Siemens Buildings uses this demo to explain how its liquid cooling hardware, controls and physics-based simulation tools fit together as a full-stack approach for high-density data centers running AI and HPC workloads. The in-rack cooler, sensors, valves, actuators and pump skid are all tied into a control layer that targets lower cost-to-compute and faster time-to-compute while keeping power usage effectiveness (PUE) within strict efficiency targets. More details on Siemens’ liquid-cooled data center platforms are outlined here: https://www.siemens.com/global/en/products/buildings/hvac/oem/liquid-cooling-solution-for-data-centers.html

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

At the heart of the rack is a closed-loop liquid cooling circuit that can be connected directly to outside heat-rejection equipment without necessarily requiring a separate cooling distribution unit (CDU). The system shows how fan-wall assisted heat exchangers capture server exhaust heat, move it through a warm-water loop via variable-speed pumps, and track performance with inline flow and temperature sensing. By supporting elevated water temperatures compatible with dry coolers or high-efficiency chillers, the design allows high-density racks to run with much lower dependence on traditional air-based cooling infrastructure.

A key part of the story is Siemens’ digital twin environment, which uses CFD and thermo-fluid simulation combined with electrical and building models to predict capacity, power consumption, and PUE before ground is broken on a new facility. Operators can evaluate full-immersion, direct-to-chip, hybrid liquid–air architectures and different heat-rejection topologies, then compare layouts, operating temperatures and control strategies inside a virtual data center model. Filmed at Supercomputing 2025 (SC25) in St. Louis, this walkthrough shows how those executable digital twins shorten design cycles and de-risk large AI data center investments.

The rack on the show floor also highlights partner technologies: ZeroKelvin cold plates for direct-to-chip heat extraction, NNDC’s expertise in data center liquid cooling architectures, and Nortek Data Center Cooling equipment for in-row or CDU-based deployments. Siemens’ control stack orchestrates these components so that pumps, valves and heat exchangers operate at optimal setpoints while maintaining safe chip junction temperatures and minimizing total energy draw. That partner model lets customers mix and match hardware vendors while still using a unified controls and simulation layer for their liquid-cooled estates.

Finally, the discussion touches on sustainability and lifecycle economics: closed warm-water loops reduce or even eliminate evaporative cooling, cutting water use and enabling deployment in regions where water is scarce but land and renewable power are available. By analyzing scenarios such as “CDU vs direct loop,” “air-assisted vs full liquid,” or “heat reuse into district heating,” the Siemens toolchain helps operators trade off capex, opex, carbon intensity and grid constraints before deploying large GPU clusters. The result is a reference architecture for liquid-cooled AI data centers that seeks to balance density, efficiency, and environmental impact over the full project lifetime.

Publishing 50+ videos from Supercomputing 2025 (SC25, St. Louis), and from other recent events, about 4 per day at 5AM, 11AM, 5PM and 11PM CET/EST.
Join https://www.youtube.com/charbax/join for early access to all my queued videos early.

Watch my full SC25 playlist:

Click the “Super Thanks” button below the video to send a highlighted comment under the video! Brands I film are welcome to support my work in this way 😁

Check out my video with Daylight Computer about their revolutionary Sunlight Readable Transflective LCD Display for Healthy Learning: https://www.youtube.com/watch?v=U98RuxkFDYY

source https://www.youtube.com/watch?v=2GL70Kyxo70

Legrand uses this booth tour to show how it wants to be a full-stack infrastructure vendor for modern data centers, from power distribution and racks to cooling, structured cabling and out-of-band management. The conversation walks through how busway, OCP ORV3 power shelves, vertical PDUs, rear-door heat exchangers, high-density fiber and remote management all fit together into one architecture for AI and cloud workloads. More details on the portfolio are on Legrand’s data center site: https://www.legrand.com/datacenter/en

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

The tour starts with overhead Starline-style busway distributing high-voltage power above the white space, so cabinets can be added, moved or decommissioned without rewiring under the floor. That same busway feeds Legrand’s OCP ORV3 cabinets and power shelves, with 1RU and 2RU shelves that convert AC into DC for IT loads and tie directly into a vertical DC busbar at the rear of the rack. Combined with the intelligent rack PDUs from the Server Technology and Raritan brands, Legrand positions itself as a major player in vertical power distribution and per-outlet metering at scale.

Cooling is handled by a rear-door heat exchanger mounted directly behind the cabinet, pulling hot exhaust air through a liquid-cooled door. In the interview they quote up to 200 kW of rack cooling capacity, which lines up with Legrand’s ColdLogik rear-door cooler portfolio aimed at high-density AI and HPC racks. Rear-door systems sit between air and direct-to-chip liquid cooling, letting operators raise supply temperatures, shrink chiller loads and still manage the thermal footprint of dense GPU nodes without completely redesigning the room.

On the connectivity side, Legrand emphasizes structured cabling and new “direct mating” fiber architectures. Instead of traditional cassettes, Infinium acclAIM style VSFF connectors mate patchcords directly to trunks, increasing density and reducing insertion loss in the link budget. The portfolio spans copper, fiber trunks up to 288 fibers, and a programmable optics line with transceivers, DACs and AOCs designed for spine–leaf fabrics and storage backbones. The goal is to keep passive infrastructure ready for 400G and 800G transitions while cleaning up patching complexity in crowded racks.

Finally they touch on out-of-band and monitoring, referring to a “ZU” or out-of-band platform that ties mission-critical IT resources together for secure remote access. That aligns with Legrand’s integration of ZPE Systems’ Nodegrid out-of-band management into its data center stack, combining serial console, remote power control and telemetry in one control plane. The interview closes by situating Legrand as a French-headquartered, globally manufactured group that delivers power, light and data to where people live and work, with data centers now a key growth area. This walkthrough was recorded at Supercomputing 2025 (SC25) in St Louis, where Legrand uses the booth as a live reference design for AI-ready data center infrastructure.

Publishing 60+ videos from Supercomputing 2025 (SC25, St. Louis), and from other recent events, about 4 per day at 5AM, 11AM, 5PM and 11PM CET/EST.
Join https://www.youtube.com/charbax/join for early access to all my queued videos early.

Watch my full SC25 playlist:

This video was filmed using the DJI Pocket 3 ($669 at https://amzn.to/4aMpKIC using the dual wireless DJI Mic 2 microphones with the DJI lapel microphone https://amzn.to/3XIj3l8 ), watch all my DJI Pocket 3 videos here https://www.youtube.com/playlist?list=PL7xXqJFxvYvhDlWIAxm_pR9dp7ArSkhKK

Click the “Super Thanks” button below the video to send a highlighted comment under the video! Brands I film are welcome to support my work

Check out my video with Daylight Computer about their revolutionary Sunlight Readable Transflective LCD Display for Healthy Learning: https://www.youtube.com/watch?v=U98RuxkFDYY

source https://www.youtube.com/watch?v=6Lz_gD73YDI

Advantech shows why it focuses on low-TDP edge AI instead of only chasing the highest data-center TDP numbers in its latest server portfolio. Product manager Alan Singh walks through short-depth rack servers that pair a single CPU with multiple SiMa.ai MLSoC accelerator cards to deliver dense video analytics at the edge, targeting smart cities, manufacturing and retail deployments that care as much about power envelopes and cooling budgets as TOPS. https://www.advantech.com/

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

Recorded at Supercomputing 2025 in St. Louis, the demo highlights an ASMB-818 server board inside the HPC-7420 4U short-depth chassis, populated with SiMa.ai cards to reach around 80 concurrent video streams of real-time AI inference for object detection and tracking. This design combines DDR5 memory expansion, multiple PCIe 5.0 slots and compact mechanical depth so the same platform can slide into shallow edge racks or wall-mount enclosures in existing facilities, without the plumbing overhead of direct-to-chip liquid cooling.

Singh then moves up the stack to larger EATX and dual-socket AMD and Intel platforms, showing how Advantech extends the same design philosophy into systems that can host up to six GPUs in a short-depth chassis, with 16 or 32 DIMM slots for high-capacity DDR5. These platforms are tuned for AI inference, data acquisition and streaming workloads at the network edge, where deterministic latency, ruggedized builds and remote-management hooks often matter more than squeezing in another couple of hundred watts of accelerator TDP.

