Goldman Sachs is pushing deeper into artificial intelligence — and what it could mean for the firm’s future.
In its 2025 shareholder letter, released Friday, Goldman’s leaders doubled down on their priorities tied to the recent refresh of their “One Goldman Sachs” initiative to streamline the bank’s services and boost revenue.
The firm described One GS as a “new operating model propelled by AI,” and highlighted six areas “ripe for disruption”: client onboarding and KYC, vendor management, regulatory reporting, lending, enterprise risk management, and sales enablement.
“This doesn’t just mean retooling our platforms,” the bank said. “It means taking a front-to-back view of how we organize our people, make decisions, and think about productivity, efficiency, and resilience.”
Goldman signaled that executing on its AI ambitions will depend heavily on talent, where a battle continues to unfold to nab top performers across Wall Street.
“Competition from within the financial services industry and from businesses outside the financial services industry, including the technology industry, for qualified employees has often been intense,” the firm said. It added that it has “experienced increased competition in hiring and retaining employees” tied to its technology initiatives and newer business lines.
That pressure is particularly acute in newer hubs central to Goldman’s operating model. The bank said 45% of its workforce is now based in strategic locations like Warsaw, Bengaluru, Hyderabad, and Salt Lake City. In those markets, Goldman often competes with firms that have a deeper local presence.
At the same time, the firm continues to draw significant interest from candidates. Goldman said it received more than one million experienced hire applications in 2025, up 33% from the prior year.
The biggest risks
In addition to shedding some light on its future AI bets, Goldman also explained what it sees as the biggest risks tied to its use of the technology that it believes shareholders ought to know.
The firm said “the legal and regulatory environment relating to AI is uncertain and rapidly evolving,” and warned that AI models may generate incorrect outputs.
Generative AI models, the firm flagged, are prone to making mistakes, which could, in a worst-case scenario, “result in the release of private, confidential or proprietary information, that reflect biases included in the data” they’re trained on. And it pointed to its use of “AI models developed by third parties” which the firm said makes it “dependent” on how those providers build their models. What’s more, bad actors could harness AI’s capabilities “to commit fraud and misappropriate funds and to facilitate cyber attacks,” it said.
Still, CEO David Solomon struck a bullish tone in the letter about his prognosis for AI. “We believe this technology is going to reshape the way we live and work,” he wrote, adding that “at the same time, there are significant questions” about the mind-bending speed of its adoption.
“With any new technology, there will be winners and losers,” the CEO added. “While there are likely to be periods of recalibration, in the long run I believe the net benefits from AI will accrue to many institutions as AI investment continues to build.”
AI has been a central focus for Solomon. The bank has partnered with developers like Cognition Labs to create unique products and rolled out its GS AI chatbot to its more than 47,000 employees.
Speaking at a conference in Europe last year, he said he wished Goldman could spend more than its roughly $6 billion technology budget, but noted at the time that investment levels were somewhat constrained by the need to deliver shareholder returns.
