Parker Harris, a co-founder of Salesforce, speaks during a keynote at the company’s Dreamforce conference in San Francisco on Sept. 12, 2023.
Marlena Sloss | Bloomberg | Getty Images
Each year, Salesforce updates its V2MOM, a planning document laying out vision, values, methods, obstacles and measures. CEO Marc Benioff has said it’s “been used to guide every decision at Salesforce” since the software company’s founding 25 years ago this week.
But in early 2023, there was a problem. ChatGPT was going viral, and Salesforce’s strategy didn’t account for it.
“The V2MOM had nothing about generative AI,” Parker Harris, who co-founded the company with Benioff, told CNBC in an interview.
It was a first for Salesforce, which had never been caught so off-guard about an emerging technology trend. If Salesforce was to become a leader in generative artificial intelligence, the company would need to quickly revise its guiding document to redirect the company — and its 73,000 employees — toward the technology that’s sweeping across Silicon Valley and making its way into every industry, from manufacturing to medicine.
Salesforce would have to go to battle with tech giants Amazon, Google and Microsoft, as well as red-hot and well-capitalized startups. But following a handful of hefty acquisitions and a run-in with activist investors that led Salesforce to disband its M&A committee, a splashy deal was likely off the table.
Salesforce would have to build. And that’s when Benioff turns to his longtime sidekick, Harris.
Well known in the software industry but largely unfamiliar outside of it, Harris has always been core to the fabric of Salesforce. In the past six years, Benioff has elevated two different top lieutenants to the role of co-CEO, but neither lasted in the job longer than 18 months. Harris, a Salesforce board member and now the technology chief of Slack, which Salesforce bought in 2021, said he’d rather avoid the limelight.
“I don’t like being front and center,” Harris said, in an interview tied to the company’s 25th anniversary, which was officially March 8. “I don’t like the articles necessarily to be written about me. I like being behind the scenes.”
Internally, Harris is in the thick of it. After generative AI made its way into the revised V2MOM last year, Harris supervised its brisk insertion into the company’s sales, customer service, marketing and commerce applications. He studied new techniques such as retrieval-augmented generation, which involves feeding information outside of an AI model’s training set to yield a better answer.
Questions swirled about whether Salesforce should spend billions of dollars to assemble its own general-purpose large language model for spitting out text in response to a few words of human input, Harris said. But the company started seeing clients use multiple LLMs.
Salesforce slashed its investment in some areas while doubling the size of its research group, which was fleshing out its own AI models. At the same time, it started drawing on models from AI startup Anthropic, as well as GPT-4, the model powering OpenAI’s ChatGPT. In September, Benioff brought OpenAI CEO Sam Altman onstage at Salesforce’s annual Dreamforce conference, which takes over a chunk of downtown San Francisco.
At past Dreamforce shows, Harris has appeared in superhero costumes, entertaining the audience of tens of thousands. But 2023 was not a time for jokes. Harris was busy repositioning the company. He chose a professional look: a checked blue suit that matched his glasses with thin blue frames.
Marc Benioff, chairman and CEO of Salesforce, right, and Parker Harris, co-founder of Salesforce, introduce Salesforce 1 Lightning during the company’s Dreamforce conference in San Francisco on Oct. 14, 2014.
Noah Berger | Bloomberg | Getty Images
In his keynote, Harris talked about the Data Cloud, a product originally called Genie that surfaces real-time information. In about 2016 he had decided to push much of Salesforce’s IT infrastructure into the public cloud, enabling tighter integration of many assets the company had acquired over the years. That helped Salesforce launch the Data Cloud.
Without the Data Cloud, Harris told CNBC, “I think we would have been in a much worse place.” It’s such a critical part of the company that Benioff mentioned it 58 times on the company’s earnings call in February.
A Robin for Batman
Despite his status as the most decorated technical leader at one of the world’s largest software companies, Harris was an English major. He earned a bachelor’s degree from Middlebury College in Vermont.
His love of computers came early, though. He told Business Insider in 2015 that he started programming on an Apple II as a kid growing up in North Carolina.
