Why is it difficult to short NVIDIA, which has surged 200% this year? Analysts say it monopolizes the AI economy.

Wallstreetcn
2023.08.17 16:32
portai
I'm PortAI, I can summarize articles.

Wall Street is becoming more optimistic ahead of Nvidia's earnings report next Wednesday, as it seems that the severe shortage of GPUs in the coming quarters will drive its performance. Many mainstream investment banks have listed Nvidia as their "preferred stock" to chase the AI boom, but some analysts are concerned that Nvidia's earnings report will be the biggest test for the AI hype. Currently, both Nvidia and Wall Street believe that second-quarter revenue will surpass $10 billion, reaching a new high.

"NVIDIA, the 'AI darling,' will announce its second-quarter earnings report after the market closes on Wednesday, August 23. Despite the stock's staggering rise of over 196% year-to-date, which is close to an astonishing 200% increase, Wall Street analysts remain confident in its future growth and recommend buying on dips.

Why is shorting NVIDIA seemingly the least popular bet right now? The reason is simple, as stated in a research report by Barclays analyst Blayne Curtis and his team on Wednesday:

"NVIDIA is the best company in the field of artificial intelligence, with no obvious competitors close behind. NVIDIA has so far monopolized the thriving AI economy, and it is hard to imagine that its earnings report won't be positive, leading to a surge in stock price."

With more and more analysts raising their target price for NVIDIA and maintaining a 'buy' rating ahead of the earnings report, it is evident that NVIDIA's performance will face a significant challenge in surpassing market expectations. However, NVIDIA doesn't seem worried.

As early as May, when the company announced its first-quarter earnings report for the fiscal year 2024, it gave a very 'dominant' guidance for the second quarter, projecting revenue to be $11 billion (with a 2% fluctuation range), which would break the previous record of $8.29 billion in revenue from the first quarter of the previous fiscal year.

Before NVIDIA provided its high guidance, analysts surveyed by FactSet had an average expectation of $7.17 billion in revenue for the second quarter, a 7% increase compared to the same period last year, while NVIDIA directly pointed out that revenue growth could reach as high as 64%.

Currently, Wall Street has raised its expectations for NVIDIA's second-quarter revenue to $11 billion and predicts earnings per share (EPS) to be $2.07, a staggering 306% increase compared to the same period last year. NVIDIA expects its adjusted gross margin for the second quarter to be 70%, higher than the previous quarter's 66.8%.

On Thursday, August 17, NVIDIA's stock price rebounded during the midday trading session.

Wall Street is increasingly optimistic ahead of the earnings report, as the GPU shortage is expected to drive NVIDIA's performance in the coming quarters

This Tuesday, Aaron Rakers, an analyst at Wells Fargo, raised NVIDIA's target price from $450 to $500, representing a 14% upside potential from Monday's closing price. The reason behind this adjustment is that "it is difficult to deny NVIDIA's outstanding position as the primary beneficiary of the AI-driven architecture data center transformation. The company still has unrecognized opportunities in platform expansion and monetization."

Tristan Gerra, an analyst at Baird, raised the target price from $475 to $570, indicating a further 30% increase. The reason for this adjustment is that "AI demand is surging across various levels for individuals, businesses, and cloud service providers."" Morgan Stanley analyst Joseph Moore suggested on Monday that the recent decline in NVIDIA's stock price over the past two weeks presents a buying opportunity, calling it the "top choice" among companies benefiting from AI. Currently, the stock has pulled back nearly 9% from its all-time high of $474.94 on July 18.

Morgan Stanley has set a target price of $500, citing the "unique supply-demand imbalance of NVIDIA's GPUs that will continue over the next few quarters." Currently, half of the demand cannot be met, and once the supply improves, its advantage in the data center will further expand.

Mainstream analysts generally acknowledge the "macro background of large-scale corporate spending shifting towards artificial intelligence." UBS analyst Timothy Arcuri also recommended buying NVIDIA on dips and raised the target price to $540, calling NVIDIA the "kingmaker" because a large amount of capital is chasing new AI software and specialized cloud infrastructure models:

"Even for more tactical investors, it seems premature to 'get off' now. Because the comparison with the performance of the same period last year will not become difficult until the end of the 2024 fiscal year, NVIDIA rarely reaches its stock price peak before the YoY comparison tops out."

