A Goldman Sachs partner leading one of the bank’s major artificial intelligence initiatives has warned that the rapid adoption of AI across Wall Street could weaken the analytical and reasoning skills of the next generation of financiers if employees become too dependent on machines.

Chris Churchman, who leads Marquee, Goldman Sachs’ digital platform for institutional clients, said AI could create a form of “cognitive atrophy” by taking over the reasoning and problem-solving tasks through which young bankers and traders traditionally develop their expertise.

“There’s a huge danger here that in the era of AI, we outsource our reasoning to these models, and we have cognitive atrophy that stops us being able to reason from first principles ourselves,” Churchman said.

His comments came during an episode of Goldman Sachs’ “Exchanges” podcast, according to a transcript provided to CNBC.

Churchman’s concern goes beyond the possibility of AI making mistakes. He noted that Wall Street could inadvertently weaken its own talent pipeline by automating the routine work that has historically served as an apprenticeship for junior employees.

“Reasoning is still important,” he said. “You still need to reason about [problems] and structure it into an argument, and now we’re delegating reasoning.”

The issue is relevant to investment banks because much of the expertise required to make complex decisions is developed through repeated exposure to real-world situations rather than formal training.

Churchman compared the potential effect of AI with previous technologies that reduced the need for people to exercise certain skills, such as navigation and memorization. If algorithms increasingly perform analysis, generate recommendations, and execute routine decisions, junior employees could have fewer opportunities to develop the judgment needed to operate independently.

That could create a long-term problem for banks. AI may increase productivity and reduce costs in the short term, but the same technology could weaken the human expertise needed to supervise those systems and make decisions when circumstances fall outside established patterns.

Wall Street firms have already been examining whether AI can reduce the number of junior bankers needed relative to senior employees. That trend could accelerate if banks discover that AI can perform significant portions of entry-level research, modelling and administrative work.

Churchman said banks therefore need to preserve an apprenticeship model in which younger employees learn alongside experienced professionals.

“You learn by doing, and a lot of knowledge is tacit, it was never written down,” he said.

Goldman needs “to make sure we don’t lose that tacit and intuitive knowledge that some of our best people have today [and] to ensure the next generation have it too,” he added.

Currency trading provides one example. Churchman, who previously ran currency trading at UBS before joining Goldman in 2021, said junior traders traditionally develop their skills by handling client pricing requests while experienced traders supervise their decisions.

“We can absolutely automate that,” he said, “but then do we get the senior traders that fully understand?”

The question marks a potential feedback loop in financial-sector automation. If AI replaces enough entry-level work, future senior employees may never acquire the practical experience that enabled previous generations to become effective risk managers. The result could be an industry with sophisticated technology but a thinner pool of people capable of understanding when that technology should not be trusted.

Churchman said AI systems should therefore be designed so that humans remain responsible for high-stakes and highly uncertain decisions rather than becoming passive operators who simply approve machine-generated recommendations.

Even Goldman Sachs has not fully resolved how to manage that transition.

The bank has invested heavily in integrating AI into its operations, and Churchman is also co-chair of its Global Banking and Markets AI working group. But he acknowledged that the firm has not yet “figured out” the appropriate balance between automation and human judgment.

The technical challenge is equally significant.

Churchman said one of the hardest problems in developing Goldman’s AI capabilities is ensuring that responses are factually accurate and can be audited. That requirement is stringent in financial markets, where a seemingly minor error in a valuation, risk calculation or market-data interpretation can have substantial financial consequences.

Marquee, Goldman’s digital platform for institutional clients, provides hedge funds and other professional investors with access to the bank’s market data, research, risk analytics and trade-execution services.

Churchman said the AI component of Marquee is currently available only to Goldman employees as the bank continues to develop the technology.

Unlike consumer chatbots, where users may tolerate occasional incorrect answers, financial AI systems have to operate under much stricter standards. Institutions need to know where an answer came from, whether the underlying information is reliable and whether the result can be independently verified.

Churchman said the AI system itself effectively highlighted this challenge during testing.

“When we challenged it hard, at least it was honest,” he said. “It was like, ‘Look, in the end, I’m better at sounding thorough than being thorough.’”

