Weather 
Artificial intelligence (AI) is charging into a new phase in 2026 – one that could reshape business operations, global competition and even which workers thrive, according to Goldman Sachs’ Chief Information Officer Marco Argenti.
In an interview with FOX Business, Argenti laid out his top predictions for the year ahead, saying 2025 marked a major turning point in AI’s evolution.
“We used to look at models as a chat that would provide questions and answers,” Argenti said. “Now a year later, we look at models as essentially entities or agents that can perform tasks on your behalf.”
DISNEY CEO DEFENDS MASSIVE AI DEAL, SAYS CREATORS WON’T BE THREATENED
Here are his top predictions for the year ahead:
AI context will greatly expand
Argenti predicts one of the biggest leaps in 2026 will be AI models capable of handling far more context – the relevant background information a system can remember and reason over.
“I think there’s going to be more and more research and more and more optimization… in how to allow models to be able to reason and ingest much larger context,” he said.
Models will soon be capable of reasoning across libraries of documents, long-running conversations and “everything that you’ve read, everything that you’ve written,” according to Argenti.
AI models will become the new operating systems
AI models will soon function like a computer’s operating system and will be able to browse the internet, access files and execute multistep tasks, Argenti predicts.
“We’re going to start to see a change in this traditional compute model, where the models are the new operating system,” he said. “So they’re going to have more and more capabilities to really be able to give applications access to intelligence and access to tools.”
Users will simply give AI a goal, similarly to entering a destination into a navigation app, and AI will take the steps to accomplish it, according to Argenti.
“It’s essentially like Waze or Maps where, instead of saying, ‘Turn right,’ ‘turn left,’ you’ll say, ‘I want to go to Boston,’” Argenti said. “And then the agent is going to figure out what are the best roads at that moment.”
AUTOMAKER GEARS UP FOR SELF-DRIVING FUTURE WITH NEW CHIP
Adaptability will become a top workplace skill
Argenti predicts that the workers who thrive will be the ones most willing to adapt.
Companies will increasingly prioritize employees who are “able to have the curiosity to rethink their expertise,” according to Argenti.
“If you’re an expert but want to stick to your old habits, you’re going to be less effective than someone that might be slightly less of an expert but is actually willing to question their own day-to-day habits,” he said. “It’s a new world. It’s like going from no computers to computers … you have to actually learn how to do things differently.”
Major industry partnerships will emerge
Argenti forecasts large-scale strategic partnerships developing across the AI sector.
“AI is going to be a game of scale, and there’s going to be a network effect of very large partnerships that are going to be forming,” Argenti said.
These “strategic alliances” will reshape the industry and create a “winner-takes-most” dynamic.
The AI race will intensify between the U.S. and China
Argenti predicts 2026 will intensify the global AI race, which is increasingly centered around competition between the U.S. and China.
“It’s going to be essentially a tale of two nations in the geopolitical context between the U.S. and China,” he said. “I think both have the chance of emerging as really powerful leaders in the AI model frontier with broadly comparable capabilities.”
While the U.S. still leads on key benchmarks, he said that “the gap is narrowing.”
Companies will face “token sticker shock”
Internal AI thinking can generate far more tokens – or units of data AI models process – than users ever realize. As models become more capable and as companies continue to ramp up their AI usage, businesses may begin to experience “token sticker shock,” Argenti says.
“In these models, you can see that they reason a lot… and during those minutes, the amount of tokens is extraordinary,” Argenti said. “As these AI pilots that everybody’s running are moving to full scale production, enterprises will face the reality of a potential token sticker shock.”
This will push companies to focus on high-value use cases and more efficient models, according to Argenti.
“In 2026, many enterprises are going to put token optimization at the center of their AI strategy,” Argenti said.
DISNEY ANNOUNCES MAJOR OPENAI DEAL, INCLUDES $1B EQUITY INVESTMENT, USE OF CHARACTERS ON SORA VIDEO PLATFORM
“Agent as a service” will take off
Argenti also predicts the growth of “agent as a service,” in which companies can essentially rent work performed by AI agents.
He predicts companies may begin using fleets of AI agents – that specialize in coding, finance, customer service, design and more – to augment their human staff.
“You could have all sorts of professions that will create sort of an ‘agent as a service’ or ‘AAS,’ instead of ‘software as a service,’ where the mental model shifts from ‘I’m renting a piece of software’ to more like, ‘I’m renting work in the form of an agent,'” Argenti said.
Energy will “more and more” become AI’s biggest roadblock
Rather than money, the real constraint on AI growth will be power, according to Argenti.
“I think it’s going to be more and more the single determinant of scalability,” he said, pointing to high demand and long timelines for building and upgrading energy infrastructure.
“Consumption of and production of tokens will exacerbate the need for power and energy, and I think that is going to get worse in 2026,” he said.