A 2U NVIDIA MGX-compliant GPU server such as the SKY-622G4 illustrates how Advantech bridges edge and core data-center requirements: dual Intel Xeon 6 processors, 32 DDR5 ECC RDIMMs, up to four dual-slot GPUs including RTX PRO 6000 Blackwell or H200/H100 NVL parts, plus redundant 6000 W PSUs and high-bandwidth networking for scale-out clusters. Around these hardware building blocks, Advantech leverages its position as a leading industrial PC vendor with extensive customization, from FDA-related and medical certifications to military-grade environmental testing, so OEMs can take MGX and edge AI reference designs and turn them into deployable, certified products in demanding vertical markets.

I’m publishing about 90+ videos from Embedded World North America 2025, I upload about 4 videos per day at 5AM/11AM/5PM/11PM CET/EST. Join https://www.youtube.com/charbax/join for Early Access to all 90 videos (once they’re all queued in next few days) Check out all my Embedded World North America videos in my Embedded World playlist here: https://www.youtube.com/playlist?list=PL7xXqJFxvYvjgUpdNMBkGzEWU6YVxR8Ga

This video was filmed using the DJI Pocket 3 ($669 at https://amzn.to/4aMpKIC using the dual wireless DJI Mic 2 microphones with the DJI lapel microphone https://amzn.to/3XIj3l8 ), watch all my DJI Pocket 3 videos here https://www.youtube.com/playlist?list=PL7xXqJFxvYvhDlWIAxm_pR9dp7ArSkhKK

Click the “Super Thanks” button below the video to send a highlighted comment under the video! Brands I film are welcome to support my work in this way 😁

Check out my video with Daylight Computer about their revolutionary Sunlight Readable Transflective LCD Display for Healthy Learning: https://www.youtube.com/watch?v=U98RuxkFDYY

source https://www.youtube.com/watch?v=ITmT4kZ3_rg

MSI uses this interview to show how its long history in graphics and client hardware is now feeding a serious push into data-center and AI infrastructure. At the booth, they focus on NVIDIA MGX-based GPU servers populated with RTX Pro 6000 Blackwell Server Edition accelerators, eight GPUs per node tied into NVIDIA CX8 networking for eight 400 GbE ports, giving dense, rack-scale inference throughput for enterprises that want to run their own language models rather than relying purely on public cloud. ([MSI][1]) These PCIe-based accelerators with 96 GB of GDDR7 per card sit in systems tuned by NVIDIA’s MGX reference design, so thermals, power delivery and PCIe lane topology are pre-balanced for sustained AI workloads. https://www.msi.com/Landing/NVIDIA-MGX

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

The conversation dives into how an 8× RTX Pro 6000 MGX server becomes a 7–8 kW box, implying 40–50 kW per rack when fully populated, and why that profile is ideal for high-volume inference, light training and fine-tuning of LLMs. By combining passive 600 W GPUs with CX8-class networking and 400 GbE QSFP ports, MSPs and large enterprises can build clusters that saturate NVLink-to-NIC bandwidth without resorting to HGX-class systems, at a lower total cost of ownership but still with data-center-grade density and reliability. This positions MSI squarely in the sweet spot for customers who want to bring agentic AI, retrieval-augmented generation and high-throughput model serving onto their own infrastructure.

A key highlight is MSI’s Grace Blackwell GB300 workstation, marketed as the CT60-class AI Station, which essentially brings DGX-level architecture under a desk. It combines an NVIDIA Grace ARM CPU with a B300 Blackwell GPU on a single module, linked over NVLink-C2C into one coherent memory space, with LPDDR5X on the CPU and HBM3e on the GPU adding up to roughly 784 GB of unified memory. That enables developers to fit very large language models into a single address space for experimentation, fine-tuning and evaluation without sharding across multiple GPUs. MSI plans to ship this as a 1.6 kW water-cooled workstation, quiet enough for an office yet packed with four M.2 NVMe slots and dual 400 GbE ports, so code and models can be developed locally and then pushed unchanged to full GB300/HDX deployments in the cloud.

For teams that don’t need GB300-class memory footprints, MSI also shows the EdgeXpert system built on NVIDIA’s GB10 Grace Blackwell “Spark” platform. This compact edge box pairs a 20-core Grace CPU with a smaller Blackwell GPU and 128 GB of unified memory, targeting about a petaFLOP of FP4 AI performance in a desktop-friendly form factor. It’s aimed at local prototyping, on-prem inference and edge deployments where developers want the same Grace-Blackwell software stack they use in the data center, but in a lower-power box that can sit under a desk. Seen together with the larger AI Station, the St. Louis SC25 booth story is really about giving AI teams a continuum from compact GB10 nodes through GB300 workstations up to full MGX server racks.

The tour closes on MSI’s broader server roadmap: OCP OpenRack V3 21-inch racks with 48 V busbars and centralized power shelves, plus DC-MHS host processor modules that let the same chassis accept Intel Xeon 6900/6700 or AMD Turin EPYC CPUs. Extended-volume air coolers (EVAC) allow them to air-cool 500 W CPUs with very low fan power, which matters once racks cross the 17 kW threshold. MSI’s representative notes that, outside NVIDIA’s Grace-based platforms, customer demand for ARM servers is still limited, so most of their modular boards focus on x86 today. But with MGX GPU nodes, Grace Blackwell workstations and OCP-ready compute sleds, MSI is clearly positioning itself as a scalable player in high-performance and AI computing rather than just a consumer PC brand.

source https://www.youtube.com/watch?v=_U-yJ7tGYK4

Micron uses this tour to argue that memory and storage, not just GPUs, now define AI and high-frequency-trading performance, explaining how its DRAM, CXL memory expansion and hyperscale NVMe SSDs are tuned to keep accelerators continuously fed with data. From 1-gamma DRAM wafers to next-generation PCIe Gen6 drives, the story is about latency, bandwidth and tiered capacity rather than only raw compute. https://www.micron.com/markets-industries/ai/ai-data-center

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

On the memory side, Micron highlights its 1-gamma DRAM node, built with EUV lithography to deliver more than 30% higher bit density, lower power and higher speed than the previous 1-beta generation, and deployed across HBM3/HBM4 stacks, DDR5 RDIMMs and high-capacity MRDIMMs. The same technology underpins record-setting STAC-A2 and STAC-M4 benchmark systems for quantitative finance, where millisecond-level gains in risk calculations and tick-by-tick analytics can translate directly into trading edge.

The discussion then moves to CXL-based disaggregated memory: a chassis populated with Micron CXL memory cards behind an XCON switch, exposing around 15 TB of pooled capacity that can be shared across four servers. This creates a near-memory, main-memory and expansion-memory hierarchy, with access latency rising from roughly 100 ns on local DRAM to about 450 ns over CXL, still around 10× faster than going out to SSDs while offering far more capacity than on-board DIMM slots. That extra tier is presented as a way to keep data-hungry HPC and AI workloads in memory longer, reducing trips to storage while matching capacity to the economics of each application.

On the storage side, Micron shows its PCIe Gen6 9650 NVMe SSDs alongside a broad Gen5 data-center portfolio aimed at AI clusters and exascale storage. The Gen6 drives push up to roughly 28 GB/s sequential read bandwidth and 5.5 million random read IOPS in EDSFF E1.S and E3.S form factors, while the new 6600 ION QLC family targets capacity tiers with 122 TB today and 245 TB per drive on the roadmap, all built on G9 276-layer 3D NAND that scales from TLC to QLC depending on endurance and cost per bit.

Filmed at Supercomputing 2025 in St Louis, the conversation closes on system-level design: how ultra-dense SSDs can collapse racks of nearline HDDs into a few EDSFF enclosures, cutting power, cooling and floor space, and how enterprise endurance ratings of multiple drive writes per day over five years intersect with real-world bandwidth limits so that drives are rarely worn out. The Micron team also touches on GPU-direct storage paths that send data from NVMe through PCIe switches straight to accelerators, sustaining hundreds of millions of IOPS per server and keeping expensive compute silicon busy instead of idle.

Publishing 50+ videos from Supercomputing 2025 (SC25, St. Louis), and from other recent events, about 4 per day at 5AM, 11AM, 5PM and 11PM CET/EST.
Join https://www.youtube.com/charbax/join for early access to all my queued videos early.

