The IoT Podcast

The next era of intelligent edge devices is taking shape - but what’s driving real-world deployment, and how are hardware and software evolving to keep pace with such a fast-moving ecosystem?

Welcome to Edge of Tomorrow – The Edge AI Debate, our spin-off series from The IoT Podcast, created in collaboration with the EDGE AI FOUNDATION. This series brings together leading voices from across the Edge AI ecosystem to debate the toughest questions shaping its future.

 

In this episode, host Pete Bernard (CEO, EDGE AI FOUNDATION) is joined by Scott Hanson (Founder & CTO at Ambiq) to explore how ultra-low power compute is enabling the next generation of intelligent edge devices. Starting with the journey from research to commercialisation, the conversation traces how advances in low-power processor design have evolved into technologies now powering hundreds of millions of devices.

It then moves into the realities of building for edge AI today, from balancing long hardware development cycles with rapid AI innovation, to the growing role of software in bridging that gap. The discussion dives into where edge AI is already delivering value at scale, including wearables, smart home systems, industrial monitoring, and medical devices. It also explores the shift from isolated use cases to widespread adoption, with an increasing number of deployed devices now running some form of intelligence at the edge.

This episode offers a grounded perspective on the rise of agentic AI, the relationship between edge and cloud intelligence, and what’s needed to support increasingly complex, real-world systems, exploring where edge AI is today and where it’s heading next.

Expect discussion on…

The role of ultra-low power compute in scaling Edge AI

Hardware vs. software: bridging development cycles

Where Edge AI is already delivering value at scale

Wearables, smart homes, industrial, and medical use cases

The shift from isolated deployments to widespread adoption

Edge vs. cloud: where intelligence actually lives

The rise of agentic AI and system-level intelligence

Chapters…

00:00 Introductions

01:30 The origin of Ambiq and ultra-low power innovation

03:00 Scaling a chip company and market growth

04:00 Hardware vs software development cycles in AI

06:00 Keeping pace with rapid AI innovation

08:30 General-purpose vs specialised architectures

10:45 Moore’s Law and efficiency gains at the edge

13:00 Real-world scale: where edge AI is being deployed

15:30 Wearables, smart homes, and industrial use cases

17:30 Edge AI in everyday life: sensing and inference

19:30 Consumer vs commercial adoption trends

21:30 The rise of “life logging” and contextual data

24:00 Societal and ethical considerations of always-on AI

25:30 Agentic AI and system-level intelligence

27:30 Edge vs cloud: where intelligence actually lives

29:30 Cost, scalability, and moving AI out of the cloud

31:30 Physical AI and robotics opportunities

33:30 Final reflections on the future of edge AI 🎧

The episode is live now on all major podcast platforms Listen here: https://linktr.ee/theiotpodcast

What’s your take? Join the debate!

About Edge of Tomorrow – The Edge AI Debate Edge of Tomorrow is our dedicated debate series bringing together technologists, researchers, and industry leaders to tackle the most pressing and controversial questions in Edge AI. From privacy and security to ethics, ownership, and governance, this series goes beyond hype to explore what it will take to build Edge AI systems responsibly. Our goal is to surface diverse perspectives, challenge assumptions, and give listeners the context they need to navigate one of the most transformative shifts in modern computing.

Why do so many IoT and embedded projects become harder to deliver as they scale, even when the technology itself works?

As IoT and embedded projects scale, the challenge shifts from simply developing the technology… to delivering outcomes reliably.

Not just technically, but commercially and operationally too.

In this episode of The IoT Podcast, Jake Gilbert sits down with Dominik Kujawski, Co-CEO & CRO at needCode, to explore how that shift is changing the way connected products are being built, delivered, and supported across the IoT industry.

Drawing on their experience across wireless connectivity, ultra wideband, BLE, and embedded systems, the conversation focuses on the realities that start to emerge as projects become more complex - from integration challenges and certification requirements, to reducing iteration cycles, avoiding costly rework, and aligning technical delivery with commercial expectations.

Rather than approaching projects as isolated technical tasks, Dominik explains why more companies are moving towards a solution-first mindset starting with the client problem, understanding the operational context, and taking greater ownership of the end result.

