TL;DR:
- Paris is becoming a thriving hub for AI startups, driven by talented individuals who have gained experience at major tech companies in Silicon Valley.
- The French AI ecosystem benefits from a large pool of skilled talent and a supportive environment for AI research.
- The availability of affordable AI engineers in Paris makes it an attractive location for startups compared to Silicon Valley.
- Three high-profile startups, Mistral, Nabla, and Dust, founded by alumni from Meta, OpenAI, and Google, are leading the charge in Paris.
- Each startup has a unique approach to building generative AI companies, with Mistral focusing on proprietary models, Nabla utilizing OpenAI’s GPT-4, and Dust exploring new use cases for LLMs.
- Paris’s AI renaissance does not require a mass return of talent from Silicon Valley but benefits from a wealth of available talent ready to innovate.
- While London has a larger share of AI talent in Europe, Paris’s growing ecosystem is attracting significant venture capital investments.
- Mistral believes that owning the underlying technology of generative AI models will be key to success, while others see potential in open-source models.
- The rise of AI startups in Paris contributes to Europe’s pursuit of tech sovereignty and offers an alternative to Silicon Valley’s dominance.
Main AI News:
Paris, known for its rich cultural heritage, is rapidly becoming a hub for groundbreaking artificial intelligence (AI) startups. Tucked away in the elegant 18th-century mansion at 22 Rue Chapon, in the heart of Paris’s trendy third arrondissement, a group of talented researchers from Meta and OpenAI are spearheading the development of the next generation of AI technologies.
These enterprising minds have honed their skills and expertise in the hallowed halls of Silicon Valley’s tech giants, but now they have chosen to spread their wings and establish their own ventures. Despite the size of the tech scene in France being a mere fraction of that in the United States, these innovators firmly believe that Paris offers the ideal environment for nurturing AI companies in the present day.
The allure of building strategic technology on home soil and reducing dependence on American tech has captivated European nations, and France is at the forefront of this movement. Alex Lebrun, co-founder of AI healthtech firm Nabla and former engineer in Meta’s AI division, shares his perspective, “I had two AI startups before, and I moved to the US for those. If you were not in [San Francisco], you weren’t ‘somebody.’ Now it’s actually really changed, and Paris is a big hotbed.”
Within the same office building, Nabla, leveraging large language models (LLMs) to assist clinicians and streamline administrative tasks, collaborates with Dust, an AI startup co-founded by former OpenAI research engineer Stanislas Polu. Just a stone’s throw away, Mistral, a generative AI startup established by former employees of Meta and Google’s AI labs, is strategizing how to utilize its €105 million seed funding effectively.
Laying the Foundation for Success
Contrary to popular belief, France’s AI renaissance does not necessitate a massive influx of prodigal sons and daughters returning from Silicon Valley, according to Polu. “I don’t think it’s about people coming back [from Silicon Valley]; I think they’ve always been here. Paris has a lot of traction these days because we have a very large pool of very strong talent in AI research,” he asserts.
Paris boasts the presence of research labs operated by Facebook and Google, as well as a surplus of highly skilled engineers and scientists, as Lebrun emphasizes. He adds that establishing the business in Paris makes economic sense, stating, “80% of our users are in the US, but we are based in Paris because AI engineers are good and affordable compared to the States.”
Nabla reveals that the current market rate salary for a junior ML engineer in Paris is approximately €60,000, while a senior ML engineer commands around €80,000 to €90,000. In contrast, Silicon Valley salaries are 2.5 to 3 times higher, according to data from the recruitment platform Wellfound (which suggests an average engineer salary of $268,000 at OpenAI).
Polu attributes France’s deep talent pool to the strong mathematical focus of the country’s engineering degrees and a government initiative called the Cifre system. This scheme incentivizes companies to hire PhD students by contributing to their salary payments, with funding provided by France’s Ministry of Higher Education, Research, and Innovation.
In Polu’s words, “Basically, you get PhD students for free. There’s been a lot of this applied thesis research happening inside the laboratories of DeepMind [owned by Google] and Meta AI, so you get a lot of these PhD students getting access to compute and being able to do actual research during their thesis.”
Attracting Top Talent from Big Tech
The presence of three prominent startups in Paris, founded by alumni from Meta, OpenAI, and Google, merely scratches the surface of the city’s potential, says Polu. “DeepMind and Meta, as I see it from the outside, seem to be having a harder time retaining talent these days,” he observes. “In the US, OpenAI is absorbing all that talent like a black hole, but they haven’t quite yet figured out how to do that in Europe. And so you end up with a pool of available talent in Paris and London — but in Paris in particular — that is ready to explore and do stuff.”
