On September 10, 2025, Wall Street’s bullish mood reached new heights as several major financial institutions raised their year-end forecasts for the S&P 500, citing the continuing boom in artificial intelligence (AI) investments and surprisingly resilient corporate earnings. But beneath this optimism, a new industry report revealed that the costs of AI are wreaking havoc on corporate margins, with most enterprises struggling to control or even predict their AI infrastructure spending.
Deutsche Bank, Wells Fargo, and Barclays all lifted their year-end targets for the S&P 500. Deutsche Bank now expects the index to hit 7,000 by the end of 2025, up from its prior estimate of 6,550. Wells Fargo increased its forecast to 6,650, a 250-point jump from its previous outlook, while Barclays raised its target by 400 points to 6,450. The S&P 500 itself set a new record high of 6,550 early Wednesday, buoyed by tame inflation data and the fervor surrounding AI-driven growth, according to CNBC and Investing.com.
What’s behind this surge of confidence? The answer, in large part, is the relentless pace of AI innovation. As Oracle’s cloud growth expectations wowed analysts this week, investors took the cue that the AI party is far from over. "The music stops when AI capex stops. Enjoy the party," wrote Wells Fargo analyst Ohsung Kwon in a note to clients, using shorthand for capital expenditures. Kwon acknowledged the frothiness in the market but insisted, "As long as AI capex remains intact, the bull market should continue." He even projected the S&P 500 could rally to 7,200 by the end of 2026, which would represent more than 10% upside over Tuesday’s close.
Yet, there are macroeconomic clouds on the horizon. President Donald Trump’s new tariffs had sparked fears that price increases could force the Federal Reserve to keep interest rates higher for longer, threatening to dampen the stock market’s momentum. However, those worries have been eased, at least for now. According to CME’s FedWatch tool, Fed funds futures are pricing in a 100% chance of a rate cut at the central bank’s meeting next week. Deutsche Bank’s chief U.S. equity strategist, Binky Chadha, expects inflation to pick up only modestly compared to the surges seen in 2021 and 2022, and he views this rise as temporary. "While we expect inflation to pick up some, we see the magnitude as modest compared to 2021-2022 and the rise likely to be viewed as temporary," Chadha wrote in a Wednesday note.
Barclays, for its part, expects the Federal Reserve to cut interest rates three times before the end of 2025 to help counterbalance the macroeconomic pressures, especially rising fears around the labor market. Venu Krishna, Barclays’ head of U.S. equity strategy, noted that corporate earnings remain "solid" and that global GDP growth is showing signs of stabilization. Krishna raised his 2026 S&P 500 target by 300 points to 7,000, driven by newfound optimism in the technology sector and the belief that AI disruption concerns within software companies are overblown. "Macro is under pressure, but we take the ‘glass half full’ view," Krishna wrote to clients.
Barclays’ latest analysis, as reported by Investing.com, also highlighted that strong corporate earnings and AI-centric growth are offsetting trade and labor market headwinds. The bank’s strategists, led by Krishna, now see the S&P 500 reaching 6,450 by the end of 2025, based on earnings per share (EPS) of $268, up from their previous estimate of $262. For 2026, Barclays raised its target to 7,000, forecasting EPS of $295. The bank expects new tax provisions to provide a further boost to GDP through 2026, and while they see some effects of tariffs spilling into next year, they described the overall impact as "less severe than initially anticipated."
Barclays also upgraded its stance on the broader technology sector, saying fears of disruption in software are "overdone" and that ongoing data center demand is keeping fundamentals strong. Financials remain a positive call, while Healthcare and Materials have been shifted to neutral. Consumer, Industrials, and Energy sectors remain negative in their outlook. Despite macro risks, including a weakening labor market and unemployment at a three-year high, Barclays maintains a cautiously optimistic stance. Their bull case projects EPS of $273 for 2025 and $309 for 2026, while the bear case foresees $260 and $278, respectively.
But while Wall Street cheers the AI boom, a sobering new report from Benchmarkit and Mavvrik, published on September 10, 2025, paints a more complex picture for enterprises actually implementing AI. The 2025 State of AI Cost Management found that 80% of enterprises miss their AI infrastructure cost forecasts by more than 25%, and a staggering 84% report gross margin erosion of 6% or more due to AI workloads. For a quarter of companies, the margin impact is even more severe—16% or higher.
"These numbers should rattle every finance leader. AI is no longer just experimental – it's hitting gross margins, and most companies can't even predict the impact," said Ray Rike, CEO of Benchmarkit. The report, based on survey data from 372 enterprise organizations, reveals that most companies lack the visibility, attribution, and forecasting precision needed to understand where AI costs originate or how they affect profitability. "Without financial governance, you're not scaling AI. You're gambling with profitability," Rike warned.
The survey found that 67% of companies are actively planning to repatriate AI workloads from the cloud, with another 19% evaluating that option, and 61% already running hybrid AI infrastructure that combines public and private environments. Yet, only 35% include on-premises AI costs in their reporting, leading to major blind spots. Data platforms were cited as the top source of unexpected AI spending (56%), with network access costs coming in second (52%), and large language models ranking fifth.
When it comes to forecasting, the situation is dire: only 15% of companies can forecast AI costs within a 10% margin of error, while 24% are off by 50% or more. Despite 94% of companies tracking AI costs, only 34% have mature cost management practices. The report also found that companies charging for AI usage show twice the cost maturity in attribution and discipline compared to those that do not.
"AI is blowing up the assumptions baked into budgets. What used to be predictable, is now elastic and expensive," said Sundeep Goel, CEO of Mavvrik. "This shift doesn't just affect IT, it's reshaping cost models, margin structures, and how companies scale. Enterprises are racing to build with AI, but when most can't explain the bill, it's no longer innovation, it's risk."
So while Wall Street’s outlook for the S&P 500 is buoyed by AI-driven growth and robust corporate earnings, the reality for many enterprises is far more challenging. The enthusiasm on trading floors is built on a foundation that, for now, is being shaken by the unpredictable costs and margin pressures of AI adoption. Until companies can gain better visibility and control over their AI infrastructure spending, the risk of profit erosion will remain a thorny issue—even as the stock market continues to scale new heights.