Nvidia’s Unspoken Problem: 40% of Revenue Comes From Companies Developing Their Own AI Chips

Jensen Huang has built a $4.6 trillion empire selling the picks and shovels of the AI revolution. But while he preaches accelerated computing, three existential threats aren’t discussed in earnings calls. They’re happening now.

Threat #1: Your Best Customers Are Building Your Replacement

Microsoft (NASDAQ:MSFT), Meta (NASDAQ:META), Amazon (NASDAQ:AMZN), and Alphabet (NASDAQ:GOOGL) represent 40-50% of Nvidia (NASDAQ:NVDA) revenue. Every single one is developing custom AI chips.

Google’s TPUs power Bard and Search. Amazon’s Trainium chips offer AWS customers cheaper alternatives. Meta’s MTIA handles inference workloads. Microsoft’s Maia chip is deploying across Azure. These are production infrastructure.

Nvidia’s response? Custom chips “complement” rather than replace GPUs. But here’s the math that should terrify investors: inference represents 80% of long-term AI compute. Training is 20%. If hyperscalers build inference chips in-house, Nvidia loses access to 80% of the addressable market. The company captured 85% gross margins selling infrastructure to customers now building their own data centers.

Threat #2: AMD Is the Alternative Nobody Wants to Acknowledge

Advanced Micro Devices (NASDAQ:AMD) stock surged 111% over the past year while Nvidia gained 28%. The market is noticing what enterprises already know: MI300X chips deliver competitive performance at 20-30% lower…