Nemotron

Arm's instruction set architecture is the foundation for most mobile devices and is rapidly gaining traction in data center CPUs optimized for AI workloads, driven by demand for power‑efficient compute.

The company's licensing model creates a resilient, recurring revenue stream, supported by a vast ecosystem of software developers and high switching costs that protect its market position.

["Dependence on smartphone market cycles","Geopolitical tensions affecting supply chain","Emerging open‑source alternatives like RISC‑V eroding pricing power"]

NVDA's GPUs are the de facto standard for AI training and inference, delivering unmatched performance per watt and benefiting from network effects of CUDA.

The CUDA platform creates high switching costs; combined with leading‑edge fab partnerships (TSMC) and a strong developer ecosystem, it sustains durable competitive advantage.

["Valuation appears rich relative to earnings.","Cyclical exposure to data‑center cap‑ex swings.","Intensifying competition from AMD and ASIC start‑ups."]