Nvidia Part II The Machine Learning Company (20062022)

In season ten, episode six of Acquired, hosts Ben Gilbert and David Rosenthal delve into Nvidia's evolution from a graphics card manufacturer to a leader in the full stack of industries ranging from gaming to enterprise data centers and scientific computing. They explore how Nvidia built an ecosystem around its powerful GPU architecture, which now goes beyond gaming, enabling the creation of digital twins and simulations for various applications. The episode highlights Nvidia's strategic bets, including their CUDA platform, which revolutionized parallel computing and deep learning, leading to their significant role in AI advancements. The discussion also touches on Nvidia's ventures into mobile with Tegra, their data center ambitions post-Mellanox acquisition, and their potential in the automotive sector with their DRIVE platform. Furthermore, they discuss Nvidia's Omniverse, a simulation platform for enterprises. Despite regulatory setbacks like the failed ARM acquisition, Nvidia's journey showcases its adaptability and the massive scale of its data and compute capabilities.

• Ben Gilbert introduces himself as the co-founder and managing director of Pioneer Square Labs and PSL Ventures.
• David Rosenthal introduces himself as an angel investor based in San Francisco.
• Both are the hosts of the podcast Acquired.

"Welcome to season ten, episode six of acquired, the podcast about great technology companies and the stories and playbooks behind them. I'm Ben Gilbert, and I am the co-founder and managing director of Seattle based Pioneer Square Labs and our venture fund, PSL Ventures. And I'm David Rosenthal, and I am angel investor based in San Francisco."

The quote serves as the opening of the podcast episode, introducing the hosts and the theme of the podcast which focuses on technology companies and their strategies.

• Ben shares a childhood memory of wondering about the predictability of fire flickering.
• He relates this curiosity to Nvidia's work in simulating the physical world.
• Nvidia's simulation capabilities extend to various fields like airflow over a wing, cell interactions for drug discovery, and climate change predictions.

"When I was a kid, David, I used to stare into backyard bonfires and wonder if that fire flickering was doing so in a random way or if I knew about every input in the world. All the air. Exactly. The physical construction of the wood, all the variables in the environment, if it was actually predictable, and I don't think I knew the term at the time, but modelable."

Ben's quote reflects his early fascination with the concept of predictability and modeling, which he ties into Nvidia's modern capabilities of simulating complex physical phenomena.

• Nvidia started as a maker of commodity graphics cards.
• They have since built hardware, software for developers, and user-facing software and services.
• The company's technology is now used for digital twins in various industries, not just gaming.
• Nvidia's architecture is crucial for tasks like drug discovery and climate change modeling.
• The scale of data and compute Nvidia handles is compared to the number of grains of sand on Earth.

"And these building blocks, listeners aren't just for gamers anymore. They are making it possible to recreate the real world in a digital twin, to do things like predict airflow over a wing or simulate cell interaction, to quickly discover new drugs without ever once touching a petri dish, or even model and predict how climate change will play out precisely."

This quote outlines Nvidia's expansion from gaming to broader applications, emphasizing the impact of their technology on various industries and the massive scale of their computational capabilities.

• Pilot is a company that offers accounting, tax, and bookkeeping services for startups and growth companies.
• It is the largest startup-focused accounting firm in the US.
• Pilot is backed by notable investors like Sequoia, Index, Stripe, and Jeff Bezos.
• The company aligns with Jeff Bezos's philosophy that startups should focus on their core product and outsource non-core activities like accounting.

"Our next sponsor for this episode is one of our favorite companies and longtime acquired partner, Pilot for startups and growth companies of all kinds. Pilot is the one team for all of your company's accounting, tax and bookkeeping needs, and in fact, now is the largest startup focused accounting firm in the US."

The quote introduces Pilot as a sponsor and highlights its focus on providing financial services to startups, allowing them to concentrate on their core business.

• Nvidia faced competition from ATI, which was acquired by AMD.
• Nvidia's focus on Cuda and other ventures led to an earnings miss and an 80% drop in stock value.
• Despite the challenges, Nvidia committed to building Cuda, a full development framework for GPU computation.
• Cuda is free to use but proprietary and exclusive to Nvidia hardware.

"But not Jensen. So instead he goes and builds Cuda and continues to build Cuda. And this is just set context. We get excited about a lot of stuff unacquired, but I think Cuda is like one of the greatest business stories of the last ten years. 20 years more?"

This quote emphasizes Jensen Huang's commitment to developing Cuda despite Nvidia's stock challenges and recognizes Cuda as a significant business story for its impact on the industry.

