Markus Holtmanns

Developer Community Manager at @NVIDIA

[NVIDIA GTC @NVIDIAGTC]

Owning an AI agent means owning more than the model.

Production agents need open models, secure runtimes, clear permissions, and tools that allow teams to specialize and optimize them for real workloads.

At GTC Berlin, Kari Briski and Justin Boitano share how enterprises can build, deploy, and optimize agents they control.

🎙️ Agentic AI in Production: From Open Models to Agents You Own: https://t.co/xWti1iwgdb
📍 #NVIDIAGTC Berlin | October 20–22, 2026
🗓️ Explore recommended sessions: https://t.co/ZJzkiMYNLJ

[Markus Holtmanns @HoltmannsMarkus]

We just released a guide on how to build a more secure sandboxed AI agent using OpenClaw and NVIDIA NemoClaw. Enable your agent with Nemotron 3 Super and run it on DGX Spark for a powerful local deployment.

Start now:  https://t.co/Po3Xe8zADX https://t.co/49XKwIsxiS

[Markus Holtmanns @HoltmannsMarkus]

Introducing NVIDIA OpenShell, an open source runtime for building autonomous, self-evolving agents more safely. 🦞

The infrastructure to run claws more safely didn’t exist, until now.
Take a Technical Deep Dive 👉

https://t.co/sA2JlBiAaQ

[Markus Holtmanns @HoltmannsMarkus]

NVIDIA NemoClaw runs OpenClaw always-on assistants with a single command, installs the secure NVIDIA OpenShell runtime, part of NVIDIA Agent Toolkit, and uses open source models for a trusted, autonomous environment.🦞

Start now: https://t.co/sE82yMBNnf https://t.co/rgOyFJO58q

[Chris 🇨🇦 @llm_wizard]

ALL DAY LIVESTREAM WED MARCH 18th 10AM PT ALL CHANNELS

[OpenClaw🦞 @openclaw]

huge shoutout to @nvidia for lending engineers to help triage our security advisories 🛡️🦞

open source security hits different when GPU companies show up to help

[Mark Heaps @lifebypixels]

👀 GTC DEV COMMUNITY - JOIN US! 🧠
@NVIDIAAIDev
Location: GTC Park
Tues, 3/17 | 10 am–12 pm
Live Agent Gauntlet 🤖

Location: GuildHouse
Tues, 3/17 | 11 am–6 pm
Hack for Impact 🦞

Weds, 3/18 | 9 am–2 pm
Hack to Create 🎨

Thurs, 3/19 | 8 am–12 pm
Vibe Hack 😎

[Markus Holtmanns @HoltmannsMarkus]

Who’s dominating the NVFP4 leaderboard? 📊
We’re wrapping up the Kernel Hackathon with a celebration during #GTC26!
Come for the GPUMode winners announcement, stay for the deep-dive kernel optimization talk and pure dev energy.

👇Claim your spot:
https://t.co/1kDZapI51z

[NVIDIA AI Developer @NVIDIAAIDev]

If you can’t make it to #NVIDIAGTC in person, this is your way inside.

We’re bringing you full-day coverage on our Developer Community Livestream packed with show floor demos, interviews with top builders, and live looks at the parts of GTC most people don’t get to see.

Plus, you get to shape what we cover. Just use #ShowMeGTC and drop your suggestions on social or in chat for our crew to show live on stream.

🗓️ March 18, 10 AM PST
Add to your calendar: https://t.co/w4lKfeSnkV

[Markus Holtmanns @HoltmannsMarkus]

30M downloads and counting for the NVIDIA Nemotron family on Hugging Face 🤗 https://t.co/zuFTsXUlGG

[Markus Holtmanns @HoltmannsMarkus]

Join NVIDIA and @DictaTools to discover Dicta‑LM 3.0, a Hebrew reasoning model, with transparent data pipelines and optimized training workflows. Watch on February 10, 2 p.m. CET on YouTube, LinkedIn, and X. https://t.co/XnTEvu8DnD

[Markus Holtmanns @HoltmannsMarkus]

The new NVFP4 precision version of NVIDIA Nemotron 3 Nano delivers up to 4× higher throughput on NVIDIA Blackwell — with up to 99.4% BF16 accuracy using our new Quantization Aware Distillation method.
https://t.co/OokZmEnp69

[NVIDIA GTC @NVIDIAGTC]

Want a chance to attend #NVIDIAGTC?

