On Tuesday Intel announced that it intends to develop its own graphics processing units (GPUs), a strategic segment largely dominated by Nvidia and now central to artificial intelligence and high-performance computing applications. CEO Lip-Bu Tan confirmed the arrival of Eric Demmers, a former Qualcomm executive, as chief GPU architect, formalising a rumour carried by the specialist press last month. Development of these new chips will be overseen by Kevork Kechichian, head of data-centre products, reflecting the priority given to that market.

The initiative marks a turning point for Intel, which is seeking to reposition itself in a rapidly changing semiconductor industry. The group is betting on strong data-center demand to shape the design of its future GPUs, with an approach focused on customers' specific needs. In parallel, Intel is highlighting growing interest in its foundry business, Intel Foundry, particularly around its 14A process technology. A ramp-up in production is expected by the end of the year.

As Intel works to catch up with rivals such as Nvidia and AMD, the announcement is part of a broader strategy of diversification and a push back into key artificial intelligence and cloud markets. The group aims to strengthen its manufacturing capabilities while tailoring its offering to the requirements of leading technology players, in an effort to once again become a must-have player in next-generation chips.