In the early 1990s, he moved to the San Francisco Bay Area and took a software-engineering job at a company called Metropolis Software, where he got to know developers Frank Dominguez and Dave Moellenhoff. The trio founded a Java consulting firm called Left Coast Software.
They were contracting at Saba Software, an online learning company co-founded by former Oracle executive Bobby Yazdani. Benioff, who was still working at Oracle under Larry Ellison at the time, told Yazdani that he had this idea to build web-based sales software. Yazdani told Benioff he needed to meet Harris, Dominguez and Moellenhoff.
“He was a very abstract thinker,” Yazdani said about Harris, in an interview with CNBC. “He had clarity around capability of what’s possible.”
In the fall of 1998, Benioff and Harris met for lunch at Kincaid’s, a seafood and steak restaurant in Burlingame overlooking the San Francisco Bay. It was an uneven match. Benioff is hard to miss at 6 feet, 5 inches tall. He’s loud and loves to talk.
Harris is scrawny and quieter. He said he’s averse to conflict. He defuses the drama, said Brett Queener, a former Salesforce executive who’s now a venture capitalist.
“Every Batman needs their Robin,” Queener said.
After the lunch meeting, Benioff had Harris, Dominguez and Moellenhoff over to his home in San Francisco’s Telegraph Hill neighborhood. They were all in.
Salesforce.com was born on March 8, 1999. Harris was 32. His parents, wife and young daughter came by corporate headquarters — a one-bedroom apartment next to Benioff’s home — to commemorate the moment, which Harris posted to YouTube eight years later.
“We are going to probably work here for six months to a year, and we’re going to just really enjoy it,” he told his father, who was behind the camera. Salesforce played the clip for employees this week during a celebration.
While Harris shared the title of co-founder with Benioff, his partner held much more of the equity. That’s why Benioff is now worth around $11 billion, with a current stake in Salesforce that exceeds $7 billion, while Harris’ holdings are worth nearly $600 million.
Though he’s relatively soft spoken, Harris has his indulgences. He’s spent money on red wine from France and Italy, works of art by Ruth Asawa and Josef Albers, a home in Nantucket and a renovation of the family home in San Francisco’s Pacific Heights.
“We really shifted it to a focus on sunlight,” Harris said.
In his office at the top of the house, he likes to put on headphones and crank up the music. He listens to the Avett Brothers, Radiohead and Miles Davis. He plays golf and surfs. A coworker said Harris is an “enthusiastic” dancer. He belongs to Middlebury’s board of trustees.
At Salesforce, Harris led the development of the platform that enables companies to build on top of its software, along with an initiative to make Salesforce work well on mobile devices. There was also the push to build the next-generation Salesforce Lightning, as well as Chatter, an enterprise social network.
He talked about AI way back at Dreamforce 2009, suggesting that the technology might one day help Chatter identify in-house experts on different topics. He admitted to his shortcomings.
“I don’t understand that area,” Harris told a group of journalists, regarding AI. “I understand we have to solve it. I have hired some people in that area that do understand it.”
Tough time in social
At the time, social was the big buzzword. Facebook was still private but taking off.
A startup called Yammer was being described as the Facebook for the workplace. A few Salesforce employees started discussing the potential for information to go viral among salespeople and customer-service agents. Benioff was intrigued. He insisted that it become the top priority.
After Harris allocated eight engineers to the new project, Benioff demanded he go bigger. Harris checked in with engineering leaders and secured a headcount of 75.
That wasn’t enough. At a briefing on the updated status, Benioff was dissatisfied, according to a meeting attendee who asked not to be named to speak candidly about the matter. Harris was silent. His face went pale. He told Benioff he’d redo the plan, the person said.
Marc Benioff, co-founder and CEO of Salesforce, sits in the audience ahead of the special address by U.S. President Donald Trump on the opening day of the World Economic Forum in Davos, Switzerland, on Jan. 21, 2020.
Jason Alden | Bloomberg | Getty Images
Harris eventually got 80% of Salesforce’s engineers to start working on Chatter. But the product never took off.
