This week we have launched a new feature that will help those sorting out tax exemptions, something that no engineer ever likes to deal with!

You can now submit your tax-exemption certificates online, removing the need to contact Ponoko.

To add a tax exemption certificate, you can visit your Ponoko profile, click “add tax exemption” and fill in your exemption details.
 

Space Forge has sparked plasma aboard ForgeStar-1, and must admire the audacity of this development because it’s the first commercial satellite to attempt semiconductor manufacturing in orbit. While many dream of space-based tech, this step proves extreme conditions for crystal growth can be tamed outside Earth’s gravity, and it hints at cleaner, higher-performance materials for power electronics and quantum systems. I find the hybrid model particularly intriguing, as returning space-grown seeds to Earth bridges ambition with practicality, and it could reshape how we think about high-end semiconductors without abandoning terrestrial infrastructure.

Read the full article here

Apple is turning heads again, because after years of relying solely on Sony, it’s bringing Samsung into the iPhone camera fold, and the move makes sense. Samsung’s hybrid-bonded CMOS sensors promise faster readouts and better light capture, and this partnership could secure Apple a more resilient supply chain without touching its TSMC logic chips. It's impressive that such high-stakes engineering decisions happen quietly behind the scenes, yet they shape the devices millions rely on, and is a reminder that innovation often comes from strategic collaboration as much as raw technology.

Read the full article here

Surgerii Robotics just raised $100 million, and it’s easy to see why their Shurui SP robot is drawing attention, because single-port surgery with snake-like instruments is no small engineering feat. While minimally invasive robots aren’t new, the ability to operate through a single 1.8 cm incision for both adults and pediatric patients sets it apart, and it hints at a future where surgical precision and patient recovery improve simultaneously. The combination of disruptive IP, CE approval, and international ambitions impressive, and it shows that high-tech innovation isn’t confined to Silicon Valley alone.

Read the full article here

NVIDIA just pushed the boundaries of physical AI, and it’s exciting because their new open models, Isaac Lab-Arena, and OSMO framework make robotics development faster and more accessible than ever. From Boston Dynamics to Caterpillar, partners are building generalist-specialist robots that can learn and adapt, and the Jetson T4000 finally brings energy-efficient AI compute to industrial and humanoid machines. I find it brilliant that NVIDIA is blending simulation, open-source collaboration, and edge computing, because this combination could transform not only industrial automation but also household and medical robotics, and it shows how AI can move off the screen and into the real world.

Read the full article here

Lego has introduced Smart Bricks, which bring lights, sound, and motion sensing to the classic building blocks, and the new system includes Smart Minifigures and Smart Tags for interactive play. Some experts have raised concerns that the added technology could reduce imaginative engagement, but the design is intended to complement traditional building rather than replace it, and the Smart Play platform integrates digital and physical elements seamlessly. The innovation demonstrates how toy-makers are incorporating sensors and AI to enhance play experiences, and it provides insight into the evolving balance between creativity and technology in children’s products.

Read the full article here

Google DeepMind and Boston Dynamics are collaborating to integrate the Gemini Robotics AI model into Atlas humanoids and Spot robots, and the partnership aims to give machines the intelligence to navigate unfamiliar environments and manipulate objects. Initial tests will take place in Hyundai factories, and the project highlights how AI can expand the usefulness of industrial robots while maintaining safety controls. The approach combines multimodal learning with real-world data, and it demonstrates a step toward general-purpose humanoids capable of performing a wide range of tasks, while also addressing the practical challenges of reliability, dexterity, and human safety in industrial settings.

Read the full article here

A recent study on humanoid robots made me pause, because it shows that AI models controlling robots can discriminate based on race, gender, disability, and even religion, and that bias compounds when multiple factors intersect. While humanoids promise to handle real-world tasks, the research reminds us that intelligence without careful safety design can produce harmful behavior, even in everyday situations. It’s striking to see how models approved actions humans would reject, and it reinforces my sense that as robotics accelerates, building safety and fairness into AI is just as critical as engineering dexterity and mobility.

Read the full article here

Neumo’s new brainwave-based driver monitoring prototype caught my attention because it goes beyond what dashcams can see, and it reads neural signals to detect fatigue, impairment, or medical issues before they become visible hazards. While traditional cameras track behavior, this system adds a cognitive layer, and the fact it installs in under ten minutes without seat modifications makes it practical for real fleets. The potential for accident prevention is significant because understanding what’s happening in the driver’s brain could help to redefine road safety, and it’s a reminder that sometimes the most critical insights aren’t what we can immediately see.

Read the full article here

Flexible topological metamaterial clothing has been developed to enhance wearable health monitoring, and it overcomes limitations in signal loss, accuracy, and movement interference that affect traditional systems. By incorporating modular topological phases and advanced biosensing networks, the garments significantly improve on-body signal transmission and allow multiple sensors to communicate reliably, and their design supports customization for monitoring metrics such as heart rate, sleep, and physical activity. The integration of machine learning further increases data accuracy and signal-to-noise ratio, and this approach demonstrates how wearable technology can evolve into a highly precise, adaptable system for continuous physiological monitoring in everyday use.

Read the full article here

TSMC has officially kicked off mass production of its 2nm process, and it’s hard not to feel the shift in the semiconductor landscape, especially when GAA transistors promise both speed and efficiency gains that the AI sector desperately needs. It’s impressive that Fab 20 and 22 are handling the scale already, and while the technology cements TSMC’s lead, it also highlights how reliant the world is on a single region for advanced chips. For anyone following high-performance computing, this isn’t just another node shrink, it’s the new foundation for AI and mobile innovation, and it’s worth watching closely.

Read the full article here

TinyML is enabling complex machine learning models to run directly on microcontrollers, bringing AI to the edge of connected devices and reducing reliance on cloud processing. Open source frameworks such as TensorFlow Lite for Microcontrollers and Edge Impulse provide the tools to compress large neural networks for low-power hardware, while specialised platforms like Apache TVM and STM32Cube.AI optimise performance for specific devices. Applications range from predictive maintenance in industrial machines to wildlife monitoring and healthcare wearables, improving efficiency, responsiveness, and privacy. By shifting intelligence to the device level, TinyML is establishing a foundation for widespread, low-power, and real-time AI deployment.

Read the full article here