May 12, 2026 · 1h 13m · a16z
Goldman Sachs Chairman on Why Finance Adopts AI Differently | a16z
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of The a16z Show, former Goldman Sachs CEO Lloyd Blankfein joins host David Haber to discuss risk management, crisis leadership, institutional culture, and the adoption of artificial intelligence. Drawing from his journey from public housing to Wall Street, Blankfein offers actionable insights on contingency planning, evaluating talent under stress, and navigating public scrutiny in tech and finance.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. How this is scored →
The host as informed peer 3.3 Guest teaching 4.0 Guest disagreement 0.3 The host pushing back 0.0
speaking balance: gold is the host, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 1:07:35 Dismissing calls to slow down technological progress
Lloyd forcefully rejects political hand-wringing from figures like Bernie Sanders, arguing that trying to pause technology or unlearn knowledge is a futile waste of time.
Hardest push from the host ▶ 10:27 Challenging the assumption on organic growth exceptionsDavid pushes back on the standard narrative of Goldman's purely organic growth by specifically raising the J. Aron acquisition as a major strategic exception.
Biggest teaching moment ▶ 11:00 Correcting the perception of the J. Aron acquisitionLloyd educates the host by revealing that Goldman internal leadership initially viewed the J. Aron deal as a disaster, finding value only by accident rather than through grand design.
The host holds their own ▶ 38:41 Articulating the firm-over-fund objective functionDavid demonstrates deep subject matter authority by presenting his own analytical framework contrasting short-term fund carry optimization with long-term firm moat creation.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The host as informed peer | Guest teaching | Guest disagreement | The host pushing back | Why |
|---|---|---|---|---|---|---|
| Episode Teasers: Contingency Planning and AI Transaction Risks | 2 | 1 | 0 | 0 | Introductory clips and light banter set a friendly tone. The host asks open questions about risk management and Lloyd's famous salad story during an active shooter event. | |
| Innate Temperament and Leadership Dynamics in Crises | 2 | 3 | 0 | 0 | Lloyd explains his crisis demeanor and offers guidance on picking board members who have experienced real crises. The host listens attentively without pushing back. | |
| Growing Up in Public Housing and Attending Harvard | 3 | 2 | 0 | 0 | David shares his own modest background in South San Diego and Harvard to build rapport. Lloyd recounts growing up in NYCHA public housing in Brooklyn with low initial expectations. | |
| The J. Aron Acquisition and Entering Goldman Sachs | 3 | 5 | 0 | 0 | Lloyd reframes David's premise about the J. Aron acquisition, explaining that Goldman initially viewed it as a disaster before discovering its entrepreneurial value. | |
| Risk Management Discipline vs. Risk Taking | 3 | 5 | 0 | 0 | Lloyd breaks down the core dual role of taking risk versus managing risk, emphasizing that effective risk management focuses on contingency planning rather than forecasting. | |
| Leadership Communication and Avoiding Hindsight Bias | 4 | 4 | 1 | 0 | David cites specific quotes from trading head Ashok, prompting Lloyd to discuss listening to junior staff and distinguishing between being wrong and being stupid. | |
| Contingency Planning and Hearing the Starting Gun Early | 3 | 4 | 0 | 0 | Lloyd illustrates risk preparedness using a track-and-field false start analogy, explaining that fast reaction stems from pre-planned contingency scenarios. | |
| Technology Adoption and Redundant Systems in Finance | 3 | 5 | 1 | 0 | Lloyd contrasts Wall Street's mandatory zero-error culture with Silicon Valley's move-fast-and-apologize ethos, highlighting why finance runs parallel legacy and new systems. | |
| SecDB Architecture and Enduring Design Lessons | 4 | 4 | 0 | 0 | David references Marty Chavez on SecDB. Lloyd contrasts partnership ownership culture—where leaders care about the whole enterprise—with corporate siloing. | |
| Capital Structure, Off-Balance-Sheet Vehicles, and Client Relationships | 4 | 5 | 0 | 0 | Lloyd details the post-Glass-Steagall capital shift and how public market expectations force a move toward off-balance-sheet fee business to smooth earnings multiples. | |
