Quick Thoughts: At the knee of the curve of disruption

AI is moving from narrative to measurable disruption, with markets now trying to price real changes in software economics, digital work and infrastructure spending rather than a speculative bubble. Jonathan Curtis argues that AI capabilities are improving faster than reported fundamentals suggest, as models deliver superior output across more digital tasks and begin to reshape productivity in software development first.

This is pressuring traditional software business models while reinforcing the case for sustained demand in AI compute and related infrastructure, particularly semiconductors, equipment makers and select utilities, industrials and materials. At the same time, investors are questioning the scale and concentration of planned AI infrastructure outlays as competition among model providers intensifies.

Curtis expects continued volatility but sees opportunity for advisors and investors who focus on beneficiaries of durable compute demand, real-world capacity constraints and early adopters able to harness proprietary data through agentic AI systems.