations pulling ahead on AI are the ones tracking a different metric entirely: tokens per business outcome. This article explains why per-token price is the wrong number to optimise, introduces tokens per resolved outcome as the unit that makes agentic AI economics legible, and sets out the three architectural patterns – task decomposition, context discipline, and evaluation-driven model selection – that create durable cost and performance advantages in agentic AI programmes.
A report from The Information revealed this week that Anthropic and OpenAI together account for roughly half of the combined $2 trillion revenue backlog held across Amazon, Microsoft, Google, and Oracle. This article examines the structural concentration risk this creates, the implications for enterprise cloud pricing and resource availability, and the practical steps CIOs can take to insulate their infrastructure strategies.
A techUK report has confirmed what CIOs have felt in their cost reviews for two years: cloud-first strategies are hitting sovereignty and cost walls. This article examines the egress tax problem, the implications of the UK Data Use and Access Act 2025, and what a genuine data-first architecture – one where compute moves to data, not the other way around – actually requires organisations to do differently.
AI data centres are projected to consume more energy than Germany and France combined by 2030, yet most enterprise AI strategies are still built on the assumption that compute is cheap and limitless. This article examines the energy and infrastructure constraints that are closing the brute-force compute era, and sets out the practical architecture shifts – task decomposition, context discipline, evaluation-driven model selection – that organisations need to make now.
88% of organisations confirmed or suspected AI security incidents this year, yet only 14.4% have full security approval for the agents they have deployed. Learn why agent adoption has created an entirely new attack surface and how to adapt your security and data architecture.
