Artificial intelligence
AI capabilities and adoption keep advancing. Depth of production use, reliability in real workflows and economics all vary. We rely on dated evidence and specific deployments more than rankings, promised returns or total spending.
Research snapshot · updated Nov 12, 2026
What limits the AI build-out
Chips, power, grid connection, cooling, construction and staff can each hold up an AI data-centre project. Which one matters most differs from project to project. We do not have a reliable general figure for how long grid connections take, so each project has to be judged on its own site, power contract and delivery schedule.
State of the art (2026)
In September 2026 the newest models listed by the main providers were OpenAI's GPT-6 Astra, Anthropic's Claude Fable 5.1, Google's Gemini 3.8 Flash and xAI's Grok 4.6. Anthropic's Claude Mythos 5.1 was open only to approved customers. Nvidia began production shipments of its Vera Rubin systems in August 2026. For a buyer, the useful questions are how reliably a model completes a real workflow, how much supervision it needs and what it costs in use. A lead on a benchmark tells you less than those three.
Core thesis
AI is becoming more capable and more widely used, but the value a company gets depends on how well it puts AI to work and what that costs. Stanford's 2026 AI Index reports that 88% of surveyed organisations used AI in 2025. Gartner forecast in July 2026 that worldwide IT spending will reach $6.37 trillion in 2026, up 14.2%. That figure covers all IT, of which AI is one part. Neither number shows which part of the industry (chips, cloud, models or applications) will keep most of the profit.
The agent transition
Agents let a model carry out work in several steps. That makes the model more useful, and it raises the need for testing, security and clear responsibility when something goes wrong. Deployment is still limited. Stanford's 2026 AI Index found that in nearly every business function fewer than one in ten surveyed organisations had deployed agents. Most enterprise work has yet to be handed to autonomous systems.
Explore the deeper analysis in CanaryIQ
Explore what is covered below. These previews show the structure; log in to read the full analysis.
Signal stack
Evidence stacked leading → lagging
8 signals
Signal stack
Evidence stacked leading → lagging
Technology-native KPIs
Metrics that predict trajectory, tracked over time
2 tracked
Technology-native KPIs
Metrics that predict trajectory, tracked over time
Landscape map
Who builds what, and who depends on whom
246 players · 9 layers
Landscape map
Who builds what, and who depends on whom
Catalyst calendar
Dated events that will move the position
3 ahead
Catalyst calendar
Dated events that will move the position
Watchlists
Companies, people and papers, each with a condition for removing it
3 · 2
Watchlists
Companies, people and papers, each with a condition for removing it
Decision frameworks
What our view means for your job
Locked
Decision frameworks
What our view means for your job
Thesis changelog
When our view changed, and why
8 updates
Thesis changelog
When our view changed, and why
Change our mind
3 disconfirming conditions
Comparable wave
The historical analogue on the S-curve
Common mistakes
What the market gets wrong right now
You've read the verdict. The file is much deeper.
The full signal stack, technology-native KPIs tracked over time, the landscape of who depends on whom, the dated catalyst calendar, decision frameworks for every desk, live watchlists and a record of every time our view on Artificial intelligence has changed. All of it is in CanaryIQ.