Sam Altman Says AI Has Already Gone Past The Event Horizon But No Worries Since AGI And ASI Will Be A Gentle Singularity

Speculating on the future of AI including artificial general intelligence (AGI) and artificial ... More superintelligence (ASI).
In today's column, I examine a newly posted blog piece by Sam Altman that has generated quite a bit of hubbub and controversy within the AI community. As the CEO of OpenAI, Sam Altman is considered an AI luminary, of which his viewpoint on the future of AI carries an enormous amount of weight. His latest online commentary contains some eyebrow-raising indications about the current and upcoming status of AI, including aspects partially coated in AI-speak and other insider terminology that require mindful interpretation and translation.
Let's talk about it.
This analysis of an innovative AI breakthrough is part of my ongoing Forbes column coverage on the latest in AI, including identifying and explaining various impactful AI complexities (see the link here).
First, some fundamentals are required to set the stage for this discussion.
There is a great deal of research going on to further advance AI. The general goal is to either reach artificial general intelligence (AGI) or maybe even the outstretched possibility of achieving artificial superintelligence (ASI).
AGI is AI that is considered on par with human intellect and can seemingly match our intelligence. ASI is AI that has gone beyond human intellect and would be superior in many if not all feasible ways. The idea is that ASI would be able to run circles around humans by outthinking us at every turn. For more details on the nature of conventional AI versus AGI and ASI, see my analysis at the link here.
We have not yet attained AGI.
In fact, it is unknown as to whether we will reach AGI, or that maybe AGI will be achievable in decades or perhaps centuries from now. The AGI attainment dates that are floating around are wildly varying and wildly unsubstantiated by any credible evidence or ironclad logic. ASI is even more beyond the pale when it comes to where we are currently with conventional AI.
In a new posting on June 10, 2025, entitled 'The Gentle Singularity' by Sam Altman on his personal blog, the famed AI prognosticator made these remarks (excerpts):
There's a whole lot in there to unpack.
His upbeat-worded opinion piece contains commentary about many undecided considerations, such as referring to the ill-defined and indeterminate AI event horizon, the impacts of artificial superintelligence, various touted dates that suggest when we can expect things to really take off, hazy thoughts about the nature of the AI singularity, and much more.
Let's briefly explore the mainstay elements.
A big question facing those who are deeply into AI consists of whether we are on the right track to attain AGI and ASI. Maybe we are, maybe we aren't. Sam Altman's reference to the AI event horizon alludes to the existing pathway that we are on, and he unequivocally implies and states that in his opinion, we not only have reached the event horizon but that we are avidly past it already. As espoused, the takeoff has started.
Just to note, that's a claim embodying immense boldness and brashness, and not everyone in AI concurs with that viewpoint.
Consider these vital facets.
First, in favor of that perspective, some insist that the advent of generative AI and large language models (LLMs) vividly demonstrates that we are now absolutely on the path toward AGI/ASI. The incredible semblance of natural language fluency exhibited by the computational capabilities of contemporary LLMs seems to be a sure sign that the road ahead must lead to AGI/ASI.
However, not everyone is convinced that LLMs constitute the appropriate route. There are qualms that we already are witnessing headwinds on how much generative AI can be further extended, see my coverage at the link here. Perhaps we are nearing a severe roadblock, and continued efforts will not get us any further bang for the buck.
Worse still, we might be off-target and going in the wrong direction altogether.
Nobody can say for sure whether we are on the right path or not. It is a guess. Well, Sam Altman has planted a flag that we are incontrovertibly on the right path and that we've already zipped down the roadway quite a distance. Cynics might find this a self-serving perspective since it reinforces and reaffirms the direction that OpenAI is currently taking.
Time will tell, as they say.
Another consideration in the AI field is that perhaps there will be a kind of singularity that serves as a key point at which AGI or ASI will readily begin to emerge and keenly showcase that we have struck gold in terms of being on the right pathway. For my detailed explanation of the postulated AI singularity, see the link here.