Jake and Dominik also discuss:

🎙️ What signals it’s time to move beyond your current business model

🎙️ How starting from the client problem (not the technical solution) changes delivery

🎙️ Why reducing iteration and rework is often where the real value sits

🎙️ What “production ready” means when certification and performance are involved

🎙️ How value is measured differently, beyond day rates or cost

The episode also gives a look at how expectations on delivery partners are changing, particularly as connected products become more intelligent, more integrated and more commercially critical.

This episode offers insight into what it actually takes to deliver successfully in modern IoT environments.

Chapters...

00:00 Introduction to needCode and Its Evolution

03:45 Moving from Capacity to Solutions

08:01 Cultural Alignment and Team Dynamics

12:07 Understanding Client Problems and Delivery Challenges

18:23 Identifying the Right Customer Profile

23:38 Navigating Enterprise vs Startup Environments

25:52 Shifting from Cost to Value-Based Delivery

27:28 The Impact of AI and Edge Computing

33:02 Certification, Interoperability, and Production Readiness

36:17 Differentiation in a Competitive Market

39:26 Real-World Case Studies and Lessons Learned

42:38 Final Thoughts and Future Outlook

The episode is live on all major listening platforms now!

Listen here: https://linktr.ee/theiotpodcast

Connect with our guest… https://linkedin.com/in/dominik-kujawski

About needCode needCode is an IoT development company focused on system integration, architecting, and deploying wireless connectivity solutions for critical applications. Their work spans ultra wideband, BLE, embedded systems, and smart connected products, helping businesses solve complex connectivity challenges and deliver production-ready solutions. Find out more: https://needcode.io/

SUBSCRIBE TO THE IOT PODCAST ON YOUR FAVOURITE LISTENING PLATFORM: https://linktr.ee/theiotpodcast Sign up for exclusive email updates: https://theiotpodcast.com/get-exclusive-access/

Contact us to become a guest/partner: https://theiotpodcast.com/contact/

What does “trust” really mean when AI lives at the edge?

Welcome to Edge of Tomorrow – The Edge AI Debate, our spin-off series from The IoT Podcast, created in collaboration with the  @edgeaifoundation  This series brings together leading voices from across the Edge AI ecosystem to debate the toughest questions shaping its future.

In this episode, host Pete Bernard (CEO, EDGE AI FOUNDATION) is joined by Rosario Cammarota (Chief Scientist, Intel Privacy Lab) and Aydin Aysu (Founder, MithrilAI) to tackle:

Can you trust Edge AI - and if so, where does that trust really come from?

As intelligence becomes distributed across edge devices, the traditional assumptions around privacy, security, and ownership start to break down.

This conversation digs into how trust changes when AI runs locally, whether hardware can (or should) be trusted, and what it means for users, builders, and regulators alike.

From cryptographic techniques like homomorphic encryption to real-world deployment scenarios, the debate explores what it actually takes to build Edge AI systems that are not only powerful, but trustworthy.

Chapters…

00:00 Introductions & debate setup

02:30 What does “trust” mean when AI lives at the edge?

07:10 Edge vs. cloud: control, risk, and responsibility

11:40 Can we secure Edge AI without trusting hardware?

17:20 The role of homomorphic encryption

22:30 Data ownership, governance, and user sovereignty

28:40 Real-world Edge AI deployments and security trade offs

34:10 Final reflections: rebuilding trust in the Edge AI stack 

The episode is live now on all major podcast platformsListen here: https://linktr.ee/theiotpodcast

What’s your take? Join the debate in the comments.

Connect with our guests…

Rosario Cammarota – Chief Scientist, Intel Privacy Lab: https://www.linkedin.com/in/ro-cammarota-a226b817/

Aydin Aysu – Founder, MithrilAI: https://www.linkedin.com/in/aydinaysu/

Moderator: Pete Bernard – CEO, EDGE AI FOUNDATION: https://www.linkedin.com/in/bernardpete/

About Edge of Tomorrow – The Edge AI Debate Edge of Tomorrow is our dedicated debate series bringing together technologists, researchers, and industry leaders to tackle the most pressing and controversial questions in Edge AI.

Edge AI is evolving quickly - what’s changing in the tooling and frameworks that support it? And where do the biggest opportunities for improvement lie?

