Lenovo's CFO explains that Chinese AI firms, facing severe chip restrictions and a cutthroat domestic market ("involution"), are forced to innovate for extreme cost efficiency. This pressure results in models that can be dramatically cheaper per token, a potential long-term competitive advantage.

Unlike past tech cycles with a single constraint, the AI boom is constrained by numerous interdependent bottlenecks at once: power, transmission, memory, optical components, and skilled labor. Solving one piece (e.g., memory supply) doesn't fix the overall systems-level challenge, making the problem uniquely complex.

Faced with restrictions on advanced NVIDIA chips, China is leveraging its electricity advantage to run vast numbers of older-generation GPUs in parallel. This hardware constraint forces a focus on software, with Chinese labs developing sophisticated algorithms and compute methods to leapfrog the hardware deficit.

Echoing Don Valentine's VC wisdom that 'scarcity sparks ingenuity,' US restrictions on advanced chips are compelling Chinese firms to become hyper-efficient at optimizing older hardware. This necessity-driven innovation could allow them to build a more resilient and cost-effective AI ecosystem, posing a long-term competitive threat.

While NVIDIA's GPUs have been the primary AI constraint, the bottleneck is now moving to other essential subsystems. Memory, networking interconnects, and power management are emerging as the next critical choke points, signaling a new wave of investment opportunities in the hardware stack beyond core compute.

OpenAI achieved a major reduction in the cost of running its models through purely software and algorithmic improvements, such as quantization and smarter caching. This demonstrates that efficiency innovation can be as impactful as acquiring more hardware, suggesting a path to overcoming compute bottlenecks without relying solely on expensive chips.

Faced with limited access to top-tier hardware, Chinese AI companies have been forced to innovate on model architecture to compete. They've developed superior techniques in memory management and multi-token prediction, making their models highly efficient and formidable competitors despite hardware constraints.

The current 2-3 year chip design cycle is a major bottleneck for AI progress, as hardware is always chasing outdated software needs. By using AI to slash this timeline, companies can enable a massive expansion of custom chips, optimizing performance for many at-scale software workloads.

The AI boom's growth has been defined by a series of shortages, from GPUs to cooling, power, and now memory chips. This reveals a pattern where solving one bottleneck creates the next one. Investors and strategists can anticipate and capitalize on these sequential constraints in any rapidly scaling industry.