For the chip design community, it has taken some time to become comfortable with cloud adoption. Four main reasons contributed to their hesitancy.
Concern over data and intellectual property (IP) security was the foundation for much of the industry’s resistance. Losing control of its IP could significantly affect a semiconductor firm’s ability to maintain its competitive advantage.
To meet their customers' demands for innovation, semiconductor companies stay in a constant race to design and develop ever more sophisticated chips. Billions of dollars are at stake in design breakthroughs, so IP security is crucial to these companies.
Second, until recently, chip makers were content to manage their computing requirements using on-site hardware, so there wasn't an urgent need to find an alternative. Also, they’ve invested heavily in on-site computing infrastructure and IT personnel. Moving to the cloud means leaving much of this infrastructure idle and possibly laying off IT employees.
Third, chip-making organizations are concerned that moving to the cloud would open them up to cyberattacks. Few organizations can afford to have their systems locked up by ransomware for an extended period. In addition, rogue code could be inserted into a chip either during development or the manufacturing process, making any system that uses that chip vulnerable to attack.
Fourth, the electronic design automation (EDA) licensing model has been traditionally have not been cloud friendly, making it difficult to adapt to cloud workloads.