Clem Delangue 🤗's Recent LinkedIn Posts

We want to make Hugging Face buckets the S3 for agents! Let us know if you can use something like that and happy to work with you on your usecase

Nvidia officially surpassed Google as the biggest organization on Hugging Face with almost 4,000 of their team members. I'm officially calling it: Nvidia is the new king of American open-source AI!

Excited to welcome Salesforce as our latest enterprise customer! Already massive contributions (180 public models like Blip) and can't wait for what they'll do next! hf.co/salesforce

What starts as a weekend project can spark a future in STEM. Together with NVIDIA AI, we're bringing robotics and AI out of the lab and into the hands of builders of all ages 🤖. See how open, hands-on learning is creating a new generation of mission-driven innovators👆 https://del.ly/6044B6ySc4

I got no other than the Godmother of AI Fei-Fei Li as a coworker at the Hugging Face office today! So inspiring! https://lnkd.in/eDXj_jaW

The Department of War just learned the golden rule of AI: Not your weights, not your brain 🧠🔒

They tell you training and running AI model costs billions. That's true for a few frontier labs. But for most real-world use cases? Dramatically lower than you think thanks to open-source. Real examples from @HuggingFace's latest analysis: - Fine-tune a text classification model: <$2k - Train a leading image embedding model: < $7k - Train Deepseek OCR: < $100k - Train a leading machine translation model: <$500k Compare that to GPT-4.5 training (~$300M est.) And the truth is that you don't need a Formula 1 car to pick up groceries. Most tasks are solved just as well by smaller, efficient, targeted models. The mistake everyone makes? Starting with "what's the best AI model?" instead of "what do I need to do?" The future of AI is not just bigger models. It's cheaper, more customized, open models solving specific problems. Explore 100+ real model costs of training and deployment yourself in the study!

After Pinterest, Airbnb, Notion, Cursor, today it’s Intercom publicly sharing that they’re finding it better, cheaper, faster to use and train open models themselves rather than use APIs for many tasks. And hundreds of other companies are doing the same without sharing. Ultimately, I believe the majority of AI workflows will be in-house based on open-source (vs API). It took much more time than we anticipated but it’s happening now!

Someone should build a truly multi-model agent that switches between hundreds of different specialized models for different tasks (including even maybe local models ultimately?) Feels like it would increase speed, affordability & powerfulness by an order of magnitude for agents and doable with inference providers on Hugging Face and Hugging Face skills!

The Korean AI Heatmap is out. Very cool and expecting a lot more from there in 2026! https://lnkd.in/epPvJy47

After Pinterest, Airbnb, Notion, Cursor, today it’s Intercom publicly sharing that they’re finding it better, cheaper, faster to use and train open models themselves rather than use APIs for many tasks. And hundreds of other companies are doing the same without sharing. Ultimately, I believe the majority of AI workflows will be in-house based on open-source (vs API). It took much more time than we anticipated but it’s happening now!

We need more competition and innovation spreading in AI, not less. Otherwise, we'll end up in a world controlled by a few companies, which would be very scary!

Got to meet Nick Frost, the cofounder of Cohere in Miami today to celebrate their awesome release of an open-source Apache 2.0 Transcribe model that could be a whisper killer and already trending on Hugging Face! Cohere deserves much more visibility in the community as one of the leaders of North American open-source!

We’ve reached superintelligence!

Local AI is free, fast & secure! So today we're introducing hf-mount: attach any storage bucket, model or dataset from Hugging Face as a local filesystem. This is a game changer, as it allows you to attach remote storage that is 100x bigger than your local machine's disk. This is also perfect for Agentic storage!! Let's go!

Fun to see that in my absence, Hugging Face became the platform not only for AI builders but also for AI agents! The future is now!

Money doesn’t buy everything. Sometimes, culture, hard-work and team spirit wins! ⚽️🇫🇷 With my family, we've been bleeding red and gold (the club colors) since 1998. I was 10 years old when I watched my hometown club, RC Lens, lift the French title for the first time. That moment made me a lifelong fan, but it's the club's culture over the decades since that makes me truly proud. Right now, Lens is sitting at number one in the French Ligue 1. They are ahead of PSG, a team with a budget 10 times larger. Here’s what most people outside of France don’t know about Lens: 🏟️ The community is unmatched: The stadium capacity is 40,000. The population of the actual city is smaller than that. On match days, the entire town literally fits inside the arena with the most festive atmosphere in the country (loud singing, beer & french fries everywhere!) 🤝 Loyalty over everything: When the club was relegated to the 2nd division, they still had higher attendance than most 1st division clubs. The fans never left. 🧠 Exceptional leadership: The masterful work from Joseph Oughourlian, Benjamin Parrot, Jean Louis LECA and everyone at the club proves that smart, humble and consistent building beats aimless spending. Lens deserves so much more international recognition. It’s a masterclass in building a faithful, passionate community and proving that heart and hard work can go toe-to-toe with billions. Can't wait to see them in the Champions League next year. Allez Lens ❤️💛!

