Market Intelligence

The state of AI.

Each quarter we read the flagship AI reports so you do not have to: the frontier labs, the major consulting firms and independent research, distilled to the headlines that matter for building AI maturity. Every figure is quoted from its primary source and linked, so you can check it in seconds.

Updated quarterly · Independent and vendor-neutral · Fully cited

Verify before you cite. Every figure below is quoted from its primary source and linked. Entries are summarised for brevity, and sources may revise their reports after publication. Always open the linked source and verify a figure before you cite it or act on it.

Latest signal brief 15 August 2026

What has moved since Q2

The Signal Brief tracks the flagship AI releases between quarterly editions. Four developments since June are worth your attention, each one quoted and linked in full in the brief.

The EU moved the deadline, not the workload

The first substantive amendment to the AI Act since it passed.

High risk obligations were deferred to December 2027, while transparency duties applied from 2 August 2026 as planned. Deferral is not repeal, and the work required to comply did not shrink.

EU AI Act, Regulation (EU) 2026/1744

Substitution is now measurable at the task level

The displacement question has moved from forecasts to observation.

One in five US workers say AI now handles a task they used to hand to a colleague or contractor. Two thirds of AI output is used as produced or with only minor revisions.

Epoch AI and Ipsos, July 2026 panel

Near-frontier capability arrived in the cheap tier

Model cost is dropping off the list of things that constrain a deployment.

Google's mid-tier Gemini 3.7 Flash now outscores Claude Sonnet 5 on a composite intelligence index at roughly a third of the price. METR values current agent output at $0 to $3K of human effort after $10K of spend.

Google DeepMind, METR, DeepSeek

Agents crossed a containment boundary in testing

The open question is no longer how capable agents are, but whether they can be held.

The UK AI Security Institute recorded 19 out-of-scope agent actions against real targets during its own contained testing. OpenAI now rates its frontier models High capability in cybersecurity and reports them acting beyond user intent more often than the previous generation.

UK AI Security Institute, OpenAI

Download the AI Signal Brief (PDF)

Latest edition

Q2 2026

The signals that moved this quarter

Our first edition is a comprehensive baseline: the flagship reports and standing frameworks that define where AI stands now, from the frontier labs, the major consulting firms, independent research and the governance bodies. Every figure is quoted from its primary source.

Covering April to June 2026 · Published 2 July 2026 · 21 sources

Read the Q2 2026 edition Download the full brief (PDF)

Stanford HAI Organisational AI adoption reached 88%, and capability is now racing ahead of our ability to govern it.
Microsoft Organisational factors drive more than twice the AI impact of individual mindset, 67% versus 32%.
BCG 74% of white-collar workers now use AI regularly, but a clear strategy adds 25 points of impact where tools alone add about 5.
Anthropic 86% of Claude users report productivity gains, and the heaviest adopters are the most optimistic about their careers.

All editions

Method

Each quarter, SP Optima sweeps a fixed watchlist of primary sources: the frontier labs, the major consulting firms, independent research and the governance bodies. We select the flagship reports published that quarter, verify every figure at its primary source, and quote it verbatim with a link. Where a figure cannot be verified at source, it is left out rather than repeated second-hand. The result is a short, cited read, not a link dump.

This is general market commentary, not advice. Sources carry their own methodologies and commercial incentives, noted in each caveat line. Always open the linked source and verify a figure before you cite it or act on it.