“We didn’t take it far enough,” Harris said. Microsoft was also hot to get into the market, snapping up Yammer in 2012 for $1.2 billion, a huge multiple for a company with a small revenue base.
Salesforce wound up buying the big prize in the space, purchasing Slack in 2021 for $27.1 billion, by far the company’s priciest deal.
But perhaps Harris’ biggest swing in his decades at Salesforce was the push to the public cloud. It wasn’t an easy choice.
“Half the engineers, the brightest people, were like, ‘We’re going to run the company if we do this,'” Harris recalled. “The great fear was that we would ruin our cost model because the cost would be much more expensive on public cloud, and then we would be able to hire less salespeople or less engineers or whatever.”
The other half of the engineering staff, Harris said, was petrified that if Salesforce didn’t move to the cloud, everyone else will “innovate faster than us.'”
Benioff didn’t have much to contribute for a change.
“Marc was like, ‘This is crazy, that these are some of the smartest people I know, and you guys can’t agree,'” Harris said.
Marc Benioff, chairman and CEO of Salesforce, left, speaks as Parker Harris, co-founder of Salesforce, center, and Kara Swisher, executive editor of Re/Code, listen during a keynote address at the Dreamforce conference in San Francisco on Sept. 17, 2015. Salesforce.com Inc. aims to cut the time its customers spend plugging data into its systems by weaving machine-learning technology from acquisition RelateIQ into its software for managing sales accounts.
David Paul Morris | Bloomberg | Getty Images
Harris saw the advantage that startups gained by outsourcing data center needs to Amazon Web Services. And he knew Salesforce had failed to build a viable platform for easily developing apps while partnering with VMware. Harris concluded that not pushing Salesforce to public cloud services like AWS would be an existential threat.
“That was a very lonely decision,” he said. But as it became a part of the V2MOM, it rippled out to thousands.
While Salesforce might have saved money when it ditched its Equinix colocation facilities, leaning more on the cloudhasn’t been cheap. Last year, after activist investors called for more profitability from Salesforce, the company signed up for longer-term cloud commitments. It agreed to spend at least $16.8 billion on infrastructure service providers as of January, up from $6.5 billion in January 2023, according to regulatory filings.
The biggest beneficiary of that spending is AWS, which is run by former Salesforce executive Adam Selipsky. Harris said Salesforce is looking at other providers as additional partners.
“Oracle has done a great job around their platform, so technically, it’s actually quite good,” he said.
‘Try to build something great’
Harris recently gave up the CTO title at Salesforce that he’d held for seven years. The company hasn’t yet named a successor.
Now he’s onto Slack.
In 2022, Slack CEO Stewart Butterfield left the company he founded in 2009. He was replaced by Lidiane Jones. She departed a year later to run dating app developer Bumble. And in January of this year, Slack co-founder and CTO Cal Henderson said he was stepping down.
“I thought, ‘I can have an impact there,'” Harris said. ‘But I can also — I would love to do that job, I would love to go back and run some engineering teams and really try to build something great.'”
Harris visited Benioff’s home in the Sea Cliff neighborhood of San Francisco, and the two co-founders were in agreement that it was the right call.
“I’m excited for this next chapter with Parker as Slack’s CTO, continuing his legacy as one of our industry’s greats,” Benioff said in an email.
Harris flew to New York to hang out with Noah Weiss, Slack’s product chief. Harris moved his desk to the Slack floor in San Francisco’s Salesforce Tower, where he’s near new unit CEO Denise Dresser. He comes in two to four days a week, and attends Monday meetings to review Slack metrics.
“People, probably fairly, had a lot of apprehension,” Weiss said.
Some of Slack’s employees suspected Harris would try to apply the Salesforce approach to Slack. But instead, Harris sought to understand how Slack had become successful.
Weiss said that at Salesforce’s new fiscal year kickoff in Las Vegas last month, Harris talked at an executive meeting about one of Slack’s product principles called prototype the path. And Harris has started writing documents in Slack’s collaborative Canvas tool.
“He’s been showing up extremely well, definitely winning hearts and minds, for sure, including mine,” Weiss said.