| Firm Over Fund Mentality and Platform Leverage | 6 | 3 | 0 | 0 | David demonstrates clear domain expertise by introducing his published essay on firm-over-fund alignment, which Lloyd validates and expands upon. | |
| Talent Development, Mentorship, and Avoiding Org Chart Bureaucracy | 3 | 3 | 0 | 0 | David asks how Lloyd instilled confidence in key mentees. Lloyd explains his drive to make people better rather than just trying to be liked. | |
| Financial Innovation: The S&P 500 Forward Cash-and-Carry Trade | 2 | 6 | 0 | 0 | Lloyd shares an anecdote about pioneering S&P 500 cash-and-carry trades for Middle Eastern clients requiring non-interest structures, demonstrating financial engineering history. | |
| Mark-to-Market Accounting and Navigating the 2008 Crisis | 3 | 5 | 1 | 0 | David brings up 2008 public backlash. Lloyd explains how strict mark-to-market accounting functioned as an early-warning risk management system during the crisis. | |
| Honoring Commitments, Career Cohorts, and Military Leadership Analogies | 3 | 4 | 0 | 0 | Lloyd discusses honoring commitments during crises, noting that peer cohorts track reputation across decades, comparing executive leadership to military command. | |
| PR Advice for Tech Leaders Navigating Public Scrutiny | 5 | 4 | 0 | 0 | David draws a sharp parallel between post-2008 Wall Street scrutiny and modern AI labs. Lloyd advises tech founders to explain their societal value before a crisis hits. | |
| Historical Parallels and Investment Dynamics in AI | 3 | 4 | 0 | 0 | Lloyd shares historical parallels to electrification and internet cycles, noting that current hyperscaler AI bets are uniquely driven by founder-principals risking their own equity. | |
| Systemic Risks, Reliability, and Societal Impact of AI | 3 | 5 | 2 | 0 | Lloyd highlights risks like the loss of operational intuition in automated systems, while pushing back against politicians like Bernie Sanders who want to freeze technological progress. | |
| Career Guidance and the Value of Broad Perspective and History | 4 | 3 | 0 | 0 | David asks for career advice. Lloyd recommends acquiring range and studying history, offering historical perspective on the late 1960s to contextualize modern political polarization. |
Statements from this episode (43)
Blankfein: Institutional risk management relies on contingency planning, not predicting the future
“Most of what we do with respect to risk is not so much predicting, it's a lot of contingency planning.”
Haber: The market is on the precipice of the largest IPOs ever
“We are on the precipice of some of the largest IPOs ever.”
Blankfein: Automated trading software introduces unprecedented scale for multi-billion dollar losses
“Before this technological age, not just AI, but in general, could you have had a mistake that could cost billions of dollars? Not really, but now you can leave a piece of software, could go out and do 70,000 transactions. The leverage in these things is themse…”
Blankfein: Active social media posting offers no value and high cancellation risk
“Being in the risk management business, I always know that everybody keeps doing that, and eventually you get, Canceled because you do, you know, you do something, you step over some invisible line that, you know, that nobody knew about. And so I realized that …”
Blankfein confirms Wall Street legend about eating salad during active shooter scare
“No, that was, yeah, that was real, but it wasn't, you know, it wasn't like I was hungry. I saw every, you know, I always used, in moments of crisis like that, I always tried to be disarming.”
Blankfein: In a Crisis, Key Leadership Task Is Keeping People Performing
“Most of the time, like at Goldman and in most, most of life, in a crisis time, the real important thing is just to get people to do their jobs, and to stop being, you know, stop, you know, don't be frozen, and don't be Don't submit to the chaos.”
Blankfein: Companies should prioritize board members who have survived real crises
“And that's why, I mean, my advice you know, when you pick board members, and by this a very, I'm turning something that's generic into a very narrow things, I think a good place to go is find people who've already gone through a crisis. Because to me, people w…”
Blankfein: J. Aron sold to Goldman Sachs at peak market valuation
“And of course, the savvy, streety guys at Jay Aaron extrapolated the value of the firm at the peaky, peaky part of its thing and sold itself to Goldman.”
