# OpenAI's Leadership Risks and Competitive Erosion

**Podcast:** FT Tech Tonic
**Published:** 2026-05-20

## Transcript

You could parachute Sam into an island full of cannibals and come back in five years, he'd be the king.
If you're Sam Altman, you don't have to be profitable to convey to investors that you'll succeed with or without them.
This is George Hammond, the FT's venture capital correspondent in San Francisco, reading a famous quote about Sam Altman, the boss of OpenAI.
So this is Paul Graham, who was the president of Y Combinator, who ultimately set Sam up as his successor.
But he said this about him.
20 years ago when Altman was in his 20s and was Paul Graham's protégé.
I think this is kind of the observation on Sam that has just stood the test of time better than anything else.
Is he always collegiate and collaborative?
Absolutely not.
Does he tend to have his way in the end?
I mean, yeah, I think that's borne out for the last 20 years.
We'll see what happens next.
As the leader of OpenAI, Sam Altman has already made his mark on the history of artificial intelligence.
The launch of ChatGPT sparked the generative AI boom and established OpenAI as the company to beat in the AI race.
But along the way, Altman ruffled plenty of feathers.
Colleagues have left to set up rival labs.
Co-founders have sued him in court.
And his own company even tried to sack him.
And on top of that, OpenAI no longer has a clear lead in the AI race.
The likes of Google and Anthropic are looming large in the rearview mirror.
So can OpenAI stay ahead of its rivals?
And if it can, is Sam Altman, king of the cannibals, the right person to lead the AI revolution?
This is Tectonic from the Financial Times.
I'm Murad Ahmed, the FT's technology news editor.
A handful of Silicon Valley companies are vying to lead the world in artificial intelligence.
In this season, I'm talking to the FT's expert tech reporters about each of them.
In this episode, Sam Altman and OpenAI.
To talk all things OpenAI, I spoke to George Hammond and also Madame Itamurgia, the FT's AI editor.
But first, I wanted to ask George about the trial that ended just last week, featuring Sam Altman and Elon Musk, one of the original founders of OpenAI.
Musk was suing Altman and others who he says stole a charity in turning OpenAI from a non-profit into a private enterprise.
On Monday, the jury ended up throwing out the case.
We spoke before that verdict, but George covered the trial with other FT reporters, and I wanted to know what he'd learned about Sam Altman from it.
So this trial is going back to 2015 and the foundation of OpenAI.
It kind of plots a course through...
this very seminal moment in 2017 and 2018, where OpenAI, having been founded as a nonprofit, was thinking about how it took on enough money to really go and compete with the likes of Google and achieve this mission of creating AGI for the benefit of all humanity, which is obviously hugely ambitious.
And the story we've heard about the various protagonists, Musk, Altman, Brockman, and others, is similar to the one that has crept out in public over the last few years, and particularly since.
Sam Altman was ousted from OpenAI in 2023.
And we've heard executives take the stand and accuse him of pitting senior leaders at OpenAI against each other, bending the truth, misleading people, having motives that are occasionally obscure or even conflicts of interest.
And we heard from board members who ousted him in 2023 because he didn't keep them in the loop on decisions as important as releasing CheckGPT.
So that has been very damaging.
And then Musk himself took the stand, kind of repeated a lot of these claims.
But his central claim in the case is that he believed from the start that this should be a non-profit and it should be open source technology that OpenAI was creating for the good of the world.
And that his co-founders in particular, Altman, basically hoodwinked him and have taken this non-profit vehicle, turned it into a for-profit for their own gain.
And that has not really borne out in the evidence.
We've seen evidence that Musk himself was willing to make it a non-profit so long as he retained some control.
He, at one point, suggested subsuming OpenAI into Tesla, which is his own for-profit car company.
So none of them are coming out of this squeaky clean.
Madhu, as EFT's AI editor, you've covered OpenAI for years.
In fact, you were the first person to suggest to me that it's a company we might want to start paying attention to.
