Project Stargate Isn’t What You Think It Is.
It is insidiously dangerous.
The $500 billion Project Stargate is an utterly unprecedented investment in the tech industry. But, to anyone keeping an eye on the reality of the AI industry, it makes zero sense. I have already covered the scaling issues, data issues, and OpenAI’s profit issue with Project Stargate in a previous article. But something didn’t sit right with me. If Project Stargate can’t further the AI industry or even help OpenAI make a profit, why are private investment firms pouring hundreds of billions of dollars into it? When you go digging, Project Stargate looks less like a tech investment and more like an insidious plot to manipulate the market. Let me explain.
Project Stargate is a joint venture between OpenAI, Oracle, investment group SoftBank, and the tech investment arm of the Abu Dhabi oil company ADNOC, MGX. The project is intended to build AI infrastructure, like data centres, in the US on a scale never seen before, with OpenAI being the exclusive customer. The investment comes mainly from SoftBank and MGX, with an initial investment of $100 billion and $400 billion extra promised over the next four years. OpenAI and Oracle will design and build the infrastructure, most likely using OpenAI’s in-house designed chips.
How this joint venture will function in terms of how it will make a profit is yet to be announced. Will OpenAI have to shoulder the entire cost of operation? Or will OpenAI not be expected to pay for everything, meaning Stargate will be run as a loss-making tax offset for these companies? We don’t know.
But, even if Stargate makes zero profit and OpenAI fails to make a better AI or even a profit from Stargate, everyone involved in this joint venture will make heaps of money!
More details show that Oracle isn’t much of an investor and more of a contractor to build and maintain these data centres. So, it will likely make a profit straight away.
What about MGX? Well, they will make money on the energy side.
Before Stargate was announced, reports showed that the tech industry was set to spend a trillion dollars on AI infrastructure by 2030 and that this infrastructure would consume at least 88 TWh of power annually by 2030. Using this as a base, we can estimate that Project Stargate will consume around 44 TWh of power per year.
Now, while the US had record growth of renewable energy in 2024, that is about to stop. Trump is set to halt federal support for clean power. But even before that was announced, many sources highlighted that the rapid increase in AI power demand was slowing the energy transition, as fossil fuel-based energy infrastructure was being kept online or even having its lifespan expanded specifically to be a cheap and reliable power source for AI infrastructure.
So, this 44 TWh per year power demand will likely be supplied mostly mostly fossil fuels, especially with Trump’s pro-oil stance. While ADNOC doesn’t export directly to the US, the extra demand will increase the global fossil fuel demand. ADNOC also owns significant US fossil fuel infrastructure, such as pipelines. On top of that, as fossil fuel production is set to decrease over the coming decades, fossil fuel prices and profits are set to dramatically increase. As such, if you run the numbers (roughly $100+ profit per barrel, 44 TWh of additional energy demand), ADNOC could be set to make multiple billions of dollars in profit per year from the increased energy demand of Stargate. They will be able to recoup their investment in Stargate within a decade or two, which is far quicker than any fossil fuel infrastructure investment.
However, Stargate has already said it plans to build and use nuclear power. Well, wouldn’t you know it, ADNOC technically owns an Abu Dhabi atomic company that is set to design, build, and supply nuclear reactors and produce nuclear fuel. Again, supplying these to the US and Stargate will make ADNOC and, in turn, MGX billions of dollars per year.
And SoftBank has a similar racket with Stargate.
SoftBank owns 90% of ARM. OpenAI’s custom AI chips are set to use ARM technology and will be produced by AMD at the new Arizona branch of TSMC. Conveniently, Trump is set to put a 100% tariff on all computer chips shipped to the US from Taiwan.
Now, 90% of computer chips, and effectively all AI chips, come from Taiwan, and this Arizona plant is one of the only semiconductor plants in the US. As a result, Trump’s tariffs will squeeze American tech customers as the supply is throttled. However, Stargate will exacerbate this issue by taking a significant portion of that supply away from the public. Naturally, ARM and AMD can price gouge the public and increase prices by 100% or more.
Through this, SoftBank could easily recoup its investment in Project Stargate in a few years, as ARM and the other tech infrastructure companies it owns increase their profits dramatically.
Okay, so what about OpenAI? How can they make money?
Thanks to DeepSeek, OpenAI now can’t make money selling AI—not that they were before DeepSeek came around, as they are still operating at a loss of multiple billions of dollars a year. Having to pay for the operation of Stargate will only exacerbate this dire situation. In short, as a business, they simply have no route to financial sustainability, let alone profitability.
Despite that, OpenAI is valued at over $157 billion. A valuation so distant from its business fundamentals that it makes a mockery of the concept of an economy. Why? This valuation is entirely based on pure, unfounded speculation propped up and fuelled by the sheer amount invested into OpenAI by giants like Microsoft. As such, Stargate could raise OpenAI’s value many times over just by virtue of the amount of money being effectively poured into it. It doesn’t matter if OpenAI can’t produce a better AI with Stargate; their value will increase because of this investment. Just as it did when, at the end of last year, Microsoft, Nvidia, SoftBank, MGX, and Thrive Capital had to bail out OpenAI at the end of 2024 with a further $6.6 billion to fend off bankruptcy. This increased its valuation from $80 billion to $157 billion, even though the AIs it had developed at that time were not better than the previous generation, and they are still light years away from profitability!
So, thanks to Stargate, OpenAI gets to become wildly valuable. Not only will this make OpenAI CEO Sam Altman grotesquely wealthy, as he is set to receive a 7% stake in the company, but it will also make SoftBank and MGX a huge sum of money, as they are both now significant shareholders in OpenAI.
Project Stargate has nothing to do with furthering AI. If OpenAI wants to be serious, it needs to focus on making smaller, cheaper-to-build models that compete with those from China. It also needs to focus on making AIs that can augment human work rather than shoddily attempt to replace it, as that is by far the best use case of AI. None of these directions require anything like Stargate.
Personally, I can’t help but come to the conclusion that Stargate is all smoke and mirrors, a trick, a sleight of hand that allows these companies to make vast amounts of money by manipulating markets through the aesthetics of innovation while functionally delivering nothing of worth. Such a scheme would be impossible in a functional economy, as it would be seen as falsely stealing money from those who do produce value, like you and me. But that isn’t the world we live in anymore. We are so separated from reality in every aspect of our lives that the largest investment companies in the world can hoodwink us like this.
Thanks for reading! Content like this doesn’t happen without your support. So, if you want to see more like this, don’t forget to Subscribe and help get the word out by hitting the share button below.
Sources: CIO, QCI, Goldman Sachs, Energy Politics, The Guardian, McKinsey, UJ, Fortune, Oil Price.com, CFO Brew, Will Lockett, TT, Insider Money, The Conversation, Oil Price.com, ADNOC, Statista, AGBI