Watch my full SC25 playlist:

This video was filmed using the DJI Pocket 3 ($669 at https://amzn.to/4aMpKIC using the dual wireless DJI Mic 2 microphones with the DJI lapel microphone https://amzn.to/3XIj3l8 ), watch all my DJI Pocket 3 videos here https://www.youtube.com/playlist?list=PL7xXqJFxvYvhDlWIAxm_pR9dp7ArSkhKK

Click the “Super Thanks” button below the video to send a highlighted comment under the video! Brands I film are welcome to support my work in this way 😁

Check out my video with Daylight Computer about their revolutionary Sunlight Readable Transflective LCD Display for Healthy Learning: https://www.youtube.com/watch?v=U98RuxkFDYY

source https://www.youtube.com/watch?v=4bQOG-emOK0

Giga Computing, the enterprise arm of Gigabyte, uses this booth tour to walk through its current data center stack, from ultra-dense CPU nodes to Blackwell-based GPU racks. The demo starts with a 3U direct liquid-cooled chassis packing ten blades, each motherboard hosting two independent nodes, yielding 20 single-socket servers in just 3U. Depending on configuration, nodes can be built around AMD Ryzen 7000/9000 or EPYC 4005 series, as well as Intel Xeon 6 6300-class processors, aimed at web hosting, game hosting and other high-density workloads where rack space and power efficiency matter. https://www.gigacomputing.com/en/

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

The tour then highlights a self-contained liquid-cooled EPYC 9005 workstation designed for quiet, desk-side AI and media workloads. CPU, memory, PSU and up to four GPUs sit on a closed DLC loop with radiator, pump and fans engineered to keep acoustic noise around 50 dB while sustaining full load. Front NVMe bays, optional M.2 boot devices, 10GbE networking and BMC remote management turn it into a compact studio or lab node for AI model development, 3D rendering or video post-production without needing data center plumbing. Filmed at Supercomputing 2025 in St. Louis, it shows how far workstation-class hardware has moved toward data center-class thermals.

On the memory side, Giga Computing shows a 48-DIMM 1U/2U EPYC 9005 platform, delivering up to multiple terabytes of DDR5 in a single node. With two DIMMs per channel and dual-socket CPUs, this class of server targets in-memory databases, caching tiers, large analytics workloads and virtualization clusters that are memory-bound rather than GPU-bound. Nearby, an 8U HGX/OAM tray separates compute and GPU tiers, supporting NVIDIA B200/B300 or AMD Instinct MI350/MI355X accelerators, with PLX switches, NVLink/NVSwitch topologies and dense front I/O tuned for training and inference clusters.

For smaller AI clusters or visualization backends, the booth introduces the NVIDIA MGX-based XL44-SX2 system populated with up to eight NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs. A built-in ConnectX-8 PCIe Gen6 switchboard with multiple QSFP 400G ports ties GPUs and network together, each CX8 ASIC wired to two GPUs and two ports, mirroring HGX-style topologies in a more compact chassis. Dual Intel Xeon 6700/6500 CPUs, dense DDR5, Gen5 NVMe bays and BlueField-3 DPU options make this platform relevant for generative AI, 3D pipelines and scientific computing in SMB data centers that don’t yet need full rack-scale Blackwell deployments.

At rack scale, Giga Computing showcases the fully liquid-cooled NVIDIA GB300 NVL72 architecture: 18 compute nodes with Blackwell Ultra B300 GPUs and Grace CPUs, nine NVLink switch trays, CDUs at the base and an OCP Open Rack busbar spine. All 72 GPUs are interconnected via NVLink so the rack behaves like a single accelerator, while facility water loops attach to the CDU heat exchangers. The tour finishes with PCIe GPU servers for H200-class GPUs, RTX PRO 6000 and Intel Gaudi 3, Xeon 6 platforms with CXL memory expansion, and AmpereOne-based servers optimized for high-core-count ARM inference. Together with the GPM management software layer for Kubernetes, Slurm and MLOps orchestration, the booth underlines Giga Computing’s push toward dense, liquid-cooled, rack-ready AI and HPC infrastructure.

Publishing 50+ videos from Supercomputing 2025 (SC25, St. Louis), and from other recent events, about 4 per day at 5AM, 11AM, 5PM and 11PM CET/EST.
Join https://www.youtube.com/charbax/join for early access to all my queued videos early.

Watch my full SC25 playlist:

Check out my video with Daylight Computer about their revolutionary Sunlight Readable Transflective LCD Display for Healthy Learning: https://www.youtube.com/watch?v=U98RuxkFDYY

source https://www.youtube.com/watch?v=uImTUVjGHFA

Carma is a media intelligence and technology company that turns global news and social data into usable evidence for PR, communications and marketing teams. By combining automated media monitoring, configurable dashboards, alerting and expert analysis, it helps organisations understand visibility, sentiment and business outcomes across print, online, broadcast and social channels, all in near real time. More information about the CARMA Insight software platform is available at https://carma.com/products-and-services/software/

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

At Web Summit Lisbon 2025, Vera Santos demonstrates how Carma has built a dedicated dashboard for the conference that refreshes every 30 minutes and tracks everything being said about the event worldwide. Organisers, sponsors and exhibitors can see in real time which outlets are covering them, which regions are driving mentions, how stories evolve over the day and how their presence compares with peers, all from a single, shared interface fed by live crawlers and analysts.

Beyond this specific deployment, the video walks through Carma’s broader service stack: continuous media monitoring, customised consultancy, campaign and reputation measurement and market research integrations. The company works with clients across financial services, automotive, ESG, FMCG and government, mapping audience journeys and evaluating how communications strategy performs across the full PESO media mix, from traditional editorial to social media and blogs.

Technically, Carma operates a hybrid pipeline that blends large-scale crawlers with human-curated reading lists to capture relevant coverage from online news, TV and radio, newsletters, blogs and user-generated content. AI models assist with language detection, topic and entity classification, sentiment analysis, tonality, influence scoring and crisis flagging, while human coders refine the edge cases so that dashboards, alerts and executive briefings remain accurate enough for board-level reporting.

The platform focuses on what is organically said about brands, products and organisations on channels such as Instagram, TikTok, YouTube, X, Reddit and local forums, separating paid placements from earned visibility. PR and communications teams use Carma’s reporting to demonstrate impact, attribute outcomes to campaigns, benchmark performance over time and adjust strategy during crises or major launches, underpinned by methodologies recognised with 31 awards at the 2025 AMEC Measurement and Evaluation Awards.([CARMA][2])

I filmed 70+ videos at Web Summit Lisbon 2025 I will publish them over the coming days/weeks into my Web Summit playlist: https://www.youtube.com/playlist?list=PL7xXqJFxvYvhGWhynTmvAvDvohtO3qjZm I publish one video (from this and from other events I recently filmed at) every 6 hours at 5AM/11AM/5PM/11PM CET/EST. Be a member on my https://youtube.com/charbax/join to get early access to all my videos (after they’ve been queued up for publishing)

This video was filmed using the DJI Pocket 3 ($669 at https://amzn.to/4aMpKIC using the dual wireless DJI Mic 2 microphones with the DJI lapel microphone https://amzn.to/3XIj3l8 ), watch all my DJI Pocket 3 videos here https://www.youtube.com/playlist?list=PL7xXqJFxvYvhDlWIAxm_pR9dp7ArSkhKK

Click the “Super Thanks” button below the video to send a highlighted comment under the video! Brands I film are welcome to support my work in this way 😁

Check out my video with Daylight Computer about their revolutionary Sunlight Readable Transflective LCD Display for Healthy Learning: https://www.youtube.com/watch?v=U98RuxkFDYY

source https://www.youtube.com/watch?v=zt6j6fPMIqk

This video tours Renesas demos focused on vision AI robotics and sensor-rich human–machine interfaces. At the booth we see gesture-controlled dexterous hands and humanoid robots driven by embedded vision models on RZ MPUs with the DRP-AI accelerator, able to recognize finger counts, interpret hand poses and classify scenes such as humans, dogs or crowded spaces for context-aware behavior. More details on Renesas vision AI can be found at https://www.renesas.com/en/key-technologies/artificial-intelligence/ai-vision

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

Recorded at Embedded World North America in Anaheim, the conversation highlights how Renesas scales the same RZ/V vision AI platform from compact camera modules to full robotic systems, combining convolutional-neural-network inference with classic computer-vision preprocessing on chip. This enables responsive gesture HMI, people and object detection, and crowd-density estimation at the edge without active cooling, in line with recent RZ/V2H and RZ/V2N devices that pair DRP-AI3 engines with quad Arm Cortex-A55 application processors for demanding workloads in deployment.

The tour then shifts to the sensor portfolio, starting from impedance-based touch sensing embedded into a robotic lawn-mower style handlebar that can detect one or two hands, slider-type interactions and virtual buttons for speed or mode control. Under the hood, Renesas capacitive and impedance sensors measure tiny changes in capacitance and electrical impedance to implement robust touch and touchless HMIs, a technique the company also showcases in hand-on and steering-wheel detection demos for automotive safety and industrial controls, all running on a unified platform.

