Performance "King Bomb" Shatters Bubble Theory! NVIDIA (NVDA.US) Q4 revenue hits new high, guidance exceeds expectations. Jensen Huang asserts that the "AI industrial revolution" has just begun.

Bloomberg News has learned that the AI chip giant NVIDIA (NVDA.US) has once again announced a performance that can be described as “explosive.” Amid recent product launches by Anthropic and Citrini’s “end-of-report” reports fueling investor panic, NVIDIA’s latest earnings continue to prove that demand driven by AI remains strong, and large-scale AI infrastructure development is still on track.

The financial report shows that for the fourth quarter of fiscal year 2026 ending January 25, NVIDIA’s revenue increased by 73% year-over-year to a record high of $68.1 billion, beating analysts’ consensus forecast of $65.9 billion. By business segment, data center revenue (responsible for its industry-leading AI accelerators and networking products) rose 75% year-over-year to $62.3 billion, also a record high, surpassing analysts’ forecast of $60.4 billion. However, other segments performed relatively weaker. Gaming revenue (formerly NVIDIA’s main source of income) grew 47% year-over-year to $3.73 billion, below analysts’ forecast of $4.01 billion; automotive revenue grew 6% to $604 million, also below the forecast of $643 million.

On the profit side, under Non-GAAP accounting standards, operating profit rose 81% year-over-year to $46.1 billion, and net profit increased 79% to $39.6 billion. Adjusted earnings per share were $1.62, beating analysts’ forecast of $1.53.

The adjusted gross margin was 75.2%, slightly higher than analysts’ expectations. NVIDIA CFO Colette Kress stated that the year-over-year improvement in gross margin was due to “reduced inventory provisions,” while the quarter-over-quarter improvement was related to the continued volume ramp-up of Blackwell chips, leading to a “better product and cost structure.”

What further encouraged investors was NVIDIA’s guidance for the first quarter of fiscal year 2027, which exceeded expectations. The company expects revenue of $78 billion (plus or minus 2%), significantly surpassing the consensus forecast of $72.8 billion, implying a year-over-year growth rate of 77%.

Analysts pointed out that NVIDIA’s total revenue and data center revenue both exceeded expectations and hit new highs in the fourth quarter, with gross margins continuing to improve as the new generation Blackwell architecture chips ramp up production. The guidance for the first quarter of FY2027, even excluding some revenue from the Chinese market, appears stronger, reinforcing the narrative of resilient AI computing demand and helping to ease market concerns about a potential AI investment bubble.

NVIDIA CEO Jensen Huang said, “Our customers are competing to invest in AI computing power—these computing factories are fueling the AI industrial revolution and its future growth.” Huang has repeatedly downplayed concerns about unsustainable AI hardware spending. He believes that replacing the world’s existing old computers with machines capable of achieving productivity leaps will take several years.

Notably, a major cloud hanging over the tech industry is the shortage of memory chips. Like most electronics companies, NVIDIA’s products rely on the stable supply of these components—short-term storage chips used in everything from smartphones to supercomputers. Supply constraints have driven up memory chip prices and put pressure on NVIDIA’s shipments this year. In response, NVIDIA stated that it has sufficient supply, “We have strategically locked in inventory and capacity to meet demand beyond the coming quarters.”

Earlier this month, NVIDIA announced that Meta Platforms (META.US) has agreed to deploy “millions” of NVIDIA AI chips over the next few years, further deepening the already close partnership between the two AI giants. NVIDIA’s main competitor AMD (AMD.US) also announced a similar long-term agreement with Meta this week. The company said the deal is worth hundreds of billions of dollars. These large transactions serve as evidence that AI demand remains robust. They aim to lock in long-term compute commitments. However, the “intimate” nature of these deals—where suppliers and customers sometimes hold stakes in each other—has also drawn criticism, suggesting that such cyclical transactions could artificially inflate demand.

Another major issue is that NVIDIA is still awaiting clear guidance on whether it can continue operations in China. China is its largest market, but NVIDIA’s ability to sell its most advanced products to Chinese customers remains restricted. NVIDIA stated that its first-quarter revenue outlook does not include any income from Chinese data center business.

However, the company mentioned in a filing that it has obtained U.S. approval this month to export “small quantities” of H200 chips to Chinese customers. NVIDIA said, “To date, we have not generated any revenue under the H200 licensing plan, and it is still unclear whether any products are permitted to enter China. The license requires H200 chips to be inspected in the U.S. before shipment to customers. Therefore, any H200 chips exported under the new license plan will be subject to a 25% tariff when imported back into the U.S.”