Forget revenue and net income: This is the critical metric to look for when Nvidia reports on August 28

On Wall Street, major data releases are a common occurrence. Federal Reserve Open Market Committee meetings, monthly inflation and jobs reports, and even quarterly Form 13F filings with the Securities and Exchange Commission provide invaluable data that investors can use to formulate an effective investment strategy .

Tomorrow (August 28), the company that played the biggest role in sending Dow Jones Industrial Average, S&P 500and Nasdaq Composite to record– Nvidia (NVDA -2.25%) — will lift its proverbial hood and allow Wall Street to examine its operating performance over the past three months. This is arguably as important to a data release as any we’ve mentioned above.

While investors are likely to focus on Nvidia’s headline numbers (revenue and net income), there’s a much more important metric to look for that will provide broader context on the health of the artificial intelligence (AI) revolution and Nvidia’s future .

Image source: Getty Images.

How did Nvidia become the most anticipated and important earnings report on Wall Street?

Before exploring this critical measure, it is imperative to provide background on how we got to this point. In other words, an explanation is needed for Nvidia’s meteoric rise from a market cap of $360 billion at the end of 2022 to a valuation of nearly $3.2 trillion as of the closing bell in August 23, 2024.

As you’ve probably rightly guessed, AI is behind this historic move. In short, Nvidia’s H100 graphics processing unit (GPU) has become the preferred choice in compute-heavy data centers. These chips are effectively the brains that power split-second decision-making, oversee generative AI solutions, and help train large language models (LLMs).

Nvidia also got an undeniable boost from its CUDA software platform. CUDA is the toolset that developers use to build LLMs and get the most out of their GPUs. CUDA works seamlessly with the H100 to keep enterprise customers loyal to Nvidia’s ecosystem of products and services.

The other key piece of the puzzle for Nvidia is enterprise demand far outstripping the supply of AI-GPUs. When demand for a good or service exceeds supply, the price of that good or service usually rises until demand falls. The cost for Nvidia’s H100 is typically between $30,000 and $40,000, which is more than double the cost of the MI300X, which is a rival chip developed by Advanced microdevices.

It’s this lack of AI-GPUs that you’ll want to keep in mind when digging into Nvidia’s fiscal second-quarter operating results, which are scheduled to be released after the closing bell on August 28.

Image source: Getty Images.

Forget about revenue and net income—this metric goes beyond both

Over the previous five quarters (ie, Nvidia’s full year of fiscal 2024 and the first quarter of fiscal 2025), Nvidia completely blew Wall Street’s highest sales and profit expectations out of the water. Given that Wall Street analysts have a history of delivering unbeatable consensus forecasts, it wouldn’t be remotely shocking if Nvidia’s revenue and net income beat expectations, once again.

But those two headline numbers tell only part of the story. If you want a more complete picture of the health of Nvidia and the AI ​​revolution, focus on the company’s gross margin. With Nvidia, I tend to focus on adjusted gross margin, which excludes factors like stock-based compensation and acquisition-based expenses.

Although Nvidia sold more of its H100 GPUs, thanks in large part to a significant increase in chip-on-wafer-on-substrate capacity from the world’s leading chipmaker Taiwan Semiconductor Manufacturingmost of Nvidia’s growth came from its otherworldly GPU pricing power. The ability to load 50% to 75% more than competing AI GPUs propelled Nvidia’s adjusted gross margin from 64.63% to 78.35% in five quarters.

NVDA gross profit margin (quarterly) data by YCharts.

Following Nvidia’s fiscal first quarter results (ended April 28), management called for an adjusted gross margin of 75.5% (+/- 50 basis points) for the fiscal second quarter. This would imply a decline of between 235 and 335 basis points from the successive quarter.

While this could just represent management’s conservation as the cost of goods rises in step with new orders, a significant decline in adjusted gross margin could also suggest a change in the AI ​​landscape.

Once again, it all revolved around the idea that AI GPUs are rare and in high demand. If Nvidia’s adjusted gross margin approaches the low end of its previous forecast, or its outlook calls for further gross margin pullback, it would signal pretty clearly that competitive pressures are beginning to be felt.

Notably, reports emerged this month that the next generation of Nvidia’s Blackwell platform will be delayed by at least three months due to design flaws and various supply chain constraints.

Despite a healthy order backlog for the Blackwell chip, there is the potential that we could see companies opt for cheaper AI-GPU alternatives that they can get their hands on more quickly. First mover advantages in the AI ​​space i am one thing, and some companies won’t want to wait months for orders to be fulfilled. This opens the door for advanced microdevices, Samsungand other hardware developers to steal valuable space in AI-accelerated data centers.

Adjusted gross margin can also tell us whether Nvidia is facing domestic competitive pressures. The company’s four largest customers, who are all members of the “Magnificent Seven,” comprise about 40 percent of Nvidia’s net sales. However, all four industry leaders are developing AI-GPUs in-house for use in their data centers. A lower adjusted gross margin forecast may reflect Nvidia losing valuable data center “real estate” to these in-house chips.

Instead of focusing on Nvidia’s sales and net income tomorrow, dig deeper and look for the adjusted gross margin figure that will give you a more complete picture of what the future holds for Nvidia and artificial intelligence.