No one seems to be coming out of this trial smelling of roses, but on Sam Altman in particular, it's brought up again a lot of complaints about his way of operating, both around this move from a non-profit to a private enterprise, and when he was temporarily ousted by his own board back in 2023.
Lots of questions about how trustworthy and honest he is.
Does that matter for OpenAI?
Is it important if OpenAI ends up being a leading AI company?
It's really interesting.
Like, you know, we talk about these personalities so much and it's because there's so few of them.
You know, if there were like 12 different CEOs, I don't think anybody could really keep track of who was lying to their, you know, lieutenant and who was cheating on their wives or whatever.
I'm just saying, you know, like the reason we're doing this is there's only really three, four, five maybe people who hold this kind of power, which makes it all the more interesting because it's so concentrated, right?
So that's the first thing.
The second thing is, you know, the ouster was the first real indication that there was something not quite right within the leadership dynamics of this company.
Well, maybe that's unfair because they obviously had the Anthropic founders quit in 2021 and go found their own company.
But I think we knew back then what we now are more convinced, which is that people didn't trust him.
And it's the same examples that we've reported on and others have as well in books and in stories, newspaper stories about, you know, him putting executives against each other, him not being fully candid about his plans, him having investments that he doesn't disclose and so on.
I guess my view on this is it's probably pretty kind of well documented now that maybe he's not the most straightforward player in a commercial sense.
But also my feeling is that we over-index too much on the individual people's personality when we think about oversight of AI.
And, you know, even if we believe that Demis is a much better person by whatever yardstick we use or he doesn't lie.
Demis Haseb is the head of Google DeepMind.
Exactly.
If we believe that he's much more trustworthy or that Dario Amadayev Anthropic is somebody that doesn't, you know, pit his executives against each other, has a much better relationship with his co-founders, which is true because, you know, all of OpenAO's co-founders have now left.
So that says something.
Even if all of that is true, I'm not sure that we should really be pinning our hopes onto any single individual to like shepherd this entire technology, given all the conversation around impact.
So I think, yes, Sam doesn't feel like the scientist mold founder who's doing this, you know, because he really has cared about this his whole life and wants to, I don't know, you know.
sort of create a world-changing technology just for the sake of it.
He certainly has a commercial lens to this, and that's part of his background, but that's okay, you know.
And he's allowed to be different to the Demis mold or the Dario mold.
I think the question really is, like, how do we control any of them?
George, you report from San Francisco.
I'm wondering what your view is, but also what's the view of...
the sorts of people that you regularly talk to, investors and entrepreneurs down in the valley.
So I think Madhu's really put her finger on it.
His reputation is very much not the same as Demis.
It's not the same as Dario Amadei, Anthropic.
But I cover, as you say, Murad, I cover venture capital from San Francisco.
So Simon's a very familiar archetype to me.
And in fact, he probably is the person who most embodies what the kind of special source of Silicon Valley is.
And it is...
as much network and access to capital as it is technical chops.
And I think no one has that more than Sam.
So what he brings to OpenAI is this ability to go out and raise ludicrous amounts of money to tap into his network of other founders, investors.
So I think the characteristics Madhu touches on are absolutely right.
And they have light and dark and they are central to the OpenAI story.
So I know that you can write Sam Altman's influence off.
And then this question of trust, I mean, there is a very obvious observation to make, which is in the trial we've heard from basically everyone who was ever regarded as a close Altman confidant has stood up and given not necessarily a wholly critical account of him as a leader, but certainly a sense that they don't altogether trust him.
And that is from Miriam Morati and Elia Sussgeber who worked alongside him.
And of course, the board who voted to oust him because they quite explicitly did not trust him.
And I think one of the things that came up in that oust and was mentioned again in the trial was that Altman decided to push out a chatbot and a model in 2023 without telling the board that the board felt was not ready for the public.
It was not yet safe enough.
It had not been tested to the fullest extent that it should have been.