Some believe that the AI singularity will be a nearly instantaneous split-second affair, happening faster than the human eye can observe. One moment we will be working stridently on pushing AI forward, and then, bam, the singularity occurs. It is envisioned as a type of intelligence explosion, whereby intelligence rapidly begets more intelligence. After the singularity happens, AI will be leaps and bounds better than it just was. In fact, it could be that we will have a fully complete AGI or ASI due to the singularity. One second earlier, we had plain AI, while an instant later we amazingly have AGI or ASI in our midst, like a rabbit out of a hat.
Perhaps though the singularity will be a long and drawn-out activity.
There are those who speculate the singularity might get started and then take minutes, hours, or days to run its course. The time factor is unknown. Maybe the AI singularity will take months, years, decades, centuries, or lengthier to gradually unfurl. Additionally, there might not be anything resembling a singularity at all, and we've just concocted some zany theory that has no basis in reality.
Sam Altman's posting seems to suggest that the AI singularity is already underway (or, maybe happening in 2030 or 2035) and that it will be a gradual emerging phenomenon, rather than an instantaneous one.
Interesting conjecture.
Right now, efforts to forecast when AGI and ASI are going to be attained are generally based on putting a finger up into prevailing AI winds and wildly gauging potential dates. Please be aware that the hypothesized dates have very little evidentiary basis to them.
There are many highly vocal AI luminaires making brazen AGI/ASI date predictions. Those prophecies seem to be coalescing toward the year 2030 as a targeted date for AGI. See my analysis of those dates at the link here.
A somewhat quieter approach to the gambit of date guessing is via the use of surveys or polls of AI experts. This wisdom of the crowd approach is a form of scientific consensus. As I discuss at the link here, the latest polls seem to suggest that AI experts generally believe that we will reach AGI by the year 2040.
Depending on how you interpret Sam Altman's latest blog post, it isn't clear as to whether AGI is happening by 2030 or 2035, or whether it is ASI instead of AGI since he refers to superintelligence, which might be his way of expressing ASI or maybe AGI. There is a muddiness of differentiating AGI from ASI. Indeed, I've previously covered his changing definitions associated with AGI and ASI, i.e., moving of the cheese, at the link here.
We'll know how things turned out in presumably a mere 5 to 10 years. Mark your calendars accordingly.
An element of the posting that has gotten the gall of especially AI ethicists is that the era of AGI and ASI seems to be portrayed as solely uplifting and joyous. We are in a gentle singularity. That's certainly happy news for the world at large. Utopia awaits.
There is a decidedly other side to that coin.
AI insiders are pretty much divided into two major camps right now about the impacts of reaching AGI or ASI. One camp consists of the AI doomers. They are predicting that AGI or ASI will seek to wipe out humanity. Some refer to this as 'P(doom),' which means the probability of doom, or that AI zonks us entirely, also known as the existential risk of AI or x-risk.
The other camp entails the so-called AI accelerationists.
They tend to contend that advanced AI, namely AGI or ASI, is going to solve humanity's problems. Cure cancer, yes indeed. Overcome world hunger, absolutely. We will see immense economic gains, liberating people from the drudgery of daily toils. AI will work hand-in-hand with humans. This benevolent AI is not going to usurp humanity. AI of this kind will be the last invention humans have ever made, but that's good in the sense that AI will invent things we never could have envisioned.
No one can say for sure which camp is right and which one is wrong. This is yet another polarizing aspect of our contemporary times. For my in-depth analysis of the two camps, see the link here.
You can readily discern which camp the posting sides with, namely roses and fine wine.
It is important to carefully assess the myriads of pronouncements and proclamations being made about the future of AI. Oftentimes, the wording appears to brazenly assert that the future is utterly known and predictable. With a sense of flair and confidence, many of these prognostications can be easily misread as somehow a bushel of facts and knowns, rather than a bundle of opinions and conjecture.
Franklin D. Roosevelt wisely stated: 'There are as many opinions as there are experts.' Keep your eyes and ears open and be prudently mindful of all prophecies concerning the future of AI.
You'll be immeasurably glad you were cautious and alert.