Employees sometimes add flair to Slack chats with a Parker Harris emoji, he said.
When it comes to keeping up with Benioff, who spends a healthy amount of time at his palatial estate in Hawaii, Harris uses other services.
“Marc is all mobile and all text and FaceTime,” Harris said.
The men talk once every few weeks. They’ll be talking more frequently, as Harris said they’re about to kick off weekly meetings on Slack and Salesforce integrations.
Harris hopes that his presence can convince Slack employees to stay after the executive exodus.
“I don’t want to talk too much about myself, but I think it is helping,” Harris said.
Elon Musk on Monday said he does not support a merger between xAI and Tesla, as questions swirl over the future relationship of the electric automaker and artificial intelligence company.
X account @BullStreetBets_ posted an open question to Tesla investors on the social media site asking if they support a merger between Tesla and xAI. Musk responded with “No.”
The statement comes as the tech billionaire contemplates the future relationship between his multiple businesses.
Overnight, Musk suggested that Tesla will hold a shareholder vote at an unspecified time on whether the automaker should invest in xAI, the billionaire’s company that develops the controversial Grok AI chatbot.
Last year, Musk asked his followers in an poll on social media platform X whether Tesla should invest $5 billion into xAI. The majority voted “yes” at the time.
Musk has looked to bring his various businesses closer together. In March, Musk merged xAI and X together in a deal that valued the artificial intelligence company at $80 billion and the social media company at $33 billion.
Musk also said last week that xAI’s chatbot Grok will be available in Tesla vehicles. The chatbot has come under criticism recently, after praising Adolf Hitler and posting a barrage of antisemitic comments.
— CNBC’s Samantha Subin contributed to this report.
Coincidentally, OpenAI CEO Sam Altman announced early Saturday that there would be an indefinite delay of its first open-source model yet again due to safety concerns. OpenAI did not immediately respond to a CNBC request for comment on Kimi K2.
In its release announcement on social media platforms X and GitHub, Moonshot claimed Kimi K2 surpassed Claude Opus 4 on two benchmarks, and had better overall performance than OpenAI’s coding-focused GPT-4.1 model, based on several industry metrics.
“No doubt [Kimi K2 is] a globally competitive model, and it’s open sourced,” Wei Sun, principal analyst in artificial intelligence at Counterpoint, said in an email Monday.
Cheaper option
“On top of that, it has lower token costs, making it attractive for large-scale or budget-sensitive deployments,” she said.
The new K2 model is available via Kimi’s app and browser interface for free unlike ChatGPT or Claude, which charge monthly subscriptions for their latest AI models.
Kimi is also only charging 15 cents for every 1 million input tokens, and $2.50 per 1 million output tokens, according to its website. Tokens are a way of measuring data for AI model processing.
In contrast, Claude Opus 4 charges 100 times more for input — $15 per million tokens — and 30 times more for output — $75 per million tokens. Meanwhile, for every one million tokens, GPT-4.1 charges $2 for input and $8 for output.
Moonshot AI said on GitHub that developers can use K2 however they wish, with the only requirement that they display “Kimi K2” on the user interface if the commercial product or service has more than 100 million monthly active users, or makes the equivalent of $20 million in monthly revenue.
Hot AI market
Initial reviews of K2 on both English and Chinese social media have largely been positive, although there are some reports of hallucinations, a prevalent issue in generative AI, in which the models make up information.
Still, K2 is “the first model I feel comfortable using in production since Claude 3.5 Sonnet,” Pietro Schirano, founder of startup MagicPath that offers AI tools for design, said in a post on X.
Moonshot has open sourced some of its prior AI models. The company’s chatbot surged in popularity early last year as China’s alternative to ChatGPT, which isn’t officially available in the country. But similar chatbots from ByteDance and Tencent have since crowded the market, while tech giant Baidu has revamped its core search engine with AI tools.
Kimi’s latest AI release comes as investors eye Chinese alternatives to U.S. tech in the global AI competition.