In this episode of Edge of Tomorrow – The Edge AI Debate, host Pete Bernard (CEO, EDGE AI FOUNDATION) is joined by Elia Schoenberger (Product Marketing, AI Division at Civa) and Nathan Francis (Business Development at AIZip) for an in-depth discussion on the role of tooling and frameworks in shaping the future of Edge AI.

While much of the attention in AI is placed on models and hardware, this conversation focuses on the layers in between — compilers, SDKs, deployment pipelines, and frameworks — and how they influence speed, scalability, and collaboration across the ecosystem.

The discussion explores how hardware constraints, model design, and tooling choices intersect, whether the industry is moving toward standardisation or continuing to prioritise innovation, and what practical steps could help simplify Edge AI development without slowing progress.

This episode offers a look at how the Edge AI stack is evolving - and what it will take to support broader, more efficient deployment in the years ahead.

Chapters…

00:00 Introductions

01:40 Why tooling and frameworks matter in Edge AI

03:25 Hardware constraints, models, and deployment realities

05:23 Fragmentation vs fit-for-purpose tooling

07:59 Power, performance, and memory trade-offs

09:18 Where Edge AI sits on the maturity curve

11:32 Standardisation, innovation, and ecosystem balance

13:32 Compilers, MLIR, and unifying the toolchain

16:02 Deployment challenges and quantisation complexity

25:17 Generative models arriving at the edge

31:36 Local-first intelligence and when the cloud still matters

32:52 What would help accelerate Edge AI development

36:13 Final reflections and closing thoughts

The episode is live on all major listening platforms now: https://linktr.ee/theiotpodcast

Connect with our guests…

Elia Schoenberger (CIVA,Inc): https://www.linkedin.com/in/eliashenberger/

Nathan Francis (AIZip): https://www.linkedin.com/in/nathanfrancis99/

About Edge of Tomorrow – The Edge AI Debate Edge of Tomorrow is our debate-led spin-off series bringing together leaders from across the Edge AI ecosystem to explore the technical, commercial, and strategic decisions shaping how intelligence moves closer to the real world.

SUBSCRIBE TO THE IOT PODCAST ON YOUR FAVOURITE LISTENING PLATFORM: https://linktr.ee/theiotpodcast Sign up for exclusive email updates: https://theiotpodcast.com/get-exclusive-access/ Contact us to become a guest or partner: https://theiotpodcast.com/contact/

What does it really take to build the next billion-dollar idea?

And why do some innovative ideas scale fast… while others quietly disappear?

In this episode of The IoT Podcast, host Tom White sits down in Dubai with one of the show’s most popular previous guests - and one of the world’s leading voices in IoT - Yasser Alsaied, President at Facilis and former VP of IoT at AWS, for an open, in-person conversation about what really drives innovation today.

They look back at predictions Yasser made in his 2023 episode and where we are now. They explore why ideas don’t win on vision alone, why startups often build the most meaningful breakthroughs and why many founders are now choosing belief over quick exits.

They also talk about the pace of change across the MENA region, what moving fast really demands from leadership teams, and why having the right people matters more than having the biggest budget.

From AI and edge to industrial systems and robotics, this episode is about what it really takes to turn an idea into something that lasts.

Chapters…

00:00 Introductions

02:48 Looking Back: Which Predictions Came True

08:05 Why AI Is Useless Without Real Data

14:30 Why Startups Move Faster Than Big Companies

20:45 When Innovation Dies After Acquisition

27:40 Robotics, Edge, and Real-World Impact

36:22 Why Tech Is Moving So Fast in the MENA Region

55:56 Speed, Execution, and Knowing When to Move

1:01:35 What It Really Takes to Build the Next Billion-Dollar Idea

The episode is live on all major listening platforms now: https://linktr.ee/theiotpodcast

Connect with our guestYasser Alsaied: https://www.linkedin.com/in/yasseralsaied/

SUBSCRIBE TO THE IOT PODCAST ON YOUR FAVOURITE LISTENING PLATFORM: https://linktr.ee/theiotpodcast

Sign up for exclusive email updates: https://theiotpodcast.com/get-exclusive-access/

Contact us to become a guest/partner: https://theiotpodcast.com/contact/

Connect with host Tom White: / tom5values

This episode is released alongside the announcement of Yasser’s strategic partnership with 5V Tech. Find out more: https://www.weare5vtech.com/welcoming-global-iot-leader-yasser-alsaied-