In the fast-evolving world of technology, the relationship between parent companies and their subsidiaries remains fraught with complexities and conflicting ambitions. A newly published biography shedding light on the life of Demis Hassabis reveals exciting details about years of secret efforts by the founders of artificial intelligence company DeepMind to break free from Google’s grip. These included promises of $1 billion in funding from prominent tech investor Reid Hoffman and secret plans to build fast financial trading algorithms to provide the financial backing necessary for the great escape.
The book “The Infinite Machine” and revealing what was hidden
Authored by journalist Sebastian Mallaby and published by Penguin Press on March 31, “The Infinite Machine: Demis Hassabis, DeepMind, and the Quest for Artificial General Intelligence” is based on three years of exceptional and unprecedented access to Demis Hassabis and his inner circle. Among the prominent revelations in its 480 pages is the complete inside story of what was known as “Project Mario,” the codename for the prolonged campaign waged by DeepMind’s leadership to gain complete legal and operational independence from its parent company.
These details expand the vision surrounding the independence pursuits previously outlined by major newspapers in 2021. However, according to Mallaby’s account, Demis Hassabis and co-founder Mustafa Suleyman pursued multiple complex paths to split from Google after growing concerned that the tech giant might one day misuse their advanced AGI research.
$1 billion and trading algorithms: The escape plan
According to the book, Reid Hoffman, LinkedIn co-founder and prominent Silicon Valley investor, secretly pledged $1 billion to fund the separation process. The goal was to structure the new entity as a “global interest company,” a non-profit-like structure bound by a charter serving humanity’s public interest rather than pursuing pure commercial profit.
Demis Hassabis‘s ambitions did not stop there; he quietly formed a secret, specialized team to develop high-frequency financial trading algorithms in stock markets, hoping these algorithms would generate sufficient financial independence to make breaking away from Google’s empire viable in practice without resorting to conditional investments.
The decisive moment and hitting the Google wall
In 2017, during a company organizational retreat in Scotland, DeepMind’s leadership confidently announced to employees that the lab would soon transform into a “global interest company,” a term carefully chosen to reflect the global and consequential impacts of the AI technologies they were developing. However, this bold announcement was later rolled back, as the book reveals that Google CEO Sundar Pichai never approved the proposed organizational structure.
The tug-of-war continued until April 2021, when Demis Hassabis frankly informed his employees that all negotiations regarding separation and independence from the parent company had completely and irrevocably ended, initiating a new phase of coexistence and integration.
The paradox of staying: How constraints became a strength
Author Sebastian Mallaby, known for his prestigious financial biographies, argues that the failed independence attempt turned out to be a major paradox; it ultimately became DeepMind’s greatest asset. Remaining under Google’s umbrella grants the lab unlimited access to unmatched supercomputing resources, alongside billions of dollars in stable annual funding.
This secure status spared them the intense pressures of rapid profit generation and commercialization—pressures weighing heavily on today’s independent market competitors like OpenAI and Anthropic, who are forced to continuously raise billions in funding to keep pace with developing their large language models.
Continuous rise and global successes
Since the merger of Google Brain and DeepMind teams into a single unit named Google DeepMind in 2023, Demis Hassabis has risen to lead all AI efforts at the company, recording consecutive successes most notably the launch of the advanced Gemini model family across Google’s products. Furthermore, his scientific career was crowned by winning the 2024 Nobel Prize in Chemistry thanks to the AlphaFold system, which revolutionized protein structure prediction.
This important book arrives at a time when the AI industry faces increasing questions about the sustainability of its business models. Mallaby summarized the journey of this brilliant scientist by stating that Demis Hassabis “dedicated his life to creating a technology he believes has the power to destroy the world”—a paradox making him, in the author’s narrative, the most important, influential, complex, and contradictory figure in the age of AI.
Frequently Asked Questions
Who is Demis Hassabis?
Demis Hassabis is a computer scientist and AI researcher, and the CEO and co-founder of Google DeepMind. He received the 2024 Nobel Prize in Chemistry for developing the AlphaFold system that predicts three-dimensional protein structures.
What is “Project Mario” mentioned in Sebastian Mallaby’s book?
Project Mario is the codename for a secret, long-term campaign led by DeepMind management, headed by Demis Hassabis and Mustafa Suleyman, aiming to gain legal and operational independence from Google and transform it into a non-profit entity.
Why did DeepMind’s attempt to separate from Google fail?
The attempt failed due to rejection by Google’s senior management, specifically CEO Sundar Pichai, who did not approve the proposed organizational structure called a “global interest company.” By 2021, all negotiations regarding this matter were halted.
How did DeepMind benefit from remaining within Google?
Despite the failed independence efforts, DeepMind massively benefited from the huge computing resources and stable annual funding provided by Google, shielding it from the fundraising pressures and quick profit demands troubling competing companies in the AI sector.
