NVIDIA, a US company headquartered in Santa Clara, California, and TSMC, a microchip manufacturer based in Taiwan, are undisputed market leaders in the dynamic field of advanced artificial intelligence (AI), albeit in two different (but related) sectors of the business.

NVIDIA designs powerful graphics processing units (GPUs) that accelerate AI computations and make tasks like deep learning and data analysis much faster. 

TSMC manufactures sophisticated microchips—including chips for NVIDIA—using state-of-the art processes.

Their collaboration facilitates rapid advancements in AI, which in turn makes progress in areas like autonomous vehicles, data centers, and scientific research possible.

From a geopolitical standpoint, it’s quite significant that one is an American company, and the other is based in Taiwan. We’ll get to why that is so. But first, let’s take a look at what makes NVIDIA  a top leader in its industry niche. This requires some background knowledge of the technologies the companies work with—so let’s dive right in. 

What makes NVIDIA special?

What sets NVIDIA apart from its competitors in advanced AI is the company’s relentless focus on innovation and specialization. NVIDIA’s GPUs are renowned for their exceptional parallel processing capabilities, which are essential for handling the