17JUL

Goldman cut 2% and refused the AI excuse

3 min read

14:01UTC

Goldman Sachs shed 2% of its staff in a quarter of record revenue, finishing at 46,200. Every incentive pointed towards crediting AI. David Solomon declined.

← Back to Fed hedges as four banks cut headcount Jump to analysis ↓

EconomicAssessed

Key takeaway

Goldman had every incentive to blame AI for its cut and declined to take it.

Goldman Sachs finished the June quarter with 46,200 staff, down 2% in three months, and disclosed it on Tuesday 14 July alongside record revenue. "Despite really robust revenue growth, headcount was down 2% quarter over quarter," chief financial officer Denis Coleman told the call.⁠¹ Chief executive David Solomon described AI as making "our people do more and be more productive", rather than as a reason anyone had left the building.⁠²

Weigh what he passed up. A firm cutting staff while revenue sets records has the most flattering explanation in modern corporate language available to it free of charge: the machines made us leaner. ResumeBuilder found 59% of employers had overstated AI's role in their own job cuts, and the Stanford estimate of roughly a million suppressed US hires a year sits in the same body of work. The pull on this beat runs almost entirely towards claiming more AI than the evidence carries. Goldman leaned the other way, and did so under no obligation to.

Hold two facts next to each other without collapsing them. Goldman Sachs Research produced the most bullish AI estimate in the literature: 1.5 percentage points of annual productivity growth over a decade, and a 7% rise in global output worth some $7tn.⁠³ Goldman Sachs the employer cut 2% of its people and would not say the technology took a single job. Neither statement is dishonest, and calling it hypocrisy would be lazy. A research desk forecasts what a technology could do to an economy over ten years. A management team reports what happened inside one firm over ninety days. The two questions are different, and only one of them carries legal exposure if the answer turns out wrong.

Which leaves the useful reading. A cut of this size, disclosed with no cause attached, passes through no labour instrument as an AI event, because nobody was required to call it one. Solomon's restraint is the honest version of the same silence that Bank of America and Citigroup kept this week, and it produces an identical hole in the record.

Deep Analysis

In plain English

Goldman Sachs cut its headcount by 2% in the same quarter it reported very strong revenue, and its chief executive, David Solomon, said AI is helping staff do more rather than replacing them outright. That framing matters because Goldman, like other Wall Street banks, is currently helping AI companies raise money, it is one of the underwriters on OpenAI's planned stock listing, while also cutting its own staff. Both things can be true at once: the bank believes in AI's commercial promise and is separately managing its own costs.

Deep Analysis

Root Causes

An investment bank's compensation ratio, staff pay as a share of net revenue, is the main lever management controls to protect margins in a fee-and-trading business.

Cutting headcount 2% while revenue grows compounds that ratio from both directions at once, a structural incentive that predates AI and would apply to any revenue gain, whatever its source.

What could happen next?

Meaning
Goldman's cut fits a 2026 pattern of headcount reduction on strong results already seen at Dell, HP, Cisco and CrowdStrike.

Source Landscape

Primary parallel: Electronic trading and decimalisation reduced floor-trading headcount at exchange-member firms through the 1990s and 2000s while revenue per trader rose, a case where a specific technology shift produced a headcount-revenue divergence investors rewarded as efficiency.

Counter-parallel: Wall Street headcount has more often tracked revenue expansion than technology substitution; firms that grew during bull markets, including the research-department expansions of the late 1990s, show that Goldman cutting staff on record revenue this quarter is the historically unusual direction, not the routine one.

Immediate: Goldman's 2% cut to 46,200 mirrors the pattern already set by Dell, HP and CrowdStrike, all of which posted strong results while shrinking staff this year.

Structural: Nvidia remains the sole large-scale counter-case, adding 6,000 staff on 85% revenue growth, underlining that headcount direction across the AI supply chain now tracks a firm's position in that chain, a chip supplier growing while service and advisory firms cut, more than it tracks AI adoption inside any one firm.

Consensus view: The Cambridge Centre for Alternative Finance's 2026 survey found the share of financial firms expecting net role reductions had doubled year on year to 24%, with 52% of finance firms now running agentic AI capable of sustained multi-step action, a sector-wide backdrop consistent with Goldman's 2% quarterly cut.

Counter-view: OpenAI's confidential IPO filing, made public on 9 June, describes a company burning close to $14bn a year even as revenue reaches roughly $2bn a month, underwritten in part by Goldman Sachs itself, a reminder that the AI infrastructure Wall Street is selling to clients has not yet demonstrated the returns being attributed to Wall Street's own cost discipline.

Key tension: Whether investment banks are cutting headcount because AI already delivers measurable productivity gains, or because record revenue and cost discipline are being rewarded by markets regardless of cause, a pattern already visible at Nvidia, Dell, HP and Cisco this year.

Sources:Benzinga

Mentions:Goldman Sachs →Denis Coleman →David Solomon →Cambridge →Citigroup →Cambridge Centre for Alternative Finance →Bank of America →Dell →Cisco →OpenAI →ResumeBuilder →Nvidia →

First Reported In

Update #17 · Fed hedges as four banks cut headcount

Benzinga· 17 Jul 2026

Read original →

Different Perspectives

Salesforce, Synopsys and TD Bank Group

Salesforce, Synopsys and TD Bank Group each filed quarterly disclosures in late August booking restructuring charges, or none at all, without naming AI as a cause. Their silence matters because Challenger's tracker shows AI as a stated reason fell to fourth place in August even as the year-to-date AI-cut total still leads at 116,175.

Singapore, South Korea, Taiwan and Indonesia

Singapore launched its Skills and Workforce Development Agency on 16 September, giving citizens six months of free premium AI tools, while South Korea ring-fenced its AI tax windfall in a new Future Response Fund. Taiwan kept funding its AI build past NT$190bn and Indonesia rewired vocational training around AI literacy, betting state-built skills beat a market-led adjustment.

ver.di, CGT Fonction Publique and CCOO

Germany's ver.di banked a 3.3% pay rise on 1 September and opened talks on a Tarifvertrag Transformation covering dismissal bans and reskilling, while France's CGT rejected Paris's AI negotiating timetable the same week. Spain's CCOO went further on 21 September, proposing to tax companies by the jobs they generate rather than wait for the next bargaining round.

BIS General Manager and Federal Reserve governors

The BIS's General Manager said on 10 September that AI displacement remains limited, even as the BIS's own survey found nearly 80% of firms plan to automate roles. Two Federal Reserve governors made the same point in July, arguing the labour-market data does not yet show a mass-firing event.

Bank of Canada, ONS and ECB

The Bank of Canada found the job-finding gap between AI-exposed and unexposed occupations widened from 2.2 to 13.9 percentage points since 2015-19, while separations barely moved. That framing, a hiring freeze rather than a firing wave, is echoed by the ECB's finding that euro-area AI use hit 52% of workers in 2026, concentrated among the university-educated.

Office for National Statistics

Deferred its Transformed Labour Force Survey beyond November 2027 and disclosed a May 2026 telephone-collection failure. The ONS carries no AI-attribution layer at all, so Britain sits outside this month's cohort of measuring states by its own admission.