That distinction could become one of the defining issues for AI adoption in finance. Generative AI systems are often capable of producing highly convincing explanations even when their underlying conclusions are wrong. In financial markets, confidence and fluency cannot substitute for accuracy.

The challenge for Goldman and its peers is therefore not simply determining how much work AI can automate. Banks must also decide which tasks should remain part of the human learning process and how much responsibility employees should retain when machines are involved.

The economic incentive to automate is strong. AI can process information rapidly, reduce repetitive work, and potentially allow senior employees to handle larger volumes of business. But Churchman warns that if junior employees lose the opportunity to develop judgment through that same work, banks could eventually face a shortage of experienced professionals capable of challenging AI systems when they fail.

China has opened applications for an 800 billion yuan ($119 billion) policy-based financing programme for local government projects, stepping up efforts to revive investment and support economic growth as a sharp contraction in fixed-asset spending raises pressure on Beijing to deliver additional stimulus.

The financing tool, announced in March, is designed to provide capital for infrastructure and strategic projects and use that funding to attract larger amounts of bank and private-sector financing.

Implementation guidelines have now been circulated to local governments, which are compiling eligible projects and submitting them to Beijing for approval, the state-backed Economic Information Daily reported.

The programme comes after China’s fixed-asset investment fell 6.7% in the first seven months of 2026, highlighting the weakness in one of the country’s traditional engines of economic growth.

The decline has been linked to tighter scrutiny of local government investment. Beijing has been trying to prevent officials from financing projects that generate inadequate economic returns, while also addressing industrial overcapacity and price competition that has contributed to deflationary pressure.

That effort has created a difficult policy trade-off. Authorities want to curb wasteful investment and local government debt accumulation, but the resulting restraint has also weakened construction activity and demand for capital goods at a time when China’s broader economy is losing momentum.

The new financing programme is intended to direct capital toward projects that have already reached a sufficient stage of preparation, rather than encouraging local governments to launch investment simply to meet spending targets.

Caitong Securities said the process from project applications to fund disbursement is likely to take at least one month, limiting the programme’s ability to generate a substantial increase in construction activity before the end of the year.

The move comes as China’s economic growth has already slowed sharply. Gross domestic product expanded 4.3% in the second quarter, the weakest quarterly growth rate in more than three years, compared with 5% in the first quarter and below market expectations.

The slowdown has increased pressure on policymakers to support domestic demand while maintaining controls on financial risks and excess industrial capacity.

The 800 billion yuan instrument is a quasi-fiscal programme that Beijing hopes will generate a much larger investment response than the initial government funding. China increased the size of the programme from 500 billion yuan in 2025, signaling a stronger policy commitment to supporting investment.

Caitong Securities estimates that the programme could ultimately support around 10 trillion yuan in total project investment if a leverage ratio of roughly 13 times is achieved.

But the brokerage expects only about 2 trillion yuan of that potential investment to have a direct impact this year. The difference reflects the time required to approve projects, disburse funds and mobilize additional financing, as well as a shortage of projects that meet Beijing’s requirements.

The programme was not used during the first half of 2026, economists said, partly because local governments faced tighter borrowing restrictions and struggled to identify enough eligible projects. The economy also began the year relatively strongly, reducing the immediate need for additional stimulus.

Those conditions have changed as investment has weakened and economic growth has slowed.

Caitong expects policy banks to accelerate bond issuance in August and September to provide financing for approved projects. Local governments are also expected to increase issuance of special-purpose bonds linked to eligible infrastructure projects.

The combined measures could increase the flow of capital into construction and strategic industries later this year, although the effect is likely to be gradual rather than immediate.

Goldman Sachs analysts estimate that if the programme is implemented in the third quarter, it could add about 0.5 percentage point to China’s GDP growth, with most of the impact likely to be concentrated in late 2026 and early 2027.

The estimates highlight the difference between the headline size of the financing programme and its near-term economic effect. An 800 billion yuan commitment may eventually leverage several times that amount in total investment, but much of the spending will occur over a longer period.