• Nvidia's initial investment in Cuda and general-purpose computing on GPUs was questioned due to scale and market concerns.
• Skepticism surrounded the potential market that would allow Nvidia to achieve the necessary scale.
• Don Valentine and other stakeholders were uncertain about the market's ability to support Nvidia's strategy.

"There was very little to believe that Nvidia was going to have the scale required to justify this investment or that there was a market to let them achieve the scale to justify this." "Where's the market that's going to enable the scale you need to run this playbook."

The quotes express doubts about Nvidia's ability to find a market large enough to justify their investment in Cuda and general-purpose GPU computing. The relevance is the skepticism about Nvidia's strategic direction and the potential lack of a market to support it.

• Nvidia faced challenges and "misadventures" in the late 2000s, with their stock getting hammered.
• In 2008, Nvidia launched the Tegra chip, a system on a chip for smartphones, which was a departure from their core competencies.
• Tegra powered several devices, including the Microsoft Zune HD, Tesla Model S touchscreen, and Nintendo Switch.
• Nvidia's Tegra efforts did not align with their core strengths and did not save the company.

"This is not what saved the company. This is more clown car style." "It was a full on system on a chip for smartphones, competing directly with Qualcomm with Samsung."

The quotes describe Nvidia's foray into mobile processing with the Tegra chip, highlighting it as a significant strategic departure that did not lead to the company's salvation. The relevance lies in showing Nvidia's attempt to diversify and the challenges faced in competing with established mobile chip manufacturers.

• GPU companies, including Nvidia, struggled with the transition to mobile.
• ATI Radeon's mobile GPU efforts were sold to Qualcomm, becoming Qualcomm's mobile GPU division.
• Nvidia's mobile graphics efforts were not as successful, and the Android ecosystem was seen as having low profitability for partners.

"These GPU companies have not been good at transitioning to mobile." "My impression of the whole Android value chain ecosystem is that there's no profits to be made anywhere."

The quotes reflect the difficulties GPU companies faced in transitioning to the mobile market and the low-profit margins within the Android ecosystem. The relevance is in the contrast between Nvidia's struggles in mobile and their eventual success in other areas.

• Nvidia purchased Icera, a mobile baseband company, in 2011, aiming to strengthen their mobile efforts.
• The acquisition did not lead to long-term success, and Nvidia eventually shut down the Icera-related operations.
• Icera's founders later established Graphcore, which became a potential Nvidia competitor.

"They shut down what they bought from Icera. They lay everyone off." "The Icera founders... go off and they found a company called Graphcore."

The quotes discuss Nvidia's acquisition and subsequent closure of Icera's operations and the founding of Graphcore by Icera's ex-founders. The relevance is in showing Nvidia's continued but ultimately unsuccessful attempts to penetrate the mobile market and the emergence of new competition.

• Nvidia experienced years of sideways growth, with occasional financial disappointments.
• A "miracle" occurred with the advent of deep learning, which became a pivotal moment for Nvidia.
• The breakthrough with AlexNet and deep learning algorithms on GPUs marked a turning point for the company's technology and market position.

"Basically a miracle happens." "Jensen did not plan Alex. Net or see it coming because nobody saw Alex. Net coming."

These quotes highlight the unexpected and transformative impact of deep learning on Nvidia's fortunes. The "miracle" refers to the unforeseen success of deep learning algorithms running on Nvidia's GPUs, which Jensen Huang, Nvidia's CEO, did not predict. The relevance is in the serendipitous alignment of Nvidia's GPU capabilities with the computational demands of deep learning.

• Nvidia's GPUs, combined with Cuda, became instrumental in the rise of deep learning and AI.
• The company's technology enabled significant advancements in image recognition, autonomous driving, and content prediction.
• Nvidia's data center segment, which includes AI and ML applications, saw substantial growth and profitability.

"This was the big bang moment for artificial intelligence. And Nvidia and CudA were right there." "The value accrues to those who help you navigate the content... someone has to make the chips and the software so that they can do that effectively."

The quotes underscore Nvidia's critical role in the explosion of AI and deep learning, facilitated by their GPU technology and Cuda platform. The relevance is in the acknowledgment of Nvidia's contribution to enabling AI applications that sort and navigate content, which has become a significant value driver in the digital economy.

• Despite the breakthroughs in AI, Nvidia's stock did not immediately reflect the company's potential.
• Skepticism persisted even as Nvidia's technology was proving crucial for AI and deep learning.
• It took years for the market to fully recognize Nvidia's value, leading to significant stock appreciation.

"The market did not realize this for years." "People were still skeptical... everyone's not rushing to buy the stock."

The quotes reflect the lag between Nvidia's technological achievements and market recognition, as well as the continued skepticism from investors and analysts. The relevance is in the delayed correlation between the company's strategic moves and their financial valuation.

