Goldman Sachs CEO David Solomon on Running a Bank in the Age of AI

Episode Summary

Executive Summary: David Solomon argues that AI will reshape work at Goldman Sachs but not produce a mass white-collar job apocalypse. He says the firm will need fewer entry-level hires over time, yet human relationships, judgment, and apprenticeship remain central. The conversation also covers Goldman’s productivity gains, private-market financing for giant tech firms, public listings, market froth, and AI’s limits with messy data and creative work.

Main Topics: AI and white-collar work at Goldman Sachs (Priority: 5/5): Solomon rejects predictions of widespread job destruction, saying AI will change workflows, displace some roles, and likely reduce headcount somewhat, but not eliminate the need for human labor or create mass unemployment. Human relationships as a durable edge (Priority: 5/5): He argues that relationship-building, emotional intelligence, trust, and direct communication remain essential in banking and will become even more valuable as AI commoditizes routine analysis. Apprenticeship, training, and junior talent (Priority: 4/5): The discussion focuses on how Goldman should train younger employees when AI speeds up access to information. Solomon says the challenge is preserving learning, not just producing answers faster. Productivity gains and operational redesign (Priority: 4/5): Solomon distinguishes between measurable efficiency in back-office processes and harder-to-quantify gains in advisory businesses, citing internal process automation as a major source of value. Capital markets, IPOs, and private company financing (Priority: 4/5): He argues that more giant private companies will eventually raise equity because their capital needs are too large for private markets alone, and this does not ‘break capitalism.’ Market concentration and valuation froth (Priority: 3/5): Solomon says markets show greed and FOMO, especially in a narrow group of mega-cap tech stocks, but current valuations are not as extreme as prior bubbles when judged against earnings. AI’s limits in messy data and creative work (Priority: 3/5): He gives examples where models produce wrong answers on public internet data and says AI will aid creative fields like music and screenwriting, but cannot replace individual voice or human taste.

Key Arguments: AI will change jobs, but history suggests technology shifts labor rather than eliminating work altogether; Solomon rejects universal basic income-style apocalypse narratives. The most valuable banking skills are still human: calling clients, building trust, reading the room, and communicating directly. AI will make people more productive, but firms must redesign apprenticeship so junior employees still learn foundational judgment instead of just receiving instant answers. Goldman expects some reduction in entry-level hiring over the next few years, but not a dramatic collapse; current hiring remains near pre-COVID levels. Operational processes such as KYC and client onboarding are easier to automate and can produce clear, measurable productivity gains. Data quality is crucial: AI works well on clean internal data but can give unreliable answers when scraping the open web. Large private companies increasingly need public capital markets because their funding requirements are getting too large for private capital alone. The current market is crowded into a handful of stocks, but the top names have real earnings and are less bubble-like than late-1990s tech. AI-generated creative output can be useful, but it lacks personal voice, emotional resonance, and proper compensation frameworks for source artists. Solomon believes more firms will issue equity to manage long-term leverage, especially when markets make equity capital cheap relative to the risks of too much debt.

Data Points: Goldman interns: 2,500 - Solomon says Goldman currently has about 2,500 interns starting. Goldman new permanent hires: Approximately the same number as interns - He says July permanent new hires are around the same level as interns. Entry-level hiring decline cited: 16% - Referenced from Stanford data in the discussion of AI and hiring. Goldman staff touching AML/KYC process before redesign: 3,800 people - Solomon cites this as an example of a process being streamlined through 1GS 3.0. Goldman staff touching AML/KYC process after redesign: A few hundred people - Projected productivity improvement after process redesign. Goldman revenue growth since investor day: About 65% - Since January 2020, according to Solomon. Goldman earnings growth since investor day: About 140% to 145% - Since January 2020, according to Solomon. Goldman capital at IPO: $6 billion - He contrasts the firm’s capital base at the time of its public offering with today. Goldman current capital: $110 billion - Used to show long-run productivity and scale growth. Goldman current employee count: 45,000 - Part of the long-term comparison with the IPO era. Goldman employee count at IPO era: 16,000 - Historical comparison used to illustrate productivity gains. Goldman stock price milestone: Over $1,000 - He notes Goldman’s stock recently crossed this level. Goldman stock price at his start: Around $200 - Solomon says the stock has risen substantially since he joined. Top 10 S&P 500 concentration: Mid-to-high 30s percent of market cap - Discussing current market concentration in mega-cap names. Top 10 forward earnings multiple: Low 30s - Solomon says current top-10 valuations are high but not extreme. Late-1990s top-10 forward earnings multiple: High 40s to 50s - Used as a bubble comparison. Other 490 S&P 500 forward earnings multiple: About 17 to 20x - Used to argue the broader market is less expensive than the headline concentration suggests. Historical 10-year Treasury yield date: 15.9% on September 15, 1982 - A marker Solomon uses to describe the early-1980s financial environment. Master’s trivia example: Tiger Woods, Jack Nicklaus, and Nick Faldo - Used to illustrate AI hallucination and data quality issues. SpaceX relationship timeline: More than 15 years - Solomon says Goldman’s relationship with Elon Musk dates back over 15 years. Goldman-led Alphabet equity raise: $85 billion to $90 billion - He says the follow-on equity offering could reach this size if the green shoe is exercised.

Pivotal Quotes: "I am hugely optimistic about this technology and the impact that this technology can have on large-scale enterprises like Goldman Sachs..." — David Solomon: His core thesis on AI’s positive long-term impact despite near-term disruption. "I don't think we're going to wake up in a world where nobody works." — David Solomon: A direct rebuttal to apocalyptic AI labor narratives and UBI scenarios. "The challenge is how do we apprentice and give people the base of knowledge, but how do we also allow these tools to allow them to get out into the world so that they can really have a bigger impact faster?" — David Solomon: His view of AI’s effect on training and junior employee development.

Implications: Listeners should expect AI to change banking workflows, not erase the need for people. Firms that combine automation with apprenticeship, relationship-building, and data discipline will gain the most. Markets may also see more huge equity raises from capital-hungry private giants.