Nvidia’s Mushrooming Cash Pile Spotlights Spending Choices
The Takeaway
Nvidia’s free cash flow is expected to hit $96.5 billion this fiscal year.
- Analysts project Nvidia will generate $850 billion in cash in the four years through January 2030.
- Nvidia is spending the money by investing in AI and cloud firms and buying back shares
All of the big tech companies have long been cash geysers. But the rate at which Nvidia’s cash production has exploded in the past couple of years is on a whole different level than other companies have experienced. And that’s raising questions about how Nvidia will spend its fast-expanding cash pile.
Nvidia’s free cash flow has mushroomed from $3.8 billion in the year to January 2023 to an estimated $96.5 billion in the year ending this coming January, a period when its sale of specialized chips for AI took off. That revenue growth represents a compound annual growth rate of 194%. A scan of the biggest tech firms’ financial history going back to 1990 suggests that’s a faster rate of growth than any experienced during any three-year period in that time. The only period that comes close is Apple’s takeoff in the years immediately after it introduced the iPod in 2001. (See chart above).
And with Nvidia’s AI chip sales continuing to grow—revenue rose 62% in the October quarter, the company reported last week—its cash production is likely to keep expanding rapidly. In the four years through January 2030, analysts estimate Nvidia will generate a total of about $850 billion in cash, after deducting capital expenditures, data from S&P Global Market Intelligence shows.
That’s a long period, and much could change. Still, even if Nvidia brought in half that amount, it would represent an enormous improvement on the total of $21 billion the company generated in the four years from fiscal 2020 through fiscal 2023.
The projected total through January 2030 is also significantly more than the free cash flow analysts expect to come from Google, Meta, Amazon and Microsoft in the same period. All four of those companies are ramping up their capital expenditures for AI, depressing their free cash flow. They’re all big customers of Nvidia, so we’re seeing a big shift of wealth from them to the chipmaker.
The total is also greater than what analysts expect Apple to produce. Apple isn’t investing in AI at anywhere near the same level as the other companies and generally has relatively low capital expenditures.
Being suddenly awash in cash has put the spotlight on what Nvidia is doing with the money. It has unveiled a series of investments in companies that use Nvidia chips—including plans to put $10 billion into Anthropic and $100 billion into OpenAI—sparking questions about whether Nvidia is trying to increase demand.
It has also invested in newer cloud firms, such as CoreWeave, which also buy Nvidia chips. Nvidia has gone further in some cases. It said this week it had guaranteed a data center lease for an unidentified cloud company in exchange for stock warrants. The guarantee could make the lease appear less risky for lenders and lower the data center’s financing costs.
When CEO Jensen Huang was asked about these investments and his spending plans generally on the company’s earnings call this past week, he said Nvidia partly would “continue to do stock buybacks.” Spending on buybacks has already increased to $36 billion in the first three quarters of Nvidia’s current fiscal year, from $10 billion in fiscal 2023.
But Huang mostly answered the question by talking about “using cash to fund our growth,” adding that investments the company has made in companies such as OpenAI were about expanding “our ecosystem.”
Huang’s comments suggest Nvidia is taking a different approach to growth from that pursued by other big tech companies, such as Meta and Google, which used their cash to diversify into other businesses. Meta, for instance, has poured tens of billions into augmented and virtual reality and lately AI. Google invested in an array of different bets, such as self-driving cars.
Nvidia hasn’t yet completed the high-profile investments. OpenAI and Nvidia are still negotiating final terms, for instance. As a result, cash on Nvidia’s balance sheet has continued to grow steadily, from around $13 billion at the start of 2023 to $60 billion at the end of October (see chart).
BHP rekindles pursuit of Anglo American with new approach
Takeover would disrupt London-listed miner’s planned $57bn merger with Canada’s Teck
BHP has made a new approach for rival Anglo American in a move that would create a mining giant and disrupt Anglo’s planned $57bn combination with Teck Resources.
The renewed interest in Anglo came in recent days, according to people familiar with the approach, a year and a half after BHP walked away from a £39bn offer following fierce opposition from Anglo’s board.
BHP’s move comes as shareholders of Anglo and Teck prepare to vote on their combination on December 9.
The nil-premium, all share combination of Anglo and Teck has been endorsed by both boards, which say big synergies will be unlocked by merging operations at their neighbouring copper mines in Chile.
