Daybreak - Friday Roundup: Adani goes nuclear and AI's talent exit

Episode Date: February 12, 2026

On February 12, 2026, Adani Power formed Adani Atomic Energy Ltd, a new unit to generate, transmit, and distribute nuclear power. This follows the SHANTI Bill opening India's nuclear sector to private firms. Both Adani Power and Tata Power, coal giants with long-life thermal plants, now lead the shift.Coal powers 74% of India's grid today. These firms profit big from it. So what happens when they control nuclear's pace too? Snigdha explores the conflict.In other news, top AI researchers have been quitting the big AI labs. Anthropic's Mrinank Sharma left over value clashes (X letter) and OpenAI's Zoe Hitzig resigned through an NYT op-ed, warning ads exploit chatbot user confessions for manipulation. Rachel breaks down how firms chase cash to scale and skip safety guardrails.Listen to our episode on AI taking over jobs here. And, here's another one we did on the previous exodus of AI talent.  Read Zoe Hitzig's opinion piece on NYT here. If you have any thoughts on this episode write to us at [e-mail] with Daybreak in the subject line. You can also leave us a comment on our website or the YouTube channel here.Daybreak is produced from the newsroom of The Ken, India’s first subscriber-only business news platform. Subscribe for more exclusive, deeply-reported, and analytical business stories.

Hi, this is Rohan Dharma Kumar. If you've heard any of the Ken's podcasts, you've probably heard me, my interruptions, my analogies, and my contrarian takes on most topics. And you might rightly be wondering why am I interrupting this episode too. It's for a special announcement. For the last few months, I and Sita Raman Ganeshan, my colleague and the Ken's deputy editor, have been working on an ambitious new podcast. It's called Intermission.

We want to tell the secret sauce stories of India's greatest companies. Stories of how they were born, how they fought to survive, how they build their organizations and culture, how they manage to innovate and thrive over decades, and most importantly, how they're poised today. To do that, Sita and I have been reading books, poring over reports, going through financial statements, digging up archives, and talking to dozens of people. And if that wasn't enough, we also decided to throw in video into the mix. Yes, you heard that right. Intermission has also had to find its footing in the world of multi-camera shoots in professional studios, laborious editing and extensive post-production. Sita and I are still reeling from the intensity of our first studio recording.

Intermission launches on March 23rd. To get alert, as soon as we release our first video. episode, please follow intermission on Spotify and Apple Podcast or subscribe to the Ken's YouTube channel. You can find all of the links at the ken.com slash I am. With that, back to your episode. Hi, welcome to daybreak. I'm Snickda and I'm Rachel. And today we are going to be doing things a little bit differently. Instead of one story, we brought you to. Sniktha is going to be looking at Adani powers move into nuclear energy. and what it means when coal giants start controlling the clean transition.

And Rachel is going to be telling you about why top researchers are walking away from companies like OpenAI and Anthropic. Stikta, why don't you go first? Sure. Okay, so yesterday Adani Power announced that it has formed an atomic energy focused unit called Adani Atomic Atomic Energy Limited. And it will generate, transmit and distribute electric power, derived from nuclear energy sources. Now, while the company did not provide any other details, this comes right after the Shanti Bill

that opened India's nuclear power sector to greater private participation. The idea is basically to meet rising electricity demands and curb carbon emissions. The government wants to increase the country's power capacity over the coming decades as a part of its clean energy push. And this makes a sense. Adani Power one of the first

privately held utilities to disclose their interest in the newly opened nuclear sector publicly. And it all adds up because with all the new trade deals with

the West, India needs to meet the latest ESG demands. Which is why coal incumbents like Adani Power, moving to zero carbon

sounds exactly like the correction that the market needs. But here is a harder and kind of obvious but a more uncomfortable question. What happens when the same companies that profit from coal

also control the pace of the transition away from it? And that is where the real catch lies. You see, coal plants have always been long-living assets. And you know what India's coal barons talking about nuclear reminds me of? The movie Gangs of Wasipur. In that film, the whole revenge saga spanning generations started about coal. Once the British colonialists left, the fight was over who controls the mines and whoever controls the mines controls the money and whoever controls the money controls the

future. Now, zoom out. For decades, India's power economy ran on coal. The biggest private generators built empires around thermal plants. They mastered the supply chains, the political relationships and the pricing structures. Coal was leverage. So now, after India opened nuclear energy to private firms, the likes of Adani Power Limited want to be the first ones to make that transition. Like I told you earlier,

coal-powered plants operate for decades. That is how long their usual life cycles last. Which is why their owners are used to depending on predictable cash flow to service their debt and reward shareholders. But if nuclear scales quickly and successfully, it directly competes with coal in the baseload segment of the market. So what incentive does a coal major have to aggressively accelerate to nuclear if it risks stranding its own coal assets?

You see, globally, energy transitions show a pattern. The incumbents try to manage the decline, not right, Russia. For example, when oil companies expanded to renewables, they also continued investing heavily in fossil fuels. Utilities in Europe invested in wind and solar, but they were also lobbying for capacity payments to keep thermal plants profitable. The risk in India is similar. If coal companies dominate both coal and nuclear, they effectively control the speed of this entire transition. A slow nuclear rollout, which, by the way, is slow anyway, keeps coal profitable for longer. Nuclear plants require massive amounts of upfront investment and long construction timelines.

