CNBC reported on 1 November 2026 that companies including Santander, Novartis and Standard Chartered are beginning to connect artificial intelligence work with revenue, costs or profitability in third-quarter earnings. The shift matters because corporate AI spending has moved from pilot budgets into capital plans, infrastructure contracts and operating targets.
The Stanford HAI 2026 AI Index put corporate AI investment at USD 581.7 billion in 2025, up 130% from the previous year. Microsoft, Amazon, Google and Meta are expected to spend USD 700 billion combined this year on AI infrastructure. Investors and boards are therefore asking for evidence that AI programmes are changing income statements, not only technology roadmaps.
Santander quantifies business value
Santander gave one of the clearer figures. The bank said AI generated EUR 194 million of business value in the first nine months of 2026, through higher revenue, lower costs and lower loan-loss provisions. Its wider results showed underlying profit of EUR 11.2 billion for the first nine months, up 16%, and revenue of EUR 46.8 billion, up 8%.
Chief Financial Officer José García Cantera said AI was reducing low-value tasks so employees could focus on higher-value work. He said the effect covered both revenue and costs, and that the bank could directly link AI to a positive contribution in the quarter. He also said AI was being embedded across bank processes, which made it difficult to isolate a single total contribution.
Santander’s own results linked the AI figure to customer service, productivity and commercial growth. The bank also said its ONE Transformation programme was contributing to lower costs excluding mergers and acquisitions, and to an efficiency ratio of 42.6%.
Other sectors report less precise effects
Novartis did not give a specific AI contribution, but Chief Executive Officer Vas Narasimhan said AI was beginning to improve the company’s profit and loss statement on a tangible basis. Novartis reported third-quarter core operating income of USD 6.04 billion, up 10% year on year at constant currencies.
Narasimhan said the company was using AI in research and development to test whether it could improve productivity. He also pointed to productivity in marketing and sales. Without a quantified contribution, the Novartis example is evidence of management attribution rather than a financial benchmark.
Standard Chartered Chief Financial Officer Manus Costello said AI was appearing across the business. The bank reported third-quarter operating income growth of 14% at constant currency, excluding notable items. Costello linked part of the bank’s operating leverage to AI tools used to improve core processes, customer relationships and income generation.
Carlsberg gave a more cautious signal. Chief Executive Officer Jacob Aarup-Andersen said the company was not seeing AI benefits at scale. He said the technology had affected internal processes, cost efficiency, commercial and pricing work, marketing and the supply chain. Carlsberg’s third-quarter revenue grew 4.9%.
Large technology companies face a different test
For large technology companies, the question is less whether AI is being used internally and more whether revenue growth can cover the cost of infrastructure. Meta said in its 21 October earnings that AI investments were delivering “substantial incremental advertising revenue”. Its total revenue rose 26% in the quarter ended 30 September.
Alphabet showed both sides of the issue. Google Cloud revenue more than doubled to USD 30.7 billion in the third quarter, above the USD 28.7 billion expected by analysts. Google Cloud Chief Executive Officer Thomas Kurian said demand was coming from AI labs, capital markets, financial services institutions and enterprises.
The wider Alphabet result was weaker. Revenue was USD 127.58 billion against analyst expectations of USD 128.22 billion, and adjusted earnings per share were USD 2.67 against USD 3.02 expected. The company reported negative free cash flow of USD 9.07 billion and capital expenditure of USD 51.36 billion in the quarter.
Alphabet kept its full-year capital expenditure forecast at USD 195 billion to USD 205 billion, after raising guidance in the second quarter. Finance chief Anat Ashkenazi told analysts that the company expected capital expenditure to increase significantly in 2027. That leaves cloud growth and AI adoption to be judged against rising depreciation, debt and operating costs.
The evidence remains uneven
The earnings season gives finance and technology leaders a small set of public reference points. Santander has provided a concrete value figure. Novartis, Standard Chartered and Carlsberg describe operational effects with different levels of precision. Meta and Alphabet show that revenue gains can coexist with large infrastructure pressure.
The broader market context is also concentrated. LSEG estimates cited for the third quarter put expected S&P 500 earnings growth at about 31% year on year, with technology and AI-related companies accounting for around two-thirds of the increase. Semiconductor companies were expected to report earnings growth of about 136% year on year, after growth of around 158% in the second quarter.
For enterprise teams, the practical issue is attribution. A bank can count reduced provisions, sales conversion or process productivity if the measurement model is agreed in advance. A pharmaceutical company may need longer cycles before research productivity appears in financial results. A cloud provider may show AI demand in revenue while free cash flow falls because infrastructure spending rises first.
Deciding whether a EUR 194 million AI value claim is comparable with unquantified gains in pharma, banking or advertising is the type of investment and governance problem ECS's CxO + Strategy topic brings to keynotes, peer panels and closed-door roundtables for decision-makers. Attendees can compare how organisations set targets, account for infrastructure and operating costs, and decide which AI work deserves further funding. Bring the measurement models and board-level questions that are difficult to settle from earnings headlines alone.
This article was written by the ECS editorial team with the assistance of Amal Agent, run.events' AI assistant, which supported the research, drafting and proofreading.