NVIDIA CEO Jensen Huang stated on the 26th (local time) that "artificial intelligence (AI) has now reached an inflection point and is doing useful work."

Huang made the remarks in an earnings statement for the second quarter of the fiscal year, covering May to July, adding, "Tokens are productive and profitable, and compute is now revenue."

He continued, diagnosing that "demand is accelerating. Around this time last year, just a single lab drove infrastructure building, but now we are witnessing a golden age of new AI labs and startups."

He also added, "Strong momentum is rising across the U.S. and globally, with multiple frontier labs scaling simultaneously, a vibrant open-model ecosystem, and physical AI beginning to operate in earnest."

"The build-out of AI infrastructure is proceeding at full speed," he said, noting that "Vera Rubin, which has now entered full-scale mass production, was created to support this exact moment."

On this day, as memory prices required for AI processor production soared, NVIDIA acknowledged that it would not be easy to meet its future gross margin targets and lowered its expectations.

Accordingly, the company provided a third-quarter gross margin guidance of 74% (±0.5 percentage points).

The second-quarter gross margin stood at 75%.

NVIDIA has previously aimed to maintain its gross margin in the mid-70% range, but it has now admitted that achieving this goal is difficult amid an "extreme" surge in memory semiconductor prices.

The company projected that its margin will bottom out at 71% to 72% in the fourth quarter (November to January of next year) before stabilizing in the 72% to 73% range starting in fiscal year 2028.