The programme also underpins the targeted nature of China’s economic stimulus. Rather than relying on a broad infrastructure spending surge, Beijing is attempting to channel financing toward projects that have already received preliminary approval and are considered viable enough to generate economic returns. That approach is partly a response to problems created by earlier investment-led stimulus. Years of rapid infrastructure expansion helped support growth but also contributed to rising local government debt, excess capacity and projects with weak returns.

Authorities are therefore trying to provide enough financing to prevent investment from collapsing without reopening the cycle of indiscriminate borrowing and construction. The challenge is that the private sector has remained cautious, while local governments have faced restrictions on debt-funded investment. That means public financing may have to carry more of the burden if Beijing wants to stabilize fixed-asset investment.

The new programme could also help unlock projects that have already been planned but stalled because of financing constraints. By providing initial capital, the policy tool is intended to reduce the amount of funding that banks and private investors need to provide upfront.

However, its effectiveness will depend on the number and quality of projects available for financing. If local governments continue to struggle to identify projects that satisfy central government requirements, a larger financing envelope may not translate into proportionately higher investment.

The weakness in fixed-asset investment also reflects a broader change in China’s growth model. Property investment remains under pressure, while policymakers are seeking to shift resources toward advanced manufacturing, strategic technology and infrastructure that can support longer-term productivity.

That transition has produced tensions of its own. China’s manufacturing sector has expanded capacity rapidly in several industries, contributing to aggressive price competition and concerns over deflation. Beijing has therefore been trying to encourage productive investment while discouraging projects that simply add capacity to industries already facing oversupply.

The 800 billion yuan programme is consequently less a return to the broad stimulus of previous cycles than an attempt to provide targeted liquidity to projects that policymakers consider economically useful.

Analysts expect financial markets’ immediate focus to be on the pace of project approvals, policy-bank bond issuance and local government special-purpose bond sales. It is hoped that faster implementation could provide a stronger floor under construction activity and related industrial demand in the final months of the year.

For the wider economy, however, the programme is unlikely to eliminate the need for broader measures if household consumption and private investment remain weak.

The Goldman estimate of a 0.5 percentage-point GDP boost suggests the financing programme could make a meaningful contribution to growth, particularly in late 2026 and early 2027. But the delayed rollout means the programme’s headline 800 billion yuan size will overstate its immediate impact on this year’s economic activity.

China is therefore entering the second half of 2026 with a familiar policy dilemma: to stimulate investment enough to stabilize growth, while ensuring that new financing does not recreate the debt, overcapacity and low-return investment problems that Beijing has spent years trying to contain.

Alibaba and Samsung Electronics shares fell sharply on Monday as investors reacted to two very different consequences of the artificial intelligence boom: Alibaba’s need to raise billions of dollars to finance surging AI investment and Samsung’s decision to return a record amount of cash to shareholders that still fell short of expectations.

Alibaba shares dropped about 8% in early Hong Kong trading after the Chinese e-commerce and cloud computing giant finalized an HK$80 billion ($10.21 billion) share placement to fund its expansion in artificial intelligence infrastructure.

Alibaba priced 710 million new shares at HK$112.70 each, an 8.4% discount to the stock’s previous close. The transaction is the largest primary follow-on offering by a Hong Kong-listed company and the world’s third-largest this year, behind share sales by Alphabet and Intel.

The sharp decline in Alibaba’s shares illustrates the immediate cost of financing the company’s AI ambitions. The placement provides Alibaba with a substantial pool of capital, but the issuance also dilutes existing shareholders and signals that the company’s AI expansion will require significantly more funding before it generates returns.

Alibaba has said all of the proceeds will be used for AI-related development, including infrastructure.

Last week, Alibaba reported that it had already spent almost half of its three-year capital expenditure programme, while its quarterly net profit plunged 75% from a year earlier, largely because of higher AI-related spending.

The company has nevertheless become more optimistic about the economics of those investments. Alibaba said it expects the payback period for its AI investments to fall to about 2.5 years from three years as demand for AI services increases. That creates a crucial test for the company: whether the revenue generated by AI services can grow quickly enough to justify the enormous upfront spending required to build computing capacity.

Alibaba is betting heavily that it can.

The company has pledged to invest 380 billion yuan ($56.54 billion) over three years in AI and cloud infrastructure. Its expansion continued last week when Alibaba Cloud opened its third data center in South Korea, bringing its network to 104 availability zones across 30 regions.

