NVIDIA's highly anticipated Blackwell "unfavorable start" but Wall Street cheers "buy on dips"

According to the Zhitong Finance APP, the AI chip leader NVIDIA (NVDA.US) encountered a "bad start" with two new cutting-edge AI chips due to design flaws and severe engineering challenges during the chip manufacturing process. These AI chips based on the breakthrough Blackwell architecture are crucial for NVIDIA to further expand its leading position in the artificial intelligence computing infrastructure market. Sources revealed that this delay affected NVIDIA's highly anticipated Blackwell architecture chip lineup announced in March, with the B100 possibly being eliminated due to design flaws and chip engineering issues, to be replaced by the slightly lower-performing B200A. Although the Blackwell delay caused NVIDIA's stock price to plummet, Wall Street institutions such as Bank of America believe it is an excellent opportunity to buy on the dip.

Sources also revealed that NVIDIA is redesigning a version of AI chips called artificial intelligence accelerators to better integrate with data center infrastructure built on the previous generation AI chip architecture, Hopper, and the H100.

However, sources indicated that the data center hardware sector targeted by the B100 is a relatively niche market, with the real large-scale market focusing on the B200 AI GPU and the market targeted by the GB200, which combines two B200 chips with NVIDIA's self-developed Grace CPU.

Furthermore, sources mentioned that due to some issues in chip manufacturing and CoWoS packaging technology support, the GB200 product combining the Grace CPU with NVIDIA AI GPU will not be rapidly launched in large quantities as previously expected by the market.

These messages revealed by sources were first reported by the Information website, reflecting the challenges NVIDIA faces in accelerating innovation. CEO Jensen Huang is speeding up the introduction of new chip design architectures and chip interconnection technologies to maintain NVIDIA's almost monopolistic advantage in the core infrastructure of artificial intelligence. The company dominates the entire AI chip market, holding over 90% market share, with AI chips driving the most core hardware behind generative artificial intelligence tools like ChatGPT. With its near-monopoly advantage, NVIDIA's performance and total market value have rapidly soared in the past two years, surpassing Apple and Microsoft at one point to become the "world's highest market value listed company."

NVIDIA has been deeply involved in the global high-performance computing field for many years, especially with its CUDA computing platform, which it single-handedly created and has gained global popularity. CUDA is a software and hardware collaborative system for high-performance computing in areas such as AI training/inference, making it the preferred choice in the field and a core force driving NVIDIA's 90% market share in AI chips. The CUDA computing platform is an exclusive parallel computing acceleration platform and programming assistance software developed by NVIDIA, allowing software developers and engineers to use NVIDIA GPUs for accelerated parallel general-purpose computing (only compatible with NVIDIA GPUs, not compatible with mainstream GPUs like AMD and Intel)

Production delays at Blackwell trigger a sharp drop in NVIDIA's stock price

However, NVIDIA has refused to comment on recent media reports regarding manufacturing issues with the Blackwell architecture chips. A company spokesperson stated that they have started widely sending samples of Blackwell to customers, and the market demand for their Hopper architecture AI chips remains very strong. The Hopper architecture AI chips currently include the H100 and the H200, which have already been mass-produced. NVIDIA also mentioned that the production pace of Blackwell architecture chips is expected to accelerate in the second half of the year.

Reports on the manufacturing challenges of Blackwell architecture chips led to a significant 6.4% drop in NVIDIA's stock price on Monday after experiencing a pullback last week. However, the broader sentiment of tech stock decline also affected the stock's performance. Meanwhile, NVIDIA's strongest competitor in the AI chip field, AMD (AMD.US), rose by 1.8% amidst the overall decline in the US stock market. Investors hope that this may be an opportunity for AMD's AI GPU to gain more market share from NVIDIA.

NVIDIA's AI chips play a core role in the hardware end of data centers of cloud computing giants like Microsoft and Google, who are NVIDIA's largest-scale customers. These cloud giants are currently spending billions of dollars globally to build massive AI data centers to meet the huge cloud computing resources needed for their global AI technology deployment.

