LightOn

Last updated

LightOn SA
Company type Public
Euronext Paris:  ALTAI
Industry Artificial intelligence
Founded2016;9 years ago (2016)
Founders
  • Igor Carron
  • Laurent Daudet
  • Florent Krzakala
  • Sylvain Gigan
Headquarters,
Key people
  • Igor Carron (Président Directeur Général)
  • Laurent Daudet (Directeur Général Délégué)
Products
Number of employees
41 (2024) [1]
Website lighton.ai

LightOn SA is a French company focused on artificial intelligence (AI), that develops generative AI solutions for enterprises. [2]

Contents

Its current product is the Paradigm platform, an on-prem/secure Enterprise generative AI platform.

In November 2024, it became Europe's first listed GenAI startup to launched an IPO on Euronext Growth Paris.

History

Early development (2016–2021)

LightOn was established in 2016 in Paris by French engineers and entrepreneurs Igor Carron, Laurent Daudet, Florent Krzakala and Sylvain Gigan. [3] The founders, with backgrounds in optics and AI, aimed to develop hardware solutions that could accelerate AI computations using optical processing. The company's first product, the Optical Processing Unit (OPU), [4] was launched in 2018. [5] [6] The approach aims to increase the efficiency of massive parallel processing in tasks that involve large-scale data such has Large Language Model. [7] [8]


In 2021, One of LightOn OPU was integrated in Jean Zay, one of the French National Centre for Scientific Research supercomputer, making it the first of the TOP500 world's most powerful computer using such hardware. [9]

Enterprise AI platform (2020–present)

Parallel to this focus, starting in 2020, LightOn trained several LLMs and released a few as Open Source [10] [11] [12] on several supercomputers. LightOn's main product is now an on-prem, secure Enterprise generative AI platform called Paradigm.

In 2024, LightOn launched an IPO [13] [14] and became Europe's first listed GenAI startup. [15]

Since the end of 2024, LightOn's solutions were implemented by various entities such as Ile-de-France region, Safran, Groupama, and the French Space Agency (CNES). The company also partnered with Orange Business and Hewlett Packard Enterprise to integrate its AI software with their cloud and other offerings. [1]

Language Models

LightOn, also released MonoQwen-Vision, the first visual document reranker [16] and in collaboration with AnswerAI, released ModernBERT, [17] a model that has been downloaded more than 20 million times. [18] [19]

In 2025, the LightOn Engineering team released several State of the Art models such as Reason-ModernColBERT, [20] GTE-ModernColBERT [21] and BioClinical ModernBERT. [22] BioClinical ModernBERT was developed in collaboration with the Dana-Farber Cancer Institute in collaboration with researchers at Harvard University, MIT, McGill University, Albany Medical College and Microsoft Research. The company also released as open source, the Ettin Suite [23] , the first state-of-the-art paired encoder-decoder models, developed with Johns Hopkins University.

Between 2020 and 2025, LightOn has trained more than 12 Large Language Models.

Advanced Retrieval Technology

To support production deployment of these models, LightOn developed a suite of open-source tools for multi-vector retrieval that enable real-time, updatable search systems. PyLate [24] simplifies multi-vector model training, while FastPlaid [25] provides a 554% performance improvement over Stanford PLAID [26] , making it suitable for enterprise RAG pipelines and recommendation systems. PyLate-rs [27] extends this capability to browser-based applications through WebAssembly compilation. These tools transform multi-vector retrieval from a research concept into production-ready technology for dynamic enterprise search and AI agents.

References

  1. 1 2 "LIGHTON | LIGHTON, a leading European player in generative AI for businesses, announces the approval of its registration document by the AMF as part of its planned IPO on Euronext Growth® Paris". 22 October 2024.
  2. "LightOn to become Europe's first listed GenAI startup with Paris IPO".
  3. Takahashi, Dean (19 December 2018). "LightOn raises $3.3 million for optics-based AI data processing". VentureBeat.
  4. "Computing with Lasers Could Power Up Genomics and AI".
  5. "Startup integrates optical processing into data center for AI operations". 28 February 2018.
  6. Hesslow, Daniel; Cappelli, Alessandro; Carron, Igor; Daudet, Laurent; Lafargue, Raphael; Muller, Kilian; Ohana, Ruben; Pariente, Gustave; Poli, Iacopo (2021). "Photonic co-processors in HPC: Using LightOn OPUs for Randomized Numerical Linear Algebra". 2021 IEEE Hot Chips 33 Symposium (HCS). pp. 1–9. arXiv: 2104.14429 . doi:10.1109/HCS52781.2021.9566948. ISBN   978-1-6654-1397-8.
  7. "LightOn Unveils LightOn Appliance, Photonic Co-processor".
  8. "LightOn at Neurips 2020".
  9. "This powerful supercomputer can now run on light instead of electric current". 25 December 2021.
  10. "PAGnol: Open Source French LLM". 19 April 2024.
  11. "RITA: Open Source Protein LLM". 19 April 2024.
  12. "Mambaoutai: Open Source State Space LLM". 21 November 2024.
  13. "France's LightOn says valued at $65 MLN in Euronext IPO". Reuters. 21 November 2024.
  14. "La start-up française d'IA LightOn veut s'introduire en bourse". Le Monde . 8 November 2024.
  15. "LightOn to become Europe's first listed GenAI startup with Paris IPO". Reuters. 8 November 2024.
  16. "MonoQwen-Vision, the first visual document reranker".
  17. "Finally, a Replacement for BERT". 19 December 2024.
  18. "All time downloads ModernBERT-Base". 14 June 2025.
  19. "All time downloads ModernBERT-Large". 14 June 2025.
  20. "Reason-ModernColBERT". 7 May 2025.
  21. "GTE-ModernColBERT". 26 May 2025.
  22. Sounack, Thomas; Davis, Joshua; Durieux, Brigitte; Chaffin, Antoine; Pollard, Tom J.; Lehman, Eric; Johnson, Alistair E. W.; McDermott, Matthew; Naumann, Tristan; Lindvall, Charlotta (2025). "A medical encoder named BioClinical ModernBERT". arXiv: 2506.10896 [cs.CL].
  23. "Ettin Suite: SoTA Paired Encoders and Decoders".
  24. "PyLate: Flexible Training and Retrieval for ColBERT Models".
  25. "FastPlaid: Bringing Multi-Vector Search to Production Scale".
  26. "PLAID: An Efficient Engine for Late Interaction Retrieval".
  27. "PyLate-rs: a lightweight tool to compute embeddings lightning-fast".