The AI industry's existential race for profits

Summary

Decoder with Nilay Patel — "The AI Industry's Existential Race for Profits"

Date: April 9, 2026
Host: Nilay Patel
Guest: Hayden Field, Senior AI Reporter, The Verge

---

Episode Overview

This episode of Decoder explores the high-stakes race for profitability within the AI industry, focusing on OpenAI and Anthropic—the two top AI startups navigating an inflection point as they approach potential IPOs. With billions invested and even bigger bets on future growth, both companies are making drastic moves to prove their business models can become sustainably profitable before the bubble bursts. The conversation centers on shifting business strategies, the costly rise of AI agents, compute scarcity, and evolving product focus, all set against the backdrop of public perception and intense industry competition.

---

Key Discussion Points & Insights

1. The Looming "Monetization Cliff" for AI Companies

• Background: AI startups like OpenAI and Anthropic have been fuelled by enormous capital investments—hundreds of billions of dollars—with expectations of breakthrough returns.
• Problem: Investment in AI infrastructure (data centers, chips) is increasing, but profits have yet to materialize. Pressure is mounting to go public and show viable business models.
• "Bubble or Boom": There’s real belief in the industry that some companies will “fail in spectacular fashion while others will succeed, but the money and opportunity are too big to ignore.” (02:00)

---

2. OpenAI vs. Anthropic: Contrasting Strategies

• OpenAI:
  • Traditionally focused on both consumer and enterprise markets, with frequent experiments and pivots.
  • Not afraid to throw “a ton of spaghetti at the wall,” per Hayden Field, i.e., multiple simultaneous projects with shifting priorities (06:21).
  • Now shifting resources to enterprise solutions, particularly code generation (Codex).
• Anthropic:
  • Seen as the “adult in the room,” focused on enterprise clients with steady, methodical growth and less distraction from side projects.
  • Sometimes explores sectors like healthcare and education but still through the enterprise lens.
• Quote:

  "Anthropic has the reputation of being the adult in the room... whereas OpenAI has the reputation of going on a ton of side quests, you know, throwing a ton of spaghetti at the wall." — Hayden Field (06:15)

---

3. The “Agent” Effect — Compute Costs and Tough Decisions

• Agents (e.g., Cloud Code, Cowork, OpenClaw):
  • These are AI-powered tools that automate complex workflows, supercharging productivity but drastically increasing compute usage ("burning tokens at a rate way faster than anticipated").
• Hard Choices:
  • OpenAI abruptly killed its high-profile video tool Sora, forsaking a $1B Disney deal to free up compute for more lucrative enterprise products (07:37).
  • Anthropic shifted users of third-party agent tools (OpenClaw) off flat-rate plans to pay-as-you-go, upsetting developers and users but seen as inevitable due to sustainability (14:50).
• Quote:

  “Agents are consuming hundreds of thousands more tokens than basic chat models…even if it’s frustrating, it makes sense, because their infrastructure isn’t built for this.” — Hayden Field (15:21)

---

4. Compute Bottlenecks: From Training to Inference

• Shift in AI Compute Usage:
  • Original focus: training ever-larger, more capable models.
  • Now: “All the compute is for inference”—serving high-demand applications versus simply building bigger models (11:00).
• Quote:

  “The compute constraints were stopping [OpenAI] from scaling appropriately…and they couldn't deliver what clients wanted unless they could get more compute.” — Hayden Field (09:34)

---

5. Pathways to Profitability—and Their Limits

• Enterprise as the Real Market:
  • Both companies’ projections for profitability (2029–2030) hinge on securing enterprise, government, and military contracts, not consumers. (17:06)
  • Consumer subscription revenues are marginal compared to enterprise deals.
• Consolidation Coming?:
  • Economists predict only a couple of LLM providers will survive; consolidation expected.
• Quote:

  “If that is ever going to happen, it’s going to be via the boring, unglamorous back of office stuff... Enterprise, military contracts, government contracts... Consumers just honestly... there is no way that stuff is ever going to add up.” — Hayden Field (17:17)

---

6. OpenAI’s Strategic and Organizational Turbulence

• Leadership Turnover:
  • Major changes at the executive level, including leaves and resignations, and confusion over strategic focus (21:23).
• Reputation Management:
  • OpenAI acquired TVPN (a podcast network) to “shape the narrative”—an aggressive move to control public perception following negative press.
• Quote:

  “It’s not the narrative. It’s literally what you are doing all the time that is getting you the bad press. You can’t just yell at us into liking you.” — Nilay Patel (23:39)
  “No AI leader is doing comms entirely right.” — Hayden Field (25:39)

---

7. Industry Perception: Tech vs. the General Public

• Distrust Runs High:
  • Public perception of AI is mostly negative, and the bigger a company’s profile, the worse its reputation (24:28).
  • Within tech, skepticism centers on business strategy and leadership statements.
• Quote:

  “When I was chatting with a firm that charts public perception... the general public really didn’t like AI for the most part… the more well known an AI company was, the worse the public perception.” — Hayden Field (24:28)

