Guillaume Lample
Co-Founder & Chief Scientist at mistral
Reviewed Updated Jun 29, 2026This profile is AI-generated. If you spot an error, please help us fix it by sharing a URL to the correct information.
Background
Guillaume Lample is co-founder and Chief Scientist (also styled Chief Science Officer) of Mistral AI, the Paris-based open-weight foundation model company he started in April 2023 with Arthur Mensch (CEO) and Timothée Lacroix (CTO) 12. He studied at École Polytechnique (X2011), where he first met Arthur Mensch, then earned a Master of Language Technologies from Carnegie Mellon University and a PhD in computer science from Sorbonne Université (UPMC) carried out at Facebook AI Research 34.
Before founding Mistral, Lample spent roughly seven years as a research scientist at Meta’s Facebook AI Research lab (FAIR) in Paris, where he worked with Timothée Lacroix on multilingual language modeling, unsupervised machine translation, and symbolic mathematics 34. He is the last-listed (senior) author on Meta’s February 2023 LLaMA paper, “LLaMA: Open and Efficient Foundation Language Models” 5. His Google Scholar page lists an h-index of 42 with approximately 67,000 citations, anchored by LLaMA, Mistral 7B, Mixtral of Experts, and earlier work on cross-lingual pretraining and named entity recognition 6.
Forbes lists Lample as a billionaire with an estimated net worth of $1.8 billion as of 2026, derived primarily from his roughly 13% founder stake in Mistral after the company’s September 2025 funding round at a $14 billion (€12 billion) valuation 27.
Stated Thesis
Lample is an operator and research leader, not a self-described investor. His public statements describe Mistral’s product and research direction rather than a personal investment thesis. In a December 2025 interview with VentureBeat tied to the Mistral 3 launch, he framed the firm’s approach around open-weight models and efficient deployment, saying “the gap between closed and open source is getting smaller, because more and more people are contributing to open source, which is great” 8. He has also publicly emphasized European access to frontier models and the practical value of smaller, fine-tuned models for most enterprise tasks 8.
No personal investment thesis statements from Lample (e.g., stage, check size, sector targeting for outside investments) could be sourced.
Inferred Thesis
Lample does not have a public personal investment portfolio, so a conventional inferred thesis cannot be computed from counted angel data. What can be inferred from his operator role at Mistral and his published research is the technical worldview he is putting capital and time behind:
- Open-weight foundation models over closed APIs. Mistral’s product line — from Mistral 7B (2023) through Mixtral, Mistral Large, Magistral, and Mistral 3 (2025) — has consistently released open-weight variants alongside commercial offerings, with Lample personally announcing several of these releases 9108.
- European AI sovereignty. Mistral’s positioning, and Lample’s own remarks, emphasize Europe-based frontier model development as a strategic necessity; Mistral’s largest backer is now Dutch semiconductor-equipment maker ASML, holding approximately 11% 1112.
- Smaller, fine-tuned models for production. Lample has publicly argued that “in more than 90% of cases, a small model can do the job, especially if it’s fine-tuned,” underpinning Mistral’s investment in compact Ministral/Mistral 3 models targeted at laptops, drones, and edge devices 8.
- Multilingual capability as a core competence. His prior FAIR research (XLM cross-lingual pretraining, unsupervised machine translation) and Mistral’s marketing of native fluency across English, French, German, Spanish, and Italian point to multilingual performance as a durable differentiator 49.
Important caveat: No independently sourced personal angel investments were found. The above is not a venture thesis; it is the technical thesis Mistral itself executes against, with Lample as a co-founder and Chief Scientist.
Portfolio
No independently sourced personal angel investments by Guillaume Lample could be verified at the time of research. Lample’s only verified equity exposure is his founder stake in Mistral AI itself.
| Company | Year | Stage | Source |
|---|---|---|---|
| Mistral AI | 2023 | Co-Founder (Seed onwards) | 113 |
If Lample later begins disclosing angel investments or board roles outside Mistral, this section should be repopulated.