Finally, walks through inductive position-sensing demos that show how PCB-coil-based, magnet-free IPS-class sensors track the relative position of metal targets and motor shafts inside a robotic joint, providing high-resolution angle feedback with strong immunity to stray magnetic fields compared with traditional resolvers. Combined with compact impedance-sensing evaluation boards and the broader Renesas embedded hardware–software ecosystem, these building blocks give engineers a coherent path from evaluation kits to production-grade humanoid robotics and collaborative automation hardware.

I’m publishing about 90+ videos from Embedded World North America 2025, I upload about 4 videos per day at 5AM/11AM/5PM/11PM CET/EST. Join https://www.youtube.com/charbax/join for Early Access to all 90 videos (once they’re all queued in next few days) Check out all my Embedded World North America videos in my Embedded World playlist here: https://www.youtube.com/playlist?list=PL7xXqJFxvYvjgUpdNMBkGzEWU6YVxR8Ga

This video was filmed using the DJI Pocket 3 ($669 at https://amzn.to/4aMpKIC using the dual wireless DJI Mic 2 microphones with the DJI lapel microphone https://amzn.to/3XIj3l8 ), watch all my DJI Pocket 3 videos here https://www.youtube.com/playlist?list=PL7xXqJFxvYvhDlWIAxm_pR9dp7ArSkhKK

Click the “Super Thanks” button below the video to send a highlighted comment under the video! Brands I film are welcome to support my work in this way 😁

Check out my video with Daylight Computer about their revolutionary Sunlight Readable Transflective LCD Display for Healthy Learning: https://www.youtube.com/watch?v=U98RuxkFDYY

source https://www.youtube.com/watch?v=6LGkp0V6i0s

Virscient is presented here as a specialist wireless engineering consultancy that takes connected products from concept to full-stack implementation, covering embedded software, PCB design, RF and antenna tuning, security, certification and mobile apps for IoT and automotive devices. The interview with Head of Sales Rick Walker walks through concrete projects that show how the team in Hamilton, New Zealand supports customers in Europe, the US and beyond with low-power wireless, cellular IoT, Bluetooth and cloud-based telemetry across multiple verticals and form factors, building on partnerships with major semiconductor vendors and a mature wireless engineering portfolio. https://www.virscient.com/

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

One focus in the video is an indoor air-quality and classroom environment monitor initially developed for New Zealand’s education sector. Virscient designed the complete system, combining a Nordic-based cellular module, Bluetooth connectivity, and a low-power e-ink display with multi-year battery life while continuously sampling temperature, humidity, CO₂ and acoustic noise levels. Data is streamed securely to a cloud backend and exposed through dashboards and a Bluetooth companion app, enabling facility managers and school leaders to assess comfort, ventilation and acoustic conditions, and to correlate those parameters with attendance, well-being and academic learning.

From there the discussion shifts to vehicular telemetry and active suspension control. For Fox’s off-road and e-bike suspension business, Virscient built the electronics and software that tap the vehicle ECU over CAN bus, stream shock and motion data in real time, and expose configuration profiles such as on-road, trail or desert with comfort, balanced and sport modes. That same data is mirrored to the cloud and to a Bluetooth mobile app, enabling parameter logging, remote diagnostics and potentially over-the-air tuning of damping strategies for pickup trucks and other off-road platforms, linking embedded firmware, RF links and backend analytics into a single data path for control and telemetry.

Another demo centres on marine lighting for luxury yachts, where bulkheads and hull construction make traditional cabling impractical. Virscient implemented a proprietary Bluetooth mesh-style network with secure commissioning, group addressing and scene control so that entire lighting zones can be reconfigured from a tablet app with one tap. By simplifying the mesh topology relative to the standard Bluetooth Mesh specification and optimising for low latency and robust RF behaviour in reflective marine environments, they achieve responsive, fine-grained control of colour, brightness and dynamic patterns across the vessel’s interior and exterior lighting.

The final part of the conversation touches on how broadly this model applies: Virscient’s 40–50 engineers work on everything from geospatial positioning equipment to dairy-farm telemetry that tracks milk production and feed intake per cow, as well as wireless audio, ultra-low-latency links and multi-radio coexistence for complex IoT products. Their licensable software platforms for embedded connectivity and wireless audio, combined with partnerships around Wi-Fi/Bluetooth combo ICs and Bluetooth Audio SoCs, allow OEMs to add Wi-Fi, Bluetooth LE, cellular IoT, GNSS, LoRa or 802.15.4 to constrained devices without building RF, protocol stacks and certification expertise in-house, shortening time to market for connected hardware across industrial, consumer and automotive deployment.

I’m publishing about 90+ videos from Embedded World North America 2025, I upload about 4 videos per day at 5AM/11AM/5PM/11PM CET/EST. Join https://www.youtube.com/charbax/join for Early Access to all 90 videos (once they’re all queued in next few days) Check out all my Embedded World North America videos in my Embedded World playlist here: https://www.youtube.com/playlist?list=PL7xXqJFxvYvjgUpdNMBkGzEWU6YVxR8Ga

source https://www.youtube.com/watch?v=ibiJXIYxOe0

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

Astra AI is building a dedicated one-on-one tutoring platform around large language models for school and university learners, turning an AI chatbot into a structured study companion across math, physics, chemistry and other subjects. Students interact through a conversational interface but always within the context of homework help, step-by-step explanations and targeted exam preparation, accessible on web, iOS and Android devices for everyday study.

At the core of Astra AI are two pillars: subject tutors and Exam Prep. Learners can talk to an AI tutor for math, physics, chemistry and more, or generate a custom exam preparation plan by selecting a subject, setting the exam date and mastery goal, and uploading their own materials such as class notes, past tests or slides. The system uses this corpus plus the student’s declared school level to tailor the difficulty, sequence of problems and explanations to their curriculum and context.

In this interview, filmed at Web Summit Lisbon 2025, the team walks through the interface: a chat-centric workspace with a math keyboard, calculator, drawing area and camera input so you can snap a photo of a problem. Students can switch between Solver Mode, which focuses on efficient step-by-step solutions, and Socratic Mode, which behaves more like a human tutor, asking guiding questions and probing understanding rather than just emitting final answers in a vacuum.

Astra AI follows a freemium model: new users can register and use the tutor with a daily question limit and then upgrade to a subscription (around €24 per month or €120 per year) for unrestricted usage. According to the team, the platform has attracted more than 600,000 learners, starting from Slovenia and now growing quickly in German-speaking markets, with specialized offerings such as Astra IB Math for International Baccalaureate exam preparation across standard and higher level tracks.

Compared with general-purpose chatbots like ChatGPT, Astra AI positions itself as an education-first layer on top of modern language models, optimizing for pedagogy rather than open-ended conversation. The discussion touches on challenges such as keeping students motivated, using light gamification, and designing UX that encourages regular returns to the app, whether on a phone, laptop or even e-paper style tablets like Daylight Computer, pointing toward a future where having a personal AI tutor is as normal as using a calculator in study.

I filmed 70+ videos at Web Summit Lisbon 2025 I will publish them over the coming days/weeks into my Web Summit playlist: https://www.youtube.com/playlist?list=PL7xXqJFxvYvhGWhynTmvAvDvohtO3qjZm I publish one video (from this and from other events I recently filmed at) every 6 hours at 5AM/11AM/5PM/11PM CET/EST. Be a member on my https://youtube.com/charbax/join to get early access to all my videos (after they’ve been queued up for publishing)

This video was filmed using the DJI Pocket 3 ($669 at https://amzn.to/4aMpKIC using the dual wireless DJI Mic 2 microphones with the DJI lapel microphone https://amzn.to/3XIj3l8 ), watch all my DJI Pocket 3 videos here https://www.youtube.com/playlist?list=PL7xXqJFxvYvhDlWIAxm_pR9dp7ArSkhKK

Click the “Super Thanks” button below the video to send a highlighted comment under the video! Brands I film are welcome to support my work in this way 😁

Check out my video with Daylight Computer about their revolutionary Sunlight Readable Transflective LCD Display for Healthy Learning: https://www.youtube.com/watch?v=U98RuxkFDYY

source https://www.youtube.com/watch?v=8do6lZEqR6E

PLLUM is presented here as a national family of Polish large language models designed specifically for public administration and Polish-speaking users, rather than as a general-purpose global chatbot. The models focus on correct Polish grammar, inflection and official terminology, minimizing negative transfer from English while staying compatible with modern transformer-based LLM architectures. They are released with open model weights so organizations can deploy them on-premise, adapt them to sector-specific use cases and integrate them into existing e-government workflows; more background on the project is available at https://pllum.org.pl/

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

Technically, PLLUM is a suite of transformer and mixture-of-experts models in the roughly 8–70B parameter range, including Mistral/Mixtral-derived variants such as PLLuM-12B and PLLuM-8x7B as well as Llama-based models and fully Polish-pretrained networks. Training relies on a large corpus dominated by Polish texts, enriched with selected Slavic/Baltic languages and some English, then refined through continuous pre-training, supervised fine-tuning, instruction tuning and preference optimization. The team evaluates linguistic and cultural competence with dedicated Polish benchmarks and focuses on high-quality “organic” data rather than unchecked web scrapes to maintain reliability in administration-heavy domains in Poland.

