# AI Chip Development: Competition and Strategy in the Industry

Asit Sharma: Companies like Amazon, Google, and others seeking to develop their own AI chips face key challenges. One critical step is creating simpler, purpose-built chips. For model training and inference, companies do not necessarily need the most advanced GPUs available. Amazon, for example, is advancing its Trainium chip, now nearing its third series. Similarly, Meta and Microsoft, which is relatively new to the chip market with its Maya chip, are focused on significantly reducing costs.

The goal for these companies is to provide more affordable services on their Cloud platforms. When users benefit from lower costs, businesses also save money, as they pay these companies directly for services. Many of these use cases do not require the high performance offered by NVIDIA’s GPUs; thus, companies are investing in custom solutions. They collaborate with prominent design firms, sending chip models to TSMC, a leading foundry, for prototypes before implementing them into data centers. Amazon’s recent conference call highlighted this strategy, with CEO Andrew Jassy noting a 30% reduction in compute costs in certain scenarios.

Mary Long: Patrick Moorhead, founder of Tech Research Form Moor Insights & Strategy, remarked in a New York Times article that “The gravy train comes to a screeching halt if Cloud companies stop spending.” This “gravy train” refers to the current benefits NVIDIA is experiencing. What can NVIDIA do today to secure its future against a potential downturn?

Asit Sharma: NVIDIA should focus on replenishing existing data center infrastructure. Even if there are declines in new builds, many current data centers require upgrades in their server and memory capacities to meet growing compute demands. The rising trend of AI usage won’t cease even if it becomes cheaper. NVIDIA can engage with customers on these upgrades rather than only relying on new builds. History shows that before generative AI took hold, there was a consistent increase in data center construction as enterprises transitioned to the Cloud. This migration is ongoing, and while some short-term slowdowns may occur, there remains a robust drive toward continued data center investment benefiting NVIDIA in the long term.

Mary Long: Jensen Huang’s projection of a $1 trillion AI infrastructure buildout stands out. He indicated that NVIDIA has already established around $150 billion of this infrastructure, suggesting a potential roadmap to the full amount in the near future. What is your perspective on this projection and its timeline?

Asit Sharma: Huang’s prediction doesn’t significantly differ from his previous estimates made three years ago. At that time, the concept of investing hundreds of billions appeared ambitious due to the uncertainty about AI adoption levels. Reflecting on earlier predictions, it seems that he may be on target with this $1 trillion vision.

Mary Long: It’s worth noting that NVIDIA’s stock has declined about 15% this year, with a drop of over 3% just yesterday, despite the advancements Huang is touting. The company’s PEG ratio currently sits under 0.3. As impressive as these developments are, the market appears to foresee potential growth stagnation. Did anything you heard in yesterday’s keynote provide reason to agree with this market sentiment? Or do you think NVIDIA’s future outlook is still plausible?

Asit Sharma: We must maintain a healthy skepticism regarding NVIDIA’s growth, especially as competitors emerge. Companies, such as Arista Networks in the networking sphere, are now vying in the same arena as NVIDIA, necessitating operational shifts. Notably, these developments signal a need for efficiency in computing solutions.

As smaller competitors explore alternative GPU designs, doubts about NVIDIA’s growth potential increase. Yet, you’re correct in pointing out the low PEG ratio. This could indicate that NVIDIA is undervalued compared to its potential future earnings. During the presentation, several elements suggested promising opportunities for growth.

For instance, NVIDIA’s Kuta—an array of software acceleration libraries—caught my attention. Previously discussed, Kuta includes libraries for high-precision math and aerospace, covering a wide range of industries. One new addition introduced yesterday mirrors the widely known Python library, Pandas, which helps with data manipulation. NVIDIA is now leveraging the principles behind data frames and expanding their application across multiple sectors.

Exploring NVIDIA’s Innovations and Industry Partnerships for Future Growth

Recent discussions highlight the evolving landscape of parallel computing and its implications for big data analysis. By leveraging advanced acceleration libraries on powerful GPU systems, expectations may shift regarding our ability to process data quickly. NVIDIA continues to innovate, suggesting that advancements in computation could lead to significant breakthroughs in science and data analytics.

Investors remain cautious about NVIDIA’s potential. Future discussions may reveal surprising growth rates, with double-digit increases possible within the next three to five years.

NVIDIA’s Role in Driving AI Across Industries

Mary Long: A key theme from the recent keynote was that artificial intelligence (AI) presents opportunities across various industries, with NVIDIA at the forefront of this transformation. Multiple partnerships have recently been announced, illustrating this point. Let’s review three standout collaborations: GM’s partnership with NVIDIA to create fully self-driving vehicles, a joint venture between Walt Disney, Google’s DeepMind, and NVIDIA to enhance humanoid robot development, and a collaboration involving Yum Brands to implement AI for order taking. Which of these interests you most, Asit?

Asit Sharma: Let’s cover all three quickly; they’re all compelling developments.

