Meta’s hefty investment in artificial intelligence includes development of an AI system designed to power Facebook’s entire video recommendation engine across all its platforms, a company executive said Wednesday.
Tom Alison, the head of Facebook, said part of Meta’s “technology roadmap that goes to 2026” involves developing an AI recommendation model that can power both the company’s TikTok-like Reels short video service and more traditional, longer videos.
To date, Meta has typically used a separate model for each of its products, such as Reels, Groups and the core Facebook Feed, Alison said onstage at Morgan Stanley’s tech conference in San Francisco.
As part of Meta’s ambitious foray into AI, the company has been spending billions of dollars on Nvidia graphics processing units, or GPUs. They’ve become the primary chips used by AI researchers for training the types of large language models used to power OpenAI’s popular ChatGPT chatbot and other generative AI models.
Alison said “phase 1” of Meta’s tech roadmap involved switching the company’s current recommendation systems to GPUs from more traditional computer chips, helping to improve the overall performance of products.
As interest in LLMs exploded last year, Meta executives were struck by how these big AI models could “handle lots of data and all kinds of very general-purpose types of activities like chatting,” Alison said. Meta came to see the possibility of a giant recommendation model that could be used across products, and by last year, built “this kind of new model architecture,” Alison said, adding that the company tested it on Reels.
This new “model architecture” helped Facebook obtain “an 8% to 10% gain in Reels watch time” on the core Facebook app, which Alison said helped prove that the model was “learning from the data much more efficiently than the previous generation.”
“We’ve really focused on kind of investing more in making sure that we can scale these models up with the right kind of hardware,” he said.
Meta is now in “phase 3” of its re-architecture of the system, which involves trying to validate the technology and push it across multiple products.
“Instead of just powering Reels, we’re working on a project to power our entire video ecosystem with this single model, and then can we add our Feed recommendation product to also be served by this model,” Alison said. “If we get this right, not only will the recommendations be kind of more engaging and more relevant, but we think the responsiveness of them can improve as well.”
Illustrating out how it will work if successful, Alison said, “If you see something that you’re into in Reels, and then you go back to the Feed, we can kind of show you more similar content.”
Alison said Meta has accumulated a massive stockpile of GPUs that will be used to help its broader generative AI efforts, such as development of digital assistants.
Some generative AI projects Meta is considering include incorporating more sophisticated chatting tools into its core Feed so a person who sees a “recommended post about Taylor Swift,” could perhaps “easily just click a button and say, ‘Hey Meta AI, tell me more about what I’m seeing with Taylor Swift right now.'”
Meta is also experimenting with integrating its AI chatting tool within Groups, so a member of a Facebook baking group could potentially ask a question about desserts and get an answer from a digital assistant.
“I think we have the opportunity to put generative AI in kind of a multiplayer kind of consumer environment,” Alison said.
Masayoshi Son, chairman and chief executive officer of SoftBank Group Corp., speaks during the company’s annual general meeting in Tokyo, Japan, on Friday, June 27, 2025.
Bloomberg | Bloomberg | Getty Images
Intel and SoftBank announced on Monday that the Japanese conglomerate will make a $2 billion investment in the embattled chipmaker.
SoftBank will pay $23 per share for Intel’s common stock, which closed on Monday at $23.66. The shares rose about 6% in extended trading to $25.
The investment makes SoftBank the fifth-biggest Intel shareholder, according to FactSet. It’s a vote of support for Intel, which hasn’t been able to take advantage of the artificial intelligence boom in advanced semiconductors and has spent heavily to stand up a manufacturing business that’s yet to secure a significant customer.
“Masa and I have worked closely together for decades, and I appreciate the confidence he has placed in Intel with this investment,” Intel CEO Lip-Bu Tan said in a statement, referring to SoftBank founder Masayoshi Son.
Intel shares lost 60% of their value last year, their worst performance in the company’s more than half-century on the public market. The stock is up 18% in 2025 as of Monday’s close.
Tan took over as Intel CEO in March after his predecessor, Pat Gelsinger, was ousted in December.
Intel has been a major topic of discussion in Washington of late, due to the company’s role as the only American company capable of manufacturing the most advanced chips.
However, Intel’s foundry business, which is designed to manufacture chips for other companies, has yet to secure a major customer, a critical step towards stabilization and expansion. Last month, Intel said it would wait to secure orders before committing to certain future investment in its foundry.
Tan met with President Donald Trump last week after the president had called for the CEO’s resignation. The U.S. government is considering taking an equity stake in Intel, according to reports.
