Intro: Each month, the Digital Diplomats cut through the clichés about the Silicon Valley and French approach to innovation to bring a clear analysis of a major tech story or trend that reveals the relative strengths and weaknesses of each ecosystem. Our goal is to give entrepreneurs the perspective they need to succeed in both markets.
Marjolaine Catil, an Investment Director at Newfund, a VC firm based in Paris and Silicon Valley. She leads Newfund’s “Road to the USA” initiative for French entrepreneurs considering U.S. expansion.*
Chris O’Brien, a Silicon Valley veteran journalist who is the founder and editor of The French Tech Journal newsletter.
Chris O’Brien: Marjolaine, there’s something extraordinary happening right now in Silicon Valley with the conversation around the impact of AI on startups. We’ve seen a wave of ultra-lean AI startups in Silicon Valley hitting massive revenue milestones with minimal headcount. It’s become such a thing that even a mainstream publication like The New York Times is talking about it.
Marjolaine Catil: You’re right. Ultra-lean startup success is now a meme, with techies excitedly sharing lists that show how companies like Anysphere, a start-up that makes the coding software Cursor, hit $100 million in annual recurring revenue in less than two years with just 20 employees, and ElevenLabs, an A.I. voice start-up, did the same with around 50 workers.
When we look at the evolution of employees needed to reach $100M ARR, productivity gains have dramatically increased across distinct technological eras:
In the early 2000s, what we can call “human-powered platforms” like LinkedIn needed 900 employees and Shopify required over 600 to reach $100M ARR. These companies operated with traditional license/subscription models requiring large sales teams, extensive customer support, and substantial internal IT departments. Scaling meant linear headcount growth.
The 2010s with “product-led companies” brought some improvements—Slack and Algolia needed around 250 employees. The automations around the product reduced reliance on sales teams while cloud-based delivery decreased infrastructure headcount. However, scaling still required substantial personnel for more complex customer interactions.
What we're witnessing with “AI-powered startups” since 2022 completely rewrites the playbook. Today's leanest AI startups achieve 0.2 employees per million in ARR, versus 3-7 employees just a decade ago—a 15-25x efficiency improvement that fundamentally transforms venture economics.
This unprecedented leap stems from merging labor and software into one market, replacing specialized human roles with specialized AI colleagues that operate across entire workflows. AI handles both organization and execution of tasks, effectively delivering human-quality service with software-level margins.
Founders in Silicon Valley aren't just asking "How do I raise funding?" but "How do I scale with fewer people?" It's a fundamental mindset shift reshaping the startup landscape that all founders need to take into account.
Chris O’Brien: I had been in Silicon Valley for a few years when the “Lean Startup” movement started almost 2 decades ago. It emerged in the wake of the excess of the dot-com boom when former entrepreneur-turned-academic Steve Blank wrote a book with a rather innocuous title: The Four Steps to the Epiphany: Successful Strategies for Products that Win. One of his students at UC Berkeley, Eric Ries, was so impressed with Blank’s customer-centric approach to entrepreneurship that he wrote his own book: The Lean Startup.
Ries and Blank introduced many of the terms that have become the lingua franca of startups: minimal viable product, split or A/B testing, build-measure-learn, and pivot, for example.
What we’re seeing now with AI, it’s like Lean Startups on Steroids.
A decade ago, if you raised a Series A, you were expected to hire aggressively. Now, hiring too fast is seen as a red flag. The new status symbol isn’t headcount, it’s revenue per employee.
And we may be just at the start. Even now companies are pushing to compress those ratios. Sweden’s Lovable reportedly hit $17m ARR in 3 months with less than 20 people. By the way, its AI-coding platform will make it even easier for other companies to do more with fewer humans.
Do AI-First Startups Have Sustainable Moats?
Marjolaine: A crucial question arises when evaluating these Full-AI models: what are their entry barriers? What prevents competitors from replicating or surpassing their performance? For investors, churn rate becomes an even more critical metric, especially as continuous R&D investments are necessary to maintain a competitive edge in the rapidly evolving AI landscape.