---

8. The Ticking Clock: IPOs, Pricing, and Product Roadmaps

• Short-Term Outlook:
  • Both Anthropic and OpenAI are racing to IPO, with leaked financials and projections (“Tortoise and the Hare” analogy—slow, steady Anthropic vs. hyperactive OpenAI).
  • Pressure means more price increases, cost-cutting, and likely further layoffs or project terminations (28:10, 34:45).
• Emerging Trends:
  • Clients weighing the costs of proprietary models vs. open-source alternatives, building internal infrastructure for cost control.
  • Price pressure may force further focus on enterprise, further reductions in consumer offerings, and potential shifts towards open-source if current products are “good enough.”
• Quote:

  “The models today are good enough to be this disruptive… there’s no reason a distilled model years from now that’s much cheaper to run couldn’t be as good... the bleeding edge is unnecessary if current models are already so disruptive.” — Nilay Patel (33:11)

• What's next?

  “Next couple of months are going to be very interesting for executive turnover, which projects get killed off, and probably some top engineers going from one lab to another… tracking who’s moving where is going to be really telling as well.” — Hayden Field (36:00)

---

Notable Quotes & Memorable Moments

• On the Inflection Point (04:15):

  “It’s kind of like time to pay the piper in a way… now as companies prepare to go public… it’s finally time to face the music and see how much money they can really make.” — Hayden Field

• On Agents Forcing Business Model Change (15:21):

  “Agents have just changed everything… they’re consuming hundreds of thousands more tokens than basic chat models.”

• On AI’s Unpopularity with the Public (24:28):

  "If you're known as a household name for being an AI company, the general public right now isn't really a fan for the majority."

• On Competitive Dynamics (37:33):

  “I feel like we could do another full hour of Decoder on just how much the AI industry is driven by Dario and Sam hating each other specifically…”

---

Timestamps for Important Segments

• [04:00] — Introduction of Hayden Field and framing the AI industry’s inflection point
• [05:29] — Distinction between OpenAI and Anthropic’s strategies
• [07:37] — OpenAI’s shutdown of Sora and shift towards enterprise (Codex)
• [11:00] — Discussion of compute shifting from training to inference
• [14:50] — Anthropic’s new pricing structure and impact on OpenClaw users
• [17:06] — Can these companies ever become profitable?
• [21:23] — OpenAI’s executive turmoil and acquisition of TVPN for PR
• [24:28] — Public perception of AI companies
• [28:10] — The "tortoise and the hare" dynamic between Anthropic and OpenAI, and what comes next
• [33:11] — Discussion on the impact of current models’ sufficiency and pricing implications
• [34:45] — Strategies for turning profitable: raise prices or expand user base
• [35:39] — Predictions for upcoming changes, leadership turnover, and further focus narrowing

---

Conclusion

This episode dives deep into the existential pressures and strategic crossroads faced by OpenAI and Anthropic as they rush toward IPOs. The conversation reveals how the industry’s newest AI “agents” have upended revenue models, accelerated compute demand, and forced both cost increases and tough business pivots. Both companies are betting their futures on conquering enterprise markets, as the consumer side proves too fickle and unprofitable at scale. Amid executive shakeups and negative public perception, the next 6–12 months promise to be a defining period for who survives, who fades, and what the true business of AI will look like.

---

This summary is crafted to give you a complete understanding of the episode’s key insights, memorable lines, and the evolving narrative within the commercial AI world.

Transcript

[Adobe Sponsor]

Support for Decoder comes from Adobe. Life is unpredictable, and that means you need your projects to adapt with whatever gets thrown at you. That means mastering the ability to pivot and collaborate with others to reach your goals. Adobe gets that, which is why they made a tool that's just as flexible as you are. PDF Spaces and Acrobat. Your PDF files are no longer static. Instead, they're living documents that flex with you and your project's needs. Learn [e-mail] do that with Acrobat

[Nilay Patel]

hello

[Neil Apitel]

and welcome to Decoder. I'm Neil Apitel, Editor in Chief of the Verge and Decoder is my show about big ideas and other problems Today, let's talk about the looming AI monetization cliff and whether some of the biggest companies in the space can become real, profitable businesses before they careen right off it. My guest today is Hayden Field, who's our senior AI reporter here at the

[Nilay Patel]

Verge, and she's been keeping close tabs

[Neil Apitel]

on both Anthropic and OpenAI, and how those two companies in particular tell us a whole lot about the AI industry as a whole in 2026. You've certainly heard a version of the monetization cliff story before. Anthropic, OpenAI and all the other big AI startups are built off the back of hundreds of billions in capital investment, and they're linked to even greater amounts of forward looking investment in data center build out chips and other infrastructure spend. At some point that return has to pay off. The profits have to materialize or the bubble pops. Maybe AGI arrives, maybe the economy crashes. Who knows?