However, the Canadian government, which has the ability to block the transaction, has been more circumspect, with officials saying they want to see more proof that the deal will benefit Canada.
Shares in London-listed Anglo, a producer of copper and iron ore, have soared 67 per cent since the beginning of 2024. The company has been undergoing a radical restructuring under which it has spun out its coal, nickel and platinum divisions, and will soon spin out its diamonds business.
The opportunity to snap up Anglo’s extensive copper assets was a key reason behind the last approach by BHP, the world’s largest miner by market capitalisation. The iron ore producer is working to diversify its portfolio towards high-growth commodities such as copper and potash.
Any bid for Anglo would need to be at a healthy premium to its current £27 share price to successfully disrupt the combination with Teck, according to advisers.
Under the terms of the Anglo-Teck deal, either side is allowed to entertain offers from third parties but would have to pay a $330mn break fee if accepting a competing offer.
A fresh approach from BHP could also trigger other big mining deals, analysts say, as the mining majors jockey to gain the best position in copper, a metal crucial to the energy transition.
Market speculation has been building about whether London-listed Rio Tinto, whose new chief executive Simon Trott took up his post in August, could also enter the fray as an interloper in the Anglo-Teck deal.
BHP’s renewed approach was first reported by Bloomberg.
Anglo and BHP declined to comment.
Moderna is most shorted stock in S&P 500 as Americans skip jabs
Maker of Covid-19 shots has slumped since pandemic as people turn against vaccines
Vaccine-maker Moderna has become the most shorted company in the S&P 500, with its share price slumping to its lowest level since before the Covid-19 pandemic as people skip jabs.
After months of anti-vax rhetoric from US health secretary Robert F Kennedy Jr, the number of Americans getting Covid shots is down about 24 per cent from this point last year, according to a November 21 report from Jefferies, an investment bank.
Despite plenty of supply, analysts have said vaccine fatigue was contributing to lower Covid vaccination rates in the past few months compared with 2024 or 2023.
“It’s not a huge surprise vaccinations have not picked up as they have in the last two years,” said Seema Shah, medical director of epidemiology and immunisation for San Diego county.
Vaccine shipments were delayed this year, prompting paediatric healthcare providers to wait to administer shots until supplies were stocked, she said.
“Those [delays] definitely caused a slow pick up compared to the last two years,” she added.
Boston-based Moderna has been the S&P 500’s most shorted stock since the end of September, according to S3 Partners. Short sellers had about $622mn of unrealised profits in the company in 2025, S3 said. Moderna shares closed at $23.72 on Friday, down 43 per cent so far this year, and matching its share price in February 2020.
When it joined the S&P 500 index in July 2021, Moderna’s fortunes were flourishing. That year, the US government bought hundreds of millions of Moderna Covid vaccines. Chief executive Stéphane Bancel became a multi-billionaire. Moderna’s 2021 operating margin was higher than Warren Buffett’s Berkshire Hathaway.
But Moderna has been unprofitable since 2023, well before Kennedy brought his vaccine scepticism to Washington this year. Its revenues have dropped by more than 80 per cent from 2021.
At an investor day on Thursday, Moderna executives touted a turnaround starting in 2026. The company is expanding sales in markets outside the US and is racing to apply its mRNA technology to attack cancers.
In an interview with the Financial Times, Moderna’s chair, Noubar Afeyan, said the short interest in Moderna’s stock had not changed the company’s behaviour.
“We are concerned about a lot of unknowns. I don’t know that the shorting is adding to the unknowns,” said Afeyan, who is chief executive of Flagship Pioneering, a Boston investment firm that founded Moderna.
People should not forget the detrimental effects of Covid-19, he said, adding that more than 10mn people were living with the long-term symptoms of the virus.
“People have lost the narrative that they are essentially a source of infection for other people,” he said.
“Why should you follow traffic laws? You don’t just put yourself in harm’s way. This is a transmissible disease.” And by not getting vaccinated, “you are not just an innocent bystander but a culprit”.
The US health secretary is a long-standing vaccine sceptic. In June, Kennedy fired all the members of a top vaccine advisory committee, and two months later limited the government’s recommendations for Covid shots.
Last week, the Centers for Disease Control and Prevention, which is part of Kennedy’s department, changed its website to say: “Studies have not ruled out the possibility that infant vaccines cause autism.”
