Goldman Sachs is putting $500m behind small-business AI—who will it reach?
The bank says its new institute will combine training, research and capital access. A survey of 2,473 programme alumni found widespread AI use but shallow integration, with important selection and response-rate limits.
The short version
- What happened
- Goldman Sachs announced a $500 million commitment for a Small Business AI Institute and the next phase of its 10,000 Small Businesses programme, intended to reach another 10,000 businesses.
- Who it affects
- Owners and employees may gain practical training and lower-cost tool access. Benefits will depend on whether small firms can turn short courses into reliable workflows without exposing customer data, accepting poor outputs or carrying unaffordable subscription and integration costs.
- What remains unproven
- The programme is newly announced; participation, completion, lending and business outcomes are not yet available.
Keep in mind
The programme is newly announced; participation, completion, lending and business outcomes are not yet available.
The survey is self-reported and limited to Goldman Sachs 10,000 Small Businesses alumni.
Only 16.5% of invited alumni responded, creating substantial non-response risk.
Comparisons with 2024 do not establish change within the same businesses or causation by AI use.
Goldman Sachs funded the programme and published both the announcement and report; partner companies have commercial interests in wider AI adoption.
Multi-source analysis · 2 linked sources. Primary analysed source date: 22 October 2026.
Read the full report · 4 min The evidence, methods, implications and limitations
The commitment joins training, research and finance
Goldman Sachs says it will commit $500 million to extend its 10,000 Small Businesses programme and create a Small Business AI Institute. The plan has three pillars: AI education, research on how small firms use the technology and AI-enabled access to capital. Google will offer an AI certificate and three months of Google AI Pro to almost 20,000 owners and employees; Anthropic and OpenAI will support training, and Meta will provide access to its Muse advertising tools. The wider commitment is intended to serve another 10,000 businesses beyond the nearly 20,000 entrepreneurs the bank says its existing programme has reached.[1]
Adoption was high, but integration remained shallow
A companion survey conducted by Babson College between 10 and 24 August received 2,473 complete or partial responses from Goldman programme alumni. Among respondents, 86% reported using AI, up from 56% in the programme's 2024 survey, and 63% of users said they used it daily. Yet only 15% described AI as fully integrated into core operations. Most spending was modest: 73% of users reported less than $100 per month, including 16% using only free tools. Marketing, content and customer communication were more common uses than deep operational redesign.[2]
The distinction that matters
What the programme promises and what the survey shows
The announcement describes planned inputs; the survey records alumni's self-reported starting point.
Programme
$500m commitment and partner training
Another 10,000 businesses intended to be served
Current use
Survey 2,473 programme alumni
86% use AI; 15% fully integrated
Evidence gap
Compare claims with measured outcomes
No independent productivity or job-impact test yet
Workforce change was uncommon in the self-reports
Eighty-four per cent reported no workforce change because of AI, while 5% said they had added jobs and 6% had reduced employees; the remainder were unsure. Seventy-four per cent cited efficiency or time savings, and 93% of AI users reported a positive business impact. These are owners' perceptions, not audited payroll, productivity, revenue or profit measures. Rural respondents also reported lower adoption—80% versus 87% outside rural areas—and more often named technical expertise as a barrier, at 47% versus 39%.[2]
The denominator matters before generalising
The survey invitation went to 14,967 alumni, producing a 16.5% response rate; 2,117 responses were complete and 356 partial. Alumni already chose and completed a Goldman-backed business programme, so they may differ from the wider US small-business population in growth plans, networks and willingness to adopt technology. The institute is also an announced programme whose reach, course completion, loan decisions and business outcomes have not yet been measured. A stronger assessment would publish participation and completion denominators, demographic and geographic reach, matched business outcomes and independent evidence on whether training improves productivity without worsening exclusion or job quality.[1][2]