[Nilay Patel]

If you've been listening to the coder,

[Neil Apitel]

you've heard me talk about this with tons of CEOs right here on the show, and a majority of them have hinted towards the bubble popping. They think some companies will fail in spectacular fashion and others will succeed, but that the opportunities, and especially the money are simply too big to ignore. The AI industry is going to do this whether we want it to or not.

[Nilay Patel]

The market depends on it.

[Neil Apitel]

And so these last few weeks have felt like a very important inflection point as both Anthropic and OpenAI have started to react to the reality of needing to go public to make money. The catalyst for all this change is the rise of AI agents. Products like Cloud Code and Cowork, the open source OpenClaw and OpenAI's Codex. They've all radically changed how these companies are thinking about their resources, and that's starting to affect how they behave, the products they support, or suddenly kill the restrictions they impose on customers and the money they're willing to burn on the way towards their next big milestone. That's because agents are valuable to customers right now, but agents also use far more compute.

[Nilay Patel]

And so the way people are using

[Neil Apitel]

agents is burning tokens at a rate way faster than these companies anticipated, and that's causing them to make hard decisions. We saw this most evidently last month when OpenAI abruptly killed its video generation app Sora, ditching a $1 billion Disney deal in the process. Why? Well, it costs too much to run, and OpenAI needs a compute for codecs. And we saw it again just last week when Anthropic decided it would no longer let CLAUDE users burn through compute resources using the Open Claw agent framework through a standard subscription plan, instead forcing those users onto a pay as you go plan, which costs substantially more. As you'll hear Hayden explain, these are glimmers of a make or break moment for the AI industry as both anthropic and OpenAI barrel towards two of the biggest IPOs in history. And the pressure on these companies to make money has never been this intense. The projection these companies have made, which just this week leaked to the Wall Street Journal, tell a story of boggling growth to the tune of hundreds of billions in revenue and profitability by the end of the decade. But the most important questions now are can these companies pull all this off? And what compromises will they make to reach that goal and avoid crashing and burning? Before we start, a quick reminder that you can listen to this episode or any episode of Decoder completely ad free by subscribing to the Verge. Just go to the Verge.com subscribe okay, Verge senior policy reporter Hayden Field and the AI monetization cliff in the race to profitability. Here we go.

[Nilay Patel]

Hayden Field, your senior AI reporter here at the Verge. Welcome back to Decoder.

[Hayden Field]

Thanks. It's great to be here.

[Nilay Patel]

I'm excited to talk to you. There's a lot going on in AI world recently. It feels like we are at a very important inflection point for this industry. What are you thinking right now?

[Hayden Field]

Yeah, we absolutely are. It's kind of like time to pay the piper in a way. You know, they've been raising a ton of money, raising a ton of hype for years. And now, you know, as companies prepare to go public and the competition is heating up more than ever and they're kind of entering all these different sectors and trying different side quests, it's finally time to really, like, face the music and see how much money they can really make and there's never been more pressure on them.

[Nilay Patel]

Also when you say them, I want to stay focused on OpenAI and anthropic which seem to be on different trajectories. Now there are obviously other big AI companies in the mix. Google exists, it's going to do whatever Google does. But Google has a big business already. It can subsidize finding product market fit with AI. It can subsidize making efficiency improvements on TPUs. It's just very different from, from in particular OpenAI Anthropic, which have to become companies. Like they have to graduate into becoming companies, particularly if they're going to go public and then they're going to have shareholders and they're going to have to show profit and loss and all this other stuff. Can you just describe how OpenAI and Anthropic are currently situated and where they might be going?

[Hayden Field]

Sure. So yeah, I mean, it's interesting because in some ways they're in the same position. They're both preparing to go public this year reportedly and kind of racing each other to do that. And they're both constantly raising a ton of money and hype. But where they differ is of course OpenAI has traditionally been really courting the consumer facing stuff and some enterprise and government, but they've really been focused equally, if not way more so on consumer. Anthropic has always been pretty focused on enterprise and they have remained pretty steady on that focus. So you know, they're not really doing as many side quests. They're not, you know, rolling out as many other experiments or projects. They're just kind of laser focused on their enterprise goals and their enterprise clients. Now are they only doing that? No, sometimes they kind of, it seems like get FOMO and they're like going into, you know, Claude for healthcare and you know, Claude for education, things like that. But it doesn't totally remove them from their goals of enterprise because like, you know, that's pretty focused on health care organizations or you know, education systems are enterprises too, as we know. So you know, they're really laser focused on this. They kind of have the reputation of being the adult in the room in some ways because they're not as perceived as like going wherever the wind blows them. They're kind of on one trajectory and they're staying really steady it seems like, whereas OpenAI kind of has the reputation of, you know, changing their focus a bunch internally and externally. People have said this. It's like, you know, going on a ton of side quests, you know, trying things you know, throwing a ton of spaghetti at the wall, like consumer, enterprise, government, everything, just seeing what works. And even Sam Altman himself has described OpenAI as kind of like betting on a ton of startups internally and just kind of seeing which one pulls ahead. But now they're finally having to realize, hey, maybe it's time to focus on the most money making endeavors here and deprioritize some of these other projects, kill them off so we can just kind of compete with Anthropic and focus on coding and enterprise.

