Wall Street · Urbans Post
While Wall Street torches a trillion dollars arguing AI spending won't pay off, a 170-year-old bank just showed its homework. Santander put a euro figure on what AI actually did for its bottom line this year, and the number is going up every quarter. Banco Santander reported on October 28 that artificial intelligence generated €194 million in business value during the first nine months of 2026.
That figure isn't a projection or a capex slide. It's the sum of extra revenue, lower costs, and reduced credit-loss provisions the bank attributes directly to AI tools already running inside its operations. In the first quarter, that number was €35 million.
By the end of September, it had climbed to €194 million, and Santander says it expects to keep growing through the rest of the year. The bank is now betting the acceleration continues. Santander has set a target of more than €1 billion in AI-driven business value between 2026 and 2028, through a combination of additional revenue and cost reductions.
To get there, it has extended AI access to all 185,000 employees worldwide, up from a rollout that started with smaller teams earlier in the year. This isn't abstract tooling sitting unused on a dashboard. In Brazil, Santander says AI has cut the time needed to process card fraud claims by 95%, with automation handling up to 90% of those claims and an error rate under 1%.
The bank also has more than 280 automated process agents running in production across areas like credit decisions, fraud detection and know-your-customer checks, and it's mixing in Microsoft Copilot alongside tools from OpenAI, Anthropic and Google depending on the task. Santander isn't alone in making this claim, and that's what makes the timing notable. Novartis CEO Vas Narasimhan told CNBC in comments published November 1 that AI is now visible in the pharmaceutical giant's actual financial results, not just its roadmap.
"I think this is the first quarter I've really seen AI start to improve actually our P&L on a tangible basis," he said. Novartis reported third-quarter core operating income of $6.04 billion, up 10% year over year at constant currencies, according to CNBC's reporting. Neither company is a tech firm.
That's the point. A bank and a drugmaker, two of the most heavily regulated, slow-moving sectors in the economy, are both pointing to AI showing up in the P&L rather than just the pitch deck. Frankly, the market hasn't been in the mood to hear it.
On October 30, the Nasdaq Composite fell 2.9% as investors questioned whether heavy AI infrastructure spending would generate returns quickly enough. Amazon and Intel were on course to erase nearly $200 billion in combined market value that day, while the Nasdaq-100 was set to lose just over $1 trillion. The Nasdaq Composite ended the week down 2%, and October was its worst month since March 2025.
That selloff reflected worries about hyperscaler capex, the money being spent to build data centers and buy chips, as well as company-specific earnings and outlook disappointments. Santander and Novartis are arguing from the other side of the ledger. They're not the ones building the data centers.
They're customers of the AI boom, and they're the ones who can actually show what they bought it for. None of this settles the capex debate. The hundreds of billions Big Tech is spending on infrastructure and the €194 million Santander says AI generated this year operate on completely different scales, and a bank saving money on fraud claims doesn't answer whether Microsoft and Meta will ever earn back what they're spending on GPUs.
But it does mean the "AI hasn't paid for itself yet" argument isn't universally true anymore. It just depends which part of the stack you're standing on, the companies building the infrastructure or the ones actually using it. Santander's next test comes with its full-year 2026 results, when the market will see whether AI-driven business value exceeds the bank's earlier full-year goal of €200 million and remains on track toward its cumulative target of more than €1 billion between 2026 and 2028, or whether the easy wins, fraud claims, call deflection, routine automation, start to run out before the harder revenue gains show up.
Also read: Aramco closes $2.6 billion deal for 49% of Saudi Arabia's AI champion Humain • CEOs mandate AI at work while barely using it themselves, survey finds • Thrive Holdings Agrees to Buy ADT for 15.3 Billion With OpenAI Backing This article is posted in AI News , check it out for more related stories.
Source: Startup Fortune
