How Microchip's Brian McCarson Is Building the Data Center Infrastructure Powering AGI

How Microchip's Brian McCarson Is Building the Data Center Infrastructure Powering AGI

In this episode, Brian McCarson (Microchip Technology) breaks down AI's four evolutionary eras: machine learning, training, inference, and AGI. In today’s conversation, Brian makes a bold case that most organizations are building infrastructure for the wrong phase. Based on nearly 25 years at Intel and now driving innovation at Microchip Technology, he explains why the real bottleneck in AI today is not compute power but the high-speed switches, retimers, and storage controllers that connect GPUs, CPUs, and memory at scale, what he calls the "nervous system" of AI infrastructure. He warns that enterprises over-investing in cloud-based training architectures are heading toward a costly redesign within 18 months, and that the winning strategy is to plan now for inference-first, agent-friendly systems that push compute as close to the endpoint as possible. 🎧 Episode Highlights [01:37]: Brian's journey from semiconductor automation to modern AI [13:41]: Inheriting and rebuilding Microchip's data center business [20:10]: How Microchip deployed AI agents to scale operations without expanding headcount [34:43]: Why Microchip positions itself as the nervous system of AI infrastructure [43:07]: The energy and supply chain constraints reshaping the future of data centers 🔑 Key Takeaways: ● Most organizations are investing in AI infrastructure for the wrong phase. The shift from the training era to the inference era demands a fundamental rethink of architecture, and companies building cloud-dependent, training-heavy systems today are setting themselves up for a costly and disruptive redesign within 18 months. Planning now for inference-first, agent-friendly systems that push compute closer to the endpoint is the strategic move that separates long-term winners from short-term optimizers. ● The real bottleneck in AI is not compute power but the infrastructure connecting it. As Nvidia accelerates its hardware refresh cycle to an annual cadence, the switches, retimers, and storage controllers that enable GPUs, CPUs, and memory to communicate at scale have become the critical constraint. Investing in high-performance interconnect technology is no longer an afterthought but a core requirement for getting full value out of expensive compute investments. ● AI delivers its greatest business value when it augments people rather than replaces them. Microchip's turnaround demonstrated that deploying AI agents as PhD-level assistants to handle menial, repetitive tasks, while keeping humans focused on high-value strategic work, drives better outcomes, stronger team morale, and sustainable growth. The companies using AI purely to cut headcount and improve a balance sheet are optimizing for the short term at the expense of long-term organizational health. 👤 About The Host: Brian McCarson Brian McCarson is Corporate Vice President at Microchip Technology, where he leads the company's data center solutions business with a focus on the high-speed interconnect infrastructure powering next-generation AI systems. He brings nearly 25 years of experience at Intel Corporation, where he built deep expertise across semiconductor innovation, factory automation, and advanced technology development. A longtime advocate for using AI to drive real business outcomes, Brian is recognized for his ability to turn around complex technology organizations and translate emerging AI trends into actionable enterprise strategy. Stay Connected: ● https://www.softeq.com ● https://www.linkedin.com/in/techris ● https://www.linkedin.com/in/brianmccarson ● https://www.microchip.com Produced by Speakerbox Media.

AI for Good: How NVIDIA’s Thorsten Stremlau Is Transforming Life for People with Disabilities

NVIDIA systems architect Thorsten Stremlau sees “AI for good” as a way to restore core human abilities like communication, independence, and participation, especially for people with disabilities. Inspired by his work with Stephen Hawking and Peter Scott- Morgan, he’s helped build affordable, AI-powered tools from eye-gaze systems to personalized language models that run on everyday devices like smartphones. He contrasts these non-invasive solutions with brain-computer interfaces from companies like Neuralink, highlighting that while implants may unlock richer interaction in the future, today’s priority is scalable tech that already improves lives. He also explores how AI can support areas beyond ALS, including heart disease, PTSD, and mobility. At its core, his work is about using AI to create a more inclusive world where disability doesn’t mean disconnection. 🎧Episode Highlights [01:24]: Thorsten’s early work with nonverbal disabled youth [07:41]: AI restoring communication, autonomy, participation [18:09]: Working with Stephen Hawking and Peter Scott-Morgan [29:55]: Low-cost eye-gaze and circular-keyboard communication [43:12]: Beyond ALS: speech, PTSD, AI prosthetics 🔑 Key Takeaways: ● AI for good is ultimately about restoring human agency, not just boosting efficiency. By focusing on communication, autonomy, and participation, AI systems can give people with disabilities the ability to express themselves, make choices, and engage with the world on their own terms, rather than being defined by their limitations. ● The most impactful assistive technologies are built on mainstream, affordable hardware instead of specialized, high-cost rigs. Eye-gaze interfaces, circular keyboards, personalized language models, and speech-decoding systems that run on laptops and smartphones dramatically expand access, making advanced assistive tech viable not just for a few patients in wealthy systems, but for millions of people worldwide. ● Non-invasive AI solutions are a powerful bridge to the future of human–computer interaction, even as brain–computer interfaces rapidly advance. By combining clever sensing (eyes, face, voice, heart rate, touch), behavior modeling, and personalized AI, we can already enable richer communication, calmer nervous systems, and more natural movement, laying the groundwork for a world where disability no longer means disconnection from work, creativity, or community. 👤 Guest Spotlight: Thorsten Stremlau Thorsten Stremlau is a Principal Systems Architect at NVIDIA and a technology leader focused on building next-generation computing platforms and customer-centered innovation. Previously with Lenovo and IBM, he holds 30+ patents and is widely recognized as a thought leader in platform security and advanced systems architecture, known for turning complex technical challenges into trusted, market-ready solutions. Stay Connected: ● https://www.softeq.com [https://www.softeq.com] ● https://www.linkedin.com/in/techris [https://www.linkedin.com/in/techris] ● https://www.linkedin.com/in/thorsten-stremlau-247930 [https://www.linkedin.com/in/thorsten-stremlau-247930] ● https://www.nvidia.com [https://www.nvidia.com] Stay inspired and ahead of the curve by subscribing to Forging the Future. Share your thoughts on this episode with the hashtag #ForgingTheFuture or tag us online!