In the interview, Philip explains how PLLUM underpins Poland’s strategy for AI sovereignty: ministries can run the models on local infrastructure, keep personal and sensitive data within the country and avoid sending internal documents to foreign cloud APIs. The same models will power assistants in governmental systems such as the widely used mObywatel mobile app, enabling citizens to ask administrative questions in natural Polish and get answers grounded in official regulations. Recorded at Web Summit Lisbon 2025, the conversation frames PLLUM as a blueprint for language-and-country-specific LLM deployment rather than a clone of generic global chatbots, emphasizing governance, licensing and data control in this discussion.

PLLUM’s development mixes open European base models with entirely home-grown training runs: the team uses Mistral, Mixtral and Llama-style foundations for some variants, while also training models from random initialization on carefully curated Polish and Polish–English corpora. On top of that, they experiment with synthetic data generation, using other open models (including Chinese systems like DeepSeek or Qwen) on their own infrastructure to expand datasets while preserving control over quality and legal provenance. The resulting chat models are published with both permissive and non-commercial licenses, and the weights are available on Hugging Face so enterprises, municipalities and research labs can build their own domain-specific assistants within this ecosystem.

Philip also highlights the scale of the consortium behind PLLUM: more than a hundred people across six scientific institutions, spanning model engineering, linguistics and large annotation teams focused on clean administrative data. The project already counts around a million public prompts from early users, with upcoming deployments expected to reach millions of citizens as chatbots are rolled into local offices and nationwide services. Future work extends beyond text: speech recognition and text-to-speech are in testing, and there is clear interest in Polish-speaking voice interfaces and, eventually, avatar-style agents, turning PLLUM into a full-stack foundation for sovereign AI services across the Polish digital landscape.

I filmed 70+ videos at Web Summit Lisbon 2025 I will publish them over the coming days/weeks into my Web Summit playlist: https://www.youtube.com/playlist?list=PL7xXqJFxvYvhGWhynTmvAvDvohtO3qjZm

I publish one video (from this and from other events I recently filmed at) every 6 hours at 5AM/11AM/5PM/11PM CET/EST.

source https://www.youtube.com/watch?v=1NmiamJPWuU

u-blox is a positioning and wireless communications specialist that builds GNSS and short-range connectivity modules so OEMs can integrate location and connectivity without becoming RF experts. Their portfolio spans automotive, industrial and consumer IoT, with modules that encapsulate RF design, firmware, certifications and security, letting engineering teams focus on system architecture, edge AI and application software instead of radio hardware details. https://www.u-blox.com/en/product/pointperfectflex

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

At the booth, the focus is on high-precision GNSS: single-band, dual-band and all-band receivers that see every relevant constellation in the sky to deliver robust, low-latency positioning for autonomous vehicles, UAVs, robotic lawn mowers and fleet tracking. Recent triple-band receivers on the X20 platform support L1/L2/L5 with RTK and PPP-RTK, enabling centimeter-level accuracy with fast convergence for dynamic platforms like robots and drones([u-blox][1]). Combined with the PointPerfect family of correction services, u-blox can take GNSS performance from meter-level to a few centimeters at continental or global scale for demanding guidance and control accuracy

On the short-range side, u-blox showcases NORA and MAYA module families that bring Bluetooth Low Energy, Wi-Fi 6 and 802.15.4 (Thread/Zigbee/Matter) into compact, production-ready packages([u-blox][3]). A Bluetooth LE channel-sounding demo with electronic dice demonstrates fine-grained ranging and indoor positioning using NORA-B2, while MAYA-W2 and MAYA-W4 tri-radio Wi-Fi 6 modules target IoT gateways, industrial controllers and embedded Linux boards such as NXP i.MX93 reference designs, where developers can prototype both application processor firmware and wireless connectivity on a single platform. This creates a coherent environment for building dense, connected sensing and control networks.

A key message in this interview is the platform concept: within a family name like NORA or MAYA, pin-compatible variants allow designers to swap between BLE-only, Wi-Fi + BLE or tri-radio modules without respinning the PCB when requirements evolve. That reduces EOL risk and BOM churn over a product’s lifetime, while manufacturing out of Europe with full production test on every module positions u-blox as a value-focused, “Western supply chain” option compared with ultra-low-cost chip vendors. Customers pay for predictable RF behavior, long-term availability and direct engineering support instead of redesign cycles and field failures, which is critical in safety-relevant or certified systems where change control is strict.

Filmed at Embedded World North America in Anaheim, this walkthrough connects those building blocks into a broader story: GNSS receivers, PPP-RTK and RTK correction services like PointPerfect Flex and PointPerfect Global, short-range modules and cloud delivery (Thingstream) combine into an end-to-end stack for “locate and communicate” use cases([u-blox][4]). Whether it’s a robotic lawn mower staying inside a virtual boundary instead of buried wire, a delivery robot navigating dense urban canyons, or an industrial asset being tracked at centimeter-level outdoors and via Bluetooth ranging indoors, the idea is to give developers modular hardware, consistent form factors and a services layer that together accelerate reliable, high-precision connected hardware.

I’m publishing about 90+ videos from Embedded World North America 2025, I upload about 4 videos per day at 5AM/11AM/5PM/11PM CET/EST. Join https://www.youtube.com/charbax/join for Early Access to all 90 videos (once they’re all queued in next few days) Check out all my Embedded World North America videos in my Embedded World playlist here: https://www.youtube.com/playlist?list=PL7xXqJFxvYvjgUpdNMBkGzEWU6YVxR8Ga

source https://www.youtube.com/watch?v=hQWxS3UjW0Q

Altium uses its Altium Designer, Altium 365 cloud platform and new SaaS lineup to cover the full PCB development lifecycle, from requirements capture and schematic design through layout, simulation, BOM management and release to manufacturing in one unified ECAD environment on Altium Develop https://www.altium.com/develop

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

In this interview from Embedded World North America 2025 in Anaheim, the team explains how Altium Develop targets smaller organizations that want shared workspaces, browser-based reviews and structured design workflows without heavy governance overhead, while Altium Agile Enterprise extends the same toolchain with stronger compliance, access control and process orchestration for regulated hardware delivery.

The demo walks through a 3D PCB view where electrical engineers, mechanical designers, procurement and requirements owners all see the same live project context, including component usage, lifecycle status and supply-chain risk, while ECAD data is connected to PLM, MRP and ERP systems so that design outputs, manufacturing documentation and change histories stay part of a single digital thread across the product lifecycle.

They also discuss simulation as an integral part of this stack: SPICE-based circuit analysis, layout-aware checks for EMI and power integrity, bus-functional models for microcontrollers and integrations with tools from partners like Ansys and Keysight so that signal, thermal and power behavior can be explored before committing to prototypes, reducing re-spins and tightening design-for-manufacturing performance.

Finally, as Altium is now part of the Renesas group, the conversation touches on how reference designs, firmware and embedded compute platforms from the semiconductor side align with cloud-native ECAD, requirements management and manufacturing release, giving embedded teams a clearer path from initial product concept through to compliant, traceable hardware ready for volume production roadmap.

