The AI landscape evolved significantly, thrusting it into the limelight for leading technology firms with the introduction of OpenAI’s ChatGPT. Before this advanced language model-based chatbot was unveiled, AI was certainly explored and considered by tech enterprises, but seldom was it prioritized. However, now, unwavering optimism about the burgeoning potential of AI continues to pervade organizations.
As we bid adieu to 2023 – notably marked by the ascendancy of AI – queries linger regarding the monetization strategies businesses will employ around this transformative technology. However, amid the clouds of uncertainty, one fact stands crystal clear – AI is here for the long haul, advancing at a stupendously swift pace.
NVIDIA Corporation (NVDA), a leading force in the semiconductor realm, has secured an iron-clad position in the AI arena thanks to its timely engagement with AI technologies. This strategic move has propelled NVDA years ahead of its competitors, enabling it to provide a comprehensive platform catering to all AI requisites – from advanced chips and processors to complex software systems.
Intel Corporation (INTC), distinguished for trailblazing semiconductor innovation and globally esteemed for its Central Processing Units (CPUs), fuels a myriad of devices spanning personal computers to expansive data centers. INTC’s Core processors are applauded for their exceptional performance and unwavering reliability. In a recent turn of events, INTC has plunged into the AI sphere with products like the AI-specialized Core Ultra and server CPUs like Emerald Rapids.
The global AI market is expected to reach $1.35 trillion, growing at a 36.8% CAGR. It would be strategically amiss for INTC not to stake its claim. However, the extent of INTC’s share within this booming market hinges upon factors like the raw processing capacity, the versatility of the tech, and the number and diversity of potential applications for its semiconductor chips.
Relations between INTC and NVDA have soured following the “AI Everywhere” event, as a harsh critique from INTC’s CEO Pat Gelsinger has elicited retaliatory remarks. The dispute originated when Pat Gelsinger criticized NVDA’s AI strategy as being “shallow and small.” This sparked a skirmish of words between the two tech titans, relentlessly escalating since then. The wrangle fueled up when INTC suggested that NVDA’s status in the sector was simply due to “luck,” provoking a robust response from the graphics giant.
NVDA reacted defensively to INTC’s comments, vehemently denying the ‘luck’ factor in their success. Bryan Catanzaro, NVDA’s Vice President and previously involved with INTC’s discontinued Larrabee project, took to the internet to express his views. He pointed to INTC’s failure to capitalize on the rising AI trend as evidence of their lack of “vision and execution.”
During the “AI Everywhere” event, INTC unveiled numerous innovative products and provided updates on their product pipeline. The highlight of the event was the introduction of the Gaudi 3 accelerator. However, Meteor Lake, an AI-optimized Core Ultra, and Emerald Rapids for cloud applications were also unveiled along with new chipsets, underscoring INTC’s commitment to fostering an extensive AI ecosystem.
It is reported that Gaudi 3 could outperform NVDA’s H100 accelerator, the current top choice for companies developing sizeable chip farms to power AI applications, despite NVDA is yet to launch its latest high-end AI chip.
Stepping into the booming AI market with gusto, INTC’s latest offerings include an upgraded version of Xeon server chips – marking the chips’ second major update in less than a year. These chips are designed to propel INTC to the pinnacle of AI innovation by offering improved performance and memory capabilities with lower electricity consumption.
INTC posits that the Xeon is the only mainstream data center processor with built-in AI acceleration. The company’s 5th Gen Xeon delivers up to 42% higher inference and fine-tuning on models boasting up to 20 billion parameters.
Let’s look at the different approaches to AI processing…
At present, divergent strategies toward AI processing are being implemented by INTC and NVDA. NVDA’s CUDA puts more emphasis on educating AI from the ground up. It involves supplying the AI model with data to let it learn – akin to educating an individual to achieve the necessary qualifications required for a job.
On the other hand, INTC leans more toward “inference.” In this approach, a pre-existing AI model adapts and acquires knowledge from an unfamiliar situation. This could be likened to assigning a task relevant to their field to someone armed with the necessary degrees but lacking experience, then observing how they apply their theoretical knowledge in a practical situation.
Probable Impacts of Approaches on the Broader AI Industry in 2024…
NVDA’s CUDA is a software platform API facilitating parallel computing with GPU hardware. This breakthrough solution simplifies the process for developers, allowing them to create software that accelerates tasks by distributing workloads across parallelized GPUs.
CUDA has proven instrumental in driving pioneering advancements within various AI sectors. Thanks to the computational firepower harnessed from CUDA, researchers can train increasingly complex models, manage larger datasets, and achieve best-in-class results in record-breaking time.
The AI industry widely embraces CUDA for various applications, including deep learning, robotics, computer vision, and natural language processing. The platform will continue to evolve, underpinning new architectures such as NVDA’s Hooper and Ada Lovelace. These innovations promise cutting-edge Tensor Cores and Transform Engines, advanced memory management capabilities, and more.
CUDA’s potential lies in boosting the performance and effectiveness of AI applications through the exploitation of GPU’s powerful parallelism and high bandwidth. For instance, it allows deep learning models to run faster on GPUs rather than CPUs or cloud servers.
Using CUDA, AI researchers and developers gain access to the enormous computational capacities of GPUs. This enables the training and deployment of AI models at revolutionary speeds, improving the efficiency of AI algorithms and decreasing the time necessary for model development and deployment. This acceleration empowers greater innovation in the field.
