Semiconductor Manufacturing Team
There’s a lot of investment going into AI chip development, alongside work on photonics, memory and energy efficiency, and a push within Europe and the UK to build more capability domestically.
Toby Nevett, Head of EU Key Accounts, and Jamie Hercus, Head of Semiconductors, Space and Defence, from Empiric’s semiconductor team sat down to take stock of what’s happening in the industry and bring together some of the questions and themes coming through in their conversations with clients and candidates.
We’ve set this out around those questions, looking at the developments shaping the market and the experience clients are asking for when they hire, particularly in chip design and verification.
What Semiconductor Experience Are Clients Looking For?
Clients are looking beyond verification engineers whose experience is solely in UVM and SystemVerilog to engineers who have worked with protocols and interfaces such as PCIe, UCIe, LPDDR, SerDes and Ethernet.
Experience with cache coherency is also coming through, alongside engineers who have worked on AI programmes and RISC-V programmes.
Why Is AI Changing Semiconductor Development?
AI is the biggest trend in the semiconductor industry at the moment, with more people using large language models and businesses looking to adopt AI. That means a significant increase in AI workloads and AI infrastructure. Advanced chips, GPUs and AI accelerators need to be developed to handle those workloads and the stress they put on hardware and data centres.
If you think of how many users are going to be using AI and how much compute power it’s going to take to run that, there has to be a drastic change in how chips are designed, developed and verified. That modernisation is needed so the chips going into data centres can handle the workload of AI – something we’ve covered in our The Data Centre Boom of 2026: Building the Backbone of AI Growth blog.
Where Is Investment Going in AI Chips?
A lot of start-ups are popping up to develop new AI chips, with the AI boom bringing substantial investment into the industry. For us, that’s where the biggest growth and investment activity is in the industry at the moment.
Some AI chip start-ups are raising hundreds of millions of US dollars to develop new chips, with substantial investment going into companies before their first products reach customers.
Why Are Power and Cooling Shaping AI Chip Development?
Data centre and AI computing are constrained by power and cooling limitations, while energy costs add to the pressure. As AI workloads increase, performance per watt becomes a really important part of semiconductor hardware and chip development. Chips need to perform better within those power and cooling constraints so they can handle the workloads being put through them.
Why Are AI Chip Start-Ups Focusing on Silicon Photonics?
Silicon photonics is another area of increased activity, with companies developing chips using optical components. These components use photons, or light, to carry information, with the potential to support the speed of AI inference tasks and AI workloads while reducing the energy consumption involved.
The scale of AI is also putting pressure on electrical interconnects, which have power and latency constraints and consume energy as they move data. Some chip start-ups are really doubling down on photonics and the optical side because that’s their approach to AI, looking to reduce those constraints and the energy consumption involved.
Why Are High-Bandwidth Memory and Verification Becoming More Important?
High-bandwidth memory and memory verification have become key themes in the work around AI processors and are coming through in what our clients are looking for. The processors being developed need memory to feed them, so companies need to get their high-bandwidth memory stacks organised and correct.
High-bandwidth memory and DRAM stacks are becoming critical because there are severe memory capacity constraints - AI is creating a real need to improve the memory capacity and high-bandwidth memory of certain chips.
Running large AI models requires a massive movement of data, which is why the quality of the high-bandwidth memory and DRAM architecture matters. That architecture affects how effectively a chip can manage AI workloads and how well those workloads can scale.
Why Is Semiconductor Sovereignty a Priority in Europe and the UK?
More European semiconductor firms are starting up, alongside a push towards greater sovereignty within the industry. Chips may be designed in Europe or the US and manufactured in Taiwan - greater sovereignty doesn’t necessarily mean moving that manufacturing away from there.
With the current geopolitical climate and rising tensions, the push is towards developing more of the AI chips and technology needed for domestic data centres. It’s becoming more self-sufficient and less reliant on companies in other nations, with UK government funding for domestic start-ups supporting that aim.
Planning Your Semiconductor Team?
Our team of contract and freelance recruiters supports clients across Europe. Semiconductor manufacturing is one of our specialist markets, so if you’re planning your next semiconductor hire, building a team or looking for specialist contract support, get in touch to discuss your requirements.
If you’re a semiconductor professional considering your next move, we’d also be interested to hear from you.
You can connect with Toby or Jamie on LinkedIn, explore our latest roles or speak to the Empiric team.
