SoftBank Group CEO Masayoshi Son in Tokyo, Japan, on February 03, 2025. Tomohiro Ohsumi/Getty Images

In the words of Jeff Goldblum’s character in the original Jurassic Park: Well, there it is.

SoftBank and Intel on Monday announced that the Masayoshi Son-led holding company would invest $2 billion in Intel stock.

The equity investment hurtles SoftBank onto the list of Intel’s top 10 shareholders—No. 6, to be precise—though it will reportedly not include a board seat or requirements to buy Intel chips.

The deal isn’t enough to rescue Intel, but it’s a welcome move. 

Intel CEO Lip-Bu Tan has been working furiously to shed costs (in lieu of a total breakup of the iconic American chipmaker) with the hope that it will buy him enough time to rally in the hypercompetitive but hugely expensive and time-consuming semiconductor business.

To underscore the issue: Intel last year recorded its first annual loss since 1986.

Meanwhile, an Intel stake gives SoftBank yet another line—a big one—into a chip business. 

In addition to playing a key role in the OpenAI-Oracle AI datacenter infrastructure project known as Stargate, SoftBank has been on a chip spending spree, agreeing to acquire—among other things—chip designer Arm (2016), Graphcore (2024), and Ampere (March 2025), founded by former Intel president Renée James.

In true Masa Son fashion, it all adds up to a splashy bid to position SoftBank at the very center of an AI revolution that’s still unfolding. 

SoftBank shares are currently trading at a record 16,115 Japanese yen, or about $109, more than double their price just two months ago. —AN

The attorney general of Texas announced Monday that he’s launched an investigation into Meta and Character.AI for deceptive marketing relating to their AI chatbots.

Ken Paxton alleges that the companies may have misled users into thinking their chatbots were “mental health tools.” 

The state’s top lawyer—and a staunch conservative who rode the so-called Tea Party wave a decade ago—is particularly focused on the uncredentialed positioning of the AI tools as “professional therapeutic tools” for vulnerable users, particularly children.

“AI-driven chatbots often go beyond simply offering generic advice and have been shown to impersonate licensed mental health professionals, fabricate qualifications, and claim to provide private, trustworthy counseling services,” Paxton said in a statement.

He added that the chatbots’ user activity tracking and targeting—which their operators disclose—could violate Texas consumer protection laws.

Both companies’ chatbots unquestionably come with visible disclaimers regarding their use; the question is whether those are sufficient for vulnerable users who may not understand or disregard them.

Both companies say their services aren’t intended for children under 13 years of age, but there are few guardrails to stop underage users. 

Meta, for example, prompts new AI users for their birth year and blocks access to those who select 2013 or later, but there’s little stopping them from selecting a different year. 

Both companies have been served with legal orders to produce documents, data, or testimony to supply the probe. —AN

Workday on Friday confirmed that it had suffered a data breach allowing hackers to steal personal information from one of its third-party customer relationship management databases.

In a blog post, the human resources software company didn’t share how much information was stolen—just that the pilfered data included “commonly available business contact information, like names, email addresses, and phone numbers.” (And notably, not customer data.)

The hacker’s motivation, by the way, was “potentially to further their social engineering scams.” Delightful.

The Workday hack comes amid a wave of similar incidents at other prominent corporations, including Adidas, Cisco, Qantas, and Victoria’s Secret.

Many of them involved CRM databases hosted by Salesforce; that company earlier this month published a notice stating that its platform had “not been compromised” and did not relate to “any known vulnerability” in its products. 

In its own June blog post, Google researchers attributed the incidents to ShinyHunters, a hacker group known to use voice phishing, or “vishing,” to trick corporate employees into granting them access to their cloud-based databases. 

The hackers “may be preparing to escalate their extortion tactics by launching a data leak site … to increase pressure on victims,” they wrote, “including those associated with the recent UNC6040 Salesforce-related data breaches.” —AN

—Chamath has a new SPAC. Palihapitiya’s American Exceptionalism Acquisition seeks to raise $250 million in a U.S. IPO. It’s focused on energy, decentralized finance, AI, and defense.

—Arm poaches Amazon AI chip director. Rami Sinno will help the chip designer develop its own chips.

—Palo Alto Networks earnings beat. Q4 revenue that topped estimates plus Q1 and FY26 guidance above expectations.

—Google boosts stake in TeraWulf. Its share of the datacenter operator jumps from 8% to 14%, bringing its total investment to more than $3 billion.

—SoftBank buys Foxconn Ohio factory. Foxconn will still operate the facility, originally intended for electric vehicles, but make AI servers for the Stargate project.

—Figure files for IPO. The New York blockchain lender hopes to ride the U.S. crypto policy wave.

—95% of corporate GenAI pilots have little or no financial impact, per MIT study.