• Cryptocurrency mining became a new use case for Nvidia's GPUs due to their parallel processing capabilities.
• The crypto boom led to increased demand for Nvidia's GPUs, but the subsequent crypto winter caused a sharp decline in sales.
• Nvidia's revenue was affected by the volatile demand from crypto miners, demonstrating the challenges of market unpredictability.

"A lot of people were going out and buying consumer Nvidia graphics cards and using them to set up crypto mining rigs." "The mining rig demand fell off... their revenue actually declined."

These quotes explain how cryptocurrency mining, an application reliant on parallel processing, created a surge in demand for Nvidia's GPUs. The subsequent decline in mining profitability led to a decrease in revenue for Nvidia. The relevance is in illustrating the impact of external market forces on Nvidia's business performance.

• Nvidia continues to utilize its own chips for training purposes.
• An estimated expenditure of $50 to $100 million was made by a single customer for Nvidia's compute cluster.
• This significant investment by one customer highlights Nvidia's strong market presence and customer reliance on their technology.

"I think they spent between 50 and $100 million that they paid Nvidia for all of the compute in that cluster." "One customer."

The quote indicates the scale of investment by a single client on Nvidia's compute technology, emphasizing the company's significant role in the industry and the high costs associated with advanced computing resources.

• Nvidia's data center segment has experienced substantial growth, tripling over two years.
• The segment's revenue has reached over $10.5 billion annually, now comparable to Nvidia's gaming segment.
• This growth reflects the widespread adoption of machine learning hardware in various industries, including cloud services and private enterprises.

"So then insanely, over the last two years, it three xed the data center segment. Three XED and is now doing over ten and a half billion a year in revenue."

The quote highlights the explosive growth of Nvidia's data center business, which has expanded to match the revenue of the traditionally dominant gaming segment, indicating a significant shift in Nvidia's market focus and success.

• Nvidia altered the user agreements for their GeForce cards to prevent their use in data centers.
• This strategic move aimed to segment the market, capitalizing on enterprises' willingness to pay for specialized data center GPUs.
• Nvidia's enterprise GPUs, such as the A100, lack video outputs, making them unsuitable for gaming but ideal for data center use.

"They're like, we really do need to start segmenting a little bit here."

The quote reflects Nvidia's strategic decision to differentiate their product offerings, creating a clear distinction between consumer gaming cards and high-performance enterprise GPUs for data centers.

• Nvidia acquired the Israeli data center compute company Melanox for approximately $7 billion.
• Melanox enhances Nvidia's capabilities in data center connectivity with high bandwidth and low latency.
• The acquisition aligns with Nvidia's ambition to dominate data center hardware communication and improve overall data center efficiency.

"In 2020, they acquire an israeli data center compute company called Melanox that I believe focuses on, like, networking compute within the data center for about 7 billion."

The quote discusses Nvidia's strategic acquisition of Melanox, which plays a critical role in strengthening Nvidia's position in the data center market by providing advanced networking solutions.

• Nvidia's vision includes a three-pronged computing approach: CPU, GPU, and DPU (Data Processing Unit).
• The DPU, influenced by Melanox technology, facilitates efficient data communication and transformation within data centers.
• This approach enables high-level abstraction in data center management, enhancing Nvidia's computing solutions' efficiency and appeal.

"Now what they're saying is you've got your cpu, you've got your gpu. Now there's a DPU."

The quote captures Nvidia's expanded vision for data center computing, where the introduction of the DPU complements the traditional CPU and GPU, creating a more robust and efficient computing environment.

• Nvidia's market valuation requires justification for future growth prospects.
• The company's growth rate and profitability are highlighted as key factors in its high valuation compared to other tech giants.
• Nvidia's strategy includes targeting a trillion-dollar market opportunity to maintain investor confidence and justify its market cap.

"They need to sort of justify that unless they are willing to have it go down. And so they need to come up with a story about how they're going after this ginormous opportunity, which maybe they are, but it leads to things like an investor day presentation of let us tell you about our trillion dollar opportunity ahead."

The quote reflects the pressure on Nvidia to present a compelling growth narrative to justify its high market valuation, suggesting ambitious market targets as part of its strategy.

• Nvidia plans to license its software independently from its hardware, catering to enterprise needs.
• This strategy allows Nvidia to capitalize on the software's industry-defining capabilities and the demand for enterprise-grade solutions.
• Licensing enables Nvidia to generate additional revenue streams without cannibalizing hardware sales.

"They've been asking for it and the reason for that is because they can't just go to open source and download all the stuff and make it work for their enterprise."

The quote indicates Nvidia's response to enterprise demand for reliable and tailored software solutions, which cannot be met by open source alternatives, thus opening up new business opportunities for Nvidia through software licensing.