Barclays analyst mentioned earlier maintains the target price of $600 and also mentioned the trend of "cloud capital spending budgets shifting towards artificial intelligence" despite the weakening of corporate spending. They expect that as the supply improves, Wall Street's expectations for NVIDIA's earnings per share for the fiscal year 2025 will be raised from $15 to over $20, making the current valuation of the stock "very cheap" compared to its performance potential.

Raymond James analyst Srini Pajjuri maintained a "strong buy" rating on Tuesday and raised the target price from $450 to $500, stating that the trend of NVIDIA's better-than-expected performance and upward guidance will continue until the 2024 calendar year. The reason is the same: "Although the trend of cloud capital spending is uneven, the demand for GPUs in the AI spending boom still significantly exceeds supply":

"Supply constraints may limit NVIDIA's short-term stock price upside, but the story of generative AI remains intact. Given NVIDIA's dominant position in the field of AI and machine learning, such a high valuation is justified. Higher gross margins and better-than-expected data center revenue are expected in the third quarter of the 2024 fiscal year, which will further strengthen NVIDIA's performance."

Piper Sandler analyst Harsh Kumar also believes that NVIDIA will maintain high-speed growth in its data center business in the third quarter of the fiscal year, and raised the target price to $500 this week, emphasizing the demand from large cloud companies and the driving force of China:

"In the third quarter, NVIDIA may record $9.5 billion in data center revenue, and the overall data center revenue for the fiscal year 2024 may reach approximately $32 billion. Due to the continuous acceleration of data center penetration and the sustained leading market share in AI applications, we expect NVIDIA's data center revenue to further grow in the fiscal year 2025. "

NVIDIA's earnings report will be the biggest test for AI hype, with the CEO stating that revenue from generative AI will be significant

However, Deutsche Bank analyst Ross Seymore remains cautious and rates NVIDIA as "hold". While he acknowledges that NVIDIA's earnings report will be "impressive" and that AI will bring more upside potential, he believes that "cyclical risks may make it difficult to predict the magnitude and slope of the company's future growth". The threshold for meeting high expectations from buyers may also be harder to surpass.

Forrester analyst Glenn O'Donnell believes that NVIDIA's earnings report next week will be "the biggest test for the AI hype". While NVIDIA is powerful, it is not invincible. The current serious problem is the inability to meet the massive chip demand from customers.

On one hand, competitors like AMD and Intel may rise and steal market share. At the same time, NVIDIA's preferred foundry, TSMC, is facing capacity constraints, and the recent surge in demand has extended the delivery time for NVIDIA's key H100 chips to 6 to 9 months. Tesla CEO Elon Musk stated in the Q2 report that they are using their own AI chips to build their own supercomputers to meet their needs.

In response to this, a Yahoo Finance column states that although NVIDIA has established a leading position through early investments in AI, the semiconductor business is known for its fierce competition, just like the former giant Intel. No company's industry-leading position is secure:

"Although NVIDIA's momentum is so strong that it is difficult to stop, it is not impossible, just difficult."

Psychologist and AI researcher Gary Marcus cites the example of the widely debunked Korean room-temperature superconductor LK-99 and says, "There are few things that have been hyped more than AI." Analysts estimate that AI will have a market worth trillions of dollars, but it is rumored that the actual revenue from generative AI is only in the hundreds of millions of dollars. "A 1000-fold increase in revenue is just speculation and may never come."

However, NVIDIA CEO Jensen Huang is optimistic about the major trend of data centers transitioning from relying on CPUs to GPUs. The value of the data center infrastructure involved is as high as $1 trillion:

"In the past, data centers relied 100% on CPUs, but increasingly large AI models require more and more computing power. In the next four to ten years, a significant portion of the upgrade and growth of data centers will shift to generative AI."

Huang previously stated at the NVIDIA Developer Conference in March that generative AI accounted for a "very, very, very small" single-digit percentage of the company's revenue in the past 12 months. However, in the next 12 months, revenue from generative AI will become "significant - it's hard to say how significant."

He predicts that AI and NVIDIA's infrastructure-as-a-service products "will expand our business model":

"

"Cloud service providers are facing challenges as the scale of artificial intelligence models is expected to increase tenfold each year. These models require a large amount of server capacity, and this demand is expected to drive Nvidia's hardware sales.

Data centers around the world are being impacted by the end of Moore's Law, and without some form of acceleration, they cannot continue to expand their computing capabilities. Therefore, Nvidia's accelerated computing solutions can help them reduce workloads that previously required thousands of CPUs to just a few GPUs."