In Their Own Words
“The gap between closed and open source is getting smaller, because more and more people are contributing to open source, which is great.”
— Guillaume Lample, interview with VentureBeat tied to the Mistral 3 launch, December 2025 8
On the role of small models in production deployment, Lample has stated that in more than 90% of cases a small, fine-tuned model is sufficient, and that smaller models bring practical benefits in cost, latency, privacy, and reliability 8.
On European access to advanced models, Lample has publicly argued that with AGI or superintelligence potentially arriving in the near future it is critical that Europe maintain independent access to frontier models rather than depend solely on US providers 8.
Lample also routinely announces Mistral product releases on X (formerly Twitter) under @GuillaumeLample, where he has posted launch threads for Mistral Large, Mistral OCR, Magistral, and other models 14.
What Founders Say
No independently sourced founder testimonials found. Lample is not an active angel investor and does not have a public record of backing external founders.
Sources
-
Mistral AI, “About Mistral.” https://mistral.ai/about/ — accessed June 2026. ↩↩
-
Forbes profile, “Guillaume Lample.” https://www.forbes.com/profile/guillaume-lample/ — accessed June 2026. ↩↩
-
Iain Martin, “How France’s Mistral Built A $14 Billion AI Empire By Not Being American,” Forbes, April 16, 2026. https://www.forbes.com/sites/iainmartin/2026/04/16/how-frances-mistral-built-a-14-billion-ai-empire-by-not-being-american/ — accessed June 2026. ↩↩
-
Guillaume Lample LinkedIn profile. https://fr.linkedin.com/in/guillaume-lample-7821095b — accessed June 2026. ↩↩↩
-
Hugo Touvron et al., “LLaMA: Open and Efficient Foundation Language Models,” arXiv:2302.13971, submitted February 27, 2023. https://arxiv.org/abs/2302.1397 — accessed June 2026. ↩
-
Google Scholar, “Guillaume Lample.” https://scholar.google.com/citations?user=H7sVDmIAAAAJ&hl=en — accessed June 2026. ↩
-
Bloomberg, “Mistral Set for $14 Billion Valuation With New Funding Round,” September 3, 2025. https://www.bloomberg.com/news/articles/2025-09-03/mistral-set-for-14-billion-valuation-with-new-funding-round — accessed June 2026. ↩
-
VentureBeat, “Mistral launches Mistral 3, a family of open models designed to run on laptops, drones, and edge devices,” December 2025. https://venturebeat.com/ai/mistral-launches-mistral-3-a-family-of-open-models-designed-to-run-on — accessed June 2026. ↩↩↩↩↩↩↩
-
Mistral AI, “Mistral AI raises 1.7B€ to accelerate technological progress with AI,” September 9, 2025. https://mistral.ai/news/mistral-ai-raises-1-7-b-to-accelerate-technological-progress-with-ai/ — accessed June 2026. ↩↩
-
TechCrunch, “Mistral closes in on Big AI rivals with Mistral 3 open-weight frontier and small models,” December 2, 2025. https://techcrunch.com/2025/12/02/mistral-closes-in-on-big-ai-rivals-with-mistral-3-open-weight-frontier-and-small-models/ — accessed June 2026. ↩
-
ASML, “ASML, Mistral AI enter strategic partnership,” September 9, 2025. https://www.asml.com/en/news/press-releases/2025/asml-mistral-ai-enter-strategic-partnership — accessed June 2026. ↩
-
Sifted, “Mistral confirms €1.7bn Series C led by ASML.” https://sifted.eu/articles/mistral-confirms-series-c-asml — accessed June 2026. ↩
-
TechCrunch, “France’s Mistral AI blows in with a $113M seed round at a $260M valuation to take on OpenAI,” June 13, 2023. https://techcrunch.com/2023/06/13/frances-mistral-ai-blows-in-with-a-113m-seed-round-at-a-260m-valuation-to-take-on-openai/ — accessed June 2026. ↩
-
Guillaume Lample on X (Twitter), @GuillaumeLample. https://x.com/GuillaumeLample — accessed June 2026. ↩