Instacart has 1.6 billion lifetime grocery orders—and is now using that data flywheel to digitize the physical store itself. In this episode, David McIntosh, Chief Connected Stores Officer at Instacart, explains how the Caper Cart—powered by an NVIDIA Jetson™ board and a sensor fusion system combining cameras, weight sensors, and location data—is bringing edge AI to the grocery aisle. He shares what's driving double-digit sales lift for retailers, how AI agents are beginning to automate store operations, and why in-store and online shopping will merge into a single unified experience within the decade.

🔬Topics covered:

How Caper Carts use NVIDIA Jetson and sensor fusion to identify items in real time

Why edge AI matters: hundreds-of-milliseconds response vs. seconds in the cloud

Double-digit sales lift from personalized in-cart recommendations

Building a grocery foundation model from 1.6B orders and live in-store data

AI agents for store ops: proactive out-of-stock alerts and CPG shelf optimization

Chapters:

00:00 – Introduction and the Connected Store vision

04:00 – Caper Cart: NVIDIA Jetson, sensor fusion, and edge AI

08:06 – How basket recognition and recommendations work

15:53 – Double-digit sales lift and what shoppers actually want

22:25 – Store ops, employees, and the real-time data flywheel

28:26 – AI agents, the grocery foundation model, and what's next

Open-weight models are closing the gap with proprietary AI — and Timothée Lacroix, cofounder and CTO of Mistral, has been betting on that since day one. In this episode, he explains why open weights accelerate enterprise adoption, how Mistral is bringing model customization into production, and what a 2.5x training speed improvement on GB200s means for the next generation of large sparse mixture-of-experts models. He also shares the open problem keeping him up at night: getting AI agent permission systems right before write access becomes the norm.

🔬Topics covered:

How open models and weights accelerate research

Mistral Forge: bringing enterprise-grade model customization to production

The Nemotron Coalition—what Mistral and NVIDIA are building together

2.5x training gains on GB200s for large sparse mixture-of-experts models

Why AI agent permissions—especially write access—is important to solve

Chapters:

00:00 – Introduction and Mistral’s origin story

04:05 – The case for open weights and why the community builds faster

09:34 – Mistral Forge: enterprise model customization in production

14:21 – What enterprise customers actually want from AI right now

18:46 – The hardest open problem: AI agent permissions and write access

Seeed Studio is a leader in open source robotics, delivering affordable NVIDIA Jetson‑powered arms that put embodied AI into the hands of millions of makers, students, and small businesses. In this episode, Seeed Studio CEO Eric Pan and Head of Robotics Elaine Wu explain how open hardware, the OpenClaw agentic framework, and NVIDIA Isaac Sim are turning robot arms into controllable, teachable agents—and what it takes to bring these physical AI tools into real‑world settings responsibly.

🔬Topics covered:

Why open source is the fastest path to accessible robotics

How the $200 SOR arm (with Hugging Face) lowers the barrier to embodied AI

Training robot arms like a dog: from months of coding to intuitive hand‑guided learning

OpenClaw on Jetson: turning natural‑language commands into robot skills

Using NVIDIA Isaac Sim and digital twins to bridge simulation and real‑world deployment

Building modular robot parts (heads, arms, wheels) instead of monolithic humanoids

Chapters:

00:00 – Welcome and introductions

02:00 – From open hardware modules to robotics and edge AI

05:00 – Why open source drives adoption and trust in robotics

09:00 – The $200 SOR arm: open source with Hugging Face

12:00 – Training robot arms like a dog: intuitive, hand‑guided learning

15:00 – OpenClaw on Jetson: text‑to‑robot control

20:00 – Isaac Sim and digital twins: bridging simulation and reality

27:00 – Modular design: heads, arms, wheels instead of humanoids

32:00 – Everyone can participate in physical AI: closing thoughts

As AI factories scale and token costs become a defining competitive variable, the way businesses measure infrastructure ROI needs to change. In this episode, Shruti Koparkar from NVIDIA's Accelerated Computing team breaks down tokenomics—the four-pillar framework of token utility, supply, demand, and monetization—and reveals why NVIDIA Blackwell's architecture delivers 50x more tokens per watt than NVIDIA Hopper, translating to a 35x reduction in token cost.

🔬Topics covered:

The four pillars of tokenomics: utility, supply, demand, and monetization

Why cost per token beats FLOPS per dollar as an infrastructure metric

NVIDIA Blackwell vs. Hopper: 50x more tokens per watt, 35x lower token cost

How extreme co-design turns spec-sheet numbers into real-world output

Jevons paradox: why lower token cost always drives more GPU demand, not less

The four business models for turning tokens into revenue

Chapters:

00:00 – Introduction and the four pillars of tokenomics

02:09 – Token value: intelligence, interactivity, and use case mapping

06:32 – Estimating token demand: users, reasoning, and agentic multipliers

10:00 – Token supply and why cost per token is the right infrastructure metric

13:12 – NVIDIA Blackwell vs. Hopper: 50x more tokens, 35x lower cost

14:52 – Extreme co-design for lowest token cost and the NVIDIA Vera Rubin platform

21:10 – How software multiplies hardware performance (8x gains in six months)

23:56 – Token monetization: pricing and business models

26:52 – Jevons paradox and the future of GPU demand

Snap processes more than 10 petabytes of experimentation data every single morning—and with NVIDIA GPU-accelerated Apache Spark on Google Cloud, Snap cut job costs by 76%, reduced memory usage by 80%, and eliminated 120 terabytes of disk spill from its pipelines.

