Information Matters Report #IM105 — Macro Briefing. May 2026.
Download the full briefing — 16 pages — here
Most analysis of agentic AI implicitly assumes one future — capabilities will keep leaping, or commoditisation will eat the model layer, or regulation will fragment the market. None of these is provably correct in 2026. All have plausible evidence behind them. Several could partially happen at once. The new Macro Briefing names eight scenarios, scores them together, and develops each as a coherent future for the agentic AI sector over the next three to five years.
Headline findings
- Eight plausible scenarios, no single forecast. We score every vendor against eight named futures — Frontier, Plateau, Verticals, Low-Cost Compute, Expensive Compute, Rewire, Inertia, Borders — producing a vector that exposes fragility a single-future analysis would miss.
- Plateau and Low-Cost Compute lead our weighting. Open-source closing on closed-source, inference price collapse, and on-device inference scaling all point to commoditisation as the dominant near-term force.
- Inertia is the under-discussed scenario. Every previous enterprise tech wave saw deployment lag the technology; we expect this one to be no different, with material implications for venture-backed challengers whose growth assumptions depend on rapid pilot-to-production conversion.
- Verticals strengthens steadily. Sector specialists in healthcare, legal, and financial services are outperforming generalists where it matters; sector-specific data partnerships become the most valuable assets in the cohort.
- Borders is real and increasingly material. EU AI Act enforcement begins August 2026; US–China divergence persists; the global market fragments into geographic spheres.
- Robustness across all eight is what enterprise buyers should optimise for. A vendor that scores well in one scenario but poorly in another is a fragile bet over a five-year integration commitment.
What the eight scenarios are for
The framework’s analytical move is to refuse the single forecast. Each of the eight scenarios is developed as an internally coherent future — the premise, the case, the dynamics of winners and losers, twelve-month indicators to watch, and what would falsify it. Vendors are then scored qualitatively against each, producing a Trajectory Profile vector across all eight. The interesting strategic implication is that vendors well-positioned across multiple scenarios are robust precisely because the combined future favours none in particular. Robustness is what enterprise buyers making multi-year integration commitments should be optimising for, even when they don’t articulate it that way.
Our weighted view in May 2026 is that Plateau and Low-Cost Compute lead, with on-device inference reinforcing both. Inertia is the under-discussed scenario most likely to compress aggressive challengers’ growth assumptions. Verticals strengthens steadily as sector specialists prove their data advantage. Borders becomes increasingly material as EU AI Act enforcement bites. Rewire happens unevenly — fast in tech-native sectors, slow elsewhere. Frontier remains a force in narrow capability domains. Expensive Compute is a tail risk we treat seriously.
The briefing closes with an indicators dashboard naming the twelve-month signals that would confirm or falsify each scenario. These are the things to watch, scenario by scenario, between now and the Q3 2026 Macro Briefing refresh.
Download the full briefing — 16 pages — here.
Report #IM105. Martin De Saulles, Principal Analyst.