Blankfein: Goldman Sachs rejected me before acquiring my firm, J. Aron
“I interviewed A lot, and being in New York, what do you go into when you're done? Whether you go to, you either become a consultant or, you know, go to Wall Street. I said, I'll go to Wall Street. And, you know, give a, you know, there I will. I will bestow my…”
Blankfein: Risk meetings should prioritize concrete action plans over debating probabilities
“And when you go around the table for those meetings, you're not so much interested in what people think about the future, where things will go. You just want to know, forget about what you think the likelihood, improbability of something happening is. What wil…”
Blankfein: Managing risk mostly means persuading hesitant employees to take more risk
“But, you know, we have a lot of risk takers. And I would say that the biggest challenge for management is the risk management side, which is really getting people to refrain from risk. Which is about a third of the time when you're in that business, and by the…”
Blankfein: Never tell reporting employees 'I already know' to prevent self-censorship
“One thing I never did, if somebody was calling to tell me something that was bothering them, that they saw an opportunity or a challenge, I never said I already know about it. Because I never wanted anybody to self-censor later.”
Blankfein: Managers must not treat wrong employees as if they are stupid
“But it's very important not, when something goes wrong, it's just, you know, when something is not right or somebody loses, it's very important not to treat somebody who's wrong like they're stupid.”
Blankfein: Evaluating past decisions using hindsight bias is managers' biggest fault
“And people make a mistake because the big fault of risk management or bosses or managers is they let after acquired information seep into their judgment of what they would have done at the time.”
Blankfein: Hindsight bias makes everyone look like a genius
“Once the present turns into the past, everybody's a genius.”
Blankfein: Contingency planning creates the illusion of prescient market forecasting
“The act of going through that thing makes you so alert and on it When things get triggered, and you so have a plan, you get off the mark so quickly that people think you did anticipate it, but when you really did is you heard the gun go off before anybody else…”
Blankfein: Financial sector leads all industries in tech adoption and spending
“No one who's a better adopter Or pays more. Sure. Except for obviously the hyperscales themselves who want to be the providers of the technology, but in terms of use of the technologies you know, the financial area is, you know, wants to be on top of it.”
Blankfein: Robinhood incorrectly claimed to offer government-insured accounts early on
“I looked at, you know, Robin Hood, great company. But early on, you know, they declared a kind of they declared that they had government-insured accounts that weren't government-insured. They had some slip-ups and a lot of apologies get made. You could do that…”
Blankfein: Tech adoption initially inflates costs before delivering long-term efficiency
“Technology in the first instance always, Augmented our costs, never detracted from it, but along, but as we go, as we went from one you know, from one lily pad to another things got better and more efficient”
Blankfein: Goldman Sachs gained a massive advantage from early risk system investments
“Our risk systems, by the way, we had a huge technological advantage because of what we invested in Early on.”
Blankfein: Goldman Sachs still uses the core structure of 30-year-old SecDB
“It was so good and so flexible that I think it's like, the system must be between 25 and 30 years old, and the core of it is still implemented.”
Blankfein: The repeal of Glass-Steagall forced Goldman Sachs to go public
“So we had to, you know, if J.P. Morgan was gonna become an advisor, we had to become a good lender and a good financier. So it meant that we had to have a bigger balance sheet. Couldn't run that on impermanent capital of a partnership. And so we had to go publ…”
Blankfein: Goldman Sachs required 25 years to preserve partnership culture post-IPO
“In an instant, legally, but it's taken 25 years to get it done in a way that it wouldn't undermine the partnership culture.”
Blankfein: Public markets penalize volatile earnings with lower valuation multiples
“In a private company, you care about the E, the earnings. In a public company, you care about P.E. And if you have volatile earnings, your shareholders don't like that. They reward you with a lower multiple.”
