OPINION: Sam Altman Defends AI’s Colossal Appetite and Mocks Musk’s Space Project

A confident, almost philosophical defense

What makes Altman’s position interesting—and unsettling—is that he doesn’t deny anything. He doesn’t downplay his industry’s carbon footprint. He doesn’t promise carbon offsets that won’t change a thing. He says, in essence: yes, it’s a lot, and it’s necessary. It’s a rare display of candor in an industry accustomed to grandiose announcements about sustainability. And this candor deserves to be heard before it is condemned. Altman’s central argument is that general artificial intelligence—true AGI, AI capable of reasoning like a human—will represent such a fundamental breakthrough in the history of the human race that the cost of building it, however high it may be, will be far outweighed by the benefits it will bring. It’s a bet on the future. A huge bet—perhaps an irresponsible one—but one that millions of people are making alongside him every day by using his tools.

Where Altman loses credibility is when he refuses to clearly articulate the mechanisms of accountability. If the resources consumed today are an investment in tomorrow, who pays the bill if tomorrow never comes? Who is accountable for the accumulated carbon emissions if AGI remains a perpetually postponed goal? Who bears the costs of water consumption if AI’s promises aren’t fulfilled by the announced deadlines? These questions remain unanswered. And it is precisely this silence that renders Altman’s defense insufficient—if not entirely false.

The privilege of those who do not suffer the consequences

Sam Altman lives in a world where the consequences of his industry’s resource consumption do not directly affect him. He does not live near a data center that draws water from his aquifer. He does not work in a lithium mine in Chile. He does not experience power outages in regions where electrical grids are strained by the growing demand from AI servers. This is not an ad hominem attack—it is an observation about the structure of power in the tech industry. Those who make decisions about resource consumption are rarely the ones who bear the brunt of the effects. And this fundamental asymmetry should be at the heart of any debate about the sustainability of AI.

I’m not asking Altman to stop building AI. I’m asking him to stop portraying himself as the only adult in the room when he refuses to be accountable to those who pay the real price for his vision. The grandeur of an ambition does not exempt one from responsibility.

The idea that drew laughter—but might deserve better

And then there’s Elon Musk. And his idea to put data centers in space. When Sam Altman laughed at it, his reaction was understandable. The image is striking: servers in orbit, powered by the unlimited solar energy of space, cooled by the vacuum of space, free from Earth’s constraints. It sounds like science fiction. It sounds like the promise of a man who’s used to promising the impossible and delivering half of it. But before we laugh too hard, we need to take the idea seriously. Because some of the most pressing problems facing the data center industry—heat, water consumption, land availability, and proximity to power grids—simply don’t exist in space.

Solar energy in orbit is indeed unlimited and constantly available, without the intermittency that complicates its use on Earth. The cold of space is a perfect natural cooler. And SpaceX, Musk’s company, has demonstrated that it can launch objects into orbit at decreasing costs. The project isn’t as absurd as it seems at first glance. It’s premature, certainly. It’s expensive, absolutely. It raises questions about latency, maintenance, and security to which no one yet has solid answers. But to dismiss it with a laugh might mean missing out on a direction the industry could seriously explore in the coming decades.

The Real Reason Altman Is Laughing

Let’s be clear about one thing: if Sam Altman is laughing at Musk’s idea, it’s not just because it’s technically premature. It’s also because Musk is now his competitor. Ever since Musk founded xAI and launched Grok, the two men have been competing head-to-head in the generative AI market. What appears to be a philosophical disagreement also has a very concrete strategic dimension. Altman has no interest in endorsing a direction that Musk might explore. And Musk, for his part, has no interest in endorsing the model of massive consumption of Earth’s resources that OpenAI embodies. Their public positions are colored by their private competition. It would be naive to ignore this.

Both men are right on some points and wrong on others. But what strikes me is that neither of them is truly speaking on behalf of those who have no voice in this debate: the communities that will live next to the data centers; the workers who will mine the minerals; the countries that will bear the externalities of their vision. These people have no one to laugh on their behalf.