Try Our AI Features

Explore what Daily8 AI can do for you:

Learn with AIFact CheckingText to SpeechAI Summary

Related Articles

Forbes

27 minutes ago

• Forbes

The Third Wave: AI And Web3 Are Changing Commerce At Amazon & QVC

Every TV show becomes a shop-able moment with Edge Video and AI When I was at AWS working on Cloud and AI, I spent time helping partners unlock the value of marketplaces—connecting buyers and sellers in more meaningful ways, simplifying transactions, and showing how digital platforms could transform traditional sales models. It taught me something powerful: when you reduce friction and deliver relevance at the right moment, commerce becomes invisible—and incredibly scalable. That same dynamic is now unfolding in the media. I'm always watching shows like Bridgerton or Emily in Paris and thinking, I wish I could buy that stunning dress or find that sleek Vespa they were riding. Or I'll be watching Outer Range or Yellowstone and wonder where to get that exact vintage jacket or rustic canoe. During a recent conversation with Joe Ward, CEO of Edge Video AI, I was introduced to Edge Video AI—a platform that turns any video stream into a living, breathing storefront. It's where watching becomes shopping and viewing becomes doing. Think of it as the next generation of monetization for content creators and media giants alike. Moving Beyond Ads and Subscriptions With AI For years, media companies have leaned on two business models: advertising and subscriptions. But both have limits. Ads fatigue users. Subscriptions are saturating the market. Viewers are seeking better experiences. Edge Video AI offers a third path: transactional media. It transforms video content into an interactive commerce layer, where AI matches relevant products to what's happening on screen. A single persistent QR code links the viewer's mobile device to the content—making everything instantly shoppable. Watching a beach scene? You might see travel excursions, swimsuits, or sunscreen appear—exactly when the viewer is most inspired. With one scan, you're browsing or buying without leaving the screen. Or check out the below from a TV show with your favorite actor. You can click and purchase that same leather jacket on your mobile device live. Watching a TV show means that you can now find that leather jacket or jeans worn by your favorite ... More actor In chatting with Robert Scoble, an AI Futurist, he told me that "we're witnessing the death of passive content. The future of media isn't just something you watch—it's something you engage with, shop from, and participate in. Real-time AI is turning every frame into a storefront and every viewer into a customer. This isn't just innovation—it's the reinvention of entertainment as an interactive, commerce-driven experience." And it works. One purchase can generate the same revenue as 300 traditional ads. Why AI Works: Relevance, Trust, and Mobile Edge succeeds by tapping into three converging trends: The model is already live with QVC, Amazon, and InTravel, and attracting attention from global players like Google. Any media company can onboard in days—Edge provides the QR layer and product feed, and the content starts working harder. This isn't just an innovation in tech—it's a shift in user experience design. Web3 Adds the Engagement Layer To Edge's AI What takes Edge to the next level is its Web3 integration. Gamified experiences—like quizzes based on what's happening in the video—reward users with points. Those points convert into blockchain-based tokens (like FAST, available on and Polygon), which can be redeemed for discounts, access, or even NFT-style digital assets. The experience transforms viewers from passive audiences into active participants—blurring the lines between media, commerce, and gaming. And it's wallet-enabled from the start. Tokens are stored in decentralized wallets like MetaMask, meaning users keep control—and brands gain loyalty through