We're partnering with our GTC community to give away Golden Tickets 🎫 including:
✅GTC Conference pass
✅VIP seating at NVIDIA CEO Jensen Huang’s keynote
✅NVIDIA DGX Spark
✅Exclusive Happy Hour at NVIDIA Headquarters
✅GTC Training Lab pass
✅NVIDIA merchandise

Follow these accounts to be the first to know each giveaway's unique entry requirements 🔽
NVIDIA team: @ctnzr, @Baxate / @NaderLikeLadder
Partners: @ollama, @huggingface / @pollenrobotics

Get contest details: https://t.co/fkDOK3S6AI
Terms & conditions apply.

[Markus Holtmanns @HoltmannsMarkus]

Congrats to our speech research team — PersonaPlex, part of the Nemotron Speech collection, just hit #1 on Hugging Face.

Huge milestone for everyone who helped bring this model from research to reality. https://t.co/4OXzmUhOqv

[Markus Holtmanns @HoltmannsMarkus]

How to Get Started with NVIDIA Earth-2 in Less Than 5 Minutes https://t.co/WGh6PKUtUA via @YouTube

[Markus Holtmanns @HoltmannsMarkus]

Run 1T parameter models at your desk 🤯

At #CES2026, we unveiled how developers can run the latest AI models on a local system, from 100B parameter models on DGX Spark to 1T parameter models on DGX Station.

Read the blog ➡️ https://t.co/wPs6Mmybvx https://t.co/BDKH0A93sp

[Markus Holtmanns @HoltmannsMarkus]

Try this step-by-step tutorial on building a voice-powered RAG agent with the newest NVIDIA Nemotron models. https://t.co/qV4ecKQBPy

[Markus Holtmanns @HoltmannsMarkus]

📣 Microsoft Azure reached 1,100,948 tokens/sec on ND GB300 v6 racks powered by NVIDIA GB300 NVL72—independently validated by Signal65.

✅ NVIDIA GB300 NVL72 delivers ~10× higher token throughput vs NVIDIA H100 to Azure customers.

🔗 Learn more. https://t.co/XAtN2hWwX7

[NVIDIA Newsroom @nvidianewsroom]

🚨NVIDIA and its partners will start shipping DGX Spark — the world’s smallest AI supercomputer.

Early recipients are testing, validating and optimizing their tools, software and models for DGX Spark.

Built on the NVIDIA Grace Blackwell architecture, DGX Spark integrates NVIDIA GPUs, CPUs, networking, CUDA libraries and NVIDIA AI software, accelerating agentic and physical AI development. 

🔗 https://t.co/vIC32ebZeJ

[Brian Caffey @briancaffey]

📢 Introducing NIM Kit, an open-source application for building with NVDIA Inference Microservices (NIMs)

NIM Kit is a local version of https://t.co/wL42vM6fd7. It brings together all of the libraries and interfaces for a variety of AI inference models including LLMs, image generation, ASR, speech enhancement, 3D model generation and more! This project is my submission for the NVIDIA GTC Washington, D.C. Golden Ticket Developer Contest 🎫

NIMs simplify AI workloads by bundling dependencies, optimizing inference engines, and serving models over HTTP & gRPC. I have been really excited to follow the development of NIMs since they were announced at GTC 2024 in Jensen Huang’s keynote! I can’t wait to see what NVIDIA GTC 2025 in DC has in store for NIMs.

#NVIDIAGTC @NVIDIAGTC @NVIDIAAIDev @NVIDIA_AI_PC @MariyaSha888 @Baxate_carter @NaderLikeLadder

NIM Kit GitHub repo linked below 🔗