lol feels like the team introduced this bug just to troll users

We just made it dramatically easier for agents to read trending research papers on HF. Let's go AI powered research!

Just talked with Deirdre Bosa CNBC about NVIDIA and everything open-source AI! Some key points: - Nvidia is the new American open-source AI king - 30% of fortune 500 are using Hugging Face and our goal is to get to the majority of them by the end of the year - Agents will be much more open-source based than chatbots (ex OpenClaw) - Agents empower all to train, fine-tune, and run their own models based on open-source - We crossed 15M AI builders on HF and hope to have as much agents using the platform by the end of the year. Agents are the new users and customers of tech platforms

What is needed for startups and medium-size tech companies to contribute to open science and open-source AI more? When thinking about open-source, people usually think about big tech or academia but in my opinion, startups and medium-sized tech companies could be massive contributors and benefit a ton from sharing AI models, datasets, research,... (in terms of visibility, hiring, ability to transition into AI,...). We're seeing that a lot in China and looks like we might start to see it in the US too as showed by the fact that the two trending models on Hugging Face have been from these types of orgs (Fal and Lighttricks) Also interesting that Airbnb hired Ahmad Al-Dahle (former Llama lead) maybe to do more of that?

A la Une du premier numéro de l'année qui paraîtra le mercredi 14 janvier: le robot Reachy Mini, symbole d'une nouvelle vague de la robotique portée par l’IA. Difficile en effet de ne pas commencer 2026 sans revenir sur la fièvre qui s’est emparée des réseaux sociaux à la veille de noël autour de ce petit robot expressif et open source initialement conçu par Pollen Robotics et "powered by" la licorne franco-américaine Hugging Face. Avec 3000 commandes qui enregistrées début juillet en moins d'une semaine, restait à transformer un prototype en un produit industriel, ce qu’ont réussi en moins de 8 mois Clem Delangue 🤗, Matthieu Lapeyre et leurs équipes. Un tour de force qui justifiait le choix de notre Une, confirmé par le tabac qu'a fait Reachy Mini au CES 206 lors du keynote de Jensen Huang NVIDIA. Egalement au sommaire : - l’Italie entre dans l’histoire de la robotique humanoïde avec Generative Bionics, dont le robot GENE.01 dévoilé dans le Nevada a lui aussi fait forte impression, parmi les quelques 70 robots humanoïdes présents (dont plus de la moitié chinois); - un focus sur NEO, l'humanoïde de la startup norvégo-californienne 1X, le premier robot humanoïde domestique commercialisé à 20 000 dollars ; - la naissance d'UMA, la startup fondée par Remi Cadene qui s'est entouré d'une véritable "dream team" de roboticiens, avec une ambition clairement affichée : incarner une voie européenne de l’intelligence artificielle. Planète Robots consacre également Dossier spécial à l'industie Usine 5.0, à l'heure comment les robots, industriels mais aussi humanoïdes, s’invitent au cœur des nouvelles organisations industrielles. L'occasion de revenir sur évènements quin abordaient ce sujet : les Ateliers de la robonumérique de l'Agglo du Saint-Quentinois et BIG de Bpifrance. Toujours au sommaire : - un entretien avec Jacques Arrighi , DG de Fenwick-Linde France, autour de la robotique de l'intralogistique; - un portrait d’✨Alice LASSALLE, cofondatrice de Nimbl'Bot et roboticienne engagée; - 4 nouveaux territoires où l'IA s'invente autrement, analysés par Qian (SUNNY) ☀️ CHEN. Bonne lecture et bonne année 2026 ! Planète Robots n°95, janvier février 2026. • 100 pages • parution:  14 janvier 2026 • 8,80 € • Version digitale disponible sur le site planeterobots.com ainsi que sur Cafeyn, Viapresse. FAULHABER Drive Systems Mona Pierrette MORET STEPHANE LANCEA Fuzzy Logic Robotics Christophe EBEMBÉ Catherine Simon Xavier Dalloz Cloé Lacombe Florence Wallet Cédric Loubiat Pierre Fauveau Anne Asensio Martin Duterte Universal Robots ipsum tek Bernard FROMENT Daniele Pucci Océane Gonzalez Hugo FONTAN Didier Ribault ⊕GESSMANN Guillaume BONIFAS

