Artificial Intelligence - Market Opportunity

Costs to train AI Models Appear To Be Declining At Twice The Rate Of Moore’s Law

GPT-3 is a large language model that uses deep learning to generate text, ranging from translation to poetry composition.From 2015 to 2020, the cost to train a GPT-3 sized model dropped at an annual rate of 65%

Wright’s Law Models The Cost Declines Of AI Hardware And Software

Informed by Wright’s Law,  AI-relative compute unit (RCU) production costs could decline at a 39% annual rate and that software improvements could contribute an additional 37% in cost declines during the next eight years. The convergence of hardware and software could drive down AI training costs dowan at an annual rate of 60% by 2030.

AI Could Automate The Tasks Of Knowledge Workers And Boost Productivity by 140% by 2030

Example: OpenAI Codex can complete 37% of coding tasks, a percentage that is likely to increase significantly during the next few years, and save developers large amounts of time.

AI Could Accelerate Productivity Gains

By 2030, AI is likely to boost the output of global knowledge workers by 9% at an annual rate. Organizations will increase spending on enterprise software by 42% at an annual rate to $14 trillion a year. 

AI Software Should Boost Demand For AI Hardware

SaaS companies spend 50%+ of their cost of goods sold (COGS) on infrastructure hosting costs. As demand for AI software grows, the demand for hardware could rise accordingly.  

AI Revenues Could Scale To $14 Trillion By 2030

By 2030, AI software companies could produce $14 trillion in annual revenue collectively.generating returns of 48% at a compound annual annual rate during through to 2030. .