News

There’s an odd duality in modern FPGA design. We want them to do more, compute more, include more features and deliver results faster and more efficiently. Yet at the same time, there are often competing demands for them to have a smaller footprint and use less power....

FPGAs no longer live inside closed, single‑purpose systems where the environments are tightly controlled and predictable. They now sit in cloud servers, edge AI devices, automotive platforms, and industrial IoT nodes, in space as well as on Earth. They are updated...

Ever since the release of ChatGPT in late 2022, I have been fascinated by the possibilities of using AI. Now it’s everywhere, and whether you write code, create content or generate images, chances are AI may be involved somewhere along the way.

FPGA design and development is always changing. A decade ago, it was about deep RTL expertise, vendor-specific flows, and deterministic, long-term deployments. Now, we are also expected to be fluent in system-level thinking and open hardware/software stacks.

I think everyone would agree that FPGA design is hard work. Sometimes it takes more time and effort than you have the resources to hand. This is where commercial‑off‑the‑shelf (COTS) products come in.

Ten years ago the robotics industry was stuck in a recurring loop. AI and machine learning provided the catalyst to change that, whilst the enabler has been FPGAs.

FPGA engineering is fascinating because everywhere you look there are different challenges to solve. In high frequency trading (HFT), FPGA performance literally comes down to nanoseconds, not microseconds. More often than not success or failure comes down to a single clock cycle.

In 1985, a radical idea changed the course of hardware design: making chips reprogrammable. That modest beginning was just a hint at what would come because for the first time it made hardware as flexible as software, allowing engineers to redefine a chip’s functionality even after manufacturing.