So I think in that, you see the approach that Sam takes, which I think is ends driven.
You know, he cares about the outcome.
He cares about winning.
He recognizes this is a commercial pursuit to some degree.
And I think he is largely responsible for opening eyes posture of raise as much money as we possibly can, locking commercial relationships before rivals can, try and get to a scale that means we're going to be impossible to unseat even if our rivals have a better technology.
And I think what we've seen recently is the approach that Dario Amadei has taken, which is build a technology that people trust and then scale it.
We've seen for the first time that the latter approach really start to take precedence.
And I think that is the thing that has...
cast a new light on this question of, do people trust Sam?
I think before it was, nah, probably not as much as they'd like, but do they take it given how good a leader he is building this company?
And now it's, do people trust Sam?
And if they don't, is that going to be commercially damaging?
And given what George said about, we've heard from his, the main people who've worked with him, Mira and Ilya Sutskova, who co-founded the company, we know that Dario, Daniela and that...
founding team, they left and it's come out since then that they didn't trust him.
Despite what I say about not over-indexing, we probably should take the signals that are very clearly like shouting in our faces about the fact that like he has caused a lot of...
friction amongst his leadership.
And that is materially damaging for a company of this size that's trying to scale up in an entirely new industry.
So while I don't want to talk a lot about personality traits and how that impacts wider things, I think we should pay attention to the fact that it's not been great, his leadership.
We can't keep saying he's been a great leader.
The people who've worked closely with him have left.
have found him to be very difficult.
And so I guess the question now is, can he take this company through to being public or the next phase without the attendant drama that comes with what seems to be his way of managing this company and some of the toxic culture traits that seem to kind of bloom around him within his company.
So I think that's a fair question.
So can I jump back in there, Maren?
I think that's completely right.
And I don't want to be misconstrued as saying I think there's no problem with Altman's leadership.
I think what he is is a kind of prototypical startup founder just at a scale that we haven't seen before.
So you see these people, launch companies in Silicon Valley, they are messy and chaotic and people leave, but they grow and they get more money.
And I think that has been the opening eye story and it's obviously has hugely bad aspects as attested to by all of these executives.
But I think where it's really becoming untenable at this stage is that OpenAI is not a scrappy startup.
They're an $852 billion company with aspirations to go public in one of the biggest listings of all time.
And public market investors are not going to look very favorably on this kind of corporate culture and conflict of interest here, trust issue here.
So it's becoming a far more pertinent issue.
But yeah, Madhu, I totally agree.
These are issues that any normal company should be.
fairly prominent.
It's just in the Silicon Valley context, they're far more regular.
Okay, well, you mentioned that OpenAI is valued at more than $850 billion by investors, and it's hoping to go public soon and raise even more money on the stock market.
So after the break, I want to find out if they're really worth so much given how much new competition it's facing and why the IPO is so important.
In business, as in science, innovation begins with a spark of imagination.
It grows with ambition and shines brightest in those who challenge the status quo.
At Imperial Business School, innovation isn't taught.
It's what defines us.
That's why we're number two in the QS World Rankings.
You'll study business alongside brilliant minds in technology, engineering and medicine to spark ideas and shape pathways for the leaders of tomorrow.
Search Imperial Business School, where science means business.
George, when I came out to see you in San Francisco, last year, and you kindly took me to meet lots of investors across the valley.
You know, the message that we were getting then is that OpenAI's lead was impregnable.
You know, they had developed a moat, which meant that even if its technology wasn't the best, it had enough users to mean that it was going to be the winner in consumer AI.
And that doesn't feel like the case anymore.
I mean, Sam Altman earlier this year announced a code red.
George, explain what the Code Red was all about and why we think that maybe this moat around the business isn't there anymore or isn't as deep as we once thought.
Yeah, there was this great phrase that we heard, Murad, when you were over visiting that ChatGPT had become the verb in the way that Google has for search.
And so the moat was impregnable.
No one could ever take on OpenAI's lead.