Two Philosophies on the Future of Technology

The clash between Sam Altman and Elon Musk is nothing new. It dates back to the very origins of OpenAI, which the two men co-founded before Musk left the board of directors in 2018, citing conflicts of interest. Since then, their paths have diverged dramatically. Altman has transformed OpenAI into a commercial enterprise backed by Microsoft, with a valuation now exceeding $300 billion. Musk has launched xAI, funded by his own resources and investors, with the stated ambition of creating an AI that is as curious as possible and less constrained by safeguards he considers too restrictive. These two paths embody two philosophies. One says: AI must be developed cautiously, with safeguards, in the service of a humanistic mission. The other says: AI must be unleashed, accelerated, and pushed to its limits without preconceived ethical considerations holding back its potential.

These philosophies are not abstract. They have real consequences for how systems are trained, for the data that is used, and for the biases that are—or are not—corrected. They determine who decides what AI can and cannot do. And in a world where these systems increasingly influence economic, medical, judicial, and political decisions, these philosophical choices affect each and every one of us.

A Personal Rift with Industrial Consequences

The rift between Altman and Musk also has a personal dimension that both men struggle to hide. Musk sued OpenAI, accusing the company of betraying its original nonprofit mission. Altman responded with calculated detachment, generally refusing to engage on an emotional level. But when Altman laughs at the idea of space-based data centers, he crosses a line. He is no longer maintaining polite neutrality. He chooses mockery. And public mockery between rivals of this stature is never trivial. It is a declaration of contempt. And contempt, in business as in politics, always ends up costing the one who expresses it.

There is something deeply human—and therefore deeply revealing—in the fact that two of the most powerful men in the tech world sometimes behave like two teenagers arguing over whose idea is better during recess. What’s at stake is the direction of humanity. The form, at times, is that of a Twitter storm. The contrast astounds me every time.

Nuclear, Renewables, and the Race for Megawatts

To understand why Altman’s argument deserves serious consideration, one must grasp the scale of the energy challenge facing the AI industry. Even the most conservative projections estimate that electricity demand from data centers worldwide could reach 1,000 terawatt-hours per year by the end of the decade—about 4% of total global electricity consumption. This figure is constantly being revised upward. To meet this demand, major tech companies are turning to every available source. Microsoft has signed an agreement to reopen the Three Mile Island nuclear power plant in Pennsylvania. Google is investing in nuclear fusion and small modular reactors. Amazon is purchasing wind and solar farms at an unprecedented rate. The AI industry is reshaping the global energy sector. And this transformation is not without risk.

The rush for energy is creating new competition between industrial and residential needs. In some regions of the United States, grid operators are warning that demand from data centers threatens grid reliability for ordinary consumers. In Ireland, data centers now consume such a large share of the country’s electricity that regulators have begun imposing moratoriums on new construction. What Altman is defending when he justifies his industry’s energy consumption is therefore also a certain vision of the hierarchy of needs: AI comes first. Everything else comes second. This hierarchy deserves to be questioned democratically.

Who Pays the Bill That AI Giants Don’t Pay

There is an economic reality that is rarely mentioned in debates about the sustainability of AI: unaccounted-for externalities. Companies that build data centers pay for the electricity they consume. But they do not always pay the cost of upgrading the power grids that their demand makes necessary—these costs are often passed on to all consumers in the form of rate hikes. They do not pay the cost of water resource depletion in the regions they are drying out. They do not pay the health costs of the workers who extract the materials needed for their infrastructure. And they certainly do not pay the climate cost of their greenhouse gas emissions, except through voluntary and often opaque mechanisms. Altman defends his industry’s consumption. But in reality, he is defending a model in which part of the actual costs are socialized while the profits are privatized. It is a very comfortable arrangement for those on the right side of the equation.

Every time I read that a major tech company has committed to being “carbon neutral” by a certain date, I feel a mix of hope and skepticism. Hope that it’s true. Skepticism that it’s just carefully packaged marketing. Recent history gives me more reasons for skepticism than for hope.

