Guess who's coming (back) to Davos

It's (mostly) not about Donald Trump – another Davos returnee has got nearly as tough a sell.

I can’t say for sure – I’m not there – but it’s hard to shake the impression that the mood at the World Economic Forum, up in the mountains of Davos, might be less cosy than usual this year.

It’s the first ever Davos event without its founder, Klaus Schwab, at the helm, after he stepped down last year amid a flurry of allegations of impropriety. But his absence feels almost like a footnote compared to the activity around who will be there, after six years away – Donald Trump is on his way to Davos, with a huge delegation in tow, and his presence will surely dominate once he arrives on Wednesday.

On paper, Davos is pretty much the epitome of everything MAGA is against. It’s a meeting of the international elite – a globalist holiday camp. It’s the respectable face of business, pushing the idea of caring and responsible global business – caring (or at least claiming to care, depending how cynical you are) about climate change, wealth inequality, gender inequality, and other diversity issues. Worse still, it’s in Europe, and MAGA is dead set against most things European.

Trump’s visit this year comes at a nadir of US-European relations, not least over the Greenland crisis. It’s sure to top the headlines, and probably not for good reasons.

But – given this occasional newsletter is tech-focused – there’s another return to Davos this year that’s worth paying attention to, and that’s former UK chancellor George Osborne. He’s got the unenviable task of convincing world leaders that OpenAI knows what it’s doing, and that they should help OpenAI do it.

A Swiss sales pitch

Osborne is just about two weeks into his new full-time job with OpenAI, and he’s got one of his most important gigs of the year very early into his tenure. His new role comes with the job title of managing director, but in practice he is there to be the head of its “OpenAI for Governments” initiative.

Osborne’s hiring serves as something of a reset for the initiative, in a role which the company is framing as much broader than the traditional lobbying or public affairs roles occupied by former politicians – such as by Nick Clegg for Meta.

Instead, OpenAI for Countries is framed as an umbrella effort to encourage governments to deploy AI, help them to prepare and train their workforces for it, and partner with education systems. The company goes as far as to say that the initiative is “one of the most important ways to ensure that AI benefits all of humanity”.

That makes Davos a very important pitching opportunity, especially because AI, and what to do with it, is very near the top of the official Davos agenda this year. The World Economic Forum expects AI to “add over $15 trillion to global GDP” in the next five years alone, but to disrupt at least one in five jobs globally as it does so. Governments are trying to coordinate their response to the new technology, and AI companies are determined to keep them on side.

Osborne has something of a tough sell here, for a whole variety of reasons. The most basic of these is that the AI love-in isn’t quite what it was. There is a growing degree of scepticism around industry claims – the era of big tech’s claims of social transformation for the greater good being taken at face value ended some years ago.

Some AI sceptics worry that OpenAI and its rivals are drawing in hundreds of billions of investments for a product that doesn’t deliver on its promises, and largely doesn’t work. A smaller but still significant grouping worries it might work too well, creating existential-level risks previously only seen in sci-fi.

A third group looks at emerging problems like AI-driven plagiarism in schools, AI slop on social media, and AI psychosis among some heavy users, and sees a need for regulation right now. Osborne’s new role is likely to place him in the middle of these debates around the world.

The Musk Factor

The immediate political backdrop around generative AI couldn’t be much worse for Osborne. Elon Musk’s xAI caused something of a global political showdown over generating sexualised imagery of women and girls – some of the latter qualifying as child sexual abuse imagery.

Musk spent days either ignoring the issue, posting nonchalantly about it, or even seeming to encourage some of the behaviour, even as governments around the world were eventually spurred to act – with some going as far to temporarily block X in their territories. As more governments threatened to do the same, xAI eventually relented, and promised to act on the issue.

The controversy seems to have calmed down for the moment, though it will inevitably resurface. But it will raise headaches for OpenAI in particular, as the company had been planning to roll out its own “spicy mode” – allowing its models to engage in sexual conversations and generate adult images – this quarter.

Generally, companies who are seeking to win over major government contracts don’t also participate in the adult entertainment industry. Trying to relaunch your government contracting division at the same time as you launch a soft porn product is basically unprecedented. Trying to do that at the same time as a scandal over non-consensual nudity is what Yes, Minister would call courageous.

Osborne has to convince governments that OpenAI is taking them seriously as a market, that it wants to be a credible enterprise product, and that it won’t create political headaches for governments – even leaving the headaches of US geopolitical tensions aside. But he has to do that even while OpenAI runs in a dozen other directions too, launching consumer products, engaging in new copyright fights over its (largely failed) Sora product, and doing who knows what else.

