As of the closing bell on Monday in the U.S. stock market, NVIDIA shares saw a significant surge, climbing by 4.9% to reach a value of $1,150 per share. This increase pushed NVIDIA's market capitalization to an impressive $2.83 trillion, marking a remarkable gain of approximately $130 billion compared to the previous trading session. This spike in market value outstripped that of Intel on the same day, which stood at about $1,289 billion at the close of the market.

The competition between NVIDIA and Apple regarding market valuation has recently been heating up. The last time NVIDIA's market cap was in striking distance of Apple's was on May 28, when only a mere $100 billion separated the two tech giants. As both companies continue to innovate and expand their reach in the technology sector, this race for market supremacy has become increasingly captivating to investors and analysts alike.

NVIDIA CEO Jensen Huang has shed light on the company's roadmap during a recent keynote speech, revealing a schedule of upcoming product iterations. The timing of this announcement coincided with a weekend market break. Huang outlined plans for NVIDIA's Blackwell chip, which has already begun mass production, introducing the forthcoming Blackwell Ultra in 2025, followed by the new architecture Rubin in 2026 and an advanced Rubin Ultra in 2027. This proposed timeline suggests an ambitious strategy of releasing updates annually, which can generate considerable excitement and anticipation within the tech community.

Advertisement

As the excitement surrounding NVIDIA's stock performance continues, the sustainability of this upward trend has become a hot topic in the tech industry. Analysts from DA Davidson, including Gil Luria, have issued warnings regarding NVIDIA's short-term outlook, emphasizing that while the immediate prospects appear bright, the long-term situation might be less rosy than anticipated. Major tech players such as Meta and Amazon are actively developing their own artificial intelligence chips or investing in partnerships with other firms, which could pose a formidable challenge to NVIDIA's continued dominance.

Currently, leading AI chips from Intel, such as Gaudi 3, Amazon's Trainium and Inferentia, Google's various TPU (Tensor Processing Units), Microsoft's Maia100, and Meta's MTIA v1 are emerging on the scene. These high-profile companies have primarily focused on developing ASIC (Application-Specific Integrated Circuits), which do not directly compete with NVIDIA's core market in GPUs (Graphics Processing Units). While ASICs tend to lack the flexibility of GPUs, they exhibit superior performance and energy efficiency in specific applications. However, transitioning large models that originally ran on GPUs to these alternative AI chips may involve substantial costs.

As these tech behemoths ramp up their AI chip research, significant developments have transpired in the industry. In April of this year, Intel announced the release of its Gaudi 3 chip, promoting its enhanced performance in training large language models over NVIDIA's H100. Meanwhile, Meta unveiled the next generation of its AI chip MTIA, which boasts a 3.5 times increase in computational density compared to its first iteration launched last year. Beyond chips, efforts to disrupt NVIDIA’s lead in chip interconnectivity have also surfaced. Eight tech giants, including Google, Meta, AMD, Intel, and Broadcom, recently formed the UALink Promoter Group aimed at establishing industry standards for interconnecting AI accelerator chips within data centers.

Breaking NVIDIA's stronghold on the GPU market may prove to be a challenging endeavor. According to a professor specializing in GPU research, the limitations imposed by Moore’s Law are nearing their peak, making it increasingly difficult to enhance GPU performance. Despite this, the professor noted that no alternative AI chip has proven to be as commercially successful as GPUs to date. The versatility of GPU architecture remains a significant competitive advantage amidst emerging rivals. Faced with hurdles in performance advancements through sheer processing production, NVIDIA is adapting by leveraging chiplet technology, minimizing data formats, embracing advanced packaging, and experimenting with novel materials to sustain its innovation momentum.

A domestic developer of large AI models shared insights, noting that the availability of NVIDIA's high-end graphics cards is relatively scarce in the domestic market. After experimenting with domestic brands' AI chips, the developer found that while the performance gap with NVIDIA might not be substantial, the usability factor often falls short. NVIDIA’s CUDA software ecosystem continues to stand out, surmounting ease of integration challenges when developing AI applications.

Apple, too, is advancing in the AI arena. With its World Wide Developers Conference (WWDC) set to take place in June, speculation is rife that AI might take center stage. Recent reports suggest that Apple has forged a partnership with OpenAI, potentially leading to the integration of ChatGPT into iOS 18. At the same time, Apple is focusing on developing smaller, endpoint models that can run effectively on devices without a constant internet connection.

Interestingly, while Apple and NVIDIA are both significant players in the AI field, they do not directly compete in most areas. NVIDIA’s AI chips see primary application in data centers and computing environments, whereas Apple benefits from a holistic and fully integrated hardware-software ecosystem through its proprietary chips. Collaborating with external firms like OpenAI to leverage powerful models while simultaneously innovating on small models for in-device usage could be indicative of Apple’s strategic direction in AI.

Furthermore, Apple’s smartphone sales in China have recently shown signs of rebounding. Data from IDC reveals that Apple’s domestic shipments fell by 6.6% year-over-year in the first quarter of this year. However, statistics from the China Academy of Information and Communications Technology indicate that foreign smartphone brands collectively shipped 3.5 million units within China in April, representing a remarkable year-over-year growth of approximately 52%. Apple, being the leading foreign smartphone brand in China, may benefit from an anticipated "supercycle" revolving around AI advancements and the unveiling of the iPhone 16. Some analysts project that Apple’s market capitalization could surpass $4 trillion by next year.

In conclusion, the demand for AI technologies continues to surge, and competition among tech giants within this arena is just warming up. As NVIDIA seeks to maintain its competitive edge amidst emerging challengers, and as Apple strategically positions itself in the AI landscape, the unfolding narrative of their rivalry presents an intriguing saga that will surely captivate investors and technophiles in the months and years to come.