[Nilay Patel]

Yeah, I think that brings me to the news this week. That really made me feel like, oh, we're at an inflection point and that is Anthropic started raising its rates for people using tools like OpenClaw. They really want you in their system on their subscription plans, using the tools and their pricing their way. And if you want to use Claude to power other systems like OpenClaw, you're going to have to pay in a different rate structure. That to me feels like they don't want you to do it at all. And then next to that, OpenAI killed Sora, which was their very buzzy video generation product that was basically a deep fake nightmare. But they also had a deal with Disney for a billion dollars, which always seemed confusing, but they canceled that deal too. Let's start with OpenAI. You're saying they're killing off all these side projects. They're trying to focus on Codex, which is fundamentally enterprise software. It is a tool for software developers to make software. Why did they kill Sora and where did this sense of focus come from?

[Hayden Field]

I think the sense of focus honestly just comes from the competition and the fact that pressure is building on them to generate more revenue than ever. They've never had more eyes on them in their balance sheet and their, their whole company history because they're preparing to go public and because they had just raised so many billions of dollars, their post money valuation right now is $852 billion. So yeah, I mean, investors are saying, okay, like what's the plan here? What's the plan for returning our money? So in order to deliver on those promises, they are having to not only devote their time and like money and staff to the projects that are gonna make the most money, but also their compute. So that's something that we saw execs at OpenAI talk about. When they killed Sora, we saw a couple internal memos go out. One of them was from Fiji, Simo, the CEO of AGI Deployment. And she said that basically the company needed to stop focusing on sidequests and just really dive fully into enterprise encoding. And yeah, I mean, compute is Super Limited. OpenAI is always, always talking about how they don't have enough compute to fulfill what they want to do or to scale appropriately. Sam Altman at Dev Day and SF in October said to reporters. I've just never seen him so stressed when he was talking about this, about how the compute constraints were stopping them from scaling appropriately and how they couldn't really deliver what clients wanted unless they just could somehow get their hands on more compute. I've never seen him more stressed out. And so, yeah, that's playing out now. Months later, Sora took up a bunch of compute and there wasn't really a big return there. And so they abruptly decided to cancel it, apparently 30 minutes after working with Disney on a related project and then just suddenly pulled the plug with no notice. So, you know, things over there it seem like are in kind of a tailspin. It's like if you're pulling the plug on a project with a huge company like Disney, 30 minutes after talking to them about, like, how it was going great, there's some, some real, some real issues there.

[Nilay Patel]

I want to come back to OpenAI, its management, which has all but turned over since the last time you were on decoder, and its sort of strategic focus. But what really strikes me about the need for compute is, I don't know, when you were on the show a year ago or a year and a half ago, all of that compute was pointed at training. We got to make bigger models. They're going to be more capable. GPT 95 will come out and it'll be digital Jesus. Or whatever it was. And the idea was that we needed bigger models with more data and the compute, the scale of compute necessary was going to get us to the capable models and AGI in some way. Now it seems like the compute is all for inference, right? There's people that want to do things with these tools, particularly in software development. And if we don't scale up the compute to meet the demand, we'll get left out in the cold because our big rival is sitting right there waiting to scoop up all those customers. Has anyone pointed that shift out explicitly that we've gone from all of the focus on training and capability in the model to. The models are pretty good. Everyone wants to use them. We need massive amounts of compute for inference.

[Hayden Field]

So in the numbers that leaked from investors in OpenAI and Anthropic this week, there are a ton of crazy charts showing just how much of their money and profit is going or lack of profit is going towards inference training. The way they break out these categories is just really interesting. It's not always apples to apples, but it's like Anthropic is going to spend 1/3 to 1/4 on model training compared to OpenAI. But OpenAI's revenue is expected to hit 275 billion in 2030. But Anthropic says its revenue will hit 150 billion in 2029. It's kind of like Anthropic seems to be spending a lot less on the same categories as OpenAI and growing slower. But OpenAI is spending a ton and then just hoping that's going to lead to a return on investment, which kind of tracks with the way that companies have been operating for the past couple of years.

[Nilay Patel]

Yeah, Anthropic seems to be much slower and more focused and constantly worried that it's going to kill everyone in the world with every successive model. OpenAI is just pedal to metal all the time.

[Hayden Field]

Yeah, but Anthropic isn't worried enough that it'll stop because they just took out their Frontier Safety pledge and said, oh, actually, we're going to stay competitive, even if we think it's a little dangerous. Sorry.

[Nilay Patel]

The race to an IPO makes a lot of people rethink a lot of their values, apparently.

[Neil Apitel]

We need to take a quick break. We'll be right back.

