Goldman Sachs CEO David Solomon on cybersecurity, the IPO market, and why AI won't replace human relationships
Key Points
- Goldman Sachs' wealth and asset management platform now holds $4 trillion in assets and generates $3.6 billion in quarterly management fees, with 60% of firm revenues classified as durable and growing at better than 10% annually.
- Solomon runs tens of thousands of daily automated vulnerability scans using frontier AI models and argues the economy's real cyber exposure lies in software supply chains and legacy tech debt, fixable through public-private partnerships within five to seven years.
- Goldman wins large IPOs by having senior leaders make tens of weekly CEO calls, treating deal involvement as a cultural commitment rather than ceremonial activity.
Summary
Read full transcript →Goldman Sachs CEO David Solomon on cybersecurity, the IPO market, and AI's limits
David Solomon runs Goldman Sachs as an institution built on human relationships first, with technology as a multiplier rather than a replacement. That framing shapes how he thinks about nearly everything — from cybersecurity to IPO strategy to AI's long-term arc.
Business mix
Goldman now manages just over $4 trillion in assets across its wealth and asset management platform, which Solomon describes as roughly the fifth or sixth largest active asset manager globally. Management fees hit $3.6 billion last quarter, up from an annualized base of roughly $4.5 billion when the platform was assembled, and are growing at better than 10%. Combined with expanded lending revenues, Solomon says approximately 60% of the firm's revenues are now "durable" — a meaningful shift from eight years ago. The target is to grow that durable base at 8–9% annually.
Goldman's wealth management business stays firmly ultra-high-net-worth and institutional: governments, large enterprises, and roughly 45,000 ultra-high-net-worth clients. Solomon does not see that changing over the next five to ten years. Where AI creates efficiency gains, he frames the choice as taking higher margins or reinvesting in an expanded footprint — more companies, more institutions, more of the wealthy individuals Goldman doesn't currently touch.
“We've been running thousands of scans, tens of thousands of scans every single day, seven days a week, twenty four hours a day, looking at the vulnerabilities, patching. I can look at you straight away and say the firm is more secure today than it was seven months ago. The technology tools make us more powerful as connectors, but that human connection is not gonna change.”
Cybersecurity
Solomon says cyber has been his top concern for years, and recent events haven't changed that. Goldman has had access to frontier AI models for roughly seven to eight months and now runs tens of thousands of automated scans every single day to identify and patch vulnerabilities. He says the firm is demonstrably more secure today than it was seven months ago, while acknowledging that no network is fully protected.
His read on the recent wave of outages is that the patch cycle itself is partly responsible — firms scanning and patching at speed are introducing instability as a byproduct. The near-term risk he flags most clearly: a significant cyber incident that saps confidence or slows economic activity. He says he wouldn't be surprised if one occurs.
The bigger structural argument he makes is that software supply chains and legacy tech debt represent the economy's real exposure. His prescription is a public-private partnership to scan, patch, and upgrade software infrastructure at scale — expensive and slow, but feasible. He thinks a materially more stable system is achievable within five to seven years.
IPO market
On winning large IPOs, Solomon is explicit that senior leadership involvement is not ceremonial. He describes making tens of calls to CEOs weekly to check in on deal teams, and frames it as a cultural commitment: companies doing something as strategically significant as going public want direct access to the top of the firm. He names John Waldron, Dan Dees, Kim Posnitt, Matt McClure, and Anthony Gutman as part of that senior coverage structure.
AI and private markets
Solomon's view on AI lands somewhere between structural optimism and deliberate caution about the hype cycle. The biggest value creation from AI, in his view, won't be in infrastructure, large language models, or hyperscalers — it will come from the application layer built over the next couple of decades, across health, food, supply chains, and categories not yet imagined.
On the private markets question around Nvidia's leadership transition — specifically whether Colette Kress will continue Jensen Huang's pattern of strategic investments in AI startups — Solomon declines to speculate. He notes that Kress is a highly capable executive who deserves time to communicate her own approach, and argues the broader AI funding ecosystem is far larger than any one company or individual. He acknowledges the market is pricing in slower capital formation, and suggests that may be healthy.
The relationship thesis
Running through all of it is a consistent argument: Goldman's competitive edge is people-to-people trust, and no technology changes that. The Lululemon IPO story — the whole team walking into the pitch in head-to-toe Lululemon gear while the competing bank stood outside in suits — is his illustration of what earning a client's confidence actually looks like. The tools get better, but the mechanism stays the same.
Every deal, every interview. 5 minutes.
TBPN Digest delivers summaries of the latest fundraises, interviews and tech news from TBPN, every weekday.