NVIDIA's Q2 earnings report falsification and inflated stock price? Wall Street analysts refute: Don't blindly believe everything you see online, guys.

Recently, conspiracy theories about Nvidia's "record-breaking" Q2 earnings report being fraudulent have been circulating on social media and investor forums, causing the stock price to drop by about 3% on September 6th and 7th, erasing all the gains since the Q2 report was released on August 23rd.

Nvidia's poor performance has dragged down the semiconductor industry benchmark, the Philadelphia Semiconductor Index, which fell more than 3% at the beginning of trading on Thursday. Its competitor in the GPU field, AMD, also experienced a drop of nearly 4%. Large tech stocks such as Apple, Microsoft, and Google, which have benefited from the AI boom, have also been consistently weak.

The reasons for Nvidia's price drop are not only related to the logic of "rising oil prices to nearly a ten-month high, resilient US economic data, which has increased market expectations of another rate hike by the US Federal Reserve in November, thereby pushing up US bond yields and putting pressure on risk assets such as the stock market," but also to the intense battle between long and short positions on Wall Street, as Nvidia's stock price reached a historical high after the earnings report and rose more than 200% this year.

Did Nvidia's costs remain unchanged while revenue doubled YoY? Channel investigations show that the company did not significantly increase AI capital expenditures.

According to the research and investment firm Macrotips Trading, although Nvidia's Q2 report far exceeded expectations and provided a stronger guidance for future revenue, there are reasons to be concerned about "suspicious disclosures and management practices," especially as a startup AI company directly supported by the company has contributed to the revenue growth. The expansion of gross profit margin, the peculiar timing of stock buybacks, and insider selling have also sent out warning signals.

The firm stated in its article that in the just-concluded Q2, Nvidia's total revenue doubled YoY to $13.5 billion, exceeding analysts' expectations by $2.4 billion. Among them, due to strong demand for GPUs supporting large language models and generative AI, its data center revenue increased by 171% YoY and 141% QoQ to $10.3 billion, surpassing market expectations by $2.3 billion:

"What is even more incredible is that Nvidia achieved astonishing growth with almost no increase in costs. Its revenue cost increased by only 6.8% YoY, but the gross profit margin reached an unheard-of 70.1%.

Nvidia has also set an ambitious target of $16 billion for Q3 revenue, far exceeding analysts' expectations of $12.5 billion. Nvidia's market value is $1.2 trillion, and never before has a company of such massive scale exceeded market expectations by such a large margin in its quarterly report.

It is surprising why Wall Street's expectations completely missed its rapid growth. In theory, more than 40 analysts covering the company have conducted channel checks in at least 10 ways before the earnings report."

For example, before the release of the earnings report, Danske Bank analyzed the second-quarter capital expenditures of major AI players and found no signs of high capital expenditures indicating such strong demand for NVIDIA GPUs."

This raises doubts about how NVIDIA's data center revenue soared "on the basis of the surge in demand for generative AI." Some market participants immediately turned their attention to Coreweave, an AI startup supported by NVIDIA.

Conspiracy theorists question NVIDIA's investment in Coreweave, suggesting collusion to boost Q2 revenue, and more big players may be involved!

The above article suggests that there is reason to suspect that Coreweave, directly funded by NVIDIA, is the main reason behind NVIDIA's revenue surge.

First, Coreweave sold $421 million in equity to investors including NVIDIA, which may give NVIDIA control over the purchasing decisions of the former. At the same time, in early August, Coreweave secured $2.3 billion in debt financing by mortgaging NVIDIA GPUs, an amount that coincidentally matches the scale of NVIDIA's data center revenue in Q2, which exceeded market expectations and included AI chip business:

"While this may be entirely legal, NVIDIA funded this startup, which then used the loan obtained by mortgaging NVIDIA GPUs to purchase the aforementioned GPUs. This practice requires NVIDIA to disclose its relationship with its customers more clearly."

Some skeptics even believe that Coreweave, the AI tech company that placed orders worth billions of dollars with NVIDIA, may not even be a real company, but rather a shell organization created and supported by BlackRock, NVIDIA's largest shareholder, in collaboration with NVIDIA.

Some say that NVIDIA has also colluded with "big players" such as Microsoft, Amazon AWS, and Inflection, the world's third-largest large-scale model startup, and that major industry players have placed more problematic orders with NVIDIA to help boost its revenue:

"An incredible theory gaining attention on X/Twitter suggests that NVIDIA's astonishing financial performance this year was driven by selling fake GPUs to false and non-existent customers.

