I keep buying NVIDIA (NASDAQ:NVDA<\/a>), and the buy button has become a habit. Every quarter sharpens the case for adding shares. This is my highest-conviction position because the world is rewiring itself around AI compute, and Jensen Huang’s company is the toll booth on the road everyone is paving. <\/p>\n AI is moving from single-shot chatbots to continuous, multi-step agentic workflows that multiply compute and memory demands on every data center. Huang put it bluntly on the last call: “The buildout of AI factories, the largest infrastructure expansion in human history, is accelerating at extraordinary speed.” Reasoning models chew through a hundred to a thousand times more tokens than a one-shot query. Every token routes through NVIDIA silicon. <\/p>\n The Receipts I Keep Coming Back To <\/p>\n Q1 FY2027 revenue landed at $81.615 billion, up 85.23% year over year, with Data Center at $75.246 billion (+92%). Data Center Networking at $14.8 billion, up 199% shows the full-stack moat in the P&L. NVLink, Spectrum-X, and InfiniBand are the plumbing agentic AI needs, and customers are paying for the whole rack. <\/p>\n Profitability: Non-GAAP gross margin of 75.0%, ROE of 101.5%, ROIC of 92.2%, and net debt\/EBITDA of 0.006. Free cash flow was $48.554 billion in a single quarter. That balance sheet funds the next architecture cycle without asking for permission. <\/p>\n Don’t wait: the analyst who called NVIDIA in 2010 just revealed his top 10 AI stocks. See the full list FREE now<\/a>. <\/p>\n Capital return signals management sees runway ahead. The quarterly dividend rose from $0.01 to $0.25, and the board added an $80 billion buyback authorization on top of the $38.5 billion already outstanding. Companies that think the story is ending do not do that. <\/p>\n Why NVIDIA Over the Obvious Alternative <\/p>\n The name a reader reaches for first is Advanced Micro Devices (NASDAQ:AMD<\/a>). I own some, but my incremental dollar lands here. The reason is the networking line. A GPU competitor can match a chip. Matching CUDA, NVLink Fusion, Spectrum-X, and the software stack that runs in every cloud and every frontier model is a different problem. Even Intel (NASDAQ:INTC<\/a>) chose to co-develop custom data center and PC products with NVIDIA using NVLink. When your rival plugs into your interconnect, that is the moat announcing itself. <\/p>\n The Risk I Refuse to Wave Away <\/p>\n China export controls are real. Q1 saw no H20 shipments to China, and Q2 guidance of $91.0 billion, plus or minus 2%, explicitly assumes zero China Data Center compute. Add $119 billion in supply commitments and hyperscaler concentration near 50% of Data Center revenue, and concentration risk is real. Demand outside China absorbs supply faster than TSMC can print wafers. The roadmap from Blackwell Ultra to Vera Rubin gives multi-year visibility into a product cycle customers have already committed capital toward. <\/p>\n Why the Buy Button Stays Active <\/p>\n At $224.09 and a P\/E near 45, This is a premium multiple, paid for a company earning $1.87 a share off a 5-for-5 beat streak while building the operating system for the next industrial revolution. As long as agentic AI multiplies tokens, and NVIDIA remains the only place they can run at scale, my next contribution goes to the same ticker. <\/p>\n Don’t wait: the analyst who called NVIDIA in 2010 just revealed his top 10 AI stocks. See the full list FREE now<\/a>. <\/p>\n<\/a> Shutterstock \/ Below the Sky <\/p>\n