We are already witnessing an intense race among Large Language Model (LLM) providers, but this competition will likely expand to other Full-AI models.
Companies are slashing prices to gain market share, with some even offering their products for free. OpenAI cut the cost of GPT-4o tokens from $36 to just $4 per million tokens, while DeepSeek and Perplexity DeepSearch have become completely free. They are also competing on the performance of their model where only the sky seems to be the limit.
As foundational models become more accessible and open-source alternatives mature, differentiation will shift from raw model performance to specialization, domain expertise, and business execution. This shift will intensify price pressure, accelerate innovation, and increase competition across AI-powered industries.
Chris: These rocketship rides to $100M ARR are impressive, but I agree there are real questions about what kind of competitive moats these companies have and whether they can demonstrate that this growth is sustainable.
Consumers and enterprises are trying to figure out all these new paradigms and which platforms can bring real meaningful value, so there’s a ton of experimentation going on. So they sign up, try something for a while, then cancel to try the next new thing.
I’m guilty of that. Looking at my phone, I have a ridiculous number of GenAI apps – ChatGPT, Microsoft Copilot, Perplexity, Le Chat (from Mistral AI), and Claude. For some, I am paying a subscription. Perplexity Pro is free via a partnership with my telecom provider. My GenAI monthly bill is higher than my cable bill used to be. Eventually, I’m going to have to prune some of these back. It’s absurd.
At the same time, if these companies know their growth is fragile, that seems like it might be even more incentive to keep headcounts low.
Of course, we’re talking a lot about Silicon Valley here. Are we seeing the same thing in France and Europe?
Different Approaches: The U.S. Full-AI Model vs. France’s Research-Driven Model
Marjolaine: When we analyzed Newfund French and American portfolio companies last year, both ecosystems recognized AI's potential for productivity gains. However, significant differences emerged in implementation and results, especially because not everything can be automatized equally through AI.
In the US, particularly in Silicon Valley, we see what I call the Full-AI Model – startups designed from day one to leverage AI for maximum automation with minimal human intervention. These companies optimize for exponential scaling with tiny teams, focusing on pure software plays with high margins and low operational complexity. Their business models prioritize revenue-per-employee and often take a "winner-takes-most" approach to market domination.
In France, we observe what might be called the Research-Driven Model – startups tackling complex problems that require substantial human expertise alongside AI capabilities: climate tech, healthcare, biotech, manufacturing, and deep research. These sectors face regulatory hurdles, physical constraints, and scientific challenges that demand specialized human talent.
Take Wandercraft which develops exoskeletons or Naarea, a French nuclear AI startup. They're not running a chatbot or a SaaS platform like Midjourney—they need physicists, engineers, and compliance specialists. The AI improves their operations but doesn't eliminate the need for human expertise.
Chris: These differences aren't just about business strategy – they reflect deeper attitudes toward AI's role in society. Silicon Valley's approach often asks: "How can AI replace human work entirely?" The French approach more commonly asks: "How can AI enhance human capabilities in tackling meaningful problems?"
Marjolaine: You're right. AI can actually be an opportunity for complex French companies to finally become scalable products.
Traditionally, deep human expertise required a service-based approach, which hindered virtuous scaling for French startups as their gross margin was capped at 30% compared to 80% for the “Software-as-a-service” (Saas). Now they can embed their unique knowledge into AI models, delivering high-quality service at scale while maintaining a strong competitive edge—something that is often harder for fast-moving American startups to sustain. The area of "Service-as-a-Software" is on, and French startups could be at the forefront of it.
Recruitment in the Age of AI: Who’s Still in Demand?
Marjolaine: This shift in productivity has a big impact on recruitment. The broader data from the US Bureau of Labor Statistics reflects a similar evolution to the one demonstrated by the trendiest startups: companies are starting leaner with an average of 3.5 workers in 2023 vs. 7.3 in 1998. And they are scaling more efficiently.