Decisions that affect everyone, made by very few people

The debate between Altman and Musk on AI’s resource consumption is fascinating for what it reveals about the governance of this industry. These colossal decisions—how much energy to consume, where to locate data centers, which power sources to prioritize—are made by a very small number of people. Not by parliaments, not by democratically elected regulatory agencies, not by the affected communities. By CEOs, boards of directors, and investment funds. This is an extraordinary form of power exercised with relative opacity. And in a democracy, this should raise urgent questions about the legitimacy of these decisions.

When Sam Altman defends the resource consumption of artificial intelligence, he speaks on behalf of a private vision of the common good. His conviction that AI will benefit humanity is sincere—I don’t doubt it for a moment. But the sincerity of a conviction does not confer democratic legitimacy upon it. Do Europeans, Americans, and residents of the Global South—who bear the externalities of this industry—have a say in this vision? For now, the answer is no. And that is the most serious structural problem posed by this exchange between two billionaires who are debating the future of global energy as if it were their personal business.

Regulation: The Missing Element in the Debate

Remarkably, in public discussions among AI industry figures about resource consumption, the word “regulation” rarely comes up. The talk is of technological solutions, improved energy efficiency, and new power sources. There is no mention of binding frameworks that would require companies to account for their consumption, limit it, and effectively offset their emissions. The European Union, with its AI Act, has taken a first step toward governing artificial intelligence in terms of rights and security. But on this specific environmental issue, regulatory tools remain largely insufficient. And while lawmakers grapple with complex concepts, data centers continue to spring up like mushrooms after rain.

We’re being asked to trust the AI industry to regulate itself on environmental issues. This is the same industry that told us social media would be good for democracy. That algorithms would be neutral. That digital surveillance would protect our privacy. I wonder how many times we’ll have to be disappointed before we demand clear rules.

When Certainty Replaces Caution

Sam Altman is a man who speaks with impressive certainty. He is convinced that AGI is coming. He is convinced that it will transform humanity for the better. He is convinced that the resources spent on building it are a justified investment. This certainty is a strength when it comes to raising funds, attracting talent, and winning over partners. It is also a potentially serious intellectual weakness. The history of major technological gambles is littered with certainties that turned out to be wrong. Advocates of civilian nuclear power in the 1950s were certain that electricity would become “too cheap to measure.” The architects of the social media revolution were certain that global connectivity would strengthen democracy. Certainty is no guarantee. It is sometimes a screen behind which blind spots thrive.

Altman’s blind spot might be the assumption that the benefits of AI will be distributed fairly enough to justify concentrating the costs on the most vulnerable. Musk’s blind spot might be assuming that his vision of maximally unrestricted AI will serve humanity better than a carefully constrained AI. Neither of them seems willing to seriously consider that their certainty might be mistaken. And that is precisely where the danger lies.

The Problem with a Single Narrative

What strikes me about this public exchange between Altman and Musk is the absence of alternative voices in the debate over the future of AI and its resources. There are researchers, activists, economists, and environmental experts raising different and more complex questions. But the public debate is dominated by the few men with the most money at stake. Their ability to capture media attention, to appear at major conferences, to be quoted first in articles like this one—that, too, is a form of power. A power that shapes the narrative. And a dominant narrative always ends up influencing policy. That is why it is crucial not to leave this debate solely to the protagonists who benefit from it most directly.

I’m thinking of all those experts from the Global South, those environmental researchers, those community activists whose analyses of the risks AI poses to their regions go unpublished. They exist. They speak out. But their voices are drowned out by the media noise generated by the exchanges between California billionaires. This imbalance is itself a governance issue.

Arrogance as a Default Posture

There is one attitude that characterizes many of the leading figures in the artificial intelligence industry: quiet arrogance. Not the loud, provocative arrogance of someone like Musk on social media. It’s the calm, well-dressed, conference-speaker-style arrogance of someone who is convinced they know the answer and don’t really need to listen to the questions. Altman is a particularly sophisticated embodiment of this. When he defends his industry’s consumption of resources, he isn’t just defending an economic position. He’s signaling a hierarchy: our vision is more important than your concerns. Our path is more necessary than your comfort. Our certainty about the future deserves more trust than your skepticism about the present. It’s a stance that can lead to technological masterpieces—and to environmental disasters. Often both at the same time.

