The recent investment of 105 million euros into French AI start-up, Mistral AI, has prompted industry commentators the world over to discuss the legitimacy of such ventures and the technology behind them. How is it that a newly formed company, just a couple of months old, can secure such a vast sum of money with no demonstrable product or technology to speak of? Is this another market bubble waiting to pop? Or is investment of this scale required to get this technology off the ground properly, for the benefit of “everyone”?
OpenAI, now world-renown, was not a company most people had heard of before November 2022, when it finally unleashed its flagship product, ChatGPT, onto the public. This launch would go on to become the most successful product launch in history, with 100 million users signing up within the first two months of its launch.
If you’ve ever used ChatGPT, you’ll likely know why there is such a buzz around this product in particular. Typing in a simple prompt and seeing it write out near-perfect responses is, for lack of a better word, magic. It’s not at all hard to see why such a product went as viral as it did, as people starting using it to write essays, books, legal documents, and everything else in-between; the use cases for ChatGPT seem to increase by the day, with seemingly infinite possibilities promised by OpenAI as it further develops its technology.
This unprecedented launch, coupled with the nature of how these systems are trained (partly on its use and feedback from users), has given OpenAI an incredible most-mover advantage over the rest of the industry. Millions of datapoints are fed into its algorithm every day enable the Company to refine its productive over time, based on user data. Despite this advantage, however, the cost of developing such a system is staggeringly prohibitive and, even with all these users (some of them paying), the Company is still recording mammoth losses.
OpenAI has received a swarth of lucrative investments from its now-partner, Microsoft. Publicly investing $1bn in the Company in 2019, it has since signed a multi-year, multi-billion-dollar investment with OpenAI going forwards, reportedly valued at $10bn.
Despite this, the Company made a loss of some $540m last year, owing mostly to development costs associated with its GPT software. These costs are mainly associated with the computing power required to develop such programmes, which is – it should go without saying – enormous.
Shifting the focus to the UK, the treasury announced it would be investing £900m into building its own AI supercomputer as part of its spring 2023 budget. As part of this initiative, the government plans to set up its own AI research body to explore the possibilities, practicalities, and risks associated with the technology. The computer, when built, will be one of the most powerful in the world, with aims to enhance the fields of cybersecurity, healthcare, and scientific research.
When examining the language used in the government’s announcement, it becomes clear that the move is more than just a desire to break ground on new technology. With the UK’s plan, as with France’s with Mistral AI, there is a large element of sovereignty and nationalist protectionism to the rhetoric.
As usual, the US is the country leading the way on AI. OpenAI, a US company with largely US backers, is the largest and best-known player in the market right now, which has restoked fears of overdependence on the states from many European nations.
For France’s part, Minstal AI’s president, Arthur Mensch, had the following to say:
“We are proud to initiate this global project from France, our home country, and to contribute, at our level, to the emergence of a credible new player in generative artificial intelligence from Europe.”
From this, as well as the fact that a majority of the investment is coming from French and European investors, it is clear that the 105 million euros invested so far reflect not just the desire to create a good product, but also the fact that France wants to create an independent and sovereign competitor to Silicon Valley.
Similarly, the UK’s £900m investment in its own supercomputer suggests that the government has been similarly swayed by this desire to create and retain its own cutting-edge industries. Adrian Joseph, BT’s Chief Data and Artificial Intelligence Officer, recently stated the following in a session with the Commons science and technology committee:
“We think there’s a risk that we in the UK lose out to the large tech companies, and possibly China, and get left behind… in areas of cybersecurity, of healthcare, and so on. It is a massive arms race that has been around for some time, but the heat has certainly been turned up most recently.”
So, factoring this line of thought into the equation, it becomes clear that these at-first seemingly large investments into relatively new fields make more sense than one might assume at first glance. Not only is it pioneering new – and potentially world-changing – technology, but this capital is also an investment into the future financial sovereignty of the countries in question.
And yet, Mistal AI’s main selling point is the credentials of its founders – hailing from both Meta and Google Deepmind’s AI research teams, both American firms. There is a slight irony to this, as the prestige afforded by these countries is, in essence, foreign. And yet, in a way, it demonstrates the need for this prestige to be dispersed more globally – the hope being that in 10 or 20 years’ time, experts from Mistral will be able to command the same legitimacy, but this time with a purely French pedigree.
It is easy to gawk at the idea of a four-week-old company receiving 105 million euros with nothing much to show for itself other than the credentials of its founders, but to do so would be to miss the bigger picture.
AI, due to its incredible potential for both good and societal harm, should not be thought about in the same way as other investments; it is not a simple product being manufactured sold. It may be better compared to something like the development of nuclear weapons. While the US was, again, the first to develop this technology, it wasn’t long before other nations had developed their own capabilities – military or otherwise. The desire for other nations to develop this technology wasn’t to outcompete the US, but also to create their own national security and energy systems.
Similarly, the use cases for AI are extremely broad. Legislation is still a long way behind, but the EU is making promising moves to become the first international body to draw up AI-based legislation. For now, it remains the case that the US leads the pack, by a substantial margin. If other countries want to stay relevant in this race, it is up to them to take the necessary – and extremely costly – steps towards AI self-reliance. In the grand scheme of things, 105 million euros is barely a drop in the ocean when it comes to AI investment capital – expect to see a lot more stories like this one over the coming years.