The “SaaS-pocalypse” Is Here: Equipt.ai’s Amanpreet Kaur on How AI Is Rebuilding Software

Episode Summary: Amanpreet Kaur, Chief AI and Technology Officer of Equipt.ai, breaks down the so-called “SaaS-pocalypse” and argues that SaaS isn’t dying, it’s evolving into smarter, AI-embedded systems of execution. Amanpreet explains how traditional SaaS created fragmentation and operational friction, and how Equipt.ai is rebuilding the stack by unifying workflows, data, and decision-making into a single AI-driven execution layer. Drawing from her experience in asset-heavy industries like energy, Aman shares how their platform replaces multiple disconnected tools while enabling real-time, guided operations from quote to cash. She also reflects on the early startup journey, from landing their first enterprise customer to navigating investor skepticism and refining their positioning. A key shift is undoubtedly taking place: the future belongs to agentic SaaS platforms that reduce human intervention and turn software into an active driver of business outcomes. 🎧 Episode Highlights  * [02:09]: SaaS solved infrastructure but created fragmentation * [04:59]: “Dumb SaaS is dying:” AI-powered execution rises * [06:17]: Replacing 10-20 tools with one platform * [10:44]: AI as “steroids” for SaaS, not replacement * [14:38]: First enterprise client before having a product * [23:13]: Embedding AI across the full workflow 🔑 Key Takeaways: * The “SaaS-pocalypse” isn’t about SaaS disappearing, it’s about a shift from systems of record to systems of execution. Traditional SaaS created fragmented workflows and heavy reliance on human coordination, while AI-enabled platforms are now unifying data and driving real-time decisions directly within operations. * AI’s real value is not replacing software, but embedding intelligence into every layer of it. From automation and optimization to predictive insights, the winning approach is combining machine learning, simple automation, and context-aware AI to reduce manual work and enable more deterministic, reliable outcomes. * Startups that succeed in this shift will focus on solving real operational problems, not just adding AI for hype. Equipt.ai’s journey shows that deep industry understanding, clear pain points, and delivering immediate value to customers matter more than technology trends, especially when building trust and scaling from early enterprise clients. 👤 Guest Spotlight: Amanpreet Kaur Amanpreet Kaur is the co-founder as well as the Chief AI and Technology Officer of Equipt.ai, where she builds AI-powered operational platforms for asset-intensive industries. With a background in energy and industrial technology, she previously led the development and commercialization of digital solutions, including enterprise-scale platforms and digital twin systems. Kaur brings deep domain expertise and a product-first mindset to redefining how businesses move from fragmented SaaS tools to AI-driven systems of execution. Stay Connected: * https://www.softeq.com/ [https://www.softeq.com/] * https://www.linkedin.com/in/techris/ [https://www.linkedin.com/in/techris/] * https://www.linkedin.com/in/amanpreet-hon-doc/ [https://www.linkedin.com/in/amanpreet-hon-doc/] * https://www.equipt.ai/ [https://www.equipt.ai/] Stay inspired and ahead of the curve by subscribing to Forging the Future. Share your thoughts on this episode with the hashtag #ForgingTheFuture [https://www.youtube.com/hashtag/forgingthefuture] or tag us online!

What If AI Worked More Like the Human Brain? ft. Chris Eliasmith of Applied Brain Research