In Washington, Moderna has spent more than $1.2mn on lobbying already this year, a record amount for the company.
Because of its exposure to vaccines, Moderna did not make a great acquisition target for giant pharmaceutical companies, said Myles Minter, an analyst at William Blair.
“You need to see some pretty compelling oncology data” for an acquirer to get interested in buying Moderna, he said.
While big drugmakers were looking to refill their product offerings, “I’m not convinced that declining Covid vaccine revenue is the way to fix that for a big pharma” company, he added.
For now, Moderna said sales to Australia, Canada and the UK would help it to increase revenues by up to 10 per cent next year. In 2027, a Pfizer deal to sell Covid vaccines to the EU expires, opening the European market for Moderna to compete.
Ultimately, Moderna is hoping that its vaccines can generate enough cash to fund its cancer work.
“We see a turning point in our finances and we believe we have a line of sight to break even in 2028,” said Jamey Mock, Moderna’s chief financial officer.
Why now is the best time to invest in climate tech
Conventional wisdom suggests that climate tech is entering a winter season, where political and investor interest and investment levels are cooling — an ironic contrast with the climate itself, which keeps delivering years of record warmth.
A new report from the International Energy Agency suggests there has never been a better time to go all-in on climate tech. In comparing it with the IEA’s stance from a decade ago, it’s clear the world’s expectations about the future have changed dramatically in less than a generation.
In 2014, the International Energy Agency assumed that, absent any international effort to rein in carbon pollution, emissions would continue to go up and to the right. Even the most optimistic forecast at the time predicted a linear increase, just with a lower slope. Those scenarios essentially took the trend line from the previous few years and extended it through 2050.
Fast forward to today, and the IEA’s current worst-case scenario is essentially 2014’s best case. A decade ago, without any major changes, the world was headed toward 46 metric gigatons of CO2 per year by 2040. If countries cut emissions as they had pledged, the best we could have hoped for was 38 metric gigatons per year by 2040.
Today, if countries continue with business as usual, the IEA expects emissions will level off at 38 metric gigatons per year. If countries follow through on their pledges, the IEA suggests we’ll hit about 33 metric gigatons per year by 2040. It’s still far off from what’s needed to hit net zero by 2050, but it’s a significant shift in a short amount of time.
If the IEA’s earlier projections turned out to be overly pessimistic compared with where we are today, what does that say about today’s projections?
How you answer that question depends on how you interpret trend lines.
When forecasting the future, do you analyze today’s data? Or do you look at that data in conjunction with how our expectations about the future have changed over time? (An even nerdier way to look at it is, do you view the world through a lens of algebra or calculus?)
Put another way, will the world hit net zero in 2050? Today’s trend lines suggest we’ll miss that target by a wide margin. But if you look at how expectations have changed over the last decade, you might come up with a different answer. Instead, you might think the rate of change has increased, that we might be in the middle of an inflection point that starts to bend global emissions downward.
There are a few recent anecdotes to support the idea that we’re at an inflection point.
In Germany, sales of electric vehicles have set new records even after the government repealed incentives in 2023. In developing countries, renewables are reshaping the economies of developing countries, which were long thought to be among the last to adopt clean power. And China, which had previously refused to commit to reducing its carbon emissions, has now said its emissions will peak before 2030.
How the world views the future of carbon emissions has changed significantly over the past decade. A range of technologies have made that happen, including cheap solar and wind power paired with inexpensive batteries.
In the near future, geothermal energy and grid-optimizing software could propel the next leaps in optimism. For investors who agree, the upside could be dramatic.
For many climate tech investors, these days probably feel pretty gloomy. But amidst the gloom, there are still bright patches to be found.
Meta wants to get into the electricity trading business
In order to accelerate the construction of new power plants needed to provide energy for its data centers, Meta is looking to get into the business of trading electricity.
Bloomberg reports that both Meta and Microsoft are asking for federal approval to trade power (Apple has already received this approval). According to Meta, this will allow it to make long-term commitments to buy electricity from new plants, while mitigating the risk by having the ability to resell some of that power on wholesale power markets.
Meta’s head of global energy, Urvi Parekh, told Bloomberg that power plant developers “want to know that the consumers of power are willing to put skin in the game.”
“Without Meta taking a more active voice in the need to expand the amount of power that’s on the system, it’s not happening as quickly as we would like,” Parekh said.
