Stanford HAI just published the AI 2026 Report, and the picture it paints is one of relentless, unsettled change — nowhere near finding any kind of equilibrium.
One thing, however, is fairly clear: artificial intelligence is not simply a tool we use. It reshapes us as we use it, and it changes the world around us — our attitudes, our workflows, investment patterns, the economy, the labour market. Everything.
Six Numbers That Say It All
A handful of figures from the Stanford report help frame the situation:
1. Investment more than doubled AI investment more than doubled in 2025 (+127.5%) and now accounts for 60% of total venture investment.
2. AI startups dominate AI startups made up 71% of all venture-backed companies in 2025.
3. Productivity surplus growing exponentially The productivity surplus tied to generative AI rose 54%, hitting $172 billion in the US alone, with the average value generated per user tripling over the same period. (So if you feel like you’re working harder than ever — you’re not imagining it.)
4. Widespread adoption (but shallow) 2025 saw AI deployed across 88% of organisations, though the use of AI agents is still in the single digits — which makes sense, because an agent is infinitely more complex than a chatbot.
5. Faster adoption than PCs or the internet In just three years, generative AI adoption reached 53% — outpacing the diffusion curves of both the personal computer and the internet.
6. Compute capacity through the roof Global computing capacity dedicated to AI has grown 3.3× per year since 2022.
The Real Problem: Who’s in Control
The point, though, is not the growth itself — it’s who controls it.
Over 90% of the models that actually matter are closed, opaque, and owned by a handful of companies — mostly American and Chinese — as is the underlying compute infrastructure.
Translation: a magnificent, colossal vendor lock-in embedded inside a monopoly. The dream of every self-respecting supervillain.
The Progress Paradox
To work faster — as you’re now expected to, because “everyone’s doing it” — you pay to access an opaque technology that learns from what you do while simultaneously making you obsolete. A technology, incidentally, controlled by four companies.
This is not a democratic revolution in access to technology. It is power concentration dressed up as innovation, and we are already neck-deep in it.
When Does Equilibrium Arrive?
Will we find a new equilibrium? Probably yes — but not in the way anyone wants to hear: it depends on how politically intolerable the concentration of power becomes.
If market fragmentation, social backlash, and disruption reach a point where the damage clearly outweighs the benefits, the system will rebalance — not out of ethics, but out of self-preservation.
The Choice
So pick a side, without illusions:
Either you build skills that survive without AI, or you accept becoming a disposable interface layered on top of systems you don’t control.
What do you choose?
Source
Stanford AI Index Report 2026 https://hai.stanford.edu/assets/files/ai_index_report_2026.pdf