Bucket Labs has developed a Cognitive Simulation (CogSim) platform that leverages digital twin technologies based on artificial intelligence (AI) to create higher quality crystals that form the electronic pathways used to build electronic circuit boards.
The startup company created its digital twin platform by collecting data from the furnaces used to create and shape crystal ingots, which it then exposed to a neural network built using machine learning algorithms that continuously retrains the CogSim platform as additional data is collected.
The overall goal is to increase the quality of the crystals in a way that also serves to reduce the total cost of creating everything from semiconductors to circuit boards, says company co-founder Cameron Khorsandi.
The CogSim platform uses neural networks to simulate how crystals are grown in a furnace using Bridgman, Czochralski, and Kyropoulos processes and techniques. It aligns simulation predictions with physical reality to optimize variables in real-time to maximize yields, eliminate grain boundaries, and perfect crystal lattices. “We’re bridging the gap between simulations and actual results,” says Khorsandi.
In contrast, traditional crystal growth relies on imperfect physical models and a lot of trial-and-error. Complex variables in temperature gradients, pulling rates, and crucible cooling inevitably lead to unpredictable yields or structural defects that create fundamental manufacturing inefficiencies, notes Khorsandi.
It’s not clear just yet what impact CogSim might have on the total cost of IT, but as IT platforms become more widely used in everything from massive AI data centers to various devices running on the network edge, the need to further reduce manufacturing costs only intensifies. The global semiconductor industry is forecasted to surpass the $1 trillion mark in sales this year, according to the Semiconductor Industry Association (SIA). Chip sales reached $791.7 billion in 2025, for a 26% year-over-year increase. According to the Futurum Group, the global data center semiconductor market alone accounted for $241 billion of revenue in 2025, with a possibility of achieving $2.2 trillion in revenue by 2030.
The challenge is that current manufacturing capacity suggests there will be a permanent shortage through at least the rest of the decade unless manufacturing methods evolve. Bucket Labs is making a case of reinventing how crystals are made as one means of achieving that goal.
Of course, any fundamental change to how core processor technologies are manufactured will naturally take time. Change tends to come slowly to manual processes that have in some cases now been relied on for decades. However, as competition continues to intensify, there should be a greater willingness to embrace innovation as nations around the world continue to invest heavily not just in chip technologies, but also the processes used to manufacture them.
In the meantime, IT organizations might want to start pressing providers of processors on not just when they plan to deliver the next most powerful processors, but also how they plan to reduce costs in a way that increases overall supply to the point where IT ultimately becomes fundamentally more affordable than it is today.
