The top 5 things that happened in the AI race this year

• The future of the AI race was laid down in 2025.
• Big Tech may hold up the entire economy with its capex spending spree.
• Meanwhile, concerns about a bubble are unlikely to subside.

2025 was unquestionably the year of AI.

Big Tech shelled out roughly $400 billion on capex, a spending spree so extensive that some economists believe it staved off an overall recession. Nvidia became the first $4 trillion company. And AI content became inescapable, seeping into everything from Hollywood to campaign ads — even Mickey Mouse is getting into AI.

It hasn't been an endless party. Seemingly every few weeks, the stock market gets spooked that music is about to stop. Only Universal's Wicked for Good was more focused on a bubble.

Here's a look back at the biggest AI storylines of 2025.

Wall Street's thought bubble

In an era of nostalgia traps, traders can't decide whether this is the Dot-Com era all over again.

Tech and AI CEOs can't agree either. In August, OpenAI CEO Sam Altman touched off concerns that a bubble had already formed. Since then, Bill Gates, Nvidia's Jensen Huang, Mark Cuban, and Mark Zuckerberg have offered their own similar or dissenting views.

The optimists' train of thought often ends up at the railroads and other breakthrough innovations that transformed the economy.

"There's been a lot of talk about an AI bubble," Huang said during Nvidia's third-quarter earnings call. "From our vantage point, we see something very different."

Even those at the forefront of AI's advancements express concern that some of their competitors are being too bold.

"There's genuine uncertainty, there's genuine dilemma, which we as a company try to manage as responsibly as we can," Anthropic CEO Dario Amodei said in early December at a New York Times event. "And then I think there are some players who are yoloing, who pull the wrist dial too far, and I'm very concerned."

Capex spending craze

The sheer size of spending is breathtaking.

JPMorgan Chase concluded that AI-related spending contributed to 1.1% of GDP growth in the first half of the year.

The spending isn't likely to slow down.

Over the last two years, Wall Street has underestimated capex growth, according to Goldman Sachs Research. Right now, Goldman said, the consensus estimate is that hyperscalers will spend $527 billion on capital expenditures next year.

Zuckerberg and OpenAI's leadership have separately suggested that the biggest risk is not spending enough.

"We want to be ahead of the curve," OpenAI President Greg Brockman said in a recent video posted on X. "And the truth is, I don't think we will be, no matter how ambitious we can dream of being right now. I think demand will far exceed what we can think of."

AI talent wars

Silicon Valley's turf wars were a lot greener in 2025. Over the summer, the AI talent wars reached another level.

Perhaps no company was as aggressive as Meta. Zuckerberg moved to poach top talent by wooing workers with tens of millions of dollars.

Altman said OpenAI offered his best employees $100 million signing bonuses. Another top OpenAI official said Zuckerberg even made homesoup for one of his targets.

OpenAI boasted that it largely fended off Meta's efforts, though ChatGPT co-creator Shengjia Zhao later joined Meta's Superintelligence Lab.

Time is a flat circle, and so are AI deals

Even the hyperscalers need some extra help to maintain their spending habits.

That's why Alphabet, Amazon, Meta, Microsoft, and Oracle issued roughly $100 billion of bonds, powering global bond sales to another record year.

The interconnected nature of many AI deals has also raised concerns among some analysts and traders. Take Anthropic's pledge to spend $30 billion on Microsoft Azure to scale up Claude's compute. As part of the deal, Microsoft will invest up to $5 billion in Anthropic. Nvidia, whose chips are at the certain of this deal and many like, also agreed to invest up to $10 billion.

Not all spenders are created alike, either. OpenAI, which is expected to be in the red by $9 billion this year, has about $1.4 trillion in spending commitments for AI data centers over the next decade. Unlike Google, Meta, and Microsoft, OpenAI doesn't have an established revenue base to fall back on, either.

It's why OpenAI CFO Sarah Friar sparked a brief firestorm when she appeared to suggest that the startup would want the possibility of a government bailout to backstop its data center bets. Friar walked back her remarks, and Altman later emphasized that OpenAI didn't believe in bailouts.

Google gains an upper hand

OpenAI has led much of the AI-model race since the release of ChatGPT in 2022.

Three years later, it was Altman who declared a "code red," just over a month after OpenAI completed its corporate restructuring, aimed at giving it more freedom to raise money to fund its AI advancements.

Google is now fighting back, and in the view of some observers, has caught up to OpenAI with the widely praised release of Gemini 3.

CEO Sundar Pichai might as well have struck Steph Curry's signature celebration when he was asked what was next for his AI team.

"I think some folks need some sleep," Pichai said of the company's staffers following Gemini 3's release.