Sifted reached out to DeepMind and Meta for comment but did not receive a response before publication.
Polu’s optimism regarding Paris may need to be taken with a grain of salt. Recent research from Sequoia reveals that London boasts over three times the share of AI talent in Europe compared to Paris (12.29% versus 3.81%). Furthermore, OpenAI has chosen the UK capital as the location for its first office outside the US.
Nonetheless, Mistral, Nabla, and Dust have successfully secured funding from world-renowned venture capital firms like Sequoia and Lightspeed, as well as prominent business figures such as Xavier Niel and former Google CEO Eric Schmidt. The backing of such luminaries underscores the significance of a Big Tech background for aspiring founders.
Diverse Approaches, Shared Ambition
While these startups share the same home city and possess experience from the world’s leading AI laboratories, their strategies for building generative AI companies differ significantly.
Mistral, at one end of the spectrum, secured substantial funding just four weeks after its launch, even before having a sizeable team or a tangible product to offer. This substantial investment will primarily be allocated to training a proprietary AI LLM, a process requiring tens of millions of euros due to the massive computational power it demands.
Nabla, on the other hand, is developing its product using OpenAI’s GPT-4 model and hopes to transition to open-source LLMs as their quality improves. By leveraging third-party models, Nabla has been able to focus on product development rather than fundamental research. Its AI clinician companion is already being utilized by doctors in hospitals across the United States and Europe, including a pilot program with the British Ramsay Group.
Dust falls somewhere in between these approaches. Like Nabla, this startup relies on third-party LLMs, with Polu asserting that models such as GPT-4 still have significant untapped potential. “It’s the realization from working at OpenAI that those models are already good enough to provide economic value, and yet they’re pretty under-deployed in the world,” he explains. “The conclusion from that is really that it’s probably as much a research problem as it is a product problem at this stage. And so the idea was to try to focus on the product problem.“
While Mistral concentrates on creating AI-powered products for enterprise clients, Dust’s focus lies in identifying novel applications for LLMs, rather than building new models. The company has already partnered with six “design partners” to explore innovative uses for generative AI in the workplace.
Polu likens the current stage to the early days of smartphone apps, where to-do lists flooded the market before realizing the true potential of platforms like Uber and Airbnb. He suggests, “I think this is where we are right now with LLMs. We are in that new paradigm, we have that new abstraction, as a new platform is emerging.”
The Quest for Sovereignty
While Dust and Nabla adopt a product-first approach to building generative AI companies, Mistral believes that the true value in the emerging generative AI market lies in owning the underlying technology — the generative models themselves.
Lebrun holds a contrasting view, predicting that free, open-source LLMs will be able to compete with GPT-4 within a year. He remarks, “I think elements are new infrastructure, like cloud [computing]. Everybody will have access to them. Your LLM is not your moat.”
However, Mistral is banking on the growing importance of European tech sovereignty. The company asserts in a pitch memo seen by Sifted that “most of the value in the emerging generative AI market will be located in the hard-to-make technology, i.e., the generative models themselves.”
Lebrun concurs that Paris’s AI renaissance is music to the ears of France’s political class, which lamented the departure of the country’s AI talent when companies like Hugging Face established themselves in the US. “Ten years ago, politicians in France said, ‘All the brains are leaving for the US!’ And I told them: ‘Wait, they leave for the US, they make money, they learn, and then they come back to start ambitious startups in France.’ It’s exactly what’s happening with Mistral, Nabla, and Dust.“
It is worth noting, albeit a tad ironic, that these rising stars of European AI are being bolstered by venture capital investments from the US. Nonetheless, the emergence of more AI companies on the continent can only be beneficial in Europe’s quest to avoid succumbing to a Silicon Valley monopoly during this era of digital disruption.
Conclusion:
The emergence of Paris as a prominent AI startup ecosystem driven by Big Tech alumni signifies a significant shift in the market. The presence of talented individuals, a supportive research environment, and access to affordable AI engineers make Paris an attractive destination. The success of startups like Mistral, Nabla, and Dust, coupled with venture capital investments, showcases the potential of the European AI market. It also highlights the ongoing debate between proprietary models and open-source approaches, as well as the importance of technological ownership and the pursuit of tech sovereignty in the AI landscape.