Prudhvi Vatala, head of engineering platforms at Snap, joins the NVIDIA AI Podcast to break down how he and his team completely modernized data infrastructure for a social platform serving nearly a billion monthly active users—using NVIDIA cuDF plugin (formerly referred to as NVIDIA RAPIDS plugin) for Apache Spark on Google Kubernetes Engine, with zero application code changes.

🔬Topics covered:

How Snap runs A/B tests at planetary scale using rigorous statistical methods like heterogeneous treatment effect detection and variance reduction

Why Snap reuses idle inference GPUs between 1–5 a.m. for batch data processing—and how it built a Kubernetes-based platform to do it

How NVIDIA cuDF delivered 3x+ speedups on join-heavy Spark jobs with no code rewrites

The full business impact: 76% cost reduction, 62% fewer cores, 80% less memory, 120 TB of spill eliminated

How a three-way partnership between Snap, NVIDIA, and Google Cloud made it possible in just 8–9 months

Chapters:

0:00 Introduction and Snap overview

3:35 What is Snap’s experimentation platform?

4:05 Why experimentation, safety, and privacy are core at Snap

4:52 How A/B testing works at billion-user scale

8:14 Discovering NVIDIA cuDF plugin

9:06 Benchmarking results: join, union, and aggregation jobs

12:00 Reusing idle GPUs overnight via GKE

13:24 Building a bottom-up GPU data platform at Snap

17:48 Results: 76% cost reduction and partnership impact

20:56 Snap’s evolution and what’s next

Learn more:

NVIDIA cuDF: https://developer.nvidia.com/topics/ai/data-science/cuda-x-data-science-libraries/cudf#accel-apache

LangChain has surpassed 1 billion downloads—and the framework that started as a weekend project is now the harness powering the next generation of production-grade AI agents. In this episode, Harrison Chase, co-founder & CEO of LangChain, breaks down the architecture behind deep agents, explains why systems like Claude Code, Manus, and Deep Research all share the same foundational pattern, and lays out what it actually takes to deploy autonomous agents responsibly in the enterprise.

🔬Topics covered:

What is a "deep agent," and why does architecture matter more than ever?

How enterprises are (and aren't) embracing autonomous agents

LangSmith: observability, tracing, and evaluation-driven development

Mixing frontier and open models (NVIDIA Nemotron) in multi-agent systems

What's next: async subagents, proactive/always-on agents, agent memory, and agent identity

Chapters:

00:00 – LangChain origin story and the deep agent architecture

01:46 – What is a deep agent?

03:31 – Enterprise trust: risk, autonomy, and iteration

04:38 – LangSmith: observability and evaluation-driven development

13:30 – Frontier vs. open models and the Nemotron Coalition

18:10 – What's next: async subagents, agent memory, and agent identity

Generative AI can predict whether a plane takes off—but does it know why? Nicolas Cerisier, VP of 3DEXPERIENCE Platform R&D at Dassault Systèmes, explains how industrial world models go beyond pattern recognition to embed the actual laws of physics, chemistry, and engineering. In this episode of the NVIDIA AI Podcast, he also breaks down Dassault's three virtual companions (AURA, LEO, and MARIE), their 25-year collaboration with NVIDIA, and a stunning real-world use case: helping NIAR rebuild aircraft designs part by part, using AI.

What if one AI brain could run every robot on the planet—a humanoid, a warehouse arm, and a dog-like inspection bot—all at once?

That's not a thought experiment. That's what Skild AI is building right now.

Deepak Pathak (CEO and Co-Founder) and Abhinav Gupta (President and Co-Founder) of Skild AI join the pod to break down Skild Brain—a universal, general-purpose AI model designed to power robots of any form factor, tackling any task, from a single shared intelligence.

What happens when you combine AI with quantum computing? NVIDIA's Nic Harrigan joins the AI Podcast to break down the state of quantum, explain why error correction is the pivotal challenge, and reveal how NVIDIA Ising—the world's first open AI model family for quantum—is changing the game.

🔗 Resources mentioned:

► Read our NVIDIA Ising announcement 

► Learn more about NVIDIA Ising 

► Learn more about NVIDIA Quantum Computing

Chapters:

0:00 Intro

0:55 What is quantum computing?

4:00 Qubits, noise, and error correction

5:26 How AI helps quantum error correction

10:57 Applications: drug discovery and materials

15:33 NVIDIA Ising announcement

20:35 Scaling quantum hardware

27:31 Algorithm development with generative AI

Enterprises are moving from AI pilots to full‑scale AI factories that turn data into trusted digital intelligence. Red Hat CTO Chris Wright and NVIDIA’s Justin Boitano unpack the "five‑layer cake" AI factory stack, from accelerated hardware and hybrid cloud infrastructure to models, agents, and production‑grade governance.

AI is reshaping electricity demand. What does increased demand, and the shape of that demand, mean for the electric grid? Ben Sooter, Director of R&D at EPRI joins the podcast to explain why most of an AI model’s lifetime energy use comes from inference rather than training, and how micro data centers located near underutilized substations can help deliver low‑latency AI services while strengthening grid resilience.