Just released – new open models, data, and tools to advance AI across every industry. 🚀 🛠️ Startups can access these models for agents, robots, autonomous vehicles, and biomedical AI, plus one of the world’s largest open multimodal datasets — all available on GitHub, Hugging Face, and build.nvidia.com. Read how these new models are accelerating every industry ➡️ https://nvda.ws/4aNAMju

OpenAI just crossed 30,000 followers on Hugging Face with very strong adoption for gpt-oss as noted by Nathan Lambert In my opinion, they have the potential to become the number one organization in open-source, excited to see what they do in 2026!

Our biggest open-source repos are getting overwhelmed by AI slop with agents creating new pull request ~every 3 minutes. Fun new challenges in an agentic world!

Cowork but with local models not to send all your data to a remote cloud!

After almost 10 years of near nonstop grind, I’m taking 2 months of paternity leave to support my hero of a wife and welcome our twin daughters. Hugging Face is in great hands with the team and Julien Chaumond acting as interim CEO. Hope to return a changed man, to an even stronger HF, and to an AI field that’s more open and collaborative than ever!

Looks like it’s confirmed Cursor’s new model is based on Kimi! It reinforces a couple of things: - open-source keeps being the greatest competition enabler - another validation for Chinese open-source that is now the biggest force shaping the global AI stack - the frontier is no longer just about who trains from scratch, but who adapts, fine-tunes, and productizes fastest (seeing the same thing with OpenClaw for example)

I’m back. The girls and their superhero of a mom are doing great 😍😍😍 What did I miss?

can people send me better emails?

Use local models in your favorite agents!

When will Reachy Mini make it to space?

Closing the Week: Strategic Insights from CES 2026 and the Path Ahead Happy Friday to our Prospero Community, 🥂 As we conclude the first trading week of 2026, the global technology landscape has reached a pivotal moment. The conclusion of CES 2026 in Las Vegas has made one thing clear: 2026 is the year AI agents stop being just a user interface and start becoming a real asset and capability. The most significant takeaway from the NVIDIA DGX Spark + Reachy demonstration isn’t just the robotics, it is the shift toward Embodied AI. This represents a practical path where perception, reasoning, and action happen through developer-controlled, open components. We are moving toward a world of "Physical AI," where intelligent agents understand and interact with the physical world in real-time. At Prospero.ai, we recognize that this evolution isn't just about hardware; it is about the massive influx of data that drives market momentum. While the world watches these agents navigate physical spaces, our community uses AI to navigate the complexities of the stock market. The Prospero Advantage in 2026: ✅ Proven Performance: Our proprietary AI/ML engine has successfully outperformed the SPY for four consecutive years, delivering institutional-grade intelligence to retail portfolios. ✅ Signal over Noise: While the market reacts to CES headlines, our community relies on 10,000+ proprietary models to separate short-term hype from sustainable long-term growth. ✅ Institutional Data for All: We distill 100M+ data points, including Dark Pool activity and Options Sentiment, into 10 actionable signals, ensuring you move with the "smart money." Strategic Partnerships in the AI Era: For those exploring embodied AI or seeking to understand what AI agents can do for your specific enterprise, we recommend connecting with the experts at TD SYNNEX Global Computing Components. Their work in scaling these capabilities is essential for anyone looking to turn AI from a concept into a core asset. Weekend Strategy: The markets are closed, but the logic remains. Use this time to review your portfolio against our Upside Breakout and Net Options Sentiment signals. Let our technology handle the data processing so you can enter Monday with a clear, objective strategy.

What can a robot be used for? 🤖 In our opinion - for important tasks only. See what Reachy Mini from Hugging Face & Pollen Robotics is helping us with first. Happy 2026!

Anyone renting their Paris apartment this summer (2/3 bedrooms, AC)?

Did we witness the first Reachy Mini backers singer band?

Knowing the world-changing work that Ahmad led at Meta, I can't wait to see what he'll do at Airbnb with the support of the legendary Brian Chesky. Time to follow hf.co/airbnb if you aren't already!