And that feels...
far harder point to make it at this stage, nine months later.
So yeah, the Code Red was, I think, really a recognition from OpenAI that Google, which had been behind in the race for a long time, was now really right back at their heels, if not ahead of them with its model capabilities, and that Anthropic was gaining ground.
And in the last six months since that Code Red was issued, we've seen OpenAI try and narrow its focus to two things, really, and one of those is its core.
ChatGPT business, which is the consumer business, and the other is its growing enterprise business.
So it's trying to find new ways to sell coding and business-focused models into enterprises.
As a result of that new focus, it's cut off a number of projects which it had committed to, including a video model that it partnered with Disney on and various other side quests, as executives at OpenAI have called them.
But I think the bigger picture is that For the best part of three years, OpenAI had enjoyed a pretty sustainable lead just because they were the first out of the blocks.
And I think what has happened with Anthropic this year has thrown that into question.
So Anthropic basically has focused more squarely on the enterprise.
It's taken a slower path to gaining commercial traction.
But when it really started to take off towards the end of last year, it has just gone up in a way that is completely unprecedented since the beginning of this year.
Anthropics, Roman News, on an annualized basis, have gone from 9 billion to now we're hearing around 45 billion.
And that has seen them outstrip OpenAI.
So there is this clear evidence that selling into enterprises has real commercial value.
And I think that's made OpenAI wake up and think, is our strategy the right one?
Madhu, is this a good move from OpenAI?
They've been focusing on consumers, most of whom don't pay for ChatGPT.
So why not compete with Anthropics and sell more AI to enterprises?
I mean, there's never been a technology like this, right?
And so it's not easy to decide which bit of it to bet on because it's such a general purpose thing.
It's just as useful to people, it seems, in their personal and private lives as it is to them at work.
And I think, you know, Anthropic started as the underdog and focused on enterprise.
But obviously, the fact that ChatGPT is so popular amongst consumers has meant that OpenAI comes in with that brand already, which it probably can sort of take with it to the enterprise as well.
But I think they probably will build up an enterprise business eventually.
And the models are all very close in terms of their capabilities.
But I do think they have this amazing...
treasure trove, which is, you know, nearly a billion consumer users of this technology.
And we've seen before, there's a history of tech companies that have offered a product for free, quote unquote, and yet become multi-trillion dollar companies.
Obviously talking about Google here, Facebook as well, Meta.
But Google is a good example here because it made this huge technology investment into search, had to build a compute capability around it.
created this huge user base of consumers and then, you know, was able to make just ridiculous amounts of money off of this newly evolved business model, which was ads.
And investors that we've spoken to, including in OpenAI, I think rightly see that as just a very obvious way.
We already know that business model works and it's still growing.
You know, Google is still going gangbusters, nearly past NVIDIA as the world's most valuable company.
And so this ad model is just growing, growing, growing.
But I think the issue for OpenAI, if they try and explore advertising, is that AI users and consumer users in particular are very mobile.
Why not just go and use Gemini from Google or Broad from Anthropic if you don't like seeing ads or if you don't want to pay for it?
So then it becomes a question of who can subsidize these models for the longest.
And then you get into this massive ticking clock that is in the background for OpenAI, which is...
They expect to be profitable by around 2030, but between now and then they're going to burn 600 billion in cash.
And that's not their cash.
That's cash they have to go and raise from investors, or if they go public, they can raise some of it there.
And that is just the context for all of this.
It's how long do they have to actually pull off a sustainable business model, which is why these competitive dynamics are so important.
Because if it's perceived that other companies are stealing a march on them, then all of a sudden they're investors who've been...
to date, very, very willing to throw tens of billions of dollars at them, might decide that actually they don't want to do that anymore.
So that brings us neatly to what I think of as the trillion dollar question for OpenAI, which is whether or not it can get to an IPO, a stock market flotation, maybe even as early as...
this year, which is what the company has indicated it wants to do.