I’m publishing about 90+ videos from Embedded World North America 2025, I upload about 4 videos per day at 5AM/11AM/5PM/11PM CET/EST. Join https://www.youtube.com/charbax/join for Early Access to all 90 videos (once they’re all queued in next few days) Check out all my Embedded World North America videos in my Embedded World playlist here: https://www.youtube.com/playlist?list=PL7xXqJFxvYvjgUpdNMBkGzEWU6YVxR8Ga

This video was filmed using the DJI Pocket 3 ($669 at https://amzn.to/4aMpKIC using the dual wireless DJI Mic 2 microphones with the DJI lapel microphone https://amzn.to/3XIj3l8 ), watch all my DJI Pocket 3 videos here https://www.youtube.com/playlist?list=PL7xXqJFxvYvhDlWIAxm_pR9dp7ArSkhKK

Click the “Super Thanks” button below the video to send a highlighted comment under the video! Brands I film are welcome to support my work in this way 😁

Check out my video with Daylight Computer about their revolutionary Sunlight Readable Transflective LCD Display for Healthy Learning: https://www.youtube.com/watch?v=U98RuxkFDYY

source https://www.youtube.com/watch?v=RoOn4ft2PnI

Toradex showcases how its new single board computer built around the Synaptics Astra SL1680 SoC turns the familiar Raspberry Pi form factor into a production-ready industrial platform, shipping out of the box with the Torizon embedded Linux stack and cloud-managed over-the-air updates. The SL1680 SBC reuses existing cases, HATs and cabling, but adds an 8-TOPS edge AI NPU, hardened Linux distribution and secure lifecycle tooling aimed at long-lived IoT deployments ([toradex.com][1]) https://www.toradex.com/single-board-computers/sl1680

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

From there the conversation dives into the developer experience around Torizon: container-based application deployment, OS customization, and tight integration with a Visual Studio Code extension that can talk directly to the boards. At Embedded World North America 2025 in Anaheim, Toradex emphasizes that the same platform used for rapid prototyping scales to fleet management in Torizon Cloud, with features such as automated CVE scanning, SBOM generation and VEX reports to support EU Cyber Resilience Act (CRA) compliance

Vertical demos make this concrete. A marine HMI for high-end yachts runs a Qt-based UI on Verdin i.MX 8M Plus hardware, while a medical-style interface uses the LVGL lightweight graphics library on the Verdin AM62P module. Nearby, a Verdin i.MX95 evaluation kit shows the same Toradex carrier design running both QNX and Linux, underlining how the hardware and Torizon ecosystem can support mixed-criticality systems where real-time tasks coexist with rich graphics and connected services.

The hardware table walks through the broader portfolio: multiple Verdin System-on-Modules, the high-core-count Achila AM69 board for compute-dense edge AI, and production-ready carrier boards like the Malo platform plus partner-built industrial PCs. A multi-camera inference demo routes several USB cameras into the Achila AM69, performing on-device object and face detection at the edge, while customer enclosures from companies like Splunk illustrate how these SoMs and carrier boards end up as sealed industrial gateways with CAN, antenna connectors and robust I/O in the field.

Security and lifecycle management are the cross-cutting themes. Toradex highlights Torizon’s Vulnerability Manager and CRA content: continuous CVE monitoring, triage of exploitable issues, SBOM-driven traceability and managed update pipelines designed so OEMs can meet the CRA’s December 2027 enforcement horizon without building their own infrastructure from scratch The T-Bot self-balancing robot running split workloads across Cortex-A Linux and Cortex-M real-time cores is a playful reminder that, underneath the regulatory and security story, these platforms are still about enabling inventive embedded products that stay secure, maintainable and updatable over many years.

I’m publishing about 90+ videos from Embedded World North America 2025, I upload about 4 videos per day at 5AM/11AM/5PM/11PM CET/EST. Join https://www.youtube.com/charbax/join for Early Access to all 90 videos (once they’re all queued in next few days) Check out all my Embedded World North America videos in my Embedded World playlist here: https://www.youtube.com/playlist?list=PL7xXqJFxvYvjgUpdNMBkGzEWU6YVxR8Ga

This video was filmed using the DJI Pocket 3 ($669 at https://amzn.to/4aMpKIC using the dual wireless DJI Mic 2 microphones with the DJI lapel microphone https://amzn.to/3XIj3l8 ), watch all my DJI Pocket 3 videos here https://www.youtube.com/playlist?list=PL7xXqJFxvYvhDlWIAxm_pR9dp7ArSkhKK

Click the “Super Thanks” button below the video to send a highlighted comment under the video! Brands I film are welcome to support my work in this way 😁

Check out my video with Daylight Computer about their revolutionary Sunlight Readable Transflective LCD Display for Healthy Learning: https://www.youtube.com/watch?v=U98RuxkFDYY

source https://www.youtube.com/watch?v=GkMvk9_qr4I

G Camp Taiwan is a government-backed global accelerator for Taiwanese startups, initiated by the Small and Medium Enterprise and Startup Administration (MOEA) and executed by the Institute for Information Industry (III). The program combines intensive training, international market immersion and investor matchmaking to help founders refine their business models, sharpen their pitch, and design realistic go-to-market and fundraising roadmaps for Europe, North America and beyond. Its “Growing, Global, Go” framework is designed to turn deep-tech and impact-driven startups into globally scalable companies that can compete on the world stage. https://www.facebook.com/gcamptaiwan/

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

In this conversation filmed at Web Summit Lisbon 2025, Walis Lin explains how each yearly cohort starts with an open call, followed by a multi-stage selection and a 2–3 day bootcamp with Silicon Valley and global mentors. Startups work on investor storytelling, unit economics, cross-border legal and regulatory topics, and channel strategy, then enter a final pitch competition. The winners earn places at large international conferences such as Web Summit, where they get access to thousands of potential partners, customers and investors in a single dense environment.

The 2025 group spans multiple domains: biotech startups working on hemostatic powder for trauma care and AI-assisted drug discovery, precision oncology companies like BloodScan Biotech focusing on circulating tumor cell (CTC) enrichment and liquid biopsy, as well as design and branding platforms using AI to accelerate digital content workflows. On the hardware side, teams such as Megasense build radar- and LiDAR-based railway intrusion detection, combining 4D sensing, embedded systems and edge AI to prevent track obstructions and collisions, while materials ventures explore novel metallurgical processes that can feed into semiconductor and advanced manufacturing supply chains to improve safety.

Other teams tackle systemic sustainability and infrastructure problems. PackAge+ drives circular packaging for e-commerce and logistics, building reusable packaging-as-a-service models that reduce single-use plastics, support ESG targets and lower lifecycle CO2 emissions through reverse logistics and data-driven reuse tracking. Cybersecurity company Rapid International focuses on digital forensics and threat monitoring platforms for enterprises, while a mattress-recycling startup turns end-of-life mattresses into new packaging materials, illustrating how the program intentionally mixes medtech, climate tech, industrial IoT and cybertech in one cohort to encourage cross-sector learning.

G Camp also acts as a long-term internationalization pipeline for Taiwan’s startup ecosystem. Since 2018 it has supported dozens of teams that went on to exhibit at Web Summit, SXSW, Slush and other global stages, with some even establishing new subsidiaries in Europe after contacts made at Lisbon. By combining Taiwan’s strengths in hardware, semiconductors and manufacturing with software, AI and circular-economy business models, the program helps founders validate products, localize for new markets and plug into international capital and corporate innovation networks, strengthening Taiwan’s position in the global startup ecosystem.

I’m publishing about 90+ videos from Embedded World North America 2025, I upload about 4 videos per day at 5AM/11AM/5PM/11PM CET/EST. Join https://www.youtube.com/charbax/join for Early Access to all 90 videos (once they’re all queued in next few days) Check out all my Embedded World North America videos in my Embedded World playlist here: https://www.youtube.com/playlist?list=PL7xXqJFxvYvjgUpdNMBkGzEWU6YVxR8Ga

Check out my video with Daylight Computer about their revolutionary Sunlight Readable Transflective LCD Display for Healthy Learning: https://www.youtube.com/watch?v=U98RuxkFDYY

source https://www.youtube.com/watch?v=ZONZK5DPDd4

Zephyr RTOS sits between bare-metal firmware and full Linux, giving developers a compact real-time operating system for connected microcontrollers when Linux would be too big. Kate Stewart describes how the Linux Foundation project has grown since 2016 into millions of deployed devices, from Chromebook embedded controllers and Intel platform firmware to deeply embedded IoT nodes, under an Apache 2.0, “for developers, by developers” model. https://zephyrproject.org/

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

She explains how the Zephyr kernel and its ecosystem—drivers, sensor and networking stacks, board support and test infrastructure—are increasingly reused beyond the core RTOS. The West meta-tool manages multi-repository projects, builds, flashing and debugging, while also generating software bills of materials (SBOMs) to help manufacturers align with requirements such as the EU Cyber Resilience Act and long-term maintenance expectations.

Recorded at Embedded World North America 2025 in Anaheim, the discussion then turns to edge AI on highly resource-constrained devices. Frameworks such as Edge Impulse and TensorFlow Lite Micro run on Zephyr to power gesture-recognition “magic wands”, satellite telemetry filters that only wake high-power FPGAs when needed, and simple on-device classifiers that pre-process sensor data before heavier cloud or gateway models.

