The two LLM titans are duking out on the pricey field of Super Bowl ads, with Anthropic’s first-ever TV ad during the big game taking aim at its arch rival. In the Anthropic ad, a scrawny guy struggling to do a pull up asks a muscular fellow how to “get a six-pack quickly.” Muscle man responds with a robotic, chatbot-like answer and then abruptly touts a special offer for workout-improving insoles. Although it doesn’t explicitly name OpenAI, it’s a clear dig at the ChatGPT maker which recently announced plans to put ads into products. (“Ads are coming to AI. But not to Claude” reads the text at the end of the Anthropic spot, which you can watch here).

OpenAI CEO Sam Altman fired back on X on Wednesday, saying that, while his rival’s TV ad was admittedly funny, it was also “dishonest.” OpenAI has previously said that its ads will be clearly labelled and served underneath ChatGPT answers, not incorporated into ChatGPT answers—”we are not stupid, and we know our users would reject that,” Altman said. Altman then showed a bit of his own spinning skills, describing Anthropic as an elitist company offering “an expensive product to rich people.” OpenAI’s embrace of ads is a matter of principle: “We need to bring AI to billions of people who can’t pay for subscriptions.”

Here’s hoping Sunday’s matchup has as much action as the pre-game show.

Today’s news below.

Alexei Oreskovic
@lexnfx
alexei.oreskovic@fortune.com

Want to send thoughts or suggestions to Fortune Tech? Drop a line here.

Google to double capex this year up to $185 billion

Alphabet plans to spend between $175 billion to $185 billion in capex in 2026, up to twice as much as the record $91.4 billion it spent in 2025, raising the stakes ever higher in the tech industry's intense battle to build AI infrastructure.

The Google parent company announced its capital expenditures plans on Wednesday along with its better-than-expected fourth quarter financial results. “We are in a very, very relentless innovation cadence, and I think we are confident about keeping that momentum as we go through 2026,” CEO Sundar Pichai said. 

Alphabet’s massive increase in AI infrastructure spending—which will go towards building data centers and loading them with computers—sets a new high water mark just one week after Meta stunned the Street by announcing plans to nearly double its capex to between $115 billion and $135 billion this year.

Investors seemed unsure how to react to Alphabet’s plans. The stock initially nosedived more than 6% in after hours trading Wednesday, then rose more than 2% as Pichai and his team spoke during the earnings call, only to dip slightly back into the red, down 0.4%.—Amanda Gerut

Meta quietly expands its giant Louisiana AI data center

One of the most gigantic infrastructure projects of the AI boom appears to be growing even larger. Meta has quietly purchased roughly 1,400 acres—an area almost twice the size of Manhattan’s Central Park—adjacent to its already-mammoth 2,250-acre Hyperion AI data center site in Richland Parish, Louisiana, Fortune has learned.

According to half a dozen people affiliated with companies working on or around the Meta site, the land purchase paves the way for a Phase 2 expansion of the Hyperion project, one of the nation’s largest AI data centers in the works. In October, Meta announced it had entered a joint venture with funds managed by Blue Owl Capital to finance, build, and operate the Hyperion data center campus—an arrangement targeting up to $27 billion in total development costs.

The apparent expansion of Meta’s already enormous project offers a window into how the AI infrastructure boom is unfolding. Hyperscalers—the Big Tech companies building out their AI infrastructure—are racing to lock up land, power, and financing for massive AI data-center campuses, often through debt-financed, politically sensitive expansions. Some, like the Louisiana project, are expanding so quickly that it may be difficult for local communities to spot or register concerns in real time. —Sharon Goldman

Pinterest cracks down on internal dissent

Pinterest fired two employees who created a tool for tracking the company’s layoffs in a move that highlights how power has largely shifted from employees back to employers in corporate America.

The internet company said last week it would lay off less than 15% of its workforce and shed office space as part of a restructuring running through the end of September that will reallocate resources to AI-focused roles and AI products. 

“After being clearly informed that Pinterest would not broadly share information identifying impacted employees, two engineers wrote custom scripts improperly accessing confidential company information to identify the locations and names of all dismissed employees and then shared it more broadly,” a Pinterest spokesperson said, calling the act a violation of employees' privacy.

But the incident also highlights a new era in which companies hold all the cards as employees have transitioned from job hopping to job hugging. During the town hall the following day, CEO Bill Ready reportedly rebuked staffers and said employees should consider finding another job if they’re “working against the direction of the company” and don’t agree with the company mission, CNBC reported. —Marco Quiroz-Gutierrez

More tech

—Andy Jassy's new shadow. Amazon CEO's technical advisor is a key role.

—Nvidia CEO says ignore AI ROI. "If you want your company to succeed, you can’t control it.”

—How the Magnificent 7 destroyed index funds: There’s nowhere to hide.

—Boring Co altered records: Nevada legislator says lawmakers will push for independent audit.

—Memory shortage bites Qualcomm. Shares slide after earnings report.

—TRM Labs notches $1 billion valuation. Crypto crime-fighting is big business.

—IonQ, the biggest quantum computing company on the stock market, disputes short-seller report.