The investment strategy places Alibaba in direct competition with global technology companies that are also committing hundreds of billions of dollars to AI infrastructure. But investors appear increasingly focused on the difference between spending on AI and generating returns from it.

Alibaba’s share-price reaction suggests that shareholders are not willing to treat higher AI expenditure as an automatic positive. The company needs to demonstrate that additional data centres, computing capacity and AI models will translate into sustainable revenue and cash flow.

Samsung Too

The same tension is visible in South Korea, although from a different angle.

Samsung Electronics shares fell more than 8% in early trading after the world’s largest memory-chip maker announced a record shareholder-return programme that investors viewed as insufficient relative to the profits being generated by the AI-driven semiconductor boom.

Samsung said it expects to return between 90 trillion won and 110 trillion won ($65 billion to $80 billion) to shareholders this year, including 30 trillion won in cash dividends in the third quarter.

The proposed distribution is five times Samsung’s previous record, set in 2020.

Yet the size of the programme failed to satisfy investors who had expected a greater proportion of the company’s AI-related windfall to be returned through share buybacks and cancellations.

This matters because dividends distribute cash to shareholders but do not directly reduce the number of shares outstanding. Buybacks and share cancellations can provide more direct support to earnings per share and the stock price by reducing the share count.

Samsung has maintained its commitment under its 2024-2026 shareholder-return policy to allocate 50% of free cash flow generated during the three-year period to shareholders.

But investors had hoped the company would go further as booming demand for high-bandwidth memory and other advanced chips drives profits across the semiconductor industry.

Rival SK Hynix, which has benefited strongly from demand for AI memory, took a more aggressive approach. SK Hynix said it plans to buy back and cancel 40 trillion won of treasury shares and allocate more than half of its free cash flow generated between 2025 and 2027 to shareholder returns.

The contrast helps explain the different market reactions.

SK Hynix shares fell about 2.5% on Monday, but Samsung declined more than 8%, while the benchmark KOSPI fell 3.1%.

“Unlike SK Hynix, Samsung Electronics did not mention the possibility of raising its existing shareholder return policy, nor did it announce a plan to cancel treasury shares that could more directly contribute to the stock price increase, which is disappointing,” Sohn In-joon, an analyst at Eugene Securities, said in a report.

Samsung’s ownership structure also complicates its ability to rely heavily on buybacks. Large buybacks could push the combined ownership of Samsung Life and Samsung Fire above regulatory limits, potentially requiring the affiliates to sell shares to bring their combined stake below 10%.

As a result, much of Samsung’s remaining shareholder-return allocation is expected to come through dividends rather than buybacks.

Kim Soo-hyun, head of research at DS Investment & Securities, estimated that of the remaining 60 trillion won to 80 trillion won, only about 10 trillion won to 20 trillion won could be directed toward share buybacks and cancellations.

Samsung Life and Samsung Fire also fell sharply, declining 9.9% and 8%, respectively, as investors adjusted their expectations around the implications of Samsung’s capital-return strategy.

Samsung said its board will decide on the remaining payouts in January 2027, with cash dividends, share buybacks and share cancellations all under consideration.

“Big capital returns, slightly below expectations,” Morgan Stanley said in a report, adding that investors will need to focus on Samsung’s next shareholder-return framework, which will take effect next year.

The contrasting reactions to Alibaba and Samsung reveal an important feature of the AI investment cycle. Companies such as Alibaba face the challenge of convincing shareholders to tolerate enormous capital expenditure today in exchange for potentially higher AI revenue in the future. For chipmakers such as Samsung and SK Hynix, the challenge is almost the reverse. AI demand is already producing substantial cash flows, and investors want a larger share of those gains returned to them rather than reinvested or retained.

In other words, the AI boom is creating two competing demands on corporate balance sheets: technology companies need unprecedented amounts of capital to build the infrastructure required for AI, while shareholders are demanding evidence that those investments will generate adequate returns.

Alibaba’s $10.2 billion placement is seen as a bet that the returns will come later, while Samsung’s record payout is an attempt to demonstrate that the returns are already arriving. The market reaction on Monday shows that investors are becoming less willing to accept either argument without evidence.