Given all these tensions and stresses on the business, how likely do you think an IPO is?
What are the hurdles that have to be jumped over to get there?
I think an IPO this year is unlikely.
And I think part of the reason for that is that we're going to see a massive IPO with Elon Musk's SpaceX in the next couple of months, we suspect.
That will take up a lot of the kind of available liquidity in the public market.
So there's a practical question.
And then I think there is the big difference between raising money in the private markets from people who you can go and sit in a room across the table from and sketch out your slightly hazy vision for a future of abundance and AI running everything and creating these new economies, which Sam is incredibly gifted at and has been incredibly gifted at since he started the company.
And then making this pitch to buttoned up.
public market funds and even retail investors who are a different breed altogether.
And we saw this probably most spectacularly with WeWork, which managed to convince again that it was going to be this kind of completely world-changing business.
Adam Neumann was another prodigious fundraiser.
But then when they came to do their public listing and they took real scrutiny from public market investors, the whole pitch fell apart very rapidly.
They pulled their IPO.
Do I think that's going to happen for opening eye?
Probably not.
They have a much more real technology and much more clear demonstration of the value of that technology.
But I think in their current form, they're going to receive a lot of scrutiny over their business model, their future plans, particularly that massive cash burn I described of 600 billion or so before they turn a profit.
So I think they're going to have a hard time persuading the public, but I certainly don't think it's impossible.
And I think, you know, if they straighten out the business a little bit as they're trying to do, and that becomes a simpler story.
then I think they could well grow public next year.
And the question is also like, you know, if they could get the money that they need in order to scale up their compute and keep making models, and if they could do that and stay private, like they would avoid all of that scrutiny, the sort of quarterly earnings cycle of stuff, surely they would choose to do that.
I think it's kind of interesting the reason that they've put it out there that they want to go public.
Obviously, part of it is capital and so on.
But I think it's also a competitive dynamic here with Anthropic.
They've been competitors from the start.
Obviously, Anthropic has spun off from them and have caught up in some senses from being the underdog and have said they are going to go public.
And I think there's a sense of OpenAI wanting to get there first and put their stake in the ground.
But I think it would be interesting to see if they can just get what they need on the private markets, whether it might just be better for them to stay, you know, like companies like Stripe have done this, you know, stayed happily private.
Well, I think the issue there is that where else can they get this sort of capital?
It feels like they've tapped out.
I mean, the current investor base is...
Microsoft, where there's been a decoupling with that company.
We're calling it a conscious uncoupling.
Very recent reference there, George.
But I mean, you've been following that relationship as well, George.
You know, they got $14 billion out of Microsoft and now they feel like they're splitting apart.
SoftBank have put in the best part of $30 billion and don't know how much more money they could possibly put in.
Yeah, I think SoftBank at this point are on the hook for closer to 60.
But exactly as you say, I think they've really exhausted a lot of these obvious supply lines for capital.
And more than that, their latest round, which was all in about 122 billion.
This is the biggest round of all time on the private markets, which suggests in itself that there's only so much further you can go.
But more than that, the investment is contingent on opening and going public or achieving artificial general intelligence, which is...
the point at which some investors think all bets are off.
So, I mean, that again sets a fuse for this company to have to go public.
And I think there's also not that many pure play AI companies on the public markets and those that there are have been unbelievably successful in the last couple of years.
Palantir is an example of a kind of AI native company, let's say, that has done incredibly well out of that trade.
So I think both OpenAI and Anthropic are looking at that.
and thinking there is a depth of support for this on the public markets.
And probably if you're the first one to go and tap it, it's going to be advantageous.
So it's a question of who's in better shape to go public.
I would expect them both to be gearing up in the next few months to try and do that.
All right.
Well, this series is about the AI race.
So I'm going to ask you to tell us where you think OpenAI might end up in the race.
Do you think it might end up winning this or could it be derailed?
I'm going to ask both of you, but Madhu first.
I think they're going to be second or third.
All right.
