I keep buying NVIDIA. Every paycheck, every pullback, every time the headlines swing bearish on AI capex, I click buy again on NVIDIA (NASDAQ:NVDA | NVDA Price Prediction). This is the single position I trust to compound retirement capital through the AI decade, and my conviction has almost nothing to do with the chips themselves.
What pulls me back to the buy button is the CUDA software ecosystem, embedded across two decades into every major AI framework, library, and developer workflow. Enterprise customers who try to leave face migration costs and operational risk they refuse to swallow. Jensen Huang described the platform on the last call as “the only platform that runs in every cloud, powers every frontier and open source model, and scales everywhere AI is produced.” That reads to me as a toll road on global AI development.
The Receipts Behind My Conviction
The financials show what a software moat looks like when it meets a demand cycle. Q1 fiscal 2027 revenue landed at $81.615 billion, up 85.2% year over year, with non-GAAP EPS of $1.87 topping the $1.7738 consensus. Data Center revenue hit $75.246 billion, up 92%, with networking alone up 199%. Net income grew 210.63%, outrunning revenue growth. That is operating leverage I can measure.
Margins tell the pricing-power story. Non-GAAP gross margin expanded to 75.0% from 60.8% a year earlier. Return on equity sits at 101.5%, ROIC at 92.2%, and debt/equity at 0.073. Free cash flow in the quarter reached $48.554 billion. Management responded by raising the dividend from $0.01 to $0.25 per share and authorizing an additional $80.0 billion in buybacks with no expiration. In Q1 alone, roughly $20.0 billion was returned to shareholders.
Then there is visibility. Total supply-related commitments stand at $119.0 billion, backed by multi-year deals with Meta Platforms (NASDAQ:META) for millions of Blackwell and Rubin GPUs, OpenAI’s 10-gigawatt deployment commitment, and CoreWeave’s 5-plus gigawatt buildout through 2030. Guidance for Q2 calls for $91.0 billion in revenue at the same 75% gross margin, and that guide excludes China entirely.
Why Not the Obvious Alternative
The name a reader reaches for first is AMD (NASDAQ:AMD). I keep passing. Nothing available to me shows an AMD data-center business growing at NVIDIA’s 92% pace, a networking line expanding 199%, or gross margins near 75.0%. CUDA is the reason. Every framework optimization, every NIM microservice, every Dynamo release lands on NVIDIA silicon first. AMD ships capable chips into a software world that already speaks CUDA. That gap is what my capital is really paying for.
The Risk I Take Seriously
The risk I take seriously is customer concentration meeting custom silicon. Hyperscalers are roughly 50% of Data Center revenue, and Amazon (NASDAQ:AMZN), Alphabet (NASDAQ:GOOGL), and Meta are all funding in-house accelerators. China has already been erased from guidance, costing the $4.6 billion that the year-ago quarter carried. What keeps the thesis intact for me is that Blackwell remains, in Huang’s words, “off the charts,” with cloud GPUs sold out. The same customers funding rival silicon are simultaneously signing multi-gigawatt NVIDIA contracts.
Why the Buy Button Stays Active
At a forward P/E of 24x against triple-digit net income growth, a fortress balance sheet, and $119 billion in booked supply, I consider that a reasonable price for the operating system of the AI economy (247’s 7 Stocks Powering the AI Boom report frames the broader stack well). As long as CUDA remains the language every serious model is trained and served in, my next buy is already scheduled.
Contact [email protected] for any questions or corrections.
Top Gaining Stocks
Top Losing Stocks