Some have begun to draw parallels between this event and the collapse of the FTX pyramid scheme, and are concerned that such a scam could bring down the entire AI/chip market."

Analysis: The conspiracy theory demonstrates the intense long-short game surrounding NVIDIA, with underlying concerns about the AI bubble bursting and dragging down the entire market.

But Bernstein analyst Stacy Rasgon strongly expressed his dissatisfaction with this conspiracy theory, stating in his latest research report this week, "Hey guys, please don't believe the childish investment arguments you randomly see on social media."

Regarding the skepticism about Nvidia's doubled revenue and only 7% increase in sales costs, Rasgon explained that this is because Nvidia's second-quarter expenses include a provision for inventory of approximately $1.22 billion, which was included in the cost of sales in the second quarter of last year. "When excluding these expenses, the sales costs in the second quarter actually increased by about 70% compared to a year ago, which is within the normal range."

The second mainstream doubt involves Coreweave, the largest Ethereum miner in North America that transformed into a GPU cloud provider. Rasgon reminded that the company mortgaged Nvidia H100 chips in exchange for debt financing led by Blackstone Group, not Nvidia's major shareholder BlackRock. The misinterpretation of this conspiracy theory itself is ridiculous, and furthermore:

"Nvidia doesn't need help from Coreweave (or anyone else) to stimulate second-quarter sales. Nvidia's products have already been distributed.

Moreover, this debt agreement was announced after the completion of the second quarter on August 3, and the debt issuer implied that the deployment of funds may not have occurred yet. (So how does this falsely inflate Nvidia's second-quarter financial report?)"

There are also analyses pointing out that Coreweave had only $371 million in self-owned funds before the August financing. How could a company of this size have the ability to jointly operate revenue with Nvidia, which exceeded $10 billion in data center revenue in the second quarter?

Currently, there is no direct evidence pointing to collusion between the two. The only criticism is that Nvidia has particularly supported its invested company during the capacity shortage.

Those who support Nvidia generally believe that the popular conspiracy theories on social media only indicate one thing, that the long and short game against Nvidia is highly competitive. The underlying logic may stem from deep fear and concern about the view that "AI will fuel the next tech bubble burst."

Rob Arnott, a well-known figure in the quantitative investment field, stated this week that Nvidia is the leader of the AI stock bubble that is about to burst, and its collapse may drag down the entire market. Nvidia's significant rise this year is a textbook example of a "major market illusion":

"Remember, whenever a bull market appears in a powerful narrative that can only drive up the stock prices of a small number of popular companies, these stocks may disappoint in the coming years. This paradigm has been played out after the 2000 tech bubble and the 2008 financial crisis.

Artificial intelligence will change our world, there is no doubt about it. But Nvidia's stock price indicates that the market believes it will not be challenged by new entrants in the same industry or regulatory agencies questioning its dominant position, which is beyond reasonable reality."

What are the doubts of those who question the authenticity of NVIDIA's Q2 earnings report? Timing of stock buybacks, unsustainable high valuation

Returning to the initial article by Macrotips Trading, unlike the deliberate anxiety-inducing skepticism on social media, this analytical piece deeply examines NVIDIA's Q2 earnings report and raises several thought-provoking points.

Firstly, NVIDIA's gross margin expansion over the past few quarters is "incredibly impressive, leaping from 43.5% in Q2 of last year to 64.6% in Q1 of this year, and further expanding to 70.1% in Q2 of this year," while its competitor AMD's gross margin remained stable YoY:

"Perhaps NVIDIA's gross margin expansion merely reflects the tremendous demand for its GPUs in the booming artificial intelligence and generative AI chip market, where customers are willing to pay double the price for NVIDIA chips, thereby doubling its revenue without significantly increasing costs.

However, we need to consider NVIDIA's revenue recognition rules for licensing and development arrangements. Unlike product sales where revenue can be recognized upon shipment and transfer of control, a portion of data center revenue may come from software licenses, allowing revenue to be recognized in advance when the software is made available to customers, even if actual payment is made later.

