At the end of March, the premier AI startup was collecting superlatives. It had just secured another $40 billion in funding, the largest private tech deal ever. That valued the company at $300 billion, which isthe highest of any startup on the planet. Its flagship product, ChatGPT, was attracting some 500 million users a week, far more than its closest competitor.
All seemed to be going great for OpenAI CEO Sam Altman, who, on top of it all, welcomed his first child a month earlier.
Then the sharks started circling.
In the last several weeks, OpenAI has faced attacks on multiple fronts, mostly from Big Tech behemoths like Meta, Google, Amazon and Microsoft. Smaller companies, too, smelled blood in the water. And rival chatbot makers, like xAI, have released buzzy new models, putting pressure on OpenAI to rush its own update.
OpenAI engineers, some of whom told media outlets they've been working 80 hours a week or more, faced burnout. The company gave them all a week off to recover earlier this month.
It's lonely at the top, as they say. Here's what the siege of OpenAI looks like.
Meta poaches OpenAI staffersAltman said Meta had tried to recruit its staffers with offers that rival superstar professional athletes.
picture alliance/dpa/picture alliance via Getty Images
It seems a top AI engineer is the new superstar athlete.
During a June episode of the "Uncapped with Jack Altman" podcast, Jack's brother Sam said Mark Zuckerberg's Meta tried to poach OpenAI's staffers with "giant signing offers."
Altman said Meta offered"$100 million signing bonuses," which he called "crazy."
"I've heard that Meta thinks of us as their biggest competitor, and I think it is rational for them to keep trying. Their current AI efforts have not worked as well as they've hoped," Altman said.
Meta CTO Andrew Bosworth later told CNBC that Altman "neglected to mention that he's countering those offers."
A week later, Meta had poached three top OpenAI researchers. One of them said on X that he was not offered a $100 million signing bonus, calling it "fake news."
OpenAI itself has poached talent from xAI and Tesla in recent weeks, Wired reported, and Altman brushed off Meta's poaching on the sidelines of the Sun Valley conference earlier this month.
"We have, obviously, an incredibly talented team, and I think they really love what they are doing. Obviously, some people will go to different places," Altman told reporters.
OpenAI's deal with Windsurf falls throughOpenAI representatives told BI its deal with Windsurf fell through.
Craig T Fruchtman/Getty Images
OpenAI took another hit this summer when its deal with Windsurf, the AI coding assistant startup, collapsed. OpenAI had agreed to purchase Windsurf for about $3 billion, Bloomberg reported.
By June, however, tensions were rising between OpenAI and Microsoft. The tech giant is OpenAI's biggest investor, and it considers Windsurf a direct competitor of Microsoft Copilot.
Microsoft's current deal with OpenAI would give it access to Windsurf's intellectual property, which neither OpenAI nor Windsurf wants, a person with knowledge of the talks told BI.
On Friday, OpenAI told BI that its deal with Windsurf had fallen through. Instead, Windsurf CEO Varun Mohan and some other Windsurf employees would join Google DeepMind.
"We're excited to welcome some top AI coding talent from Windsurf's team to Google DeepMind to advance our work in agentic coding," Google's spokesperson told BI. "We're excited to continue bringing the benefits of Gemini to software developers everywhere."
Tensions with MicrosoftOpenAI and Microsoft signed an agreement that defines artificial general intelligence as a system that can generate $100 billion in profits.
Justin Sullivan via Getty Images
The failed Windsurf deal was just another in a string of disagreements that have fueled tension between OpenAI and its largest investor.
The deal between OpenAI and Microsoft is unsurprisingly complex. At the heart of the dispute is revenue splits and equity, of course, but also the very definition of artificial general intelligence. AGI is broadly considered AI that matches or surpasses human intelligence, but in terms of the deal between OpenAI and Microsoft, AGI is defined as $100 billion in profit.
That's a lot of potential revenue.
Under the deal, once OpenAI reaches that benchmark, Microsoft loses its share of OpenAI's revenue. Microsoft would understandably like to revise that line.
As BI's Charles Rollet wrote earlier this month, the tension is made worse by the fact that Microsoft CEO Satya Nadella isn't as sold on AGI's transformative power as all the people developing it at OpenAI. He also doesn't think it's coming anytime soon. He called AGI "nonsensical benchmark hacking" on a podcast earlier this year.
OpenAI delays release of new modelAltman said Friday that he would postpone the release of his much-anticipated new model.
Sebastian Gollnow/picture alliance via Getty Images
Back in simpler times, at the end of March, as Altman was basking in the glow of the world's most valuable startup, he said the newly secured funding would allow OpenAI to "push the frontiers of AI research even further."
He then announced that OpenAI was close to rolling out its first open-weight language model with advanced reasoning capabilities since GPT-2 in 2019.