A healthy democratic debate on AI requires that this stance be constantly challenged. Not to stifle innovation—AI has real and immense potential. But to ensure that innovation does not come at the expense of those who have had no say in the choices that affect them. That is the role of the press, researchers, regulators, and also columnists like me: to state clearly that Altman’s defense is insufficient, even if it isn’t entirely wrong. To say that Musk’s laughter is telling, even if his idea isn’t entirely absurd. To say that this debate deserves more than two billionaires mocking each other.

The pivotal moment no one wants to name

We are at a pivotal moment. Artificial intelligence is establishing itself as a fundamental infrastructure of our societies, just like electricity or the internet. And like any major infrastructure, it will shape power dynamics, create dependencies, and produce winners and losers. The decisions being made now—on energy, governance, and the distribution of benefits and costs—will have effects that will be felt for decades to come. It’s not too early to be concerned. It’s not premature to demand accountability. And it’s not naive to think that we deserve better than to watch two giants fight over the direction of our shared future.

I am not against AI. I am in favor of AI that answers to someone other than its creators. For AI whose costs are shared as equitably as its promised benefits. For a debate on its future that includes voices other than those of the people who profit from it the most. This may be a naïve position. But it is the only one I can honestly defend.

Efficiency, frugality, redistribution: the paths not taken

The debate as framed by Altman and Musk is a false dichotomy. On one side, the massive consumption of Earth’s resources. On the other, servers in space. But there are alternatives that neither of them takes seriously enough. The energy efficiency of AI models has made remarkable progress in recent years. Smaller, more targeted models consume a fraction of the energy used by large, general-purpose models for specific tasks. Research into alternative architectures—neuromorphic chips, photonic chips, and hybrid symbolic AI approaches—could significantly reduce energy consumption if it received investment comparable to that allocated to large language models. Frugal AI is a legitimate field of research with promising results. But it doesn’t make the headlines. That’s because it doesn’t generate the same market valuations. And in an industry driven by investment funds, market valuations are often more persuasive than environmental imperatives.

There is also the issue of geographic redistribution. Data centers could be located in regions with access to abundant and inexpensive renewable energy, where water stress is minimal, and where the local economy would benefit from skilled jobs. Scandinavia, with its hydroelectric power and naturally cold temperatures, is already a popular destination. But other regions—in Canada, North Africa, and parts of Latin America—could host AI infrastructure under more sustainable conditions. This conversation is possible. It’s simply not a priority for those who prefer to build where power grids are already in place and permits are easier to obtain.

Digital Sobriety as a Real Option

And then there’s a concept the industry fears almost as much as regulation: digital frugality. The idea that certain AI applications may not be worth the resources they consume. That generating an artistic image with DALL-E or Midjourney has an environmental cost that deserves to be explicitly disclosed to the user, much like calories on a food label. That not all uses of AI are equally legitimate given their environmental footprint. This is a difficult conversation to have in an industry that has made unlimited growth its religion. But it is necessary. And it will be inevitable, as the environmental impacts of AI become impossible for the general public to ignore.

A day will come—perhaps sooner than we think—when people will ask on their electricity bills what portion goes to the AI data centers that sold them a subscription. On that day, Altman’s defense will be judged by how much it cost each household. And I bet it will seem less convincing.

A dispute that reveals our own contradictions

It would be convenient to conclude that all of this is the problem of distant billionaires, disconnected from everyday reality. But we’re all involved. Every time we use ChatGPT to draft an email, Grok to analyze a news feed, or Gemini to generate a presentation, we’re fueling the demand that justifies the consumption Altman advocates. We are the consumers who give these companies the commercial legitimacy to keep building bigger and bigger. Our digital behavior is a form of voting. And so far, we’re voting overwhelmingly for more AI, faster—without asking ourselves too many questions about what it costs the planet.

This complicity isn’t a reason to feel excessive guilt—we don’t have clear information on the environmental costs of our digital habits, largely because the industry doesn’t communicate them in a way that’s easy to understand. But it is a reason to be more demanding. To demand transparency on actual energy consumption. To demand that regulators take action. To demand that political candidates take clear stances on the governance of AI and its resources. This debate does not belong to Altman and Musk. It belongs to all of us.

