It was once a conspicuously intimate and almost reverential friendship, the kind Silicon Valley likes to mythologise until it collapses into litigation and sworn hostility.
A decade ago, Sam Altman visited Elon Musk’s Tesla factory and described him with near-admiring reverence as “an extremely talented person,” adding that “if he is into rockets he would remember from nut to end product how to complete that very complicated machine”.
That sentiment now reads like a relic from a different era of Silicon Valley, when founders still imagined themselves as part of a shared civilisational project rather than rivals in a struggle for dominance over Artificial Intelligence (AI).
Today, they operate as adversarial states in permanent breakdown, with one lawsuit following another and former alliances recast as evidentiary ammunition.
Last Monday (18) in Oakland, California, a jury delivered a decisive win for OpenAI CEO Altman in Musk’s $ 150 billion lawsuit – the amount often described as one of the largest damages claims in tech litigation history. The court ruled that Musk had waited too long to bring the case, dismissing it on statute-of-limitations grounds rather than on its merits.
What has followed in Musk vs. Altman is not merely a dispute over corporate structure, but a public autopsy of the myth that AI would be built by disinterested agents. The courtroom record reads less like a contractual disagreement and more like an extended confession of how power actually accrues in frontier technology: through shifting alliances, retrospective moralisation, and the constant rewriting of founding intent.
Musk’s central allegation is deceptively simple. He says OpenAI “stole a charity”. Musk’s claim is that capital was solicited under the guise of altruism and then converted into a vehicle for private enrichment. The legal system, however, has refused to even reach that question. A jury in Oakland disposed of the matter in under two hours, not because the facts were irrelevant, but because Musk had “failed to file his lawsuit within the time frame required by law”.
Dual narratives
Yet the deeper argument Musk has been attempting to make is not legal at all. It is metaphysical: that a promise was made about the governance of intelligence itself, and that promise was broken.
In his telling, OpenAI was supposed to remain “a non-profit Artificial Intelligence research company” dedicated to advancing “digital intelligence in a way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return”. That language now reads like an artefact from a prelapsarian Silicon Valley, before trillion-dollar model training runs turned ‘humanity’ into a branding device rather than a constraint.
Altman’s counter-narrative is more prosaic, and more economically credible. OpenAI, he argues, had no viable route to survival without embracing capital markets. The shift towards a for-profit subsidiary, capped returns, and later public benefit corporation status is framed not as betrayal but adaptation. The inconvenient truth underpinning this position is that AI at frontier scale is structurally incompatible with donation-based funding. As one account of the restructuring noted, “it is impossible to court investors without the promise of returns”.
Meanwhile, Musk’s own history within OpenAI complicates any attempt to position him as a purely betrayed benefactor. Internal accounts presented in court describe him attempting to fold OpenAI into Tesla, effectively converting a nominally independent research lab into a corporate division under his control.
Greg Brockman recalled that Musk “wanted us to feel indebted to him in some way”. In another exchange, Musk told Co-Founders that without “a dramatic change in execution and resources,” OpenAI’s probability of competing with Google DeepMind was “0%. Not 1%. I wish it were otherwise”.
A question of control
The disagreement was never about whether to win, but about who would hold the steering wheel while winning. That unresolved question of control animates some of the more striking testimony.
Brockman described a 2017 meeting in which Musk became visibly enraged during discussions about governance, saying: “I actually thought he was going to hit me, physically attack me.” Musk, in turn, presented himself as the only actor sufficiently serious to manage the risk, proposing structures in which he would “unequivocally have initial control of the company” before authority would later be delegated.
Altman’s courtroom posture has been more evasive, but no less revealing. Under cross-examination he conceded, when asked whether he always tells the truth: “I’m sure there is some time in my life when I have not.” That banal honesty is precisely what makes the broader spectacle uncomfortable. It is not that Altman claims moral perfection; it is that the system in which he operates does not reward it.
Closure vs. abdication
The litigation has therefore become a referendum on narrative authority in the AI industry. Musk’s lawyers have attempted to recast OpenAI as a misappropriated philanthropic entity, arguing that early donations were made under the assumption of non-profit permanence.
OpenAI’s defence has been procedural and structural: even if intentions were as Musk describes, the legal and commercial environment evolved in ways that made those intentions obsolete. As one legal argument summarised, the nonprofit’s structure was always going to require “more money than they could ever hope to raise through donations alone”.
What is striking is how little space there is in this dispute for a stable concept of mission. Every invocation of ‘benefit humanity’ is immediately followed by a financing constraint; every appeal to safety is countered by competitive urgency. In this sense, the courtroom has not clarified the meaning of OpenAI’s founding charter so much as demonstrated its semantic elasticity.
The most corrosive effect of the trial, however, is not doctrinal but institutional. It has exposed how governance of advanced AI is already effectively being negotiated through litigation rather than legislation. The question of whether OpenAI is “a nonprofit dedicated to humanity or a corporation dedicated to its shareholders,” as one account put it, is not being resolved by regulators or democratic oversight, but by procedural arguments about timing and standing.
This is not incidental. The jury’s decision did not engage with the merits at all. It found only that Musk had waited too long to bring his claim. In legal terms, that is closure. In political terms, it is abdication.
The moral language of technology
The fact is that the Musk-Altman conflict persists beyond the courtroom. It is not merely a disagreement between two Founders. It is a struggle over who gets to define the moral language of a technology that neither side fully controls. Musk’s accusation of theft and Altman’s defence of necessity are both, in different ways, retrospective rationalisations of a system that was always going to converge on scale, capital concentration, and corporate form.
In one of the more revealing early emails, Altman wrote that building OpenAI required something like a “Manhattan Project for AI”. The historical analogy is often invoked as shorthand for urgency, but it carries a darker implication.
The Manhattan Project was not governed by consensus, philanthropy, or public accountability in any meaningful sense. It was governed by a presumption of fear and a drive for strategic superiority rather than consensus or public accountability. Attempting to reproduce that model inside venture-backed private institutions was always going to generate contradiction.
Even Musk’s most theatrical claims – such as describing Microsoft’s investment as turning OpenAI into something “essentially captured” – reflect not only critique but recognition of how capital dominance operates. His own later ventures in AI, framed as corrective alternatives, reproduce the same structural logic.
What the litigation has done, then, is not resolve a dispute but formalise a recognition that the governance of AI is already post-legal in practice. Courts can decide timing. They can allocate damages. They can enforce procedural limits. What they cannot do is reconstruct a coherent moral foundation for an industry that has already outgrown the institutional language used to describe it.
The result is a public record in which everyone appears simultaneously sincere and self-interested, principled and opportunistic. Musk describes betrayal. Altman describes necessity. Brockman describes sacrifice through “blood, sweat, and tears”. None of these formulations is fully persuasive on its own, yet together they form the grammar of contemporary technological power. Thus, the tale of our future is unfolding.
(The writer is an author based in Colombo)
(The views and opinions expressed in this article are those of the writer and do not necessarily reflect the official position of this publication)