This policy of recognizing revenue in advance, even before receiving payment, can explain why NVIDIA's accounts receivable increased by $3 billion QoQ to $7.1 billion in Q2. Moreover, Q2 accounts receivable benefited from $1.25 billion in payments made by customers before the invoice due date. All of these factors contributed to a staggering $6.3 billion QoQ increase in NVIDIA's overall quarterly revenue."

Secondly, the article questions the timing of NVIDIA's stock buybacks. With $4 billion remaining from the previously approved buyback plan at the end of Q2, the company's board approved a new buyback plan of $25 billion:

"The choices made by NVIDIA's management are perplexing. The financial report shows that in Q1, they did not repurchase any shares, and in Q2, they repurchased 7.5 million shares for $3.28 billion. Furthermore, between July 31 and August 24 of this year, they repurchased 2 million shares for $998 million.

Calculations show that 7.5 million shares worth $3.28 billion, equivalent to an average repurchase price of $437 per share, and 2 million shares worth $998 million, equivalent to an average share price of $499.

In the second quarter of NVIDIA's fiscal year, which ended on July 30th, only a few days had trading prices higher than $437. In order to maintain an average price of $437, the company had to repurchase stocks at the end of the second quarter, when the stock was already close to its historical high. Similarly, between July 31st and August 24th, only on August 24th did the stock price exceed $499, indicating that most repurchases definitely occurred on that day.

(Interestingly, NVIDIA almost chose to repurchase stocks when the price was close to its historical high), but most companies repurchase stocks through plans such as volume-weighted average price over a longer period of time to avoid excessive impact on the stock price."

Thirdly, the article also questions that while the company was continuously repurchasing stocks at historical highs, insiders at NVIDIA were accelerating the sale of their holdings:

In the past 6 months, insiders sold $234 million worth of stocks, with the CFO selling $2.3 million on August 28th and $2.5 million on May 30th, and CEO Jensen Huang exercising stock options to sell $117 million on September 1st.

Finally, the article states that NVIDIA's current valuation is "sky-high," with a trading price of up to 16.7 times the future enterprise sales value (Fwd EV/Sales) and a trailing P/E ratio of 117:

"Even if analysts' predictions are correct, that NVIDIA can increase its revenue to $111 billion in the fiscal year 2027 (calendar year 2026), the company's current enterprise value of $1.2 trillion still exceeds 10 times the EV/Sales ratio.

While artificial intelligence and NVIDIA GPUs are likely to change the world as we know it, valuing NVIDIA at least 10 times its projected sales in the next three years would create a bubble even more shocking than the dot-com bubble of 2000.

As insiders accelerate their selling, the risk of holding NVIDIA stocks seems to be increasing. Personally, I am uncomfortable holding shares of a company whose valuation is based on expected revenue more than 10 times in three years. I wouldn't short NVIDIA stock, but current holders should sell."

NVIDIA Executives and Industry Insiders Fear No Doubt: AI Market Size Could Reach $600 Billion, Trend Undeniable

But NVIDIA executives seem unfazed by skepticism in the market.

At a technology conference hosted by Goldman Sachs this week, Manuvir Das, Vice President of Enterprise Computing at NVIDIA, predicted that the long-term potential market value of artificial intelligence could reach as high as $600 billion.

This will be driven by $300 billion in chips and systems, $150 billion in generative AI software, and $150 billion in omniverse enterprise software, all powered by NVIDIA's "accelerated computing" initiative.

Das stated that NVIDIA is capitalizing on an inevitable industry trend, where ultimately, enterprise operations will become digitized and efficiency will be improved in ways previously unimaginable:

"The simplest way to look at NVIDIA is that we made a big bet, and we won. We've been brewing this for decades."

Tom Traugott, Senior Vice President of Digital Infrastructure Strategy at EdgeCore, a digital infrastructure solutions provider, also noted that under the AI boom, critical infrastructure is undergoing changes that could have far-reaching effects:

"Companies will have to improve their existing systems to meet the demands of AI, thereby driving the demand for NVIDIA chips that support more efficient GPUs in data centers.

A company may have AI chips to distribute, but the next critical question is whether they have the data center capacity to make these chips useful. AI is driving data center construction to be larger and denser than ever before."

Contrary to online conspiracy theories, according to FactSet data, out of 51 Wall Street analysts covering NVIDIA, 47 analysts have given a "buy" rating, 4 analysts have given a "hold" rating, and none have given a "sell" rating. The average target price is $649.22, with a potential upside of 42%.