participation, not interruption. The AI Business Model Built for Scale Edge Video AI is quietly reshaping the media landscape by embedding real-time shopping directly into broadcast and on-demand video. Unlike traditional models that rely on manual tagging or interruptive overlays, Edge's fully automated, patent-backed system enables commerce to happen seamlessly within video content itself. Edge Video AI is already live with Intravel and DangerTV, and is rolling out across two major national media networks and five regional or mid-sized platform has integrations with over 21,000 vendors and access to more than 15 million products, including from giants like Amazon and Viator. The company's approach is designed for simplicity—deployment takes just hours, and it layers onto existing content without altering production workflows. Media partners see revenue shares ranging from 35 to 55 percent, along with optional setup fees and GDPR-compliant privacy protections. For an industry under pressure to find new revenue models, Edge Video AI offers a compelling case: commerce that doesn't interrupt, but integrates. It's a model designed to be additive, not disruptive—meaning it layers onto existing content without requiring major workflow changes. It's also built for scale: from major broadcasters to niche creators, anyone with video can now drive direct revenue from it. Web3 and AI Moves From Intention to Action What makes this moment different from past shop-able media attempts is the convergence of maturity in three spaces: It's the next evolution of how we connect attention to action. In a discussion with Joe Ward, the CEO, Edge Video AI, he said that "we believe content should work harder—for creators and for viewers. Our AI doesn't just watch video, it understands it, and instantly connects it to products people care about. Whether you're streaming on your phone or wearing AR glasses, the experience is seamless, contextual, and rewarding. This is where storytelling meets commerce—and where creators finally get paid when their influence drives action." Joe Ward, CEO of Edge Video Ai As someone who's helped build platforms at the intersection of technology and commerce—from enterprise ecosystems to decentralized identity—I see Edge Video AI as part of a broader shift from attention-based economies to intention-driven experiences. The Downsides Of AI But as with any technological leap, there are trade-offs worth considering. When every moment of entertainment becomes a moment of monetization, we risk blurring the line between storytelling and salesmanship. There's a fine balance between offering convenience and fostering consumption fatigue—especially when AI determines what we see and when. While shop-able content can feel seamless, it may also erode the escapism that draws people to media in the first place. If every sunset scene becomes a travel ad and every stylish outfit a product pitch, creators and audiences alike may start to feel the weight of commercial saturation. The challenge moving forward will be ensuring that utility doesn't come at the cost of authenticity. The New Revenue That AI Brings Helping partners realize the value of marketplaces at AWS gave me a front-row seat to how digital experiences can create exponential outcomes. What I'm seeing now in the media world feels just as pivotal. Edge Video AI isn't just turning content into a storefront. It's turning engagement into revenue and viewers into participants. It's building a path where content creators can scale without more ads or more paywalls—and where consumers can enjoy, interact, and transact in one fluid motion. The marketplace of the future isn't a website or an app—it's wherever our attention lives. And for media, that future just got a lot closer. Did you enjoy this story about AI and Web3's impact on revenue for shopping? Don't miss my next one: Use the blue follow button at the top of the article near my byline to follow more of my work.