On Friday evening, generally a good time to unveil bad news, Altman soberly told the world that OpenAI's new model would be delayed — again.
"We need time to run additional safety tests and review high-risk areas," Altman said on X. "We are not yet sure how long it will take us."
He then apologized and assured everyone that "we are working super hard!"
It marked the second delay in a month, pushing the timeline indefinitely beyond earlier promises of a June launch.
Open-weight AI models offer a middle ground between open-source and proprietary systems by sharing only the pre-trained parameters of a neural network but not the actual source code. OpenAI products, unlike some of its competitors, like Meta's Llama and the Chinese AI chatbot, DeepSeek, and despite the company's name, are not open source.
The new model's delay comes days after Elon Musk's xAI launched a major update to its chatbot, Grok. While that update came with some significant trouble, forcing xAI to ultimately apologize, the chatbot boasts advancements in vision and voice that are resonating with users.
Iyo sues IOJony Ive, the famous Apple designer, and Altman struck a deal in May to work together.
LoveFrom
In May, OpenAI announced a partnership with io, the design company founded by the famous former Apple design chief Jony Ive. Together, the two stars would develop future AI consumer devices.
The deal was valued at about $6.5 billion. The announcement included a photo shoot of the two men that wouldn't have been out of place in a Vogue spread and a highly produced video in which Altman and Ive sit and chat in a wine bar drinking espresso.
A month later, OpenAI removed all mentions of the collaboration from its platforms. Another company, iyO, a Google spinoff, had filed a trademark complaint. The names io and iyO were too similar, the suit says, and by all accounts, the new io collaboration would be developing products similar to ones iyO had planned.
US District Judge Trina Thompson ruled that iyO's case is strong enough to move to a hearing this fall. She ordered Altman, Ive, and OpenAI not to use the io brand and take down mentions of the name.
OpenAI denied the claims and said it was reviewing its legal options.
OpenAI announced on July 9 that, despite the lawsuit, it had completed the deal to acquire io and posted a statement on its website.
"We're thrilled to share that the io Products, Inc. team has officially merged with OpenAI. Jony Ive and LoveFrom remain independent and have assumed deep design and creative responsibilities across OpenAI," the statement said.
Amazon is making a movie about AltmanAmazon is developing a movie about OpenAI and Altman called "Artificial."
Kevin Dietsch via Getty Images
The coming film, "Artificial," produced by Amazon Studios, is all about Altman.
And it's not a wholly flattering account, said Matt Beloni, a reporter at Puck who said he has seen a recent draft of the script.
Belloni said the drama recounts the period in 2023 when Altman was fired and then rehired as CEO. It also follows OpenAI cofounder Ilya Sutskever, who was also at the center of that drama and who left the company months later.
At the heart of the tension over those few days was a disagreement between Altman and some top OpenAI execs over the company's commitment to its mission to develop AGI safely.
A string of engineers working on alignment, an AI industry term for ensuring the tech is developed safely, left the company afterAltman's reappointment (Microsoft, incidentally, played a key role in helping Altman survive). While many OpenAI employees rallied around Altman, others described him to the press at that time as "manipulative."
Belloni reported that the film has parallels to "The Social Network," the 2010 biographical drama about Facebook and CEO Mark Zuckerberg.
That film gained critical acclaim and likely damaged Zuckerberg's public persona. Zuckerberg called "The Social Network" inaccurate and "hurtful."
According to Belloni, the version of the script he read depicts Altman as a "master schemer" and a liar.
OpenAI won't go down without a fightAltman at the Sun Valley Conference in 2018.
Drew Angerer/Getty Images
Despite all the competition, OpenAI is still the leader in the space and is making its own moves that will likely worry rivals.
It is planning to launch a new AI-powered web browser, for instance, that could compete with Google Chrome, the current industry leader. The browser will embed ChatGPT and feature an AI agent that can handle tasks like booking reservations and filling out forms.
It also secured a $200 million contract to provide AI support to the US military. OpenAI will help develop capabilities to "address critical national security challenges in both warfighting and enterprise domains," the Pentagon said in June. OpenAI earlier partnered with Palmer Luckey's defense tech firm, Anduril.
OpenAI is also forming more playful partnerships. Last month, Mattel announced it was working with OpenAI to bring AI to its iconic doll, Barbie.
By using OpenAI's technology, Mattel will "bring the magic of AI to age-appropriate play experiences with an emphasis on innovation, privacy, and safety," the California-based toy manufacturer said in a press.
Altman, for his part, is at least publicly optimistic.
"I have never seen growth in any company, one that I've been involved with or not, like this," Altman said at a TED conference in Vancouver in April. "The growth of ChatGPT — it is really fun. I feel deeply honored. But it is crazy to live through."