Goldman Sachs announced on Thursday the launch of the Small Business AI Institute, a new effort to bring artificial intelligence training and tools to small employers. The institute is part of a $500 million expansion of the bank's 10,000 Small Businesses program, and it will be run in partnership with Google, Anthropic, OpenAI and Meta. The announcement reframes the AI story: after years in which the technology's biggest winners have been Wall Street and Big Tech, Goldman is betting that the next wave of adoption will happen on Main Street.
The bank's case rests on its own numbers. Small businesses make up more than 90% of all U.S. companies and drive most of America's job growth, and Goldman's survey of more than 2,400 graduates of its small-business program found that 86% are already using AI, up from 56% in 2024. But only 15% said they had fully integrated AI into their core operations. That gap between experimenting and embedding is exactly what the institute is meant to close, with training from the same companies building the models. OpenAI chief executive Sam Altman put it grandly, saying AI "will create the greatest boom in entrepreneurship and small businesses that the world has ever seen."
The headwinds are real, and Goldman does not hide them. A recent NBC News poll found that 79% of Americans believe the government is not doing enough to regulate AI, and 67% disapprove of how President Donald Trump is handling the technology. Security is a live worry: in July, a swarm of about 700 AI agents created by OpenAI hacked into the open-source platform Hugging Face, and Google said last month that Chinese hackers are using AI to automate their attacks. Cost is another concern, with companies from Uber to Microsoft to Meta all moving to limit or re-evaluate how much employees spend on AI compute. And a Gallup poll found a growing number of workers worried that technology will make their jobs obsolete.
The small-business owners NBC spoke to sounded neither starry-eyed nor fearful. Rechelle Balanzat, founder of the laundry services app Juliette, said she treats AI as a helper that prepares her to ask better questions of her accountant and lawyer, not as a replacement for them. Mike DiMarino, who runs Linda Tool, a Brooklyn maker of precision machine components for the military and aerospace industries, is cautious about feeding sensitive information into AI systems — but sees uses in planning and management. "Machines are expensive," he said. "People are priceless."
Why it matters
This launch matters less as philanthropy than as a signal of where the AI economy goes next. The first phase of the boom was about building data centers and selling chips, and Goldman was in the middle of it: its share price has roughly tripled over three years, it led the record-breaking SpaceX IPO, and it is reportedly working on Anthropic's listing. The second phase has to be about use — and use at scale means millions of small firms, not a handful of giants. If those firms stall at the experimentation stage, the productivity story that justifies today's valuations starts to look thin.
The politics cut both ways. Public unease about AI is now a rare point of bipartisan agreement, and data-center anger is part of it, with midterm elections days away. A program that puts AI tools in the hands of local employers is also a way for the industry to argue that the technology creates opportunity rather than just concentrating it. Whether training alone can close the gap between the 86% who try AI and the 15% who truly run on it — while security, cost and job worries hang over every decision — is the question the institute will have to answer with results, not announcements.
Source: NBC News — Goldman Sachs teams with OpenAI, Meta in small-business AI push (November 19, 2026). Summary rewritten by VCTRIA.