Coal plants, by contrast, are already built. They generate revenue now. So, think about it. If a company must choose between spending billions on a reactor that may produce power in 10 years or running an existing coal fleet at high utilization at a time of rising demand, what is the logical outcome? Immediate financial gains would favor coal, right? India has seen coal generation surge during periods of high electricity demand,

especially during heat waves. And that is our current reality. When the grid tightens, coal is still the fallback. For shareholders who are fond of, focused on quarterly earnings, the temptation to prioritize immediate returns over long-horizon nuclear bets is stronger. Also, let us not whitewash the truth. Coal conglomerates are politically and economically influential. If they expand into nuclear, they become stakeholders on both sides of the energy transition, fossil and low-carbon, which is basically like playing both

sides. They may support nuclear incentives while also advocating for coal subsidies or capacity payments. They may argue that nuclear needs tariff guarantees while simultaneously resisting policies that accelerate coal retirement. And they may lobby for grid rules that ensure coal plants still remain the primary backup. In effect, the same forms will shape both the old system and the new one. Also, and here's the most important part, if you look at how transitions like these have played out globally, you will notice that it has a lot to do with public risk sharing. For example, in the UK, Sizewell C financing model is a nuclear project that basically allows for construction cost recovery from consumers via the energy bill surcharges. If India adopts similar mechanisms to make private nuclear viable, the public, which is a public,

which means you and I, may have to absorb part of this risk of this whole transition. Meanwhile, it will be the companies who will retain long-term operating profits once plants are running. So for coal majors, it is like a win-win, because coal profits will remain private, but the risk of the transition to nuclear could become partially socialized. That tension is politically sensitive, especially in a country where, electricity affordability is crucial. So, does that mean we should be averse to nuclear energy? No.

The issue here is about who builds it. If coal incumbents truly pivot capital and capability towards nuclear, it could accelerate India's clean energy build out. But if nuclear remains a strategic hedge, while coal continues to bear profits, the transition is going to be slow. Meanwhile, for you and I, the consumers, the stakes are tangible, whether it is air quality, power grid stability, or how the financial risks of electricity tariffs are distributed. And with that, I'm going to hand it over to my co-host, Rachel.

Thanks, Nika. So here's what's been up. Over the last week, two top researchers at two of the most successful AI companies in the world walked away from what they were building. Brenanke Sherma, a safety researcher at Anthropic, left the organization this Monday, citing unspecified concerns about how AI was making people less human. He released a letter that announced this on X. And then, a couple days ago, Zoe Hitsick, another researcher at Open AI, announced her resignation from the company through an opinion piece in the New York Times.

For both of them, the running theme seems to be that both companies do not seem to be holding up the same values they did before. Now, Mrenang's reasons were largely unspecified, but he said in his letter, and I'm quoting here, that he had repeatedly seen how hard it is to truly let our values govern our actions and that it was including at Anthropic, which he said, constantly faces pressures to set aside what matters most. Meanwhile, Zoe spelled out her reasons more explicitly.

She said that she was seeing Open AI head down the same route as Facebook did years ago, when she saw that they were launching ads for the free and go tiers of chat GPT. She wrote that people tell deeply intimate and personal details to AI chat pods. Basically, OpenAI has ended up building an archive of very sensitive information that has never been built before. See, the whole reason this archive even exists is mostly because people believe they're talking to an entity that has no ulterior motive. But this is also the same archive that would be fueling the ads on the platform, and that's coming from a company that very much does have ulterior motives and a whole host of investors to answer to.

So, her biggest concern was this. Advertising built on this archive creates a potential for manipulating users in ways we don't have the tools to understand, let alone prevent. So what we're seeing here is yet another turning point in the use and development of AI. It's no longer just about who has the largest compute power or even who has the largest funding round. It's about who can deliver on the most basic promises of AI first, which is firstly that AI can take over work and roles that previously humans used to do, and secondly, that it can be massively profitable.

And to deliver on those promises, certain corners have to be cut. For the most part, AI has been lagging behind on both fronts. Yes, it can take on certain tasks that humans use to do end-to-end, like research or copywriting. But singular tasks don't equal jobs, which means that while AI could reduce singular workloads on a case-by-case basis for individuals, it's still not able to take on entire jobs entirely unsupervised. By the way, we cover this more in depth on an episode of Daybreak just last week. I'll link it in the show notes. Now, as for the profitability bit, it's now a very well-known fact.

that AI companies that solely depend on AI as a product are not profitable yet. See how OpenAI went from a non-profit research company to a for-profit company quite recently. There's also the fact that the general belief is that AI is a bubble. So the best way to keep investor trust and interest going is to make sure that all these theories are proved wrong. Take Open AI, for example. It has been burning cash incessantly and not even turned a single dollar in profit. Hence, the introduction of the ads model, because it needs to be making profits somehow, and that profit is not coming from its 800 million user base.

And for Anthropic? Well, the company only has a fraction of Open AIs users, though the recent launch of Claude Co-work and Opus as Enterprise First Tools might slowly bridge that gap. But still, Anthropic is at a stage where it needs to put scale first, because they're also in talks for racing another round of funding that would put its entire value at $350 billion. And like I said, it's impossible to build such vast scale and in very short amounts of time without cutting corners.

For any major tech development, research has to be done, guardrails have to be built in, and the full capability of these models have to be understood before the products are launched. But we're still at a stage where every new release is developing and growing in real time. which is also why we have employees like Mrenank and Zoe leaving these companies because they're now saying in real time that values cannot keep pace with investor expectations.

And that's all for today. If you enjoyed listening to this episode, do tell us what you think at podcast at the ken.com. We will be back on Monday with more. Thank you for listening to Daybreak.