How are French founders reacting to this trend? We were discussing this recently with Mathias Frachon, co-founder of The Product Crew, a French recruitment agency. He told me: “France is merely following the US trend with a time lag; I don't see any fundamental difference. The majority of founders now want to reach $100M ARR with a maximum team of 50 people.”
While France follows the US-led trend with some delay, the direction is clear—the age of efficient growth demands fewer, but increasingly exceptional, people. It will mean scarcity of these profiles and startups will need to pay them a lot more. According to Mathias: “There's a fundamental rejection of traditional management and large teams in favor of small, highly intelligent groups that leverage AI to become ultra-productive."
Startups seek ultra-senior developers who are autonomous, "plug-and-play," and combine sector expertise, top-tier academics, and a strong business impact—rewarded with generous pay when they fit the bill.
Startups are now looking for humans who can bring extra value on top of AI and this approach creates a bifurcated talent market: extraordinary opportunities for exceptional senior talent who can drive outsized impact, alongside diminishing prospects for those earlier in their careers.
Chris: In the US, companies are no longer trying to reassure the public about the impact of AI on jobs. In fact, in the U.S. they’re so confident that it’s an asset that they’ve become quite brazen about it. There were those ads from Artisans in San Francisco, for instance, and that job listing from Firecrawl:
French companies will have to adapt if they want to compete globally. But that will be a touchy subject here because a pillar of the generous government support for La French Tech is that it will be an engine for job growth.
The government and associations like France Digitale like to emphasize the potential for job creation when they talk about AI. Last year, a report submitted to President Macron by the Committee on Generative Artificial Intelligence argued that AI will increase employment with proper training and education.
From a cultural perspective, it would be tough to imagine a French startup bragging about how few employees it needed to reach $100M ARR. And given how much money flows to venture firms here via the Fund of Funds operated by state bank Bpifrance, it could be problematic if French VCs become too vocal about doing more with fewer employees.
So what’s the advice for a French founder today? How should they think about AI-driven efficiencies as they scale?
AI-driven efficiency: What Founders Should Consider
Marjolaine Catil: For French founders navigating this evolving landscape, here's my practical advice on approaching AI-driven efficiency:
If you're building a pure AI/software startup:
- Embrace the ultra-lean model from day one – design your organization to maximize automation of core functions.
- Focus on revenue-per-employee – meaning stay lean with less than 30 people up to $10M, then target $1M per employee
- Consider the US market for faster scaling and funding access, but maintain R&D in France to leverage local talent.
- Be ready to pivot your product regularly as there is less defensibility.
If you're in deep-tech or industrial AI:
- Use French advantages – research funding, talent pools, and government support.
- Your team will necessarily be larger – but automate everything you can.
- Articulate your hiring strategy clearly to investors – explain why specific domain experts and researchers are essential.
- Ensure your focus presents high entry barriers to AI: access to data, complex distribution, and physical constraints.
Chris: At its core, this shift represents a fundamental rethinking of how value is created in the digital economy. The new question for founders isn't just "How do I hire the best people?" but "Do I need people at all for this function?"
For French entrepreneurs specifically, I'd add:
- Don't feel pressured to mindlessly copy Silicon Valley's ultra-lean approach if your sector genuinely requires human expertise.
- But also don't hide behind European cultural differences to justify inefficient team structures.
- Find your unique balance between leveraging AI for efficiency and building human teams with the judgment to direct that technology.
- Recognize that French startups can compete globally by combining technical excellence with strategic deployment of AI.
The new question for founders isn’t how do I hire the best people? It’s do I need people at all? But even if AI is changing how startups scale, not every industry follows the same playbook.
Marjolaine: I believe we're still in the early stages of understanding how AI will transform startup operations and team structures. The companies that thrive will be those that find the right balance between leveraging technology for efficiency and building human teams with the judgment and creativity to direct that technology.
This article is part of the Digital Diplomats series, bringing insights from both sides of the Atlantic to help entrepreneurs navigate the global tech ecosystem. For more French tech news, contact Chris O’Brien or subscribe to The French Tech Journal. For more information on startup fundraising, contact Marjolaine Catil, Investment Director at Newfund.