AI Ordering in Fast Food: Pros and Cons

Mary Long: Starting with the drive-thru, do you think AI ordering is the way to go? Many fast-food chains, like McDonald’s, have rolled out AI solutions only to backtrack afterward. This raises questions about efficiency at all costs, especially when considering other models like Starbucks, where too much efficiency may deter customers. How do you see the future of Yum Brands and NVIDIA’s partnership shaping up?

Asit Sharma: AI technology continues to improve. While I value human interaction in transactions, I think an AI that occasionally makes mistakes—like misinterpreting an order—could enhance that human-like experience. If it gets my order right most of the time but creates a dialogue, it might improve customer engagement.

GM’s Venture into Self-Driving Technology

Mary Long: Next, let’s discuss the self-driving initiative. GM has faced challenges with its autonomous vehicle technology, including recently halting funding for its Cruise robotaxi service. Were you surprised that GM is renewing efforts towards fully self-driving vehicles?

Asit Sharma: I was surprised initially, but upon reflection, GM has hinted at this direction. We observed their shift in focus towards enhancing software capabilities and providing exceptional driving experiences. They aim to integrate advanced software into their vehicles rather than pursuing a massive fleet of autonomous cars.

Disney’s Robotics Ambitions

Mary Long: Finally, let’s look at the partnership involving Disney, DeepMind, and NVIDIA. How does Disney fit into this collaboration?

Asit Sharma: Disney has a long-standing reputation in robotics, particularly with its animatronics in theme parks. Their research facilities near Burbank focus on improving robotic motion and human-like expression. This expertise allows Walt Disney to merge well with DeepMind’s AI capabilities and NVIDIA’s strength in simulation and virtualization, making it a sensible partnership.

Anticipating Innovations at Quantum Day

Mary Long: Looking ahead, there’s an upcoming event, Quantum Day. What specific developments are you eager to see there?

Asit Sharma: NVIDIA has announced advancements in COTA libraries for quantum applications. Quantum computing needs to address significant challenges, such as error correction during computation. I’m particularly interested in how NVIDIA’s software might enhance accuracy at the quantum level. This could be a pivotal step forward.

Mary Long: Thank you, Asit Sharma, for sharing your insights on NVIDIA and for making these complex topics accessible.

Asit Sharma: It’s been a pleasure, Mary. Thank you!

Listener Questions and Investment Insights

Mary Long: We appreciate your questions. In our next segment, we’ll address listener inquiries related to fundamental analysis, the use of AI in healthcare, portfolio management, and the role that customer service experiences should play in investment strategies. Let’s dive in!

Advertisement: Schwab trading is now enhanced by Ameritrade, providing specialized support for all trading needs, from complex questions to simple strategy checks. Enjoy 24/7 professional assistance through phone, email, and chat. Learn more at schwab.com/trading.

Mary Long: Our listener mailbag reveals a plethora of questions about getting started in investing, and we have analysts ready to provide insights tailored to your interests.

Key Insights for Beginner Investors: Evaluating Stocks Effectively

In response to inquiries from novice investors, listener Cody King asked: “For new investors, what fundamentals should I consider when investing in a Stock? Should I focus on metrics like the PE ratio, recent quarters’ earnings, or current news headlines? What essential information should I look at to evaluate whether a Stock is a worthwhile investment?” To answer these questions, we consulted Fool analyst Asit Sharma.

Asit Sharma: Cody, that’s a fantastic question. When selecting stocks, I focus on two key elements. First, I aim to uncover new ideas, or as I like to say, “turn over stones.” Here, various metrics, including the PE ratio, serve as useful tools for comparing companies. A high PE ratio may indicate a company is experiencing or is expected to experience significant revenue growth. However, let’s take a step back to address your main concern: identifying the critical information for determining if a Stock represents a winning investment.

For me, the most crucial factor is the company’s narrative—the story behind it. Understanding this narrative allows investors to analyze the supporting financial data more effectively. As a starting point, I recommend reviewing the MD&A (Management’s Discussion and Analysis) section found in a company’s quarterly (10-Q) or annual (10-K) reports. These documents, available through the SEC’s EDGAR database, provide essential insights into what the company does, its recent performance, and management’s outlook.

This section offers required disclosures from the SEC, guiding investors to grasp how the management team presents the company’s products or services and their future direction. If you understand the story well, analyzing other factors like earnings results, industry comparisons, and news impacts become much more meaningful. Once you have a solid grasp of the company narrative, these additional elements will help you decide whether to add it to your portfolio or pass on the investment.

Mary Long: We often discuss numerous stocks on the show, some of which are genuine investment opportunities while others are simply noteworthy public companies. Listener Sumit Maru has posed a question about our “radar stocks.” They wrote, “I listen to the show religiously every day. What should everyday investors like me think about the picks highlighted by the team every Friday?” To clarify, we turn to our Friday show host, Dylan Lewis.