Yahoo

27 minutes ago

• Yahoo

Can Archer Aviation Make You an eVTOL Millionaire?

Archer Aviation is aiming to debut its eVTOL product shortly -- after it gets FAA approval. While the company is raising a lot of money, it is also experiencing a large burn rate. Its questionable unit economics and premium valuation should keep investors away. 10 stocks we like better than Archer Aviation › Electric air taxis -- otherwise called electric vertical takeoff and landing (eVTOL) aircraft -- have entered the investor zeitgeist. Many companies are aiming to build this innovative transportation vehicle, and perhaps none more aggressively than Archer Aviation (NYSE: ACHR). The start-up has raised billions of dollars and is aiming to sell its vehicle for air taxi networks around the world. Investors have hyped up this pre-revenue stock, driving up the price by 226% in the last 12 months alone. Will Archer Aviation stock help make you a millionaire? The story is more complicated than you think. Here's why Archer Aviation is a high-risk stock that investors should be heedful of buying today. Archer Aviation is a manufacturer building an eVTOL called the Midnight. Powered by electric batteries, this aircraft will have a pilot and four passengers that can take off vertically like a helicopter, but with minimal noise disruptions. It will be able to travel through the air to point-to-point destinations, which will be operated by Archer Aviation and its partners such as United Airlines. It is also working with partners internationally to build air taxi networks, including Abu Dhabi Aviation. The promise of eVTOL air taxis is enabling passengers to avoid traffic on busy routes, such as going from Manhattan to the airport. Archer is betting there will be insatiable demand for this product, and likely at a premium price. Many cities around the globe are ensnared in persistent traffic jams. The Midnight aircraft promises to relieve the pressure. That is, if the aircraft is approved by the Federal Aviation Administration (FAA). The FAA is 15% through its final verifications on the Midnight, which has already gone through a successful test flight with one of its propellers not working. More testing is to be done, but Archer Aviation looks to be on the path to FAA certification at some point in the near future, albeit without an exact date. In the meantime, the company is not generating any revenue and burning cash. Management has raised a lot of money in the capital markets, most recently an $850 million stock offering at a $10 share price. The company now has $2 billion of liquidity, which will give it a few years of runway at its current annual burn rate of $450 million in free cash flow. The Midnight aircraft is reportedly being sold for around $5 million a unit, or will be used directly by Archer Aviation with one of its aviation partners. This is nowhere near the cost of a commercial airplane, but the Midnight Aircraft is a bit different. It still has a pilot, who will need training and a salary, but only four passengers. In order to make up for pilot, energy, and depreciation costs, these air taxi tickets will need to be fairly expensive. Let's assume that a typical Midnight air taxi gets used 10 times a day with a full passenger load. That is 40 tickets sold to customers for a short 10-minute trip. At $100 a pop, that turns into $4,000 in daily revenue. Multiplied by 365 days a year -- this is assuming full capacity year round -- and you get $1.46 million in annual revenue. It would take a taxi operator close to four years to break even on a $5 million purchase before considering operating costs like hiring a pilot. Call me skeptical that Archer Aviation can sell a $100 ticket on its air taxis. It may have to be closer to $250 a ticket, putting it close to the price of an airline ticket. How many people are willing to pay up to skip the traffic? Time will tell. To be clear, there is a lot of growth potential for Archer Aviation and other eVTOL stocks. Customers would line up out the door to turn an hour-long trip in traffic to a 10-minute ride above the city. That is, if the routes can be operated profitably while selling tickets at a reasonable price, which is still a major uncertainty with this business. The hype around eVTOL stocks has reached close to fever pitch. Even though it generates zero dollars in revenue, Archer Aviation now has a market capitalization of $6.5 billion -- and this is before considering the dilution coming from its most recent capital raise. The company is burning $450 million in free cash flow a year in order to scale up its manufacturing capabilities and has not sold a single product to a customer. Management can talk all it wants about bringing point-to-point air taxi networks to cities around the world, and even for the Los Angeles Olympics in 2028 (which it plans on doing). But with the unit economics discussed, Archer Aviation is trying to thread the needle in order to scale up its operations and generate a profit for all stakeholders. Even if it does end up selling a bunch of air taxis every year, a $5 million price point would only get you to $500 million in revenue at 100 annual units. Profit margins will likely be slim and not get the stock to a reasonable price-to-earnings ratio (P/E). At a large market cap for a pre-revenue stock, Archer Aviation does not look like a millionaire maker for your portfolio. On the contrary, it looks like a good way to lose a million dollars. Before you buy stock in Archer Aviation, consider this: The Motley Fool Stock Advisor analyst team just identified what they believe are the for investors to buy now… and Archer Aviation wasn't one of them. The 10 stocks that made the cut could produce monster returns in the coming years. Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you'd have $664,089!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you'd have $881,731!* Now, it's worth noting Stock Advisor's total average return is 994% — a market-crushing outperformance compared to 172% for the S&P 500. Don't miss out on the latest top 10 list, available when you join . See the 10 stocks » *Stock Advisor returns as of June 9, 2025 Brett Schafer has no position in any of the stocks mentioned. The Motley Fool has no position in any of the stocks mentioned. The Motley Fool has a disclosure policy. Can Archer Aviation Make You an eVTOL Millionaire? was originally published by The Motley Fool

Washington Post

28 minutes ago

• Washington Post

No one has made fusion power viable yet. Why is Big Tech investing billions?