Meet the competition

Osborne, then, has something of a tough sell. But it gets more difficult when you remember that he’s not acting alone – and weirdly, he’s competing against other British politicians, too.

One of those is former prime minister Tony Blair, who according to the excellent Democracy For Sale is, through the Tony Blair Institute, trying to develop his own AI products to sell to governments.

The …ambitious… initiative seems to have failed to have secured the backing of the TBI’s billionaire backer Larry Ellison, and honestly seems somewhat confused: when companies actually at the cutting edge of building AI models are offering these services, why should anyone go with a barely modified model produced by a think tank? Osborne, and OpenAI, probably don’t feel too worried on this front.

Anthropic, though, is a different story. Where OpenAI has aggressively chased the consumer market, Anthropic has been much more focused on the business-to-business and enterprise markets. CEO Dario Amodei estimates revenue for 2026 will “land somewhere between” $8 billion and $10 billion, and its market share in coding is now estimated at 54% – with business customers including Netflix, BMW, Citi, and pharmaceutical company Eli Lilly.

Anthropic even has a former British Conservative politician of its own on its roster, in the shape of former UK prime minister Rishi Sunak – though his role is described as part-time and “advisory”, in contrast to Osborne’s full-time role with OpenAI. Sunak’s pitch can look more focused, doesn’t have to explain away so many controversies, and comes with more enterprise endorsements.

Analysts suggest Anthropic’s more focused strategy, on companies actually able and willing to spend on AI at scale, is winning out over OpenAI’s bid to get retail customers paying for its product – especially given doing so means going toe-to-toe against Google and Meta.

Hiring Osborne to persuade governments to sign major deals with OpenAI might represent a major shift in strategy for the company – but some think it might already be too late.

Jon Reed, industry analyst and co-founder of diginomica, said he believed Osborne could provide good insight to OpenAI on how policy works, and how governments make decisions to procure major systems, but this would struggle to overcome other issues – firstly, that the public sector will likely spend less than private players, and secondly, that OpenAI is simply not best placed to persuade them to spend – especially given their role in ongoing controversies.

“I thought George Osborne’s comments on OpenAI and, I quote, ‘their mission to ensure the power of artificial intelligence is developed responsibly’ really stretched credibility,” he adds. “OpenAI has done an inarguably effective job at popularizing AI but ‘responsible AI’ has never dictated their actions, as a search on ‘OpenAI and suicides’ or ‘OpenAI and copyrights’ will surface multiple pending litigations from the move fast and break things playbook.”

What it all means

Most people reading this probably don’t care how George Osborne fares in his first big test of his new job. Some (reasonably!) won’t care too much about which AI company is doing best in which domain – though knowing what the different companies are trying to do, and how they explain their strategies, is always useful information.

If you just care about how AI is doing, and what it’s doing to us, this year’s Davos pitches matter for one basic reason – AI has successfully convinced investors that it’s the future, and it’s managed to attract huge sums of equity and debt investment. No new industry has ever been so successful as them at doing that.

What they’re finding more difficult is finding customers as enthusiastic as their investors. They have set expectations for their growth sky-high. They are now absolutely reliant on that growth – and given the sums of money at stake, it has to come soon. They cannot just wait it out if development is slower than anticipated. The cash crunch is simply too big.

When it comes to creating products people actually want to use and pay for, AI is struggling both with regular consumers and with enterprise (governments and big businesses).

Turning the potential of a new technology into actual reliable products is always difficult, and always takes longer than companies think it will. Reaching the mass market beyond enthusiastic early adopters always takes a lot of time and money.

Consumer AI is vastly better than sceptics who only tried it briefly a few years ago believe, but it’s also nowhere near as good or reliable as its most enthusiastic advocates claim. It’s a niche product.

More importantly, individual subscribers choosing to pay $20 a month are never going to repay the kind of investment going into AI. The tech giants need big corporations to invest in AI in a big way, and so far the technology simply doesn’t seem to be good enough to justify that. AI CEOs are having to claim it is boosting productivity in ways that can’t be measured, even as businesses that tried to use AI to replace staff admit it hasn’t really worked, at least so far.

Without AI investment, the US economy didn’t grow last year. AI stocks are propping up global stock markets. There are trillions of dollars at stake. And all of that needs AI to have a convincing business case, with real customers willing to spend serious money. That’s what’s at stake with these Davos pitches – and if Osborne and co aren’t successful (or if the products don’t work), it could be a bumpy year indeed.

And that’s before President Trump opens his mouth. What a year.