Betting right here on a specific number.
I think Google will come out ahead.
Just, you know, they can play a slower game.
There's fewer risks, less drama, you know, major technical chops.
etc.
They're building their own infrastructure.
We've discussed before kind of what Google's advantages are here.
Yes, it's true.
They aren't necessarily a clear number one today, but I think that's how it will shake out.
Between Anthropic and OpenAI, I think that's kind of interesting.
OpenAI has been clearly the leader.
But I do think there's been significant change.
And I think, you know, culture sounds like a soft word, but it actually means people.
And in this industry, there isn't an unlimited pool of people who know how to do this stuff.
And we have seen a lot of movement from opening eye towards anthropic.
Can the brand around being a sort of safety first, responsible company, can that win out versus first mover advantage?
And I do think right now they are neck and neck on that.
And unless OpenAI can pitch a good pitch on why people should go work for them, what they stand for, what kind of company they want to be, I think they will find that they'll lose out.
George, they're the most...
It's so over, we're so back company on the planet.
I mean, to give you a more contemporary reference, Maren.
The story about open AI is always binary.
It's always they're going to win it all or they're about to implode.
I suspect the truth is somewhere in between, although I can see the case for both sides of the binary.
There's one investor I spoke to who said, you know, they could be the Netscape of AI in reference to the browser wars in the 90s where Netscape.
had the advantage and then got outflanked by Internet Explorer and wound up being a kind of total relic that none of us use.
That is a possibility for OpenAI.
And I think the way it happens is exactly what Manu described, that Google comes and eats the consumer market and Anthropik takes the enterprise market and kind of leaves them in no man's land.
I don't think that is going to happen.
I don't think they're going to vanish into obscurity.
But I think the real risk for OpenAI is that their business model...
and more generally the business model of these frontier labs just doesn't quite make sense.
So they are having to pay far more to train the next generation of models than they are making day to day.
They have to find a way of increasing their pricing or reducing their costs.
And Madhu mentioned advertising earlier, that was the thing that worked for the search giants, or you might have an Uber Lyft.
duopoly that emerges and they both have pricing power and stop subsidizing their services.
Any of that is possible, but if you're opening an eye and you're going up against particularly Google, but now Meta as well, these companies that have enormous amounts of free cash flow and can keep financing and subsidizing their models, then how do you ever win that race?
Madhu, George, thank you very much.
Thank you.
Thank you.
That was Madhami Tamerge, the FT's AI editor, and George Hammond, the FT's venture capital correspondent in San Francisco, talking to me about Sam Altman and OpenAI, thanks to both of them.
Next week on Tectonic, Elon Musk was an original co-founder of OpenAI.
The world's richest man runs electric car maker Tesla and rocket builder SpaceX, and he desperately wants to be the top dog in AI through his own lab, XAI.
But is it a serious competitor or has he already lost?
And before we go, another reminder that we're running a short survey on Tectonic listeners.
This is your chance to tell us what you think of the podcast and you'll get the chance to win a pair of Bose wireless headphones.
Find the link to the survey in the show notes.
Tectonic is hosted by me, Murad Ahmed.
The producer is Edwin Lane.
The executive producer is Topher Forges.
Sound design by Breen Turner.
And Sam Giavinko, original music by Metaphor Music.
Cheryl Brumley is the FT's global head of audio.
In business, as in science, innovation begins with a spark of imagination.
It grows with ambition and shines brightest in those who challenge the status quo.
At Imperial Business School, innovation isn't taught.
It's what defines us.
That's why we're number two in the QS World Rankings.
You'll study business alongside brilliant minds in technology, engineering and medicine to spark ideas and shape pathways for the leaders of tomorrow.
Learn more.
Search Imperial Business School.
Who knows the issues like security and compliance right, allows growth.
Who waits for long, breaks it out.
That's why many startups start early on Wanta and grow up.
And if it's going to an audit, be in the morning and not only in months.
Now start on Wanta.com.