[Adobe Sponsor]

Support for Decoder comes from Adobe for every big idea, your Documents folder tells a story. Let's say you've just finished pulling together a brief, so you hit export on final version PDF, but then you open the file and you immediately notice a typo. Several versions later, you're exporting final v4.actual final draft PDF. Adobe Acrobat can save you the digital clutter with PDF spaces. It takes your documents and turns them into a living project that you can engage with, get insights from, and collaborate with others on. You can gather all your files into one workspace and have a whole conversation with your AI assistant about it and ask questions to get deep insights about your project. You can even invite people to your PDF space and let them add files, comments, notes, and more. You could doodle in the margins or even turn your project into your own personal podcast episode. Acrobat lets you generate an audio overview of your project in just one click. Learn more at Adobe.com do that with Acrobat.

[Neil Apitel]

We're back with the Verge of senior AI reporter Hayden Fields talking about anthropic and AIs race to profitability.

[Nilay Patel]

The other piece of news, right, this is again, pricing compute usage. Anthropic changed its pricing structure to make using CLAUDE with openclaw much more expensive in different ways. So if you have a Claude Pro or Mac subscription, you can't just hook it up to openclaw and go, you've got to buy tokens on top of that subscription. That in some ways felt inevitable in other ways obviously made a lot of people very upset. It made the developers of openclaw upset. They said they could only delay the decision by a week. What's going on there?

[Hayden Field]

So when I talked to an economist about this this morning, he said that agents have just changed everything. And I talked to another couple of tech leaders about this and they said agents are consuming hundreds of thousands more tokens than basic chat models have been. So it does make sense, even if it's frustrating, because it seems like the way Anthropic put it was, hey, our infrastructure isn't built for this. We didn't plan for this. And yeah, I mean, if you're using openclaw with Claude, obviously you basically have an agent on your behalf prompting CLAUDE for you and delivering those prompts back to Claude and saying, no, do this, no do this. So it's way more than a human would be able to do. And Anthropic's point was, hey, we only really built this for humans to be able to prompt Claude right now, unless it's for our own products, obviously it's a money grab. Of course they don't want a third party tool doing the same stuff that they want their own.

[Nilay Patel]

Right. They want people using coworkers.

[Hayden Field]

Yeah, of course. So it's like basically they just want to keep it like a walled garden. They want it to be a moat. That's kind of the only advantage that AI companies have right now is just trying to keep their users engaged and keep them on their own platform, build a moat of some sort. Because otherwise, like, you know, everyone loves to switch between whatever model is the best that day, that week. So yeah, this is just Anthropic looking to like deepen its moat. And also, you know, it compute is constrained everywhere, so they don't really want a third party tool prompting Claude way, way, way, way, way more than a human would be able to.

[Nilay Patel]

They don't seem to mind when it's cowork doing the prompting.

[Hayden Field]

Right.

[Nilay Patel]

Because you're inside of Claude and then they can monetize you in whatever ways that they come up with to monetize you.

[Hayden Field]

Exactly. They want to monetize it if they can.

[Nilay Patel]

Is there a path either in enterprise for Anthropic or consumer, for OpenAI to actually make a dollar in profit? That seems to be the big question everybody has.

[Hayden Field]

Yeah, that's a really good question. And it's something that when I was chatting with economists, they didn't feel that it would be that easy or likely, but they think that one or two LLM providers will come out on top and the rest will have to consolidate. So yeah, there's a chance for sure. Anthropic and OpenAI are both projecting some form of profitability in 2029, 2030. Anthropic said maybe this year it'll also like be slightly in the green and then go back in the red and then go back in the green for 2028, 2029. But yeah, I mean, I think that they've all realized that if that is ever going to happen, it's going to be via the boring, unglamorous back of office stuff. Enterprise, military contracts, government contracts, all that stuff. Because consumers just, honestly, yeah, they'll maybe Pay for a $200 a month subscription if they're a power user, but there is no way that stuff is ever going to add up to the amount of money that's involved in these enterprise or government contracts.

[Nilay Patel]

This seems to me like the point at which OpenAI faces just a fork in the road. They were made to be a consumer business, as you've pointed out. It seems very much like they wanted to bite off some of Google's business. Google search, one of the greatest successes in business history, maybe the greatest business that you can run in world history. They haven't really succeeded. Right. They might have shifted some search behavior to ChatGPT, but they haven't taken meaningful

[Neil Apitel]

dollars away from Google.

[Nilay Patel]

Google just keeps doing better and better and they keep lacing Gemini into everything that they make. And eventually the idea that you would open ChatGPT to do a search when Google's going to deliver you something substantially the same, that's going to get harder and harder for OpenAI to compete on, are they just pivoting away from consumer entirely? They hired all those meta people to do ads and the ads have come to approximately nothing from what I can tell.

