NVDA Q2 Guidance of $91B Clears Consensus by $5B-$7B, Silencing Peak-Capex Bears

Nvidia's first-quarter blowout earnings and aggressive second-quarter guidanceCompany-issued forecasts of future financial performance. have effectively silenced bears worried about peak AI capex. The company posted $81.6 billion in revenue, up 85% year-over-year, crushing the $79.2 billion consensus. More importantly, data-center revenue hit $75.2 billion, up 92%, signaling unrelenting demand from cloud providers racing to deploy Blackwell chips in production.

The Q2 guidanceCompany-issued forecasts of future financial performance. of $91 billion stands far above Street expectations of $84-86 billion and notably excludes any contribution from China data-center compute, a material omission that underscores the company's confidence in core demand. Multiple hyperscalers, Microsoft, Alphabet, Amazon, and Meta, have publicly raised AI capital expenditure budgets in recent weeks, and Nvidia's guidance proves those commitments are translating into actual orders. The company also announced an $80 billion new buybackA company repurchasing its own shares from the open market. and maintained its massive dividend, signaling management conviction on future cash flows.

The narrative removes a key bear case that had circulated for weeks: that Nvidia's customers were front-loading orders without intent to deploy at scale. Instead, the data suggest that AI infrastructure buildout is hitting an inflection point. Energy-intensive data centers require the largest chips to achieve efficiency targets, and Blackwell's performance-per-watt advantage creates genuine scarcity value. Semiconductor peers like AMD and Broadcom have seen strong order signals too, but Nvidia's dominance in high-end accelerators remains unchallenged.

Some skeptics still point to valuation: Nvidia trades at a 33x price-to-earnings multiple despite 80%+ revenue growth. On a forward basis, even doubling earnings would leave the stock at roughly fair value by historical tech standards. The real risk is a slowdown in hyperscaler capex if AI inference, which requires far fewer chips than training, accelerates faster than expected, or if regulatory action in China tightens further.