Valued at a market cap of nearly $2.7 trillion, Nvidia has returned more than 20,000% to shareholders in the past decade. Today, the chip maker is among the largest companies globally, while the tech stock is down 30% below its all-time highs. Despite its massive size, Nvidia continues to grow at an enviable pace. Its chips are used to build, train, and deploy artificial intelligence models such as ChatGPT and Claude.
During the recent GPU Technology Conference, Huang emphasized that the company is at the forefront of three key AI infrastructure areas: cloud data centers, enterprise IT, and robotics systems.
Huang stated that Nvidiaâs $39.3 billion Q4 revenue (up 78% year-over-year) signifies the beginning of AIâs transformation. Nvidia expects revenue to reach $43 billion in the first quarter of its fiscal 2026 (ending in January), fueled by strong demand for new Blackwell architecture, which delivered $11 billion in revenue last quarter despite being in the early stages of deployment.
Bears out there, please note that this is where we continue to stock up and hold Nvidia shares.
Huang highlighted a fundamental shift in computing paradigms, stating that âaccelerated computing is the way forwardâ and that traditional data centers not built with accelerated computing âare being built wrong.â He expects â100% of the worldâs data centers will be acceleratedâ before the decade ends.
A key insight from Nvidiaâs presentations involves what Huang calls âAI factoriesâ â specialized facilities dedicated solely to AI processing that will become essential to every industry. âJust as they have car factories, they have AI factories to power those cars,â Huang explained. âEvery company that has factories will have an AI factory with it.â
Nvidiaâs competitive advantage stems from its comprehensive approach: âWeâre not just building a chip, weâre building networking and switches... weâre basically building systems components,â Huang noted.
Nvidiaâs full-stack solutions provide superior performance, with Blackwell offering up to 25x higher token throughput and 20x lower costs than previous generations.
Comments