[Hayden Field]

Yeah, it's still early days for that, but I mean, it is, it's, it's funny, I don't, I'm really interested to watch the ad stuff play out. But no, they're not pivoting from consumer. It just looks like they're trying to really front load their resources, their compute and their staff into the enterprise encoding efforts. So consumer it's already built. It's not that crazy to keep it maintained and just keep rolling stuff out. But it seems like most of their efforts are still going to go towards catching up and closing the gap in enterprise encoding, especially because reputationally they also have a gap to close there. Anthropic, for better or worse, has the reputation of being pretty trustworthy, pretty brand safe. That's what a lot of startups were telling me when I talked to them about this a few months ago. They all were afraid of like the risk associated with especially XAI and somewhat OpenAI as well. Anthropic. They felt pretty safe using it. They didn't feel like they'd be on the hook for reputational risk. So yeah, OpenAI has to close the gap in terms of like actual usage. Anecdotally, what people prefer, the hype that's involved with Claude code and all that that brings, and then also the reputational stuff that Anthropic kind of has going for it. Just because of its steady, slow growth.

[Nilay Patel]

That steady state for Anthropic is kind of reflected across the company, not just in product development or strategy, but in terms of employee retention or the ability to attract people from across the AI industry. A lot of people just head towards Anthropic. I would contrast that to OpenAI, which you're pointing out has a different reputation and then just in the last week or two weeks feels like it's turned over its entire executive team. Fiji Simo, who you mentioned, is CEO of AGI Deployment. Her title like a minute ago was CEO of Applications and they switched to the AGI deployment, which I don't understand at all. And now she's out on medical leave. I wish her well. They have other executives who are out on leave. Their head of marketing just left. Right before all those people left, they bought a podcast called tvpn. What's going on over there? Is this like a stable company right now?

[Hayden Field]

It is. It is pretty crazy right now. I think they're going through a huge strategic shift and it's just. It's a question of whether this is going to be like every other shift we've seen in the past, with them kind of going all in on one thing, going all in on another thing, going all in, in parallel on five things. Is this really gonna last? That's kind of what people are asking and we don't know. I mean, I think if they are entirely focused on coding and enterprise yeah, it makes sense that there'd be a lot of upheaval, but they're also really into building this super app. And Greg Brockman just took over charge of that while CMO is out. And then on the business side, their cso, their CFO and their CRO are going to take charge. And then their CMO just stepped down due to health reasons. Their head of communications stepped down in January and there's still no replacement there. And that, I think, is part of why they bought TVPN. There's been a lot of bad press about OpenAI in the last few months. They've had a lot of public controversy and a lot of just drama playing out, and that is not good for their quest to have a reputation as a company that enterprises and the government can trust. And so I think that's part of why they bought tvpn. They said that literally they wanted to help shape the narrative, AKA help control the public narrative playing out about AI. And so what better way than to hire the people that are being watched three hours every weekday who are talking about it? Plus, TVPN is now going to help with OpenAI's comms and marketing in their free time. So, yeah, I mean, it's really an aqua hire situation.

[Nilay Patel]

Yeah, I have a lot of thoughts on that. I think a lot of the other AI companies who put their executives on TVPN certainly have a lot of thoughts about that that I've heard. We'll set that aside. The idea that it's just a marketing problem or a narrative problem. The last time I heard that was Uber executives complaining that Uber only got bad press. And most of the media that I knew at the time was like, no, you just keep doing stuff.

[Hayden Field]

Yep, yep.

[Nilay Patel]

It's not the narrative. It's literally what you are doing all the time that is getting you the bad press. You can't just yell at us into liking you. It's not quite the same with OpenAI. They are darlings in a very specific way. But the bad press or the perceived bad narrative, in my opinion, has everything to do with their strategic confusion.

[Hayden Field]

Right, Right.

[Nilay Patel]

The media is not out to get AI. When you do the polls, the broad public in the United States is like, AI is less popular than ice. There's a gap in how people feel about these tools even though they're exposed to them all the time. Is that all on OpenAI? Is that across AI in general? To me, it feels like it's very much OpenAI. In particular, Sam Altman running around all the time saying, I might destroy the world by accident if you don't give me all this power.

[Hayden Field]

Yeah, I think it's really interesting to like look at how the general public feels versus how people that are in tech feel because they kind of feel they feel different ways, but there is some overlap, but it's for different reasons. So I remember when I was chatting with a firm that charts like public perception of things in the last couple months, they said that, yeah, the general public really didn't like AI for the most part. They broke it out by generation, by gender, all sorts of different things. But for the most part, yeah, the general public was not a fan and they noticed that the more well known an AI company was, the worse the public perception of them, just because they were more aware that it was an AI company. So OpenAI had a worse reputation by this firm's standards than Anthropic per se. But as Anthropic's public perception was on the rise, opinion of it was going down. So that was just interesting. It's like if you're known as a household name for being an AI company, like the general public right now isn't really a fan for the majority, at least within tech, people are looking at the business angle and how these companies are conducting themselves. What the CEOs are saying, maybe Sam Altman isn't a household name for the average person in America. I think a lot of people know who he is. But I will be talking to people in the wild and say Sam Altman and they're like, who's that? So when I say CEO of OpenAI, they get it. But yeah, I think it's interesting just comparing the tech reactions to the general public's reactions because in the tech industry, yeah, people are really raising an eyebrow at opening eyes business strategy. And Sam Altman going around saying like, oh, I couldn't raise my child without chatgpt or oh, if you, if your job gets replaced by AI, maybe you should think about switching jobs. And Dario Amadeus said kind of sus things as well. So I mean no AI leader is doing comms entirely right. But it is interesting to see kind of, yeah, the difference between the general public's perception versus people in tech.