Dylan Lewis: Great question! The radar Stock segment is an opportunity for our analysts to discuss market trends or companies they anticipate will be in the news shortly. Often, analysts provide previews of companies set to report earnings soon, or they highlight companies that may not excite them but warrant attention for listeners, especially if they are part of their portfolio. At times, radar stocks are featured for positive reasons—perhaps they foresee promising developments. However, they may also highlight companies facing challenges, serving as a reminder for investors to monitor factors that could impact their investments.

Mary Long: With the buzz surrounding artificial intelligence, Dana from Cincinnati wants to know how healthcare firms are using this technology and if she might discover compelling Stock picks in this arena.

Asit Sharma: Healthcare companies are leveraging AI in numerous innovative ways, particularly in drug design. For instance, the Nobel Prize in Chemistry 2024 recognized the creators of AlphaFold, which predicts protein structures based solely on amino acid sequences. This advancement allows for easier three-dimensional visualization of drug targets—an enhancement over previous methods. Once you reach this stage, AI can assist in creating molecules that fit precisely into those targets. While rational drug design has existed for years, the complexity of these initiatives has significantly increased with AI integration. As research moves into clinical trials, companies targeting previously deemed “undruggable” conditions are drawing attention. Successfully binding to specific targets can revolutionize drug design.

Mary Long: You can gain insights about companies not only from market reports but also from your personal experiences with their services. Listener Yuan Lu raised a question for Fool analyst Jason Moser regarding PayPal’s customer service, noting, “Why has its customer service become aggressively mediocre?” In the past, reaching out for help was straightforward, but recent experiences have shifted to initiating a chat without follow-up support.

Assessing Customer Experience in Stock Investment Decisions

When evaluating stocks, an important question arises regarding the significance of customer experience on buying or selling decisions. How much weight should investors assign to their experiences with a company’s customer service as they navigate the market?

Jason Moser: This is a thought-provoking question that doesn’t have a single answer—it varies for each investor. Personally, I favor PayPal due to its reliability; I rarely contact their customer service, which I consider fortunate. Poor customer service experiences can certainly lead investors to reconsider their ownership of stocks. For example, Comcast has built a reputation for subpar customer service. While I don’t use Comcast myself, the negative feedback makes one question their investment sentiments. Investors need to weigh whether a company can improve aspects of its service against the strengths that may already exist. Remember, no investment is flawless.

In my case, I still hold shares in PayPal, having owned them for years. It continues to lead the digital financial landscape. Are there areas for improvement? Definitely. With new leadership under Alex Chriss, I’m optimistic about the strategies being implemented. Market recognition of these changes seems to be growing. Ultimately, each investor must assess the importance of customer experience in their decision-making—it is a valid factor to consider.

Mary Long: Once you identify the stocks you want to purchase, determining the size of your investment in each is the next step. A listener on X, Jorso, asked whether beginners should buy small shares in many companies or save up for larger purchases in fewer companies. Let’s turn to Asit for clarity.

Asit Sharma: For investors at any level, including beginners and even seasoned ones, starting with smaller stakes in multiple companies is advisable. Though experienced investors may apply this concept differently, the essence remains. Early on, it is critical to take smaller positions to learn about various businesses as they evolve, gradually adding more investment in those that show solid performance. This strategy affords better probabilistic outcomes than concentrating large investments in a few stocks right away, especially without a definitive edge in the market.

Over time, even Warren Buffett—known for his concentrated investments—has diversified through gradual scaling. Of course, his capital can make a more significant impact than what everyday investors can allocate. Nonetheless, diversifying by investing in several companies with smaller amounts allows for learning and adapting in the market. As I see it, taking time to identify winners and adjusting your portfolio accordingly is a prudent strategy.

Mary Long: We appreciate receiving questions from our listeners! If you have a question you’d like us to address, reach out on X or email us at [email protected]. We’ll share more listener insights in tomorrow’s show. See you then, Fools.

As a reminder, participants on this program may have interests in the stocks they discuss, and the Motley Fool may hold formal recommendations. Therefore, don’t base your buy or sell decisions solely on this information. This is Mary Long from the Motley Fool, thanking you for listening. We’ll see you tomorrow.

Suzanne Frey, an executive at Alphabet, is on The Motley Fool’s board of directors. Randi Zuckerberg, a former director at Facebook and sister to Meta Platforms’ CEO, also serves on the board. John Mackey, previous CEO of Whole Foods Market, is yet another board member. Asit Sharma has interests in Amazon, Arista Networks, McDonald’s, Nvidia, PayPal, and Walt Disney. Dylan Lewis has no positions in the stocks mentioned. Jason Moser holds shares in Alphabet, Amazon, PayPal, Starbucks, and Walt Disney. Mary Long does not have positions in the mentioned stocks. The Motley Fool’s holdings include and recommend Alphabet, Amazon, Arista Networks, IBM, Meta Platforms, Nvidia, PayPal, Starbucks, and Walt Disney. The organization recommends shares in Comcast and General Motors, among others. Refer to their disclosure policy for specifics.

The views and opinions expressed herein are of the author and do not necessarily reflect those of Nasdaq, Inc.