In May, NVIDIA CEO Jensen Huang stated at an earnings call that the Blackwell architecture AI chips are about to enter full production and are expected to deliver a certain scale of shipments to cloud providers in the fourth quarter of this year. He predicted at the time that the demand for the new series and its predecessor Hopper architecture products will continue to exceed supply. "This year we will see sales of Blackwell architecture," Huang said during an earnings call with analysts. The company plans to release its next quarterly earnings on August 28th, Eastern Time.

NVIDIA's next-generation architecture AI chips - the Blackwell-based AI chip series - will see a new boost in ultra-high performance. Tech giants such as Amazon, Dell, Google, Meta, and Microsoft will heavily deploy Blackwell AI GPUs in their latest data center AI server systems. Wall Street analysts widely speculate that the demand for NVIDIA hardware from these tech giants will far exceed market expectations. Recently, industry insiders revealed that due to the strong global demand for NVIDIA's upcoming mass-produced Blackwell architecture AI chips, NVIDIA has significantly increased its chip foundry orders with TSMC by at least 25%.

NVIDIA's hottest AI chips, the H100/H200 GPU accelerators, are based on NVIDIA's breakthrough Hopper GPU architecture, providing more powerful computing capabilities compared to the previous generation, especially in floating-point operations, tensor core performance, and AI-specific acceleration. More importantly, the AI GPU performance based on the Blackwell architecture far exceeds that of the Hopper architecture. In the GPT-3 LLM benchmark with 175 billion parameters, the inference performance of the Blackwell architecture GB200 is 7 times that of the H100 system And provides training speed 4 times that of the H100 system.

Bullish NVIDIA remains the main theme on Wall Street, with strong calls to "buy on dips"

Although delays may affect NVIDIA's chip manufacturer TSMC's chip manufacturing process and capacity planning, forcing TSMC's tense CoWoS capacity to shift to chip giants such as AMD. However, most Wall Street analysts have taken a calm approach and still insist on the bullish trend of NVIDIA's stock price.

Based on NVIDIA's upcoming launch of the new generation Blackwell AI GPU and the strong demand for NVIDIA H100/H200 AI GPUs by the end of the year, some Wall Street analysts expect this to stimulate a new round of performance and stock price growth for NVIDIA. Therefore, they are more determined to be bullish on NVIDIA's stock price trend within the next 12 months, believing that NVIDIA's stock price is poised for a new round of growth, and stating that the recent plunge is an excellent opportunity to "buy on dips" for NVIDIA.

"Given the acceleration of innovation, capacity disruptions may continue to occur." Analyst Matt Ramsay from TD Cowen stated in a research report.

He emphasized that even if there is a delay in delivery for a few weeks, it may not have a significant negative impact on NVIDIA's explosive revenue growth rate this year or long-term performance growth. However, Ramsay stated that this largely depends on how NVIDIA ultimately resolves these issues and how quickly they can deliver chips to major customers. The analyst reiterated a "buy" rating on NVIDIA and reaffirmed a 12-month target price of $165 (NVIDIA's stock price fell 6.36% to $100.45 on Monday).

Morgan Stanley recently stated that NVIDIA hopes to "redesign" the Blackwell architecture AI chip to further improve the stability of the chip and servers equipped with NVIDIA AI chips. It is expected that due to yield issues with CoWoS-L packaging, the B200A will switch to CoWoS-S packaging. Morgan Stanley emphasized that although the Blackwell architecture chip may be temporarily suspended from production at TSMC for up to two weeks due to redesign, the fourth-quarter expansion of TSMC's chip manufacturing and packaging capacity is expected to recover the delayed capacity and ultimately deliver Blackwell to cloud giants on time. Morgan Stanley reiterated a target price of up to $144 for NVIDIA within 12 months.

Bank of America emphasized that the recent decline in NVIDIA's stock price provides an excellent opportunity to buy on dips, reaffirming a target price of $150 and being the preferred investment target in the chip industry. Regarding reports that NVIDIA's next-generation Blackwell chip is delayed, Bank of America stated that Blackwell may not be included in performance until the fourth quarter, and even if there is a delay in capacity, it may not affect NVIDIA's Blackwell-related revenue. NVIDIA and TSMC may be able to solve chip design and engineering issues in the short term. At the same time, Bank of America emphasized that the strong lifecycle of NVIDIA's previous generation Hopper product will continue to drive NVIDIA's high-speed performance growth, and Bank of America stated that the challenge is not in market demand but on the supply side