Still, despite the excitement about DeepSeek, the privately-held company has yet to announce a major upgrade to its R1 and V3 model. Meanwhile, Manus AI, a Chinese startup that emerged earlier this year as another DeepSeek-type upstart, has relocated its headquarters to Singapore.
Over in the U.S., OpenAI also has yet to reveal GPT-5.
Work on GPT-5 may be taking up engineering resources, preventing OpenAI from progressing on its open-source model, Counterpoint’s Sun said, adding that it’s challenging to release a powerful open-source model without undermining the competitive advantage of a proprietary model.
“Kimi-Researcher represents a paradigm shift in agentic AI,” said Winston Ma, adjunct professor at NYU School of Law. He was referring to AI’s capability of simultaneously making several decisions on its own to complete a complex task.
“Instead of merely generating fluent responses, it demonstrates autonomous reasoning at an expert level — the kind of complex cognitive work previously missing from LLMs,” Ma said. He is also author of “The Digital War: How China’s Tech Power Shapes the Future of AI, Blockchain and Cyberspace.”
Co-founder and chief executive officer of Nvidia Corp., Jensen Huang attends the 9th edition of the VivaTech trade show in Paris on June 11, 2025.
Chesnot | Getty Images Entertainment | Getty Images
Nvidia CEO Jensen Huang has downplayed U.S. fears that his firm’s chips will aid the Chinese military, days ahead of another trip to the country as he attempts to walk a tightrope between Washington and Beijing.
In an interview with CNN aired Sunday, Huang said “we don’t have to worry about” China’s military using U.S.-made technology because “they simply can’t rely on it.”
“It could be limited at any time; not to mention, there’s plenty of computing capacity in China already,” Huang said. “They don’t need Nvidia’s chips, certainly, or American tech stacks in order to build their military,” he added.
The comments were made in reference to years of bipartisan U.S. policy that placed restrictions on semiconductor companies, prohibiting them from selling their most advanced artificial intelligence chips to clients in China.
Huang also repeated past criticisms of the policies, arguing that the tactic of export controls has been counterproductive to the ultimate goal of U.S. tech leadership.
“We want the American tech stack to be the global standard … in order for us to do that, we have to be in search of all the AI developers in the world,” Huang said, adding that half of the world’s AI developers are in China.
That means for America to be an AI leader, U.S. technology has to be available to all markets, including China, he added.
Washington’s latest restrictions on Nvidia’s sales to China were implemented in April and are expected to result in billions in losses for the company. In May, Huang said chip restrictions had already cut Nvidia’s China market share nearly in half.
Last week, the Nvidia CEO met with U.S. President Donald Trump, and was warned by U.S. lawmakers not to meet with companies connected to China’s military or intelligence bodies, or entities named on America’s restricted export list.
According to Daniel Newman, CEO of tech advisory firm The Futurum Group, Huang’s CNN interview exemplifies how Huang has been threading a needle between Washington and Beijing as it tries to maintain maximum market access.
“He needs to walk a proverbial tightrope to make sure that he doesn’t rattle the Trump administration,” Newman said, adding that he also wants to be in a position for China to invest in Nvidia technology if and when the policy provides a better climate to do so.
But that’s not to say that his downplaying of Washington’s concerns is valid, according to Newman. “I think it’s hard to completely accept the idea that China couldn’t use Nvidia’s most advanced technologies for military use.”
He added that he would expect Nvidia’s technology to be at the core of any country’s AI training, including for use in the development of advanced weaponry.
A U.S. official told Reuters last month that China’s large language model startup DeepSeek — which says it used Nvidia chips to train its models — was supporting China’s military and intelligence operations.
On Sunday, Huang acknowledged there were concerns about DeepSeek’s open-source R1 reasoning model being trained in China but said that there was no evidence that it presents dangers for that reason alone.
Huang complimented the R1 reasoning model, calling it “revolutionary,” and said its open-source nature has empowered startup companies, new industries, and countries to be able to engage in AI.
“The fact of the matter is, [China and the U.S.] are competitors, but we are highly interdependent, and to the extent that we can compete and both aspire to win, it is fine to respect our competitors,” he concluded.