NVIDIA has publicly projected that global AI infrastructure spending will reach $1 trillion by 2027, a figure consistent with hyperscaler capex trends and data center buildouts underway at Microsoft, Google, Amazon, and Meta. The company reported FY2026 revenues of $215.9B, up 65.5% year-over-year, with gross margins of 71.1% and net margins of 55.6% — a profitability profile almost without precedent at this revenue scale in semiconductors.

The $1T demand projection matters because NVDA currently captures a disproportionate share of AI training and inference workloads through its H100 and Blackwell GPU families. If the TAM projection holds, NVDA's addressable market is still meaningfully larger than its current revenue run-rate, which is the core bull narrative.

The friction point is valuation and timing. At these revenue levels and growth rates, NVDA is almost certainly a consensus long across institutional portfolios — the enrichment data doesn't show a significant price-target gap that suggests the street is behind the curve. The $1T figure is a management projection, not a third-party estimate, and management has incentive to frame the opportunity optimistically. Competitive encroachment from AMD, custom silicon from hyperscalers (Google TPUs, Amazon Trainium), and export control headwinds to China remain structural overhangs.

What to watch: any signs that hyperscaler capex guidance softens, or that Blackwell supply-demand dynamics shift. A deceleration in data center segment revenue growth would be the most direct read-through that the $1T narrative is getting ahead of itself. Conversely, sustained 40%+ revenue growth into FY2027 would vindicate the projection and keep estimates moving higher.