What the Coming Years Will Clarify

In the coming years, several questions will be answered, making this debate even more crucial. The promise of AGI will—or will not—become a reality. The climate goals of major tech companies will be judged by their actual results—not by their commitments. Regulators in Europe, the United States, and elsewhere will either succeed or fail to implement binding frameworks for AI sustainability. And consumers—that is, us—will decide whether we want to continue using services whose environmental costs are hidden, or whether we demand greater transparency. These questions remain unanswered. But the way Altman and Musk are positioning themselves today offers valuable clues as to what we can expect.

I’ve been covering this industry long enough to know that companies rarely change their behavior simply because of arguments. They change when the rules change. When the costs they externalize are passed back to them. When consumers withdraw their trust. We’re still a long way from that tipping point. But I believe it will come. And I hope we don’t wait for a catastrophe to reach it.

Call for a debate that lives up to the stakes

What we need is not a referee between Altman and Musk. What we need is to raise the level of the debate. To take it out of the hands of commercial actors alone and open it up to truly democratic deliberation. That means serious congressional hearings, with independent experts—not just industry lobbyists. It means media outlets covering AI’s resource consumption with the same intensity as they cover its mind-boggling capabilities. It means universities training specialists capable of explaining the environmental impacts of AI in terms accessible to the general public. And it means, perhaps above all, informed citizens who understand that the question is not “are you for or against AI,” but “what conditions do you set for its development.”

AI will continue to develop. Its consumption of resources will continue to grow, at least in the short term. But the pace, the conditions, the mechanisms of accountability—all of this remains to be defined. And these definitions belong to us collectively. It is not a utopian ideal to think this way. It is simply the definition of democracy applied to one of the most complex challenges of our time.

An open-ended conclusion because the debate has only just begun

The exchange between Sam Altman and Elon Musk on AI resources and space-based data centers will not be the last of its kind. It will be followed by many others, as the stakes intensify and the choices become more pressing. What matters is that we—citizens, readers, users—be present in this debate. Let’s not let it unfold without us. Let’s ask our own questions, set our own standards, and vote with full knowledge of the facts for representatives who will make sound decisions regarding the governance of this technology. The future of artificial intelligence will not be shaped solely by Altman, Musk, or the few hundred people who run the industry’s major companies. It will also be shaped by what we, collectively, accept or reject. And that responsibility, however uncomfortable it may be, is ours.

I’ll leave you with this thought: Thirty years from now, our children will look back on this period when AI took off. They’ll look at the choices that were made regarding energy, resources, and governance. And they’ll judge us not on our intentions—everyone had good intentions—but on our courage to hold people accountable. Let’s be courageous.

The Lesson from This Exchange, Beyond the Spectacle

Ultimately, what the exchange between Sam Altman and Elon Musk on AI’s resource consumption reveals is an uncomfortable truth: the most important decisions regarding the technological future of our species are made by an extraordinarily small number of people, who are accountable to no one but their shareholders. Altman is right on some points—AI has real benefits, and its development requires significant resources. Musk raises valid questions about the limits of Earth-based solutions—even if his answers may seem premature. But neither of them speaks on behalf of the millions of people who will live with the consequences of their choices.

This lack of representation is the fundamental problem. And Altman’s laughter at Musk’s idea, as entertaining as it may be for the media, should not make us forget that what is at stake in these exchanges is our shared future—a future that is, unfortunately, being treated as a private matter.

What We Must Demand Now

If this article has convinced you of just one thing, let it be this: demand transparency. Demand that AI companies publish clear and verifiable data on their energy, water, and material consumption. Demand that your elected representatives raise these questions in parliamentary committees. Demand that the media you follow cover this topic with the depth it deserves. Artificial intelligence will be one of the most transformative technologies of this century. It deserves a debate commensurate with its importance—a debate that goes beyond the squabbles among billionaires to reach each and every one of us, where we live, where we breathe, and where we pay our electricity bills.

The question isn’t whether AI will change the world. It’s already changing it. The question is who decides how. For now, the answer is: not you. Not me. But that can change. If we decide to.

Signed, Jacques Pj Provost