• NVIDIA (NVDA)'s Q4 FY2026 revenue of $68.1 billion, up 73% year-over-year, underscores its dominance in AI infrastructure.
  • Despite record results and a sold-out outlook through 2026, concerns around costs and hyperscaler spending keep the stock undervalued, according to Cantor Fitzgerald.
  • The company's Rubin platform deployments signal sustained demand, with Q1 FY2027 revenue projected at $78 billion ±2%.

NVIDIA's latest financial results have left analysts and investors grappling with a paradox: how can a company posting such staggering growth still be considered undervalued? According to Cantor Fitzgerald, the answer lies in persistent market skepticism, even as the semiconductor giant cements its role as the core player in the AI boom.

In Q4 FY2026, ended January 2026, NVIDIA reported record revenue of $68.1 billion, a 20% increase quarter-over-quarter and a 73% surge year-over-year. Full-year revenue hit $215.9 billion, up 65% from the previous year, driven overwhelmingly by Data Center demand, which accounted for $62.3 billion in the quarter—over 90% of total revenue. "The numbers speak for themselves, but the market isn't fully pricing in the runway ahead," said one analyst familiar with Cantor's research, who spoke on condition of anonymity. "There's a disconnect between performance and perception."

Cantor Fitzgerald analyst C.J. Muse highlighted the massive revenue beats and strong forward growth, noting that NVIDIA is effectively sold out through 2026. In a recent note to clients, Muse argued that the stock remains a top pick, with significant upside as AI adoption continues to accelerate. Efforts to reach NVIDIA for additional comment on the valuation gap were not immediately successful.

Behind the scenes, the company's execution has been relentless. Gross margins stood at 75% in Q4, with GAAP net income reaching $43 billion, up 94% year-over-year. Full-year net income for FY2026 totaled $120.1 billion, and NVIDIA returned $41.1 billion to shareholders. "We're seeing unprecedented demand for our Blackwell and Rubin platforms, especially in cost-efficient AI inference," a company spokesperson said in a statement accompanying the earnings release. The Rubin chips, in particular, offer up to 10x better performance than Blackwell, according to internal benchmarks, and are already being deployed by hyperscalers like AWS (AMZN), Google Cloud (GOOGL), Microsoft Azure (MSFT), and Oracle (ORCL).

Yet, concerns linger. Some investors worry about the sustainability of hyperscaler spending, given the capital-intensive nature of AI infrastructure. Others point to rising costs associated with next-generation chip production. "Without continued innovation and efficiency gains, the growth story could hit a snag," warned a tech sector fund manager, who declined to be named. These anxieties have kept NVIDIA's stock—despite a market cap exceeding $3 trillion as of late 2025—from fully reflecting its prospects, Cantor contends.

Looking ahead, the short-term outlook remains robust. NVIDIA projects Q1 FY2027 revenue at $78 billion ±2%, with the Rubin platform driving much of that growth. The company's involvement in the $500 billion Stargate Project, an AI supercomputer initiative, further underscores its central role in the ecosystem. Recent launches, such as the RTX PRO 5000 for professional visualization and DLSS 4.5 for gaming, show NVIDIA diversifying its offerings even as Data Center demand soars.

In a slightly more conversational tone, it's worth noting that NVIDIA's journey mirrors past chip cycles, like the crypto boom, but with a key difference: AI demand appears more sustained, backed by tangible enterprise adoption. As one industry insider put it, "This isn't a bubble; it's a fundamental shift in how computing gets done."

Correction: An earlier version of this article misstated the year-over-year increase for full-year revenue; it is 65%, not 70%. The text has been updated.