Today, betting against NVIDIA is like betting against Microsoft in the early days of personal computers or betting against Apple after the release of the iPhone.

A year ago, executives in the Technology industry were enthusiastic about the idea that a new generation of AI capabilities on mobile phones and PCs would stimulate demand for new Consumer devices. At that time, the hype around AI drove various optimistic expectations. Although the market was optimistic about the construction of Datacenter and NVIDIA's (NVDA.O) ongoing opportunities, caution was still needed regarding Consumer devices, as was said at the time.

Recently, Earnings Reports from Micron Technology (MU.O), HP Inc (HPQ.N), and Dell (DELL.N) confirmed this. Their performance showed weakness in traditional PC and Smart Phone Businesses. Meanwhile, trends for NVIDIA and AI Datacenter infrastructure saw significant growth. According to the Wall Street consensus, NVIDIA is expected to achieve a 112% revenue growth in the fiscal year ending January 2025, reaching $129 billion, while Intel's (INTC.O) sales (as a representative of traditional Technology infrastructure) are expected to decline by 3%.

Looking ahead to the entire year of 2025, AI on local devices may continue to struggle, while cloud AI's rise in Datacenter will reach new heights. Cloud AI is more capable and efficient than local AI, and with every device connected to the Internet, there is little necessity for running AI locally.

Companies must upgrade their infrastructure to keep pace with the AI-driven new era. The traditional model of relying on CPUs to extract static files has become outdated, while AI-based GPU computing can generate better insights and analyses in real-time.

The shift from a CPU-centric model to a GPU-dependent parallel system is now widely recognized. Intel's interim CEO Michelle Holthaus stated at a recent Wall Street investor meeting that this represents "the biggest shift we've seen in decades."

There will be other AI growth drivers in 2025. Further development of AI agents will drive trends as they can complete multi-step tasks; the expansion of multi-modal AI models will enable them to handle images, videos, and audio, not just text; advancements in "timeliness computing" will allow AI models to perform research, reasoning, and reflection more comprehensively, providing higher quality answers.

All these advancements will require more NVIDIA GPU computing power, which means that the growth of AI will continue to be attributed to NVIDIA.

Although AI-related businesses among vendors (such as Server manufacturers and memory manufacturers) will also benefit, these companies are largely facing pressure from a decline in traditional markets. By 2025, the trend of transitioning from CPU-based computing to GPU-based computing may accelerate, as enterprises have limited spending on IT.

One company's performance may be a variable: AMD has a chip that technically competes with NVIDIA's GPUs. However, the issue lies in its lack of software maturity on the NVIDIA CUDA platform, and it cannot provide a complete end-to-end datacenter solution, including optimized Hardware, systems, and scalable large GPU clusters. Similarly, large cloud platforms will attempt to compete with NVIDIA through the latest custom AI Chips, but they still struggle to match NVIDIA's decade-long Software advantage.

Today, betting against NVIDIA is akin to betting against Microsoft (MSFT.O) or Apple (AAPL.O) early in the personal computer revolution after the release of the iPhone. In 2024, NVIDIA surpassed everyone, and it is expected to maintain this momentum in 2025.