Had a wonderful visit with His Excellency this week in Palo Alto! Thank you for the visit and the partnership all these years!

🦾 @SambaNovaAI & @Stanford's new ACE framework is giving AI memory & the ability to learn from every success & mistake. This is a shift from static models to self-improving systems. At SambaNova, we're building the infra to make this a reality. Paper: arxiv.org/abs/2510.04618

Honored to welcome @KanishkaNarayan MP, the UK’s Parliamentary Under-Secretary of State for AI and Online Safety, to SambaNova today. As he begins his vital new role, it’s inspiring to see the UK’s ambition and commitment to driving tech prosperity.

AI efficiency is important. Today, Google is sharing a technical paper detailing our comprehensive methodology for measuring the environmental impact of Gemini inference. We estimate that the median Gemini Apps text prompt uses 0.24 watt-hours of energy (equivalent to watching an average TV for ~nine seconds), and consumes 0.26 milliliters of water (about five drops) — figures that are substantially lower than many public estimates. At the same time, our AI systems are becoming more efficient through research innovations and software and hardware efficiency improvements. From May 2024 to May 2025, the energy footprint of the median Gemini Apps text prompt dropped by 33x, and the total carbon footprint dropped by 44x, through a combination of model efficiency improvements, machine utilization improvements and additional clean energy procurement, all while delivering higher quality responses. See the blog or technical paper for more about our methodology and ongoing efforts. Blog: cloud.google.com/blog/produc… Link to detailed paper: services.google.com/fh/files…

🐳 @deepseek_ai V3.1 on SambaCloud.... Stay tuned.

Really good read! Training using RL for pass@1 does not guarantee pass@k, I have been thinking of improving draft models for SD using #reinforcementlearning , assuming that GRPO training will help pass@k and drive overall acceptance rate. But this paper shows thats not necessary. h/t to @_Alex_Borghi_ for pointing to this! His depth of knowledge in RL is quite amazing, always has a paper that can point you in the right direction :).

Initial reaction to the AI Action Plan is net positive – it highlights its focus on U.S. leadership, innovation, and scale, signaling a promising step toward accelerating progress and supporting AI advancements.

SambaManaged, the 1st turnkey AI inference solution that lets you go from zero to AI in just 90 days. 🚀 ✅ Unmatched efficiency – More with less ✅ Rapid deployment – Weeks, not years ✅ Openmodel flexibility – Your rules, your AI ✅ Modular & scalable – Grow without limits ✅ Fully managed or self-service

We did it 🏆 SambaNova just won the @AI_Breakthrough Awards 2025 AI Chip Solution of the Year and we couldn't be more proud of our team. Designed for the AI era. Built different. 🦾

AI is getting smarter, but is it getting greener? 👀 🌱 Join @RodrigoLiang at @LDNTechWeek as he chats about scaling AI without scaling emissions! #LTW #LTW2025 #LondonTechWeek

"To lead, firms must move beyond red tape, empower people & drive innovation that delivers real returns." During his chat with @TomMackenzieTV, @RodrigoLiang discussed the future of AI in finance & how we're committed to helping the UK scale sustainably. Watch now 👇

Touchdown, London 🇬🇧 We're here at @LDNTechWeek where @RodrigoLiang will be dropping some AI knowledge on multiple panels. It's going to be a great week 🚀 #LTW #LTW2025 #LondonTechWeek

📣 Our AI platform is now available in @awsmarketplace. Seamlessly purchase & deploy our fast inference services alongside existing infrastructure in #AWS. ✅ Get super fast & more efficient inference ✅ Secure private connectivity via AWS PrivateLink ✅ Existing AWS billing

You asked — we answered 🦙 64K Context Length is now available on Llama 4 Maverick! 🔜 128K Context Length is on the way!

Redefining AI hardware for the next era! Discover how our intelligent chip & full-stack AI platform unlock new possibilities for developers-running open-source models with up to 5 trillion parameters. ➡️ Listen to the podcast here: bit.ly/4dakJeS @humanX_io

So fast it's almost unbearable 🐻 Our recently launched @Alibaba_Qwen's Qwen3 32B on SambaNova Cloud is officially the fastest! 🚀 282 t/s — verified by @ArtificialAnlys! 🌍 Multilingual mastery 🧠 Smarter reasoning 📈 Built for complex agent apps

It's an honor to be a finalist for the @EYnews 2025 Entrepreneur Of The Year Bay Area. This is a testament to @SambaNovaAI's relentless drive to innovate and create impact. Grateful to my team, partners, and supporters who make this journey possible.

📣 Announcing @Alibaba_Qwen's Qwen3 32B now on SambaNova Cloud! 🐻 Superior multilingual capabilities across 100+ languages 🐻 Advanced thinking & reasoning for complex agent apps 🐻 One of the best open source models for MCP Check it out now on ➡️ cloud.sambanova.ai/

⚡️ Welcome to the SambaNova Cloud Lightning Digest, your quick read on the latest dev updates! TLDR: 🦙 Llama 4 Scout & Maverick are Live! 🦾 New Models: E5 Mistral & Whisper 🔑 Multiple API Key Support 🗓️ Upcoming Events

A picture's worth a thousand words 📸 And SambaNova Cloud's blazing fast image analysis can use native multimodal #Llama4 to write a rap faster than you can retake that photo 🦾