At CES 2026, we sat down with Chris Eliasmith, CTO of Applied Brain Research, to discuss how brain-inspired AI is enabling fast, low-power voice interfaces that run directly on edge devices. Drawing on research modeling the hippocampus, his team developed new neural network architectures that significantly improve efficiency and accuracy for tasks like speech recognition and text to speech. These advances allow devices such as AR glasses, robots, and wearables to respond to voice commands in under 300 milliseconds, creating interactions that feel natural and conversational. Eliasmith also explains the tradeoffs between model size, accuracy, and power consumption, and how running AI at the edge can reduce costs and reliance on the cloud. He ultimately envisions a future where complete AI agents run locally on small devices, making technology simpler and more accessible for everyday users. 🎧 Episode Highlights: ●[01:59]: Introducing ultra-low-power voice AI at the edge ●[03:27]: Why 300ms latency is critical for natural conversations ●[09:06]: Brain-inspired neural networks modeled after the hippocampus ●[15:02]: Tiny AI chips for AR glasses, robotics, and wearables ●[20:25]: Cutting cloud costs with local speech processing ●[27:54]: The future of full AI agents running at the edge 🔑 Key Takeaways: ● By modeling neural networks after how parts of the brain like the hippocampus process time-based information, researchers can build AI systems that achieve higher accuracy with far fewer parameters. This approach allows models to process speech and other signals more efficiently, making advanced AI practical even on small, resource-constrained devices. ● For voice interfaces to feel natural, responses must happen within roughly 300 milliseconds, the same timing humans expect in conversation. Designing AI systems that meet this latency requirement changes how models are built and deployed, pushing developers to prioritize real-time performance rather than relying on slower cloud-based processing. ● Low-power AI that operates directly on devices reduces reliance on internet connectivity, lowers operational costs, and improves responsiveness. As models become efficient enough to run locally, entire AI agents could operate on wearables, robotics platforms, and AR devices, simplifying technology and making intelligent interfaces accessible to more users. 👤 Guest Spotlight: Chris Eliasmith Chris Eliasmith is the Director of the Centre for Theoretical Neuroscience at the University of Waterloo and holds the Canada Research Chair in Theoretical Neuroscience. He is also the CTO and co-founder of Applied Brain Research, where he works on low-power AI technologies for machine learning, robotics, and edge computing. Eliasmith is the co-inventor of the Neural Engineering Framework, the Nengo software platform, and the Semantic Pointer Architecture, and is the author of How to Build a Brain (Oxford University Press) and Neural Engineering (MIT Press). Stay Connected: ●https://www.softeq.com/ [https://www.softeq.com/] ●https://www.linkedin.com/in/techris/ [https://www.linkedin.com/in/techris/] ●https://www.linkedin.com/in/chris-eliasmith/ [https://www.linkedin.com/in/chris-eliasmith/] ●https://www.linkedin.com/company/applied-brain-research/ [https://www.linkedin.com/company/applied-brain-research/]

The Race to Ultra-Efficient, Low-Power AI with Edge Impulse and Nordic Semiconductor

At CES 2026 in Las Vegas, Brandon Shibley of Edge Impulse and Thomas Soderholm of Nordic Semiconductor join Chris to explore the exciting shift of AI from the cloud to the edge. Brandon shares how streamlined, low power machine learning models are unlocking new possibilities in health wearables, industrial inspection, and agriculture by bringing fast, responsive intelligence directly onto devices. Thomas highlights how Nordic’s latest ultra low power chips with built-in neural processing are making this next wave of innovation possible. Together, they paint an optimistic picture of AI that is faster, smarter, and more accessible, running right where data is created. 🎧 Episode Highlights ●[01:02]: Why AI is moving from cloud to edge ●[08:08]: Wearables and health monitoring on-device ●[11:08]: Industrial and agricultural vision at the edge ●[22:25]: Bluetooth Low Energy and ultra low power design ●[28:02]: New Nordic chips with built-in neural processing 🔑 Key Takeaways: ●Edge AI is about efficiency, not scale for the sake of it. Smaller, purpose-built models running directly on devices can reduce latency, preserve privacy, and dramatically lower power consumption while still delivering high impact outcomes in health, agriculture, and industrial settings. ●Hardware innovation is unlocking the next wave of on-device intelligence. Ultra low power chips with integrated neural processing units and advanced Bluetooth Low Energy connectivity make it possible to run meaningful machine learning workloads on wearables and battery-driven products. ●The future of AI is distributed by design. Instead of relying entirely on massive cloud models, intelligence will live closer to where data is created, balancing performance, cost, and connectivity to create scalable and practical real world solutions. 👤 Guest Spotlight: Brandon Shibley Brandon Shibley is a Founder at Edge Delivery and Senior Staff Engineer at Edge Impulse, a Qualcomm company, where he helps developers and enterprises build and deploy machine learning models on edge devices. With a background spanning CTO, founder, and innovation leadership roles, he has led full-stack IoT and edge computing strategies across industrial, robotics, and embedded systems markets. Brandon specializes in bringing intelligent software closer to the physical world, enabling scalable, low power AI solutions that run directly on devices. Thomas Soderholm Thomas Soderholm is the Vice President of Business Development at Nordic Semiconductor, where he helps drive the company’s strategy in ultra low power wireless connectivity and edge AI. With deep roots in Bluetooth Low Energy innovation, he works at the intersection of hardware, software, and connectivity to enable smarter battery-driven devices. Thomas focuses on advancing integrated solutions that bring efficient machine learning and secure connectivity to wearables and connected products worldwide. Stay Connected: ●https://www.softeq.com/ ●https://www.linkedin.com/in/techris/ ●https://www.linkedin.com/in/shibley ●https://www.linkedin.com/company/nordic-semiconductor/ Stay inspired and ahead of the curve by subscribing to Forging the Future. Share your thoughts on this episode with the hashtag #ForgingTheFuture or tag us online!