OpenAI has five years to turn $13 billion into $1 trillion (techcrunch.com)

OpenAI is generating roughly $13 billion a year—about 70% of that reportedly from ChatGPT Plus subscribers paying $20/month—out of an 800 million user base where only ~5% pay. At the same time the company has contracted more than 26 gigawatts of compute capacity from partners including Oracle, Nvidia, AMD and Broadcom and has signaled plans to spend over $1 trillion across the next decade. To bridge the massive gap between current revenues and planned capital and operating outlay, OpenAI is pursuing a five‑year playbook: government contracts, shopping and video services, consumer hardware, and even selling compute via its Stargate data‑center initiative. For the AI/ML community this matters on multiple fronts. Technically, 26 GW of committed capacity underscores an enormous uptick in demand for GPU/accelerator cycles, power and cooling — pressure that will influence cloud pricing, hardware supply chains and where large models get trained and served. Strategically, OpenAI’s move toward vertical integration and diversified revenue (products, services, hardware, and cloud supply) could reshape industry economics and vendor relationships: many companies already depend on OpenAI’s APIs for critical contracts, creating concentrated systemic risk if performance or capacity falters. The next few years will test whether OpenAI can convert subscription momentum and product expansion into sustainable economics while scaling unprecedented infrastructure.