Etron Technology founder Nicky Lu believes the explosive growth of artificial intelligence is doing something many in the chip industry have been waiting to see again: pushing semiconductor innovation forward fast enough to reignite the momentum behind Moore’s Law. Even as manufacturers grapple with a lingering global memory shortage, Lu argues that the AI wave is forcing breakthroughs in design, manufacturing, and performance that are reshaping the future of computing.
AI has become the ultimate stress test for modern hardware. Training and running large AI models demands enormous computing power, and that pressure is driving chipmakers to find new ways to deliver higher performance and better efficiency. According to Lu, this intense demand is helping the industry move past the slower pace of progress that has fueled years of debate about whether Moore’s Law was fading. Instead of relying on a single improvement, the rebound is being powered by a mix of advances across the semiconductor stack, as companies race to build faster AI accelerators and the supporting infrastructure around them.
However, the surge in AI is also amplifying one of the industry’s biggest pain points: memory supply. Lu points to a continuing memory crunch that could stretch through the first half of the year, creating challenges for device makers and data center operators alike. AI workloads depend heavily on high-bandwidth memory and other advanced memory solutions to keep powerful processors fed with data. When memory is constrained, it can slow deployments, raise costs, and complicate planning for companies trying to scale up AI capabilities.
The situation highlights a growing reality across the semiconductor market: compute power alone isn’t enough. AI performance depends on the entire system working in sync—processors, memory, packaging, and manufacturing capacity. When demand spikes in one area, such as memory, it can ripple across the entire supply chain. That’s why the current shortage matters, even in a period when innovation is accelerating.
Lu’s outlook captures a key turning point for the industry. AI is not only creating demand for more chips, but also changing what “progress” looks like in semiconductors. The renewed pace of innovation suggests that Moore’s Law-style gains may continue through a combination of better architectures, improved production techniques, and next-generation memory technologies—while the memory crunch serves as a reminder that supply constraints can still shape how quickly the AI era expands.
In the months ahead, industry watchers will be tracking two stories at once: whether memory supply loosens after the first half of the year, and how far AI-driven semiconductor innovation can go in restoring the rapid advancement that defined earlier eras of computing.