INTC’s Inference is the process used for making predictions through a trained model to make a prediction. The company’s primary focus is vested in the inference market. Gelsinger commented in the event, “As inferencing occurs, hey, once you’ve trained the model … There is no CUDA dependency. It’s all about, can you run that model well?”
This evolving realm presents considerable competition for INTC. Nevertheless, the firm’s leadership considers the inference market to be a thriving ground for the futureAgra Investment. The inference model aims to provide momentum to the expansive AI market by promoting cost efficiency and deployment of AI models across diverse platforms and devices. This would result in stimulating innovation and fostering collaboration among numerous stakeholders within the AI ecosystem.
With the adoption of AI reaching unparalleled heights, the demand for ground-breaking methods to train these AIs will be vital to conserve time and resources. While it is premature to conclude whether INTC’s strategy will outmaneuver CUDA, the recent launch of INTC’s Meteor Lake CPUs featuring an integrated Neural Processing Unit (NPU) reflects the firm’s determination to incorporate AI profoundly into its products.
We shall now delve into some other factors to ascertain which stock holds an advantage:
Past and Expected Performance
NVDA’s revenue has grown at 44.8% and 29.3% CAGRs over the past three and five years, respectively. Its tangible book value has increased at CAGRs of 49.4% and 25.7% over the same periods.
On the other hand, INTC has yet to translate to top-line improvement, as its revenue has shown negative CAGR growths of 12.2% and 5.3% over the past three and five years, respectively. However, its tangible book value has increased at CAGRs of 22.5% and 14.6% over the same periods.
Analysts expect NVDA’s EPS for the fiscal year ending January 2025 to reach as high as $19.94 from the $12.30 expected in Fiscal 2024 (ending January 2024). For the fiscal year ending January 2024, NVDA’s revenue is expected to reach $58.77 billion, up 117.9% year-over-year, while for the fiscal year 2025, analysts expect its revenue to reach $90.66 billion. For the fourth quarter, the company expects its revenue to be $20 billion, plus or minus 2%.
Analysts expect INTC’s revenue for the fiscal year (ending December 2023) to come in at $54.07 billion, indicating a decline of 14.2% year-over-year. The consensus EPS estimate of $0.95 for the ongoing year indicates a 48.2% year-over-year decrease. For the fiscal year ending December 2024, its revenue and EPS are expected to increase 13% and 100% year-over-year to $61.10 billion and $1.91, respectively.
Financials
NVDA’s net revenue for the fiscal third quarter ended October 29, 2023, increased 205.5% year-over-year to $18.12 billion. Also, its non-GAAP net income and non-GAAP EPS came in at $10.02 million and $4.02, up 588.2% and 593.1% year-over-year, respectively.
During the fiscal third quarter that ended September 30, 2023, INTC’s revenue amounted to $14.16 billion, down 7.7% year-over-year. Its net income came at $297 million, while non-GAAP EPS increased 10.8% year-over-year to $0.41.
Bottom Line
NVDA showcases remarkable potential in the AI field, reaping significant profits from the sector. However, despite impressive revenue growth in the last quarter, boosted by a surge in AI GPU sales, NVDA continues to grapple with intricate macroeconomic difficulties. Sustained export limitations have negatively impacted sales to organizations in countries like China. Given the present regulatory environment, NVDA foresees a significant downturn in sales.
As the U.S. intensifies its export control measures, NVDA is proactively working to develop chips tailored for the Chinese market. However, the U.S. is expected to tighten these controls even more tightly to curb strides in new technologies that could favor Beijing.
A delay in the deployment of chips specifically optimized for China’s market has resulted in an NVDA stock dip over the last month. While the recent slide may seem like an attractive investment opportunity, investors must be mindful of the possibility of further short-term volatility.
It is crucial to note that NVDA shares are currently trading at 20.59x sales and 39.83x earnings; valuations that imply any missteps have the potential to affect the company’s market position significantly. Given the ongoing market instability and associated risks, it may be judicious for prospective investors to wait for a better entry point into the stock.
A notable concern for INTC stems from its primary focus as a PC hardware entity. Evidenced by the 2021 historic surge in PC sales, mainly attributed to COVID-19 and a rising interest in PC gaming, INTC relayed a substantial dependency on this market.
However, as we navigate into 2023, the PC market has dramatically retrenched compared to the extraordinary circumstances of 2021. Adding insult to injury, it appears NVDA has been outpacing INTC within the data center arena since the launch of ChatGPT.
An emergent trend of building accelerated computing server instances and a marked preference for H100 GPUs has seen a decline in the demand for CPU-heavy, compute-optimized instancesSimla Wealth Management. This act has placed INTC in a challenging position since the Xeon processors, once a market favorite, are now struggling to find takers.
While INTC’s Gaudi has indeed enhanced AI/ML performance to a certain degree, there’s no denying its noticeable lag behind NVDA. The comparative analysis drawn from third-party benchmarks and NVDA’s unprecedented winning streak at the last MLPerf inference and training benchmarks only reinforces this fact.
INTC certainly boasts commendable computing expertise and continues its endeavors to gain mastery in process technology. Alas, Critics argue it has been sluggish in its execution compared to its counterparts. Rising costs associated with fabrication work and a tarnished reputation – being labeled as expensive, slow, and difficult to collaborate with – have acted as further impediments to its growth. Nevertheless, currently, INTC appears confined to providing chip fabrication services for ARM chips to fabless chip designers.
Indore Investment