[Neil Apitel]

We have to take another quick break. We'll be back in just a minute.

[Adobe Sponsor]

Support for decoder comes from Adobe. For every big idea, your Documents folder tells a story. Let's say you've just finished pulling together a brief. So you hit export on final version PDF but then you open the file and you immediately notice a typo. Several versions later, you're exporting final v4 actual final draft PDF Adobe Acrobat can save you the digital clutter with PDF spaces. It takes your documents and turns them into a living project that you can engage with, get insights from, and collaborate with others on. You can gather all your files into one workspace and have a whole conversation with your AI assistant about it and ask questions to get deep insights about your project. You can even invite people to your PDF space and let them add files, comments, notes, and more. You could doodle in the margins or even turn your project into your own personal podcast episode. Acrobat lets you generate an audio overview of your project in just one click. Learn more at Adobe.com do that with Acrobat.

[Neil Apitel]

We're back with Verge senior AI reporter Hayden Field discussing the AI industry's make or break year and what might happen next.

[Nilay Patel]

This kind of brings us to the make or break moment, right? These companies are both headed towards IPOs. Their financials have all leaked. We can see their cap tables, we can see their revenue projections. And it feels a little bit like Tortoise in the hare, right? You've got Anthropic committing to being an enterprise company, being a solution for software development. That has kind of change the entire nature of software development. You have OpenAI with Codex that thinks that it can eat a piece of that market as well, shutting down consumer applications that aren't working. Maybe it will figure out ads, maybe it will actually bite off a piece of Google Search, who knows? But they're still just hopping all over the place. It's just pedal to the metal at OpenAI all the time, and Anthropic just continues moving along its roadmap. How do you think that plays out? Not over the course of the IPO timeline, but in the short term, do you think OpenAI can recapture a sense of focus?

[Hayden Field]

I think it's going to try very hard and I think it will be able to. The question is just can it hold on to that focus? I've seen them change their strategy just like this in the past, and usually it just falls by the wayside a couple months later or a year later. So what I'm wondering is how long they can hold onto this. Anthropic has, you know, committed to one thing, stuck with it for the most part. OpenAI, whenever they commit to something like a year later, things shift. You know, teams will be disbanded, they'll have a reorg. Anthropic's also seen a ton of change, but it's always had this one goal and kind of stuck with it. And I've been tracking both these companies for so long that it's like you can kind of see the trends there. So maybe this is a big step change for OpenAI and they're really going to pivot and maybe that's why we're seeing all this executive restructuring and the side projects being killed and trying to really force them to commit. But it is interesting because I don't know the rate at which they'll be able to catch up with Anthropic when it comes to enterprise encoding, because I've even seen anecdotally a bunch of startup founders switching entirely over to Anthropic when they used to be testing both. We did a piece of a couple months ago about how cloud code was having a moment and most of the founders and company leaders I talked with in a ton of different sectors preferred anthropics products for this stuff. So, I mean, to OpenAI's credit, it has been working really hard on getting its coding models up to date. More improvement. We've seen people start to prefer that sometimes. So, I mean, they've made a huge advancement here. It's just, can they close the gap? That's what we'll have to wait and see on.

[Nilay Patel]

Yeah, it really feels like software development is product market fit for these tools, and that's a very lucrative market. That is a lot of jobs that might go away or change substantially in some meaningful way. And everything else is like, wait and see. Like anything that kind of looks like software development might get product market fit. But software development is such a big category. That's in particular what a bunch of these companies are going to focus on. I'm very curious as the pressure on turning these companies into real businesses goes up as they get closer to an IPO and Anthropic has to monkey with pricing even more to make sure that they're running an actual business and not a token subsidy operation. And OpenAI has to cut down on more projects. If that pressure, that pricing pressure, that monetization pressure, changes the companies in any meaningful way. You've been thinking about this and even talking to people about this, do you see glimmers that that's about to happen?

[Hayden Field]

I think so, just because, I mean, the pressure is building. When I was talking to economists, there's no way that things can continue the exact way they are now. I was chatting with a couple execs this morning about how the price has kind of been passed on to enterprise clients and how that's shifted. So a lot of company leaders are thinking about moving to open source instead, or at least moving to open source for a lot of their simpler tasks or queries and kind of building their own evals to see what it makes sense to pay top dollar for for either anthropic or OpenAI more complex, more powerful models, which ones they can kind of route to the simpler models, and which ones they can just go open source for. So it's interesting that, you know, in order to kind of combat these pricing shifts and just the amount that they're paying these two labs every month, they're starting to just build their own internal infrastructure and tests and evals, just being like, okay, let's really budget here, like, what do we really have to pay top dollar for? And what can we kind of like, you know, skimp on? But it'll have around the same type of answer. So that's what I think is interesting. Like this cottage industry of charting this stuff out internally and then just keeping that close to the vest and using that as a guide.

