Demis Hassabis, CEO of Google DeepMind, became such a source of anxiety for Elon Musk that the Tesla founder repeatedly questioned whether the AI researcher was "evil" and warned his inner circle that OpenAI was on a "path of certain failure" compared to Hassabis's operation. Court testimony and trial documents from the ongoing Musk v. Altman lawsuit reveal just how deeply Hassabis's achievements and ambitions lodged in Musk's mind, shaping critical decisions about OpenAI's structure and strategy.

Why Did Musk Fear Demis Hassabis So Much?

The anxiety began early. In 2014, when DeepMind was running out of funding, Musk made a pitch to Hassabis to join a different venture. Hassabis declined, choosing instead to accept Google's acquisition offer, reportedly valued between $400 million and $650 million. That decision set the stage for years of competitive tension. Musk had warned Hassabis about the risks of handing artificial general intelligence (AGI) research to a corporation, but Hassabis prioritized the computational resources Google could provide at the scale his ambitions demanded.

According to testimony from OpenAI president Greg Brockman during the trial, Musk brought up Hassabis "many, many times" in OpenAI's early years, describing himself as "very consistent and fixated" on the man. At one AI-focused dinner with Brockman and Sam Altman, Musk's first question was blunt: "Is Demis Hassabis evil?". A separate dinner with Hassabis left Musk so rattled that he emailed Brockman and OpenAI cofounder Ilya Sutskever, describing the encounter as "extremely alarming".

By 2016, Musk's concerns had crystallized into a sports metaphor that captured his desperation. "I feel like they are playing the Super Bowl and we are playing the Puppy Bowl," he wrote. "Unless we want to have our ass handed to us, we need to step up our game dramatically". The comparison wasn't hyperbole in Musk's mind; it reflected a genuine belief that OpenAI was structurally outmatched.

What Did Hassabis Actually Accomplish to Justify the Fear?

Musk's anxiety proved prescient. Under Hassabis's leadership, DeepMind delivered breakthrough after breakthrough. In 2016, AlphaGo defeated Lee Sedol, one of the world's strongest Go players, in a five-game match broadcast globally. The victory signaled that AI had reached a new level of capability, capable of mastering one of humanity's most complex games through self-learning and pattern recognition.

But Hassabis didn't stop there. Immediately after the Go match, he pivoted to what many consider an even more consequential problem: protein folding. AlphaFold, DeepMind's AI system, solved a 50-year-old challenge in biology by predicting the three-dimensional structure of proteins from their amino acid sequences. The breakthrough had immediate applications in medicine and drug discovery, earning Hassabis the 2024 Nobel Prize in Chemistry, making him the first AI researcher to receive the honor.

The scale of Hassabis's ambition extended beyond individual projects. Alphabet committed roughly $175 billion to AI infrastructure in 2026 alone, providing DeepMind with computational resources that dwarfed what OpenAI could access independently. Hassabis also founded Isomorphic Labs, a for-profit spinoff focused on AI-driven drug discovery, further expanding DeepMind's reach into real-world applications.

How Did Musk's Fear Reshape OpenAI's Strategy?

Musk's fixation on Hassabis led to increasingly desperate proposals. By January 2018, Musk was convinced that OpenAI was "on a path of certain failure relative to Google" and that "immediate and dramatic action" was needed or "everyone except for Google will be consigned to irrelevance". He and OpenAI cofounder Andrej Karpathy even suggested folding OpenAI into Tesla to give it better resources.

Shivon Zilis, an OpenAI board member at the time and now a mother of four children with Musk, made a personal plea to him. "There is a very low probability of a good future if someone doesn't slow Demis down," she wrote. "Slowing him down is the only nonnegotiable net good action I can see". She relayed rumors that members of AI's inner circle were meeting in London coffee shops without cell phones, fearing Hassabis might spy on their communications.

By November 2018, Musk had lost confidence entirely. "My probability assessment of OpenAI being relevant to DeepMind/Google without a dramatic change in execution and resources is 0%," he wrote. He pivoted his strategy, announcing plans to pursue AGI research through Tesla instead, writing that the company had "cash flow on the order of billions of dollars per year to build hardware that hopefully has at least a dark horse chance to keep Google honest".

Steps to Understanding the AI Competition That Shaped OpenAI's Future

• The 2014 Acquisition: Google acquired DeepMind for between $400 million and $650 million, giving Hassabis access to computational resources that Musk's alternative offer could not match, fundamentally shifting the competitive landscape.
• AlphaGo's 2016 Victory: DeepMind's AI defeated world champion Lee Sedol at Go, a game with more possible positions than atoms in the universe, proving that machine learning could master intuitive, strategic thinking.
• AlphaFold's Protein Breakthrough: Hassabis's team solved the 50-year-old protein-folding problem, earning him the Nobel Prize in Chemistry and demonstrating AI's potential to accelerate scientific discovery in medicine and biology.
• Structural Disadvantage: Musk believed OpenAI's nonprofit structure, designed to prevent corporate control of AGI, actually handicapped it compared to DeepMind's access to Google's $175 billion AI infrastructure budget.
• The Pivot to Tesla: Losing confidence in OpenAI's ability to compete, Musk shifted his AGI ambitions to Tesla, betting that hardware and manufacturing could provide the edge that software alone could not.

What emerges from the trial documents is a portrait of competitive anxiety that shaped one of tech's most important institutions. Musk's fear of Hassabis wasn't paranoia; it was a rational assessment of resource disparity and execution speed. Hassabis had Google's backing, world-class researchers, and a track record of delivering breakthroughs. Musk had a nonprofit structure designed to prevent exactly the kind of aggressive resource deployment that Hassabis was executing.

By March 2019, the last mention of Hassabis in the trial exhibits came from a cryptic message from Sam Altman to Musk: "Have some mild Demis updates to share," Altman wrote. Musk agreed to discuss it by phone, but the conversation was never detailed in court documents. By then, the competitive dynamic had already shifted. OpenAI would eventually secure massive funding from Microsoft, transforming its resource position. But the anxiety that Demis Hassabis inspired in Musk's circle had already left its mark on how OpenAI was structured, funded, and strategically positioned.

Hassabis, meanwhile, continued his work in London, pursuing what he describes as his ultimate goal: solving intelligence itself. In a recent documentary aired on South Korean television, Hassabis declared, "My goal is to solve intelligence, and then use that intelligence to solve everything else," predicting that the era of AGI could begin within the next decade. Whether that prediction proves accurate or not, the trial documents make clear that Musk took it seriously enough to reshape OpenAI's entire strategy around the fear of falling behind.

Hassabis, meanwhile, continued his work in London, pursuing what he