Goldman Sachs says small businesses are embracing AI, but fewer than 1 in 5 are good at actually integrating it

More than three-quarters of small business owners are using AI and over 90% say it’s working, but only 14% have embedded AI across core operations, according to a Goldman Sachs survey of 1,256 participants. Nearly 70% expect AI to help grow revenue, yet barriers—lack of technical expertise, crowded tools, and data privacy concerns (about half cite privacy)—and consumer skepticism limit deployment. A Chamber report found 58% of small firms used generative AI in 2025 while under 25% apply it to revenue-driving tasks like supply-chain optimization or customer identification. Adoption is broad but shallow, implying constrained near-term revenue upside without more training, implementation support, and clearer ROI evidence.

Analysis

The headline adoption numbers mask a structural two-speed market: a small cohort of digitally native SMBs will internalize AI-driven revenue engines, while the majority will outsource implementation and training. That bifurcation creates a multi-year services annuity opportunity—paid training, managed-inference, compliance tooling and verticalized models—where unit economics compress for model vendors but expand for go-to-market specialists who own customer touchpoints. Expect cloud compute demand to creep up in a lumpy fashion: elevated API/inference spend from a minority of accounts, gradual upsell from the majority as managed solutions lower operational friction.

Winners are therefore likely to be SMB-facing SaaS and payments platforms that can bundle AI as a recurring feature plus dedicated privacy/security vendors that neutralize small-business fiduciary risk; pure-play LLM vendors without SMB distribution or privacy-certified stacks will commoditize. Financial intermediaries with SMB distribution (commercial banks, payments lenders, and advisory outfits) can cross-sell financing and subscription training, turning one-off AI spend into multi-year revenue streams. Conversely, vendors that rely on large cap pilots to prove ROI will see slower unit growth and higher churn among SMB customers.

Key catalysts are measurable revenue uplift cases from mid-market cohorts and targeted public funding/training programs (6–24 months); triggers to reverse the trend include high-profile SMB data breaches or new consumer protections that raise compliance costs substantially and quickly. Near-term timing is tactical: expect meaningful monetization signals within 6–18 months from vendors who sell turnkey AI integrations, but durable margin expansion for integrators could take 18–36 months as churn and re-training costs settle.

Key Decisions for Investors

• Long GS (12-month call or stock): size 1–2% notional. Rationale: GS can monetize SMB training and lending programs; use a call or call spread to capture a 12–18 month monetization runway. Risk/reward: premium-limited loss with asymmetric upside if GS converts programs into recurring fee income; downside if SMB adoption stalls or macro credit costs spike.
• Long SHOP (Shopify) or equivalent SMB commerce exposure (6–18 month calls): size 1–2% notional. Rationale: merchants that actually embed AI in storefront optimization and content generation will show measurable GMV lift; Shopify is levered to that cohort. Risk/reward: high volatility/optionality—limited premium loss for multi-bagger upside if SMB monetization accelerates.
• Long CrowdStrike (CRWD) 9–12 month call spread (hedged exposure): size 0.5–1% notional. Rationale: surging privacy concerns among SMBs should drive spending on security and privacy-compliant AI controls. Risk/reward: call spread limits cost while capturing upside if cybersecurity budgets reaccelerate; downside is compressed if SMBs delay AI projects.
• Portfolio hedge via small QQQ put position (3–6 months): size 0.5% notional. Rationale: protects against a rapid, sentiment-driven drawdown if consumer/legislative pushback or a high-profile breach triggers a sudden rotation out of tech. Risk/reward: small drag in benign scenarios, outsized pay-off in a regulatory or panic de-risking event.