[Nilay Patel]

Yeah, I mean, that's kind of why I was asking about inference at the beginning. If the models today are good enough to be this disruptive to software development, there's no reason that a distilled model a few years from now that's much cheaper to run or you can run locally wouldn't be as good. And that the bleeding edge model is unnecessary because it's so expensive. And we haven't run this industry long enough to know how those pricing dynamics play out. But it feels like the additional capability from the next model, the next model, next model is unnecessary if the current models are already so disruptive to at least the industry of software development. And I'm just very curious to see how that pricing plays out, because incentives to keep burning money on training go down if the products are good right now. And I haven't really seen the labs talk about it, but you can see the bigger companies that have to be much more tightly run, like Google starting to understand, oh, we can deploy a lot of different models for a lot of different uses and lower costs across the board.

[Hayden Field]

Right. And it's funny too, that these labs do promise like, oh, as the years go on, prices are going to drop. Don't worry, we're going to offer these models and access cheaper and cheaper, but that doesn't really square with what's going to have to happen for them to turn a profit, especially one that investors won't like, roll their eyes at. So that's what's going to be interesting. There's a lot of tension here that we'll have to track over the next six months or so because you know, this is going to be the big year for paying the piper.

[Nilay Patel]

Yeah. I mean, there's only two ways to do it, right. You can increase the number of people paying the amount they're paying now, or you can increase the price. Which one do you think it's going to be?

[Hayden Field]

It's funny because that's exactly what an economist was telling me this morning. He's like, well, you either got to expand to basically the entire general public globally or you're going to have to raise the prices a lot. And then even then it may not square with what you actually need. So we'll see. It's going to be a really interesting year.

[Nilay Patel]

Yeah, like I said, it feels like this past week, real inflection point. As we saw Anthropic starting to play with pricing in a way that shifted user behavior, I think somewhat meaningfully. And then OpenAI realizing it needed to get into the more lucrative part of the market and abandon some things that got it a lot of attention but ultimately had no path towards money. What do you think happens next here?

[Hayden Field]

I think we're going to see OpenAI further, you know, kind of consolidate resources into these two focuses and lose a couple of other side projects. So we've heard maybe ATLAS is going to go by the wayside probably even though they haven't said this. Some safety research, I would guess, or at least doing what they tend to do, which is reassigning people on a certain safety research team into other departments. And so they say they're not actually, you know, diminishing the research, but who knows whether they are or not, you know. So I think we'll see some changes to its research Org and maybe people studying long term risks. They've got to keep some of them around because it's good pr. But you know, I could see them definitely kind of like on the DL reassigning some of those people and then, yeah, I mean it's all about just devoting more compute to the things that are going to make the most money so that they can make investors happy. And so whatever ways they can do that, that's what we're going to see. Probably maybe another funding round before they go public. OpenAI is also reportedly trying to go public before Anthropic. That's something Sam Altman is apparently really serious about. But he's apparently the information reported sparring with its CFO Sarah Fryer about that. She apparently doesn't think it's ready to go public as quickly. So yeah, we're going to see maybe some more leadership turnover, who knows. But I think that the next couple of months are going to be very interesting for executive turnover, which projects get killed off, and also probably some top engineers going from one lab to another or vice versa. And tracking who's moving and to where and what team they moved from is going to be really telling as well.

[Nilay Patel]

I feel like we could do another full hour of Decoder on just how much the AI industry is driven by Dario and Sam hating each other specifically. I don't know if that's today, but you're going to come back and do that one soon. Aidan, thank you so much for being on Decoder.

[Hayden Field]

Thanks so much.

[Neil Apitel]

I'd like to thank Hayden for taking the time to join Decoder, and thank you for listening. I hope you enjoyed it. Let us know what you thought about this episode or really anything else at all. Drop us a line. You can email us atdecoder the verge.com we really do read all the emails. Or you can hit me up directly on threads or bluesky. We're also on YouTube. You can watch full episodes at Decoder Pod and we have a TikTok and an Instagram. They're also at DecoderPod. They're a lot of fun. If you like Decoder, please share it with your friends and subscribe wherever you podcasts. If you really like the show, hit us with that five star review. Decoder is a production of the Verge and part of the Box Media Podcast Network. The show is produced by Kate Cox. Nick Stat. This episode was edited by Xander Adams. Our editorial director is Kevin McShane. The Decoder Music is my Brickmaster cylinder. We'll see you next time.

[Adobe Sponsor]

Support for Decoder comes from Adobe. Life is unpredictable, and that means you need your projects to adapt with whatever gets thrown at you. That means mastering the ability to pivot and collaborate with others to reach your goals. Adobe gets that, which is why they made a tool that's just as flexible as you are. PDF Spaces and Acrobat. Your PDF files are no longer static. Instead, they're living documents that flex with you and your project's needs. Learn [e-mail] do that with Acrobat.