Kate argues that observability for AI systems at the edge will require an “AI BOM” alongside the SBOM, capturing models and training data so failures can be debugged as rigorously as code. She also outlines Zephyr’s neutral governance and membership model, where project members fund shared CI, hardware farms and show presence, while any company or individual can still upstream code under the same technical steering processes.

The roadmap she sketches includes functional safety certification (IEC 61508 for industrial applications, ISO 26262 for automotive and interest in avionics-style DO-178 use), longer five-year LTS branches and tighter security workflows around coordinated disclosure, Bluetooth hardening and CVE-tracked fixes. The conversation closes with global community growth, from Zephyr tracks at Open Source Summit events to real products in agriculture, wind energy and even Antarctic wildlife tags, showing how a tiny RTOS can quietly span industries, regions and use cases.

Zephyr at Embedded World North America 2025: Tiny RTOS for Millions of MCUs, CRA and AI BOMs
Zephyr at Embedded World North America 2025: Neutral RTOS Governance, Edge AI at the Edge

I’m publishing about 90+ videos from Embedded World North America 2025, I upload about 4 videos per day at 5AM/11AM/5PM/11PM CET/EST. Join https://www.youtube.com/charbax/join for Early Access to all 90 videos (once they’re all queued in next few days) Check out all my Embedded World North America videos in my Embedded World playlist here: https://www.youtube.com/playlist?list=PL7xXqJFxvYvjgUpdNMBkGzEWU6YVxR8Ga

This video was filmed using the DJI Pocket 3 ($669 at https://amzn.to/4aMpKIC using the dual wireless DJI Mic 2 microphones with the DJI lapel microphone https://amzn.to/3XIj3l8 ), watch all my DJI Pocket 3 videos here https://www.youtube.com/playlist?list=PL7xXqJFxvYvhDlWIAxm_pR9dp7ArSkhKK

Click the “Super Thanks” button below the video to send a highlighted comment under the video! Brands I film are welcome to support my work in this way 😁

Check out my video with Daylight Computer about their revolutionary Sunlight Readable Transflective LCD Display for Healthy Learning: https://www.youtube.com/watch?v=U98RuxkFDYY

source https://www.youtube.com/watch?v=MaRuKJfjRQc

Peridio’s Avocado OS is an embedded Linux distribution built on the Yocto Project, designed specifically for teams shipping physical AI products and connected devices. Instead of treating the OS as a generic base image, Avocado OS treats it as part of the product, focusing on deterministic builds, reproducible runtimes, secure boot, disk encryption, and OTA-friendly layouts so fleets stay consistent from prototype to volume deployment. The goal is to give embedded teams a developer-friendly environment that still behaves like a production-grade Linux stack for edge AI hardware. https://www.peridio.com/avocado-os/

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

In this interview, the team explains how Avocado OS sits between Ubuntu/Debian and vanilla Yocto: you keep the binary-package ergonomics during development while still generating immutable images for production. That means you prototype and debug in a familiar Linux environment, then freeze the exact stack into a deterministic runtime without a six-month porting exercise. Because the same tooling supports hardware-in-the-loop testing, retargeting to new boards, and manufacturing modes, it is aimed at shortening the path from R&D lab to automated factory manufacturing.

The demo on the show floor represents a small industrial line: dual robotic arms pick and place candy onto a conveyor according to a user’s preference profile. An LLM and computer-vision pipeline infer pick paths offline at the edge, combining perception and simple task planning. Under the hood, Avocado OS is running across several nodes: a Seed Studio reTerminal DM, an NVIDIA Jetson AGX handling heavy vision workloads, and an Advantech smart camera with a Jetson Orin NX module. Together they illustrate how the same embedded Linux distribution can span HMI, real-time control, and high-throughput inference in a single line.

The interface itself is implemented as a WebKit-based HMI, serving a JavaScript application locally on the device. That approach lets teams reuse standard web development skills for industrial touchscreens and kiosks instead of committing to a proprietary HMI stack. Beyond Jetson, Avocado OS targets platforms like NXP i.MX and Raspberry Pi, and can coexist with microcontroller-based subsystems for ultra-low-power tasks, making Linux the control plane for heterogeneous edge AI hardware stack.

On the business side, Avocado OS is fully open source, with Peridio offering commercial LTS, OTA update orchestration, fleet observability and device management for long-lived products. That combination of immutable, security-hardened images (secure boot, dm-verity, LUKS), controlled rollouts, and remote diagnostics is aimed at operators who need to manage thousands of industrial robots, smart cameras, or HMIs over many years. As more edge AI workloads move to Linux, Peridio positions Avocado OS and its management layer as a way to keep development agile while keeping deployed fleets predictable at the edge.

I’m publishing about 90+ videos from Embedded World North America 2025, I upload about 4 videos per day at 5AM/11AM/5PM/11PM CET/EST. Join https://www.youtube.com/charbax/join for Early Access to all 90 videos (once they’re all queued in next few days)

Check out all my Embedded World North America videos in my Embedded World playlist here: https://www.youtube.com/playlist?list=PL7xXqJFxvYvjgUpdNMBkGzEWU6YVxR8Ga

This video was filmed using the DJI Pocket 3 ($669 at https://amzn.to/4aMpKIC using the dual wireless DJI Mic 2 microphones with the DJI lapel microphone https://amzn.to/3XIj3l8 ), watch all my DJI Pocket 3 videos here https://www.youtube.com/playlist?list=PL7xXqJFxvYvhDlWIAxm_pR9dp7ArSkhKK

Check out my video with Daylight Computer about their revolutionary Sunlight Readable Transflective LCD Display for Healthy Learning: https://www.youtube.com/watch?v=U98RuxkFDYY

source https://www.youtube.com/watch?v=hPS2H_fZdZk

Arduino’s new UNO Q brings a dual-brain architecture to makers and embedded developers: a quad-core Qualcomm Dragonwing QRB2210 application processor running Debian Linux alongside a low-power STM32U585 microcontroller for deterministic, real-time control. This pairing lets you mix Linux apps, Python, containers and local AI with classical Arduino sketches, GPIO and ISR-driven tasks on one board, keeping hard-real-time loops isolated from heavier user-space workloads. Entry pricing starts at $44 for the 2 GB model, with a 4 GB variant planned. https://www.arduino.cc/product-uno-q

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

In the demo, Arduino shows typical heterogeneous workflows: MCU-side RTOS handling fast actuation and sensor IO, while the QRB2210 hosts web services, on-device inference and media pipelines. The QRB2210 integrates an Adreno-class GPU and dual ISPs, enabling camera-centric use cases and GPU-assisted inference (no discrete NPU), making it suitable for edge vision, gesture recognition and local model evaluation without mandating a cloud round-trip.

A major software piece is Arduino App Lab, pre-installed on UNO Q. It introduces “Bricks”—modular components for data storage, messaging, audio/image classification and cloud connectors—so you can compose pipelines that tie sketches, Python and Linux services together, or push datasets to Arduino Cloud. The same environment exposes a CLI for packaging and launching apps, and integrates with Arduino Project Hub to share reproducible builds with the community.

Hardware continuity matters here: UNO Q retains the classic UNO form factor and shield pinout for broad first- and third-party compatibility, while adding high-speed headers for displays, cameras and additional sensors. Schematics and gerbers are published under open licensing, preserving Arduino’s open-source model and easing a path from shield-based prototyping to carrier- or chip-down designs around QRB2210 when projects mature.

Context for the moment: this platform follows Arduino’s announcement that it is joining the Qualcomm family, with both companies stating Arduino will keep brand identity and multi-vendor support as they target education, industrial IoT and edge-AI developers. In practical terms, that means broader access to Qualcomm silicon, toolchains and camera/graphics stacks for Arduino’s reported 33 million-strong user base. Filmed at Embedded World North America, this conversation frames UNO Q as a bridge between classroom-friendly sketches and production-grade Linux robotics, vision and IoT systems.