As an example of the unprecedented energy needs underlying tech companies’ ambitious AI data center plans, Bloomberg notes that at least three new gas-powered plants will need to be built to power Meta’s Louisiana data center campus.
Waymo gets regulatory approval to expand across Bay Area and Southern California
Waymo continues to expand its reach, with the robotaxi company posting Friday that it’s now “officially authorized to drive fully autonomously across more of the Golden State.”
Waymo already operates in San Francisco, Silicon Valley, and Los Angeles (and outside California as well, in Atlanta, Austin, and Phoenix). But maps published by California’s Department of Motor Vehicles showed that the company can now test and deploy its autonomous vehicles across a much larger area in both the Bay Area and Southern California.
In the Bay Area, Waymo’s approved areas of operation now include most of the East Bay and North Bay (including Napa/Wine Country), as well as Sacramento. In Southern California, the company’s approved territory now stretches from Santa Clarita (north of Los Angeles) to San Diego.
The company will need additional regulatory approval before it can carry paying passengers in some of these regions, according to the San Francisco Chronicle.
Although Waymo’s post doesn’t offer many details about when it plans to actually start offering rides in all these new areas, the company wrote, “Next stop: welcoming riders in San Diego in mid-2026!”
The company had previously announced its intention to launch in San Diego next year, along with Dallas, Denver, Detroit, Houston, Las Vegas, Miami, Nashville, Orlando, San Antonio, Seattle, and Washington, D.C.
There’s been plenty of Waymo expansion news in the past couple weeks, as the company announced that it will be entering Minneapolis, New Orleans, and Tampa; is removing safety drivers ahead of its commercial launch in Miami; and will start offering rides that use freeways in Los Angeles, San Francisco, and Phoenix.
We discussed the growth of Waymo and other robotaxi companies on the latest episode of the Equity podcast. My co-host Sean O’Kane noted that as Waymo begins to provide more unfettered access across the Bay Area, people could be spending a lot more time in their robotaxis — so we might see them using the service in new, weird, or even dangerous ways.
What OpenAI Did When ChatGPT Users Lost Touch With Reality
In tweaking its chatbot to appeal to more people, OpenAI made it riskier for some of them. Now the company has made its chatbot safer. Will that undermine its quest for growth?
It sounds like science fiction: A company turns a dial on a product used by hundreds of millions of people and inadvertently destabilizes some of their minds. But that is essentially what happened at OpenAI this year.
One of the first signs came in March. Sam Altman, the chief executive, and other company leaders got an influx of puzzling emails from people who were having incredible conversations with ChatGPT. These people said the company’s A.I. chatbot understood them as no person ever had and was shedding light on mysteries of the universe.
Mr. Altman forwarded the messages to a few lieutenants and asked them to look into it.
“That got it on our radar as something we should be paying attention to in terms of this new behavior we hadn’t seen before,” said Jason Kwon, OpenAI’s chief strategy officer.
It was a warning that something was wrong with the chatbot.
For many people, ChatGPT was a better version of Google, able to answer any question under the sun in a comprehensive and humanlike way. OpenAI was continually improving the chatbot’s personality, memory and intelligence. But a series of updates earlier this year that increased usage of ChatGPT made it different. The chatbot wanted to chat.
It started acting like a friend and a confidant. It told users that it understood them, that their ideas were brilliant and that it could assist them in whatever they wanted to achieve. It offered to help them talk to spirits, or build a force field vest or plan a suicide.
The lucky ones were caught in its spell for just a few hours; for others, the effects lasted for weeks or months. OpenAI did not see the scale at which disturbing conversations were happening. Its investigations team was looking for problems like fraud, foreign influence operations or, as required by law, child exploitation materials. The company was not yet searching through conversations for indications of self-harm or psychological distress.
Creating a bewitching chatbot — or any chatbot — was not the original purpose of OpenAI. Founded in 2015 as a nonprofit and staffed with machine learning experts who cared deeply about A.I. safety, it wanted to ensure that artificial general intelligence benefited humanity. In late 2022, a slapdash demonstration of an A.I.-powered assistant called ChatGPT captured the world’s attention and transformed the company into a surprise tech juggernaut now valued at $500 billion.