• Nvidia's automotive segment has historically been small and stagnant.
• The company is pivoting towards electric vehicles (EVs) and autonomous driving as a potential growth area.
• Nvidia aims to provide comprehensive hardware and software solutions for automakers, positioning itself as a key player in the future of automotive technology.

"What Nvidia used to do for automotive is what everyone used to do for automotive, which is make fairly commodity components that automakers buy and then put in there."

The quote illustrates Nvidia's shift from providing generic automotive components to offering complete solutions for next-generation vehicles, indicating a strategic move towards a more integrated role in the automotive industry.

• The Nvidia Omniverse is not to be confused with a metaverse; it is a platform for 3D simulation and digital twin technology.
• This technology can create highly detailed and functional simulations of real-world environments, such as a digital twin of Earth.
• The Omniverse represents Nvidia's investment in advanced simulation capabilities, which has applications across various industries.

"So a good example of it is this Earth two, this digital twin of earth that they're creating."

The quote introduces the concept of Nvidia's Omniverse as a sophisticated simulation tool capable of creating detailed digital replicas of physical entities, showcasing Nvidia's innovative edge in the simulation space.

• NVIDIA's sophisticated climate models serve as proof of concept for enterprises.
• The company pitches to enterprises the idea of using simulations in the Omniverse for planning and testing before real-world deployment.
• NVIDIA's Omniverse is positioned as a powerful enterprise solution for modeling changes to real-world assets.
• The Omniverse allows for the training of models and running simulations in an efficient and user-friendly way, potentially reducing the need for real-world testing.

"When you want to make a tweak to a model, you're not just going to deploy those to all the robots. You kind of want to run that in the omniverse first, and then when it's working, then you want to deploy it in the real world."

This quote emphasizes the practicality of testing changes in a simulated environment before implementing them in reality, which could lead to more efficient and less risky updates to systems.

• The enterprise metaverse is designed for simulating applications, not primarily for human interaction.
• The majority of the enterprise metaverse may operate without human presence, although there will be user interfaces for human interaction when necessary.

"This is what an enterprise metaverse is going to be. This is not designed for humans. Humans may interact with this. There will be UI, you'll be able to be part of it."

The quote clarifies that the enterprise metaverse is mainly focused on simulation rather than human-centric experiences, although there will be provisions for human interaction.

• NVIDIA's bull case is presented as capturing a significant market share in a future trillion-dollar industry.
• The bear case involves the potential for competitors, such as AMD, to challenge NVIDIA's dominance in high-end gaming graphics and data center markets.
• Another bear case is the possibility of large customers, like Tesla and Google, developing their own silicon solutions, reducing reliance on NVIDIA.

"The bare case is that there's a TikTok rather than a durable competitive advantage for Nvidia."

This quote suggests that NVIDIA's current success might not be sustainable if competitors manage to create similar or superior products, thus eroding NVIDIA's competitive edge.

• NVIDIA faces competition from companies like AMD in gaming and potentially new entrants in enterprise AI data center markets.
• Large customers might consider designing their own hardware to better suit their specific needs, posing a threat to NVIDIA's business with them.
• Companies like Cerebrus are taking a different approach with highly specialized hardware, which could disrupt the market if successful.

"Is there someone else who is coming along with a completely different approach to accelerated computing, whether accelerating workloads off the GPU onto something new, like a Cerebrus or like a graph core that is going to eat their lunch in the enterprise AI data center market?"

This quote highlights the uncertainty in the market regarding whether new approaches to computing will outpace NVIDIA's current offerings, potentially leading to a shift in market leadership.

• Google is creating its own silicon for data centers and differentiating its cloud services with TPUs.
• Google's approach to using TPUs exclusively within Google Cloud could be a strategic move to gain market share in the cloud industry.
• NVIDIA also offers GPUs in various cloud services and is starting its own cloud, which adds another layer to the competitive landscape.

"Google is sort of counterpositioned against Nvidia here, where they're saying, we want to differentiate Google Cloud with this offering."

The quote indicates that Google is positioning itself as a competitor to NVIDIA by offering unique hardware solutions within its cloud services, aiming to attract customers with potentially more cost-effective options.

• NVIDIA's valuation is considered rich, with its future growth prospects being a key factor in justifying its current market price.
• The debate centers on whether NVIDIA will continue to grow at its current rate or if recent growth is a temporary surge due to external factors like the pandemic or cryptocurrency trends.

"It depends if you think their growth will continue. Are they a 60% growing company year over year over year for a while? Then they're not richly valued."

This quote captures the crux of the investment debate around NVIDIA, where the justification for its high valuation hinges on the company's ability to maintain its rapid growth trajectory.