I’m publishing about 90+ videos from Embedded World North America 2025, I upload about 4 videos per day at 5AM/11AM/5PM/11PM CET/EST. Join https://www.youtube.com/charbax/join for Early Access to all 90 videos (once they’re all queued in next few days) Check out all my Embedded World North America videos in my Embedded World playlist here: https://www.youtube.com/playlist?list=PL7xXqJFxvYvjgUpdNMBkGzEWU6YVxR8Ga

This video was filmed using the DJI Pocket 3 ($669 at https://amzn.to/4aMpKIC using the dual wireless DJI Mic 2 microphones with the DJI lapel microphone https://amzn.to/3XIj3l8 ), watch all my DJI Pocket 3 videos here https://www.youtube.com/playlist?list=PL7xXqJFxvYvhDlWIAxm_pR9dp7ArSkhKK

Click the “Super Thanks” button below the video to send a highlighted comment under the video! Brands I film are welcome to support my work in this way 😁

Check out my video with Daylight Computer about their revolutionary Sunlight Readable Transflective LCD Display for Healthy Learning: https://www.youtube.com/watch?v=U98RuxkFDYY

source https://www.youtube.com/watch?v=NMUftqFIQ-g

Trenz Electronic explains how a German design house can take an FPGA concept from schematic to series production under one roof. They design and assemble the modules, run bring-up and validation, and ship board support packages with reference designs so teams can plug in a SoM, boot a project, and iterate without re-spinning base silicon. U.S. customers work via Concurrent EDA as value-add distributor and systems partner, especially for turnkey deployments that combine Trenz hardware with domain-specific IP. https://www.trenz-electronic.de/

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

The discussion centers on adaptive compute for edge workloads, using AMD Versal devices to run image classification, translation, and text output pipelines on the AI Tiles while offloading control to programmable logic and embedded compute. Engineers pick FPGAs for deterministic latency, massive parallelism, and the ability to retarget dataflow without the thermal and memory overheads of general-purpose CPUs. In practice, customers prototype quickly on carrier boards, then scale with footprint-compatible SoMs across a few dozen product families to match cost, I/O, and power envelopes to deployment needs that evolve over time like machine vision

Demand for Versal has outpaced any prior Trenz family launch, and interest is compounding with Versal Gen 2 roadmaps that promise an order-of-magnitude more compute per watt for embedded AI. Trenz’s approach is to start early with pre-production silicon, validate designs before mass availability, and ship modules that boot cleanly on day one. Concrete examples visible in their catalog include a Versal AI Edge SoM around the VE2302 and a TE0950 evaluation board class, pairing fast DDR, OSPI/eMMC boot, and multi-lane GTY connectivity to stand up real applications like perception and text generation at the edge without cloud round-trips like latency

A recurring theme is lifecycle management. Trenz supports platforms for as long as silicon is available—often 20–25 years in industrial lines—and prepares form-fit-function successors when parts approach end-of-life. During supply shocks, they guide customers to nearest-equivalent AMD families (e.g., migrating designs validated on older Spartan or 7-series parts onto Zynq/Artix or Versal footprints) and provide design files or last-time-buy strategies as needed. That obsolescence playbook keeps fielded systems stable while preserving pinout and timing margins across revisions like continuity

Finally, the team highlights why “everything in one building” matters: same-day feedback loops between layout, FPGA, and test reduce bring-up risk and compress time-to-market. Customers can A/B different FPGA families on the same baseboard, quantify throughput vs. power on real workloads, and lock the bill of materials with confidence. While many end uses are under NDA, examples span medical instrumentation, industrial control, test and measurement, defense, and even space-rated subsystems—evidence that configurable compute still earns its place when determinism and longevity matter more than generic throughput like resilience

I’m publishing about 90+ videos from Embedded World North America 2025, I upload about 4 videos per day at 5AM/11AM/5PM/11PM CET/EST. Join https://www.youtube.com/charbax/join for Early Access to all 90 videos (once they’re all queued in next few days) Check out all my Embedded World North America videos in my Embedded World playlist here: https://www.youtube.com/playlist?list=PL7xXqJFxvYvjgUpdNMBkGzEWU6YVxR8Ga

Click the “Super Thanks” button below the video to send a highlighted comment under the video! Brands I film are welcome to support my work in this way 😁

Check out my video with Daylight Computer about their revolutionary Sunlight Readable Transflective LCD Display for Healthy Learning: https://www.youtube.com/watch?v=U98RuxkFDYY

source https://www.youtube.com/watch?v=RwXuj4fi_HA

Advantech outlines a coherent edge-AI roadmap spanning NVIDIA Jetson Thor, Qualcomm Dragonwing IQ9, AMD Ryzen Embedded, Intel Core Ultra and NXP i.MX families, with systems and dev kits positioned for on-prem inference, robotics control and machine vision. Their AIR-075 and EPC-class edge computers target high-throughput LLM and multimodal inference at the edge, while new carrier boards accept Thor-series modules and expose high-bandwidth camera and field I/O for deterministic pipelines. The emphasis is on keeping data local for latency, privacy and cost control, with SKUs tuned for FP4/FP8 AI math, CUDA/TensorRT acceleration, and mixed-precision workloads across GPU, NPU and CPU paths. https://www.advantech.com/

—
HDMI® Technology is the foundation for the worldwide ecosystem of HDMI-connected devices; integrated with displays, set-top boxes, laptops, audio video receivers and other product types. Because of this global usage, manufacturers, resellers, integrators and consumers must be assured that their HDMI® products work seamlessly together and deliver the best possible performance by sourcing products from licensed HDMI Adopters or authorized resellers. For HDMI Cables, consumers can look for the official HDMI® Cable Certification Labels on packaging. Innovation continues with the latest HDMI 2.2 Specification that supports higher 96Gbps bandwidth and next-gen HDMI Fixed Rate Link technology to provide optimal audio and video for a wide range of device applications. Higher resolutions and refresh rates are supported, including up to 12K@120 and 16K@60. Additionally, more high-quality options are supported, including uncompressed full chroma formats such as 8K@60/4:4:4 and 4K@240/4:4:4 at 10-bit and 12-bit color.
—

From robotics to medical imaging and industrial inspection, Advantech highlights a broad set of use cases that benefit from tight sensor-to-compute integration and time-sensitive control loops. The robotics inference and development kit bundles GMSL/MIPI camera inputs, synchronized machine-vision capture and a modular carrier so teams can prototype AMRs, autonomous shuttles or surgical assistance systems with a clear path from POC to deployment. Developers can start with pre-integrated drivers and SDKs, then scale to custom carrier boards or ruggedized enclosures without rewriting the software stack for production readiness

A second track focuses on ARM-based edge AI using Qualcomm’s Dragonwing IQ9 platform inside compact AIR-055 class boxes. This brings on-device AI up to triple-digit TOPS with Hexagon NPU operators, high-efficiency Kryo CPU cores and CSI-2/GMSL camera ingress for vision AI in tight power envelopes. Paired with the Edge Impulse toolchain, the kits accelerate dataset iteration, operator selection and deployment of quantized models for tasks like anomaly detection, OCR, pallet counting, or worker-safety analytics in facilities with constrained connectivity and strict data-retention policies

Advantech also shows x86 and heterogeneous options when discrete GPUs aren’t practical. Ryzen Embedded 8000 demonstrates CPU, integrated GPU and NPU concurrency for multimedia pipelines and classic CV, while 12th-to-14th-gen Intel systems add MXM form-factor GPUs when PCIe slots are limited. For customers standardizing on SOMs, Advantech’s ecosystem covers solder-down OSM modules and “goldfinger” boards, easing thermal design and lifecycle management across performance tiers and chassis constraints in field deployments

Filmed at Embedded World North America, the discussion underscores how partners like NVIDIA and Qualcomm inform Advantech’s reference designs, software enablement and early access programs. The through-line is platform breadth plus vertical focus: camera-rich dev kits for autonomy, air-gapped inference boxes for regulated environments, and configurable I/O for brownfield industrial lines, all aiming to shorten the distance from lab demo to robust, maintainable edge systems in production

I’m publishing about 90+ videos from Embedded World North America 2025, I upload about 4 videos per day at 5AM/11AM/5PM/11PM CET/EST. Join https://www.youtube.com/charbax/join for Early Access to all 90 videos (once they’re all queued in next few days) Check out all my Embedded World North America videos in my Embedded World playlist here: https://www.youtube.com/playlist?list=PL7xXqJFxvYvjgUpdNMBkGzEWU6YVxR8Ga

This video was filmed using the DJI Pocket 3 ($669 at https://amzn.to/4aMpKIC using the dual wireless DJI Mic 2 microphones with the DJI lapel microphone https://amzn.to/3XIj3l8 ), watch all my DJI Pocket 3 videos here https://www.youtube.com/playlist?list=PL7xXqJFxvYvhDlWIAxm_pR9dp7ArSkhKK

source https://www.youtube.com/watch?v=71XNPrM3wGQ