Arm Holdings CEO Rene Haas told CNBC’s Jim Cramer on Wednesday that moving some AI functions away from the could help reduce energy usage.

Over time, he suggested, a large number of multi-gigawatt data centers won’t be sustainable.

“You look to yourself, well, what are the kind of things that need to happen? I think there’s two vectors to it,” Haas said. “One is low power, the lowest power solution you can get in the cloud. Arm really contributes there. But I think even more specifically is moving those AI workloads away from the cloud to local applications.”

While he said AI training will likely always happen in the cloud, running AI, called inference, can happen locally — meaning on the chips inside people’s phones, computers and glasses. History has shown “we always go to hybrid models around computing,” according to Haas.

He suggested that hybrid dynamic will play out when it comes to AI, which will help alleviate huge power investments.

Chip designer Arm’s technology powers devices made by a number of major Big Tech players, including Microsoft and Amazon. Semiconductor giant Nvidia has a major stake in Arm and actually attempted to acquire the company in 2020.

Arm and Meta on Wednesday said they would expand their partnership to “scale AI efficiency across every layer of compute – spanning AI software and data center infrastructure,” according to a press release. Arm stock saw gains following the announcement, finishing the day up 1.49%.

Haas told Cramer that the partnership with Meta is “largely around data centers, but more broadly…around software and the software stacks associated with it.” He also discussed Arm’s involvement in Meta’s new Ray-Ban Wayfarer glasses, saying the AI for the technology is running both in the cloud and locally.

“For example, when you say, ‘hey, Meta,’ into those glasses, that’s not happening on the cloud, that’s actually happening in your glasses, and that’s running on Arm,” Haas said.

Sign up now for the CNBC Investing Club to follow Jim Cramer’s every move in the market.

Disclaimer The CNBC Investing Club Charitable Trust owns shares of Meta, Microsoft, Amazon and Nvidia.

Salesforce shares rose as much as 5% in extended trading on Wednesday after the software vendor issued new financial targets for the next few years.

The company said it now expects revenue of over $60 billion in 2030, above the $58.37 billion consensus among analysts polled by LSEG.

The guidance excludes impact from the pending acquisition of data management company Informatica. The $8 billion deal, announced in May, is slated to close in the fiscal fourth quarter or in the first quarter of the 2027 fiscal year.

“We have had some lower-stage growth for a while,” Robin Washington, Salesforce’s chief operating and financial officer, said during an investor briefing at the company’s annual Dreamforce conference in San Francisco. “That is reaccelerating.”

She company called for an organic year-over-year revenue growth rate above 10% in the 2026 through 2030 fiscal years. The growth rate has been under 10% since mid-2024.

Investors have been concerned, in part because of the rise of “vibe-coding” tools for automatically generating software with a few words of human input. Industry observers have predicted that artificial intelligence services might threaten longstanding software providers. Microsoft CEO Satya Nadella said in April that AI is creating up to 30% of new code at the company.

“There’s a certain amount of, let’s just say, nonsense that’s out there,” Salesforce CEO Marc Benioff said on Wednesday. “Like, for example, that these products are writing all the software, and that is not what’s happening.”

As of Wednesday’s close, Salesforce’s stock had fallen 29% for the year, while the Nasdaq has gained 17%.

To increase revenue, Salesforce is counting on its Agentforce software for automating customer service and other business processes, said Washington, who also sits on Salesforce’s board. The company introduced Agentforce last year as a way for brands to add chat-based customer service agents that connect large language models to internal data.

“Investors continue to ask why Agentforce adoption has been slower than anticipated,” analysts at RBC Capital Markets wrote in a note to clients earlier this month.

Salesforce executives are hoping product enhancements will attract more business.

The company on Monday released Agentforce Voice, which allows clients to have agents answer customer service calls. On Tuesday, Salesforce announced larger partnerships with AI model developers Anthropic and OpenAI, bringing their latest models to Agentforce.

At Dreamforce, Salesforce pointed to Agentforce adoption at FedEx, Pandora, PepsiCo, Williams Sonoma and other companies.

WATCH: People don’t understand Agentforce is part and parcel of Salesforce, says CEO Marc Benioff

As U.S. states start to react to growing constituent concerns around the risks associated with artificial intelligence use, Tennessee Sen. Marsha Blackburn said moving forward with a federal preemption standard is “imperative.”

Earlier this week, California Gov. Gavin Newsom signed a series of bills focused on those concerns — while also vetoing some strict AI conditions legislators hoped for — requiring safeguards around chatbots, labels around the mental risks of social media apps, and tools that require age verification in device maker app stores.

In addition, Utah and Texas have also signed laws implementing AI safeguards for minors, and other states have indicated similar regulations could be on the horizon.

“The reason the states have stepped in, whether it’s to protect consumers or protect children, is because the federal government has, to date, not been able to pass any federal preemptive legislation,” Blackburn said at the CNBC AI Summit on Wednesday in Nashville. “We have to have the states standing in the gap until such time that Congress will say no to the big tech platforms.”

Blackburn has long been a proponent of legislation around children’s online safety and regulation of social media, introducing the Kid’s Online Safety Act in 2022 that aims to establish guidelines to protect minors from harmful material on the platforms. The bipartisan legislation has passed the Senate with an overwhelming majority, and Blackburn said while big tech companies have worked to hold up the legislation from passage in both chambers, “We are hopeful the House is going to take it up and pass it.”

But the concerns that the Act was aimed to address as it relates to social media have now cascaded alongside the rise in AI, Blackburn said.

“One of the things we’ve heard from so many people involved in this is that you have to have an online consumer privacy protection bill so that people have the ability to set those firewalls and protect the virtual you, as I call it,” she said, adding that “once an LLM scoops [your data and information], then they are using that to train that model.”

Blackburn is also focused on several other ways of safeguarding the information that AI is using, including a bill focused on how AI can use your name, image or likeness without your consent.

“We have to have a way to protect our information in the virtual spaces just as we do in the physical space,” she said.

With the fast advancement of AI, Blackburn acknowledged that regulation would require a focus on “end-use utilizations and legislate that framework in that manner and not focus on a given delivery system or a given technology.”

That also means reacting to the ways that AI companies change their products. Earlier this week, OpenAI CEO Sam Altman said the company will be able to “safely relax” most restrictions now that it has been able to mitigate “serious mental health issues,” adding that the company is “not the elected moral police of the world.”

Blackburn said that legislators are increasingly hearing from “parents who know what is happening to their children and that they can’t un-experience or unsee something that they have been through with these chatbots or in the virtual world or the metaverse.”

“I have talked to so many people who are now saying kids are not going to get cell phones until they’re 16, and many parents believe that is just like driving a car,” she said. “They’re not going to allow their kids to have that because we as a society have to put rules and laws in place that protect children and minors.”