DEVENS, Massachusetts — Inside a cavernous factory in a quiet Boston exurb, workers wearing hard hats and safety glasses swarm around giant magnets powerful enough to lift an aircraft carrier. In another building — where some work proceeds in strict secrecy — the magnets are being assembled in a spaceshiplike vessel designed to contain a magnetic field in temperatures that will soar to tens of millions of degrees. The plan is to squeeze atoms together and create energy from fusion, a potentially limitless and cheap source of power that scientists have been chasing for decades. The reactor under construction here at Commonwealth Fusion Systems is one of at least 43 private-industry ventures or partnerships in the United States and allied countries that are racing to commercialize fusion power. It's a prize that has eluded scientists for so long, many still believe it can't be done, at least not anytime soon. But tech companies and investors are pouring billions into these companies, encouraged by breakthroughs they contend have placed a sustained fusion reaction tantalizingly within reach. China also factors into their urgency, with a government-sponsored effort there that is putting the West at risk of losing the global competition. Scientists dreaming of fusion are no longer toiling in the shadows. They are being courted by governors, billionaires and tech behemoths eager to get in on the ground floor of what they see as a transformative, carbon-free fusion economy. 'A lot of people thought we were chasing ghosts,' said Michl Binderbauer, at TAE Technologies, which has partnered with Google to build a fusion reactor in Southern California and is one of Commonwealth's top rivals. Now more than $8 billion in mostly private money has been invested in fusion start-ups, most of it in the past four years. 'This really has the ability to change the world,' said Genevieve Kinney, a partner at General Catalyst, a venture capital firm that late last year led a $900 million round of funding for a young company called Pacific Fusion. 'If it happens, the outcomes are massive. It could replace many of the technologies we use today.' The current and former Energy Department secretaries are boosting its promise. 'Fusion has hit that tipping point where things are going to happen fast,' Energy Secretary Chris Wright, a student of fusion decades ago at MIT who ultimately became an oil and gas CEO, said at a conference in Washington this month. While Trump officials have scaled back support for wind and solar energy, Wright has touted fusion because if harnessed, it would produce power without regard to weather or time of day. Fusion also is supported by one of Wright's predecessors, Ernest Moniz, a nuclear physicist himself who was once skeptical it could be commercialized. To be sure, nobody is promising a miracle overnight. The industry is grappling with the huge challenge of sustaining a fusion reaction, a massive, costly undertaking that could require materials that have yet to be invented. The most optimistic companies talk about getting power on the grid within the next decade, but they caution that electricity from early plants will be very expensive and limited. Skeptics warn that it could take at least another decade or two. But federal and state officials are already beginning to plan for the day fusion power becomes reality. The views of some fusion skeptics began to shift after government scientists in late 2022 used giant lasers to generate a reaction that produced more energy than went into creating it. That reaction lasted just a fraction of a second. But it proved that fusion was achievable, shifting the quest to an engineering challenge to create a lasting reaction, contain it and channel it into usable power. The forecasts for electricity demand around the world in the coming decades dwarf what experts say energy companies can deliver using current technologies. Much of the demand is driven by the intense energy needs of the artificial intelligence industry, motivating some of Silicon Valley's most powerful companies to embed themselves in the fusion moonshot, engaging their AI machinery in the effort to get fusion power out of the lab and onto the power grid. TAE, for example, is so closely collaborating with Google on its work that the Silicon Valley tech giant has a virtual control room on its campus enabling it to engage with the fusion firm's experiments. 'They have access to all the data,' said Binderbauer, a physicist who founded TAE more than a quarter-century ago. 'It's like a marriage. There are very few secrets left. The upshot is that they are partnering deeper with us.' It's a radically different landscape than when Binderbauer launched the company. Now, Moniz sits on TAE's board, and Google and Chevron are major investors. OpenAI CEO Sam Altman is the executive chairman of West Coast fusion firm Helion Energy, which has inked an agreement to supply Microsoft with electricity if it gets a plant up and running. Worries that China will win the fusion race are also giving U.S. firms a boost. China is building what experts believe will be one of the most powerful fusion reactors in the world, bigger than the U.S. government facility in Berkeley, California. As with U.S. efforts, the project is focused on advancing nuclear weapon design, as fusion reactors can be used to simulate the conditions of a nuclear explosion. China is now investing substantially more public funds in fusion than the United States is. The risk is that fusion power could be one more U.S. energy innovation, like solar panels and electric-car batteries, that stalls out here amid a lack of public investment, enabling China to monopolize the industry and its supply chains. 