SoftBank, meanwhile, has become an increasingly large player in the global chip and AI markets.
In 2016, SoftBank acquired chip designer Arm in a deal worth about $32 billion at the time. Today the company is worth almost $150 billion. Arm-based chips are part of Nvidia’s systems that go into data centers.
SoftBank was also part of President Trump’s Stargate announcement in January, along with OpenAI and Oracle.
The three companies committed to invest an initial $100 billion and up to $500 billion over the next four years in the AI infrastructure project. Two months later, SoftBank led a $40 billion investment into OpenAI, the largest private tech deal on record.
“This strategic investment reflects our belief that advanced semiconductor manufacturing and supply will further expand in the United States, with Intel playing a critical role,” Son said in a statement.
Nikesh Arora, CEO of Palo Alto Networks, looks on during the closing bell at the Nasdaq Market in New York City on March 25, 2025.
Jeenah Moon | Reuters
Palo Alto Networks reported better-than-expected quarterly results and issued upbeat guidance for the current period. The cybersecurity software vendor said Nir Zuk, who founded the company in 2005, is retiring from his role as chief technology officer.
The stock rose about 6% in extended trading.
Here’s how the company did compared to LSEG estimates:
Earnings: 95 cents adjusted vs. 88 cents expected
Revenue: $2.54 billion vs. $2.5 billion expected.
Revenue in the fiscal fourth quarter rose 16% from about $2.2 billion last year, the company said in a statement. Net income fell to about $254 million, or 36 cents per share, from about $358 million, or 51 cents per share, in the year-ago period.
The company also issued upbeat guidance for the fiscal first quarter. Earnings per share will be between 88 cents and 90 cents, Palo Alto said, topping an 85-cents estimate from StreetAccount.
For the full year, Palo Alto said revenue will range from $10.48 billion to $10.53 billion on adjusted earnings of $3.75 to $3.85 per share. Both estimates exceeded Wall Street’s projections.
Palo Alto said that for the fiscal first quarter, remaining purchase obligations, which tracks backlog, will range between $15.4 billion and $15.5 billion, surpassing a $15.07 billion estimate.
Last month, the company announced plans to buy Israeli identity security provider CyberArk for $25 billion. It’s the largest deal Palo Alto has made since its founding, and most ambitious in an acquiring spree that ramped up after CEO Nikesh Arora took the helm of the company in 2018.
Shares sold off sharply after the news broke and have yet to recover previous highs. The stock is down about 3% this year as of Monday’s close.
“We look for great products, a team that can execute in the product, and we let them run it,” Arora told CNBC following the announcement. “This is going to be a different challenge, but we’ve done well 24 times, so I’m pretty confident that our team can handle this.”
Lee Klarich, the company’s product chief, will replace Zuk as CTO and fill his position on the board.
Satellite internet service Starlink, which is owned and operated by Elon Musk‘s SpaceX, appeared to suffer a brief network outage on Monday, with thousands of reports of service interruptions on Downdetector, a site that logs tech issues.
The outage marked the second in two weeks for Starlink. SpaceX did not immediately respond to a request for comment.
The network’s July 24 outage lasted for several hours, with SpaceX Vice President of Starlink Engineering Michael Nicolls blaming the matter on “failure of key internal software services that operate the core network” behind Starlink.
That outage followed the launch of T-Mobile‘s Starlink-powered satellite service, a direct-to-cell-phone service created to keep smartphone users connected “in places no carrier towers can reach,” according to T-Mobile’s website.
SpaceX provides Starlink internet service to more than six million users across 140 countries, according to the company’s website, though churn and subscriber rates are not publicly reported by the company.
Read more CNBC tech news
The SpaceX Starlink constellation is far larger than any competitor. It currently features over 7,000 operational broadband satellites, according to research by astronomer Jonathan McDowell.
On Monday, Musk’s SpaceX successfully launched another group of satellites to add to its Starlink constellation from the Vandenberg Space Force Base in Southern California.
SpaceX is currently aiming to increase the number of launches and landings from Vandenberg from 50 to about 100 annually.
On Thursday last week, the California Coastal Commission voted unanimously to oppose the U.S. Space Force application to conduct that higher volume of SpaceX launches there.
The Commission has said that SpaceX and Space Force officials have failed to properly evaluate and report on potential impacts of increased launches on neighboring towns, and local wildlife, among other issues.
President Donald Trump recently signed an executive order seeking to ease environmental regulations seen by Musk, and others, as hampering commercial space operations.