Blankfein: Off-balance-sheet vehicles lower unit revenue but boost ROE and P/E
“We shifted a lot of that to off-balance sheet vehicles, and by the way, means you have to do more of it. Because instead of earning a hundred-cent dollars, you're earning twenty-cent dollars with lower risk, and a higher P.E., and a higher R.R. Return on equit…”
Haber: Funds maximize short-term carry while firms build compounding competitive advantages
“The objective function of a fund is how do I generate the most carry with the fewest people in the shortest amount of time possible? And a firm, you know, you have to deliver exceptional returns, which is sort of a prerequisite for doing that well. But I think…”
Blankfein: Subordinating short-term ego to a firm yields greater long-term power
“We just said that if you subordinate it in the short term or during, at key times, in favor of a platform, you can exploit that platform, again, professionally, because the firm would have much more heft and power and authority.”
Blankfein: Goldman Sachs partnership grants a baseline presumption of intelligence
“Saying that, you know, look, I was a partner at Goldman. I'm not saying this is exclusive to Goldman. But saying your partner, at least people will, the presumption has shifted that you're not a dummy unless you prove you're a dummy, as opposed to other people…”
Blankfein: Goldman Sachs's first S&P cash-and-carry trade was a record $100 million
“And the first order that came in, and this was like in, back when this was real money, was for a hundred million dollars worth of this. That was by far the biggest trade ever.”
Blankfein: Young entrepreneurs win because they lack attachment to tradition
“Where the entrepreneurs are advantaged by their lack of attachment to history and tradition and the old way of doing things. Where the iconoclast is the in your business, the iconoclast and the young guy. Not only celebrate, they're the you know, they're the t…”
Blankfein: Mark-to-market accounting serves as a critical risk management early warning system
“Mark-to-market is not just a P&L system, it's a risk management system. Because we, that was our early warning that something was amiss here.”
Blankfein: Goldman Sachs fully hedged its 2008 AIG exposure with collateral agreements
“We were fully hedged because we had bought credit protection. Right. But we also had a collab. So we go, we, a single A credit, got a collateral agreement with AIG AAA.”
Blankfein: Goldman Sachs honored its Chrysler loan commitment during 2008 crisis
“There was a time we had this loan outstanding, you know, to Chrysler. I remember the CEO then at Chrysler calls me up. And are you gonna honor that commitment? And I said, yes. And I think it was due at a certain, and I said, and he said, can you do that now? …”
Blankfein: Early-career crisis conduct shapes long-term reputation with future industry leaders
“Your cohort is gonna going to run for all the important institutions 35 years from now, or 30 years from now, or 20 years from now. And you're gonna going to make your reputation with those people 30 years from now Believe it or not, are gonna be how they reme…”
Blankfein: Crisis defense is the worst time to build public trust
“And when you're being defensive and people are trying to kill you, it's not the best time to try to make friends with the public.”
Blankfein: Goldman Sachs took Tesla public before profitability was optional
“We took Tesla public before, at a time when, and this sounds like a quaint time, when companies didn't go public until they made money.”
Blankfein: Staying understated creates severe risks for influential market players
“I think there's no, you know, being modest and understated carries a lot of disadvantages. And I think you have to explain the role you Are in the market, so that there's some appreciation of what you do.”
Blankfein: AI hyperscalers are led by founder-owners risking their own equity
“The people who are, you know, the big hyperscalers are firms that are dominated by founding shareholders who are putting their own money where their mouth is. These aren't professional managers Making bets on the future with other people's money. This is their…”
Blankfein: The large language model market will eventually consolidate to two winners
“The world may not need 10 large language models. Maybe it needs four will be winners, and two will be very big winners, and the other two will get by, and maybe it'll get reduced over time to two.”
Blankfein: LLMs create systemic risk by removing institutional intuition and audit trails
“You know, one of the, you know, when you get, when you go into some of these large language models, you don't know the thought process. You lose intuition in these things. It used to be when I started out in the business, You know, people be shrieking each oth…”
Blankfein: AI automation will not create permanent mass labor displacement
“Turn back the clock at the beginning of the 20th century. More than half the country was in agriculture. Guess what? A single digit percentage is today. People found stuff to do. We'll find stuff to do.”
Blankfein: Halting AI development is futile because knowledge cannot be unlearned
“You're not gonna get people to be stupider than they are or unlearn things they've already learned. You can wish that atomic, that the atom had never been split, because maybe the adverse consequences of atomic bombs are worse than the benefits of nuclear powe…”