'The winner in the fusion race will be the country that can build these plants at scale and do it around the world,' said Jimmy Goodrich, a nonresident fellow at the University of California Institute on Global Conflict and Cooperation. He said China is well positioned, as it is vastly outpacing the United States in building traditional nuclear fission reactors — which power today's legacy nuclear plants using technology that splits atoms, rather than fusing them — with 27 under construction compared with zero in America. 'The speed and scale at which they are moving is remarkable,' Goodrich said. 'They can apply that to fusion, and we are left in the dust.' Germany, Japan and Britain are also racing to build the world's first fusion power plant. In the United States, companies are jockeying with one another, sharing some scientific findings and technologies but also making bold claims that their specific approach is superior and most likely to succeed. TAE claims to have the 'cleanest and safest approach to commercial fusion power.' It is conceptually similar to that of Commonwealth Fusion's magnet configuration, called a tokamak, but is designed to use different fuel and operate at lower temperatures. Commonwealth arguably has a leg up, having brokered a deal with Virginia to locate its first fusion plant near Richmond, with the aim of selling 400 megawatts of power by the early 2030s. It is enough electricity to power a sizable data center. The firm spent hundreds of millions of dollars on its Massachusetts magnet factory, which also helps supply the experiments of other fusion companies. Among them is Type One Energy, which in February signed an agreement with the Tennessee Valley Authority, the nation's largest public utility, to build a 350-megawatt fusion power plant called Infinity Two on the grounds of a retired coal-powered generating station. Infinity Two would be powered by what is called a stellarator, which the company says will be able to sustain a fusion reaction without needing to invent new materials to handle the heat and energy intensity involved, because it will operate at lower temperatures. Equipment breakdown is one of the biggest challenges fusion faces, as generating energy for even a few seconds can destroy the machinery creating that energy. 'If you have a promising approach but you still need to invent new materials, the hard reality is you are not going to be putting fusion energy on the grid in 10 years,' said Christofer Mowry, CEO of Type One. Other companies aren't using magnets, instead taking the giant-laser approach used by the U.S. government at Lawrence Livermore Laboratory in Berkeley, where scientists have eight times since late 2022 generated a fusion reaction that expelled more energy than it consumed, known as 'ignition.' The costs are so high and engineering challenges so extreme that one of the most prominent U.S. fusion experts, Harvard physicist and former White House science adviser John Holdren, said in an interview that 'it is extremely unlikely we will see fusion power on the grid much before 2050.' It took scientists 70 years to reach ignition, Holdren said, and developing the engineering capabilities required to sustain that reaction is just as difficult. 'We are just miles short of the conditions a practical reactor would require,' he said. Victor Gilinksy, a former member of the Nuclear Regulatory Commission, has also warned that companies are vastly downplaying the huge hurdles they have yet to overcome. Michel Claessens, former communications director for ITER, an international effort to advance fusion science, says the industry is misleading the public with its promises that fusion energy is within sight. But scientists engaged in the chase say those views are outdated. 'Investors who spend even a cursory amount of time looking into this are coming away thinking there is a path here,' said Bob Mumgaard, an MIT scientist who co-founded Commonwealth Fusion Systems. Still, fusion energy is now where the auto industry would be if it had unlocked the formula for building an internal combustion engine before metal had been invented, said Greg Piefer, CEO of Shine Technologies, a fusion firm in Wisconsin. That makes it a risky business. Shine is using fusion neutrons to develop products such as imaging machines and medical isotopes, so it can stay solvent while trying to unlock commercial electricity. Piefer is acutely aware that no fusion company is going to profitably sell electricity before it reaches what is known as scientific 'break-even' — the point at which the fusion reaction generates more energy than is needed to ignite it. The only place in the United States that has happened is at the government facility in Berkeley — which is the size of three football fields and uses a laser pulse that for a billionth of a second shoots more energy than the entire U.S. power grid 2,500 times over. 'It is pennies worth of heat for millions of dollars in,' Piefer said. 'There are still a lot of factors to overcome.'