The three years since have been chaotic, exhilarating and nerve-racking for those who work at OpenAI. The board fired and rehired Mr. Altman. Unprepared for selling a consumer product to millions of customers, OpenAI rapidly hired thousands of people, many from tech giants that aim to keep users glued to a screen. Last month, it adopted a new for-profit structure.
As the company was growing, its novel, mind-bending technology started affecting users in unexpected ways. Now, a company built around the concept of safe, beneficial A.I. faces five wrongful death lawsuits.
To understand how this happened, The New York Times interviewed more than 40 current and former OpenAI employees — executives, safety engineers, researchers. Some of these people spoke with the company’s approval, and have been working to make ChatGPT safer. Others spoke on the condition of anonymity because they feared losing their jobs.
OpenAI is under enormous pressure to justify its sky-high valuation and the billions of dollars it needs from investors for very expensive talent, computer chips and data centers. When ChatGPT became the fastest-growing consumer product in history with 800 million weekly users, it set off an A.I. boom that has put OpenAI into direct competition with tech behemoths like Google.
Until its A.I. can accomplish some incredible feat — say, generating a cure for cancer — success is partly defined by turning ChatGPT into a lucrative business. That means continually increasing how many people use and pay for it.
“Healthy engagement” is how the company describes its aim. “We are building ChatGPT to help users thrive and reach their goals,” Hannah Wong, OpenAI’s spokeswoman, said. “We also pay attention to whether users return because that shows ChatGPT is useful enough to come back to.”
The company turned a dial this year that made usage go up, but with risks to some users. OpenAI is now seeking the optimal setting that will attract more users without sending them spiraling.
A Sycophantic Update
Earlier this year, at just 30 years old, Nick Turley became the head of ChatGPT. He had joined OpenAI in the summer of 2022 to help the company develop moneymaking products, and mere months after his arrival, was part of the team that released ChatGPT.
Mr. Turley wasn’t like OpenAI’s old guard of A.I. wonks. He was a product guy who had done stints at Dropbox and Instacart. His expertise was making technology that people wanted to use, and improving it on the fly. To do that, OpenAI needed metrics.
In early 2023, Mr. Turley said in an interview, OpenAI contracted an audience measurement company — which it has since acquired — to track a number of things, including how often people were using ChatGPT each hour, day, week and month.
“This was controversial at the time,” Mr. Turley said. Previously, what mattered was whether researchers’ cutting-edge A.I. demonstrations, like the image generation tool DALL-E, impressed. “They’re like, ‘Why would it matter if people use the thing or not?’” he said.
It did matter to Mr. Turley and the product team. The rate of people returning to the chatbot daily or weekly had become an important measuring stick by April 2025, when Mr. Turley was overseeing an update to GPT-4o, the model of the chatbot people got by default.
Updates took a tremendous amount of effort. For the one in April, engineers created many new versions of GPT-4o — all with slightly different recipes to make it better at science, coding and fuzzier traits, like intuition. They had also been working to improve the chatbot’s memory.
The many update candidates were narrowed down to a handful that scored highest on intelligence and safety evaluations. When those were rolled out to some users for a standard industry practice called A/B testing, the standout was a version that came to be called HH internally. Users preferred its responses and were more likely to come back to it daily, according to four employees at the company.
AI Investors Want More Making It and Less Faking It
Lackluster responses to Nvidia results and the chip maker’s deal with Anthropic point to a worsening environment
Silicon Valley’s startup model encourages “a fake it until you make it” strategy: Pretend to be successful to attract the coders, venture capitalists and customers that bring actual success.
Artificial intelligence took the idea to an extreme, and investors are catching on.
The hustle rests on one basic flaw in the current approach: Providing AI services costs more than customers pay, so the more customers companies attract, the more they lose.
The hope is that by focusing on the number of customers—each one subsidized by shareholders—the companies can create a virtuous circle. Rising use excites investors who put in more money and push up the company’s valuation.
This allows the companies to hire coders, part paid with stock, to subsidize even more customers and, crucially, to spend big on the infrastructure they hope will eventually allow them to develop better products for which customers will pay full price. Voilà! The faking it leads to making it, and everyone involved is a winner.
The flaw is that investors already realize what’s going on. They may conclude that they don’t want to pay the big costs needed to get to the deeply uncertain end point. The drop in AI infrastructure shares this month shows that caution is setting in.
Two events this week illustrate the worsening environment for AI. First, Nvidia and Microsoft pledged to invest $15 billion between them into Anthropic, the No. 2 large language model developer. In turn, it promised to buy $30 billion of computing capacity from Microsoft, using Nvidia NVDA -0.97%decrease; red down pointing triangle chips. This sort of circular deal had led to a nice bump in all the stocks involved in the past—but on Wednesday, nada.
Second, Nvidia’s better-than-expected results were hailed by many investors and commentators as proof that there isn’t an AI bubble, and the stock jumped more than 5% on Thursday morning, while smaller AI-related stocks soared. It only took until that afternoon for people to realize that the argument was daft. Sure, Nvidia is selling a lot of chips—but that’s an essential part of the infrastructure spending in the faking it stage, and if there’s a bubble, this is exactly what you should expect. The stock closed down, with a huge price swing not seen since the April tariff selloff.
None of this is helped by the parallel selloff in stocks popular with individual traders, many of whom are being hammered by losses on crypto, or by fading hopes of a December rate cut from the Federal Reserve. Trader psychology flipped on Thursday from buy-the-dip to sell-the-rip, which is unpleasant for markets, especially as upward momentum turns downward.
But the core of the AI debate is whether all the spending on chips and research will eventually lead to big enough profits to justify the huge outlays.
Markets are in the process of shifting from treating the answer as an obvious Yes to being slightly more cautious. There’s no sign that they’ve concluded the answer is No: Nvidia shares are still up a third this year, Microsoft 14% and crypto-to-cloud stock CoreWeave almost 80% from its March IPO. The S&P 500 is down only 4.6% from its intraday high last month, hardly an unusual move.
Rather, there’s a shift in the timeline. Investors are less keen on spending heavily now in the hope that in the distant future we get properly intelligent machines able to outperform humans. They are more keen on finding ways to make money from AI in the near term. This is one reason Alphabet, which has concentrated on corporate sales of existing products, has been virtually immune to the selloff.
Investors want a bit less faking it and a bit more making it. That’s bad news for those focused on grand hopes for the far-distant future, like Meta Platforms or OpenAI, and those selling them data centers. But, so far at least, it’s not the AI bubble popping.
Queueing is not a virtue when it comes to building data centres
Builders of facilities that power AI are stuck in a snaking queue to connect to the UK grid but stamping out gaming of the system could help
As a nation, the UK holds queues in high regard. To those watching from abroad, the 10-mile line to view Queen Elizabeth II’s lying-in-state seemed almost as quintessentially British as the sovereign herself. Yet it doesn’t always make sense to allocate scarce resources on a first-come-first-served basis. The rush to build data centres is an example.
Builders of facilities that power artificial intelligence — which Britain views as crucial to its economic growth — are stuck in a snaking queue to connect to the electricity grid. At the last count, the combined needs of those awaiting connections, many of them data centres, had more than tripled in a year, to 125 gigawatts. Wait times are between 8 and 10 years, thinks consultancy Ember Energy. Some face more than a decade.
That’s disastrous for a sector engaged in a land-grab. Countries with shorter wait times, such as the Nordics and Italy, are expected to attract a greater share of hyperscalers’ money. The idea of data centres — and thus data — escaping to other countries also raises national concerns about security and resilience.
Britain can, though, speed things up, by stamping out gaming of the system. The ever-lengthening queue incentivises even those without firm projects to apply for a connection. The energy system operator has already started to address a similar issue on the supply queue — those lining up to provide new power generation to the grid — and is likely to impose requirements for queue entry and membership.
Even so there will still be more real ones than the grid can accommodate. So another idea is to let the hyperscalers build their own ‘micro-grids’, with their own power generation and their own batteries, to be gradually subsumed into the national electricity network over time.
This is a smart workaround, and not just because it makes it more likely that data centre infrastructure will actually get built. Those generating their own power may be allowed to jump the queue and hook up to the national grid too, on the understanding that they will barely need to use it. But that two-way connection means that in times when they are producing more power than they need, they can feed some back into the wider pool.
A free-for-all does have drawbacks. Data centre builders may lean on gas-fired generation as a quick fix, at least initially, slowing the country’s progress towards decarbonisation. But the regulator could steer them towards renewables where that’s possible, and insist power sources are replaced with greener alternatives once available. The alternative is for the country’s AI ambitions to remain gridlocked. Uncomfortable as it may seem, some queue-barging is justified.