Conference Presentation
The $10 Trillion AI Revolution: Why It’s Bigger Than the Industrial Revolution
- Sequoia anticipates the AI revolution will be a transformation comparable to or exceeding the Industrial Revolution, identifying a $10 trillion opportunity in the cognitive sector with details to be shared in a four-section presentation.
- Investment themes are expected to unfold over the next 12 to 18 months, targeting the maturation of complex systems that combine general-purpose and highly specialized components to create applications by startups that have yet to be formed.
- Software as a Service spending is projected to exceed $650 billion, with AI expected to follow a similar expansion logic to grow the U.S. services market and potentially expand it tenfold, leading to the inclusion of many large, standalone public AI services companies in the S&P 500.
- Work processes are shifting toward over 100 percent leverage on tasks with reduced certainty regarding income manifestation, while portfolio companies forecast a minimum 10x increase in flops per knowledge worker, with optimistic scenarios suggesting consumption could reach 1,000 to 10,000 times the current level.
- AI agents are predicted for specific tasks such as individual customer interactions, requiring human correction for errors, and real-world measurement is expected to replace academic benchmarks like ImageNet as the gold standard, aided by reinforcement learning's recent prominence.
- Physical world AI is expected to mature through hardware and process applications, including accelerating manufacturing, while persistent memory is deemed critical for AI to penetrate broader work functions.
- Seamless communication protocols are anticipated to drive major applications in research, price finding, and purchase execution, with AI video expected to be available within a year and AI voice accelerating enterprise uses in logistics and over-the-counter trading.
- Security presents a significant opportunity across development, distribution, and consumption layers, with forecasts of hundreds of AI security agents for every human and agent, while open source models are considered essential to compete with state-of-the-art foundation models and ensure broad accessibility beyond well-funded giants.
- The time required to build a cognitive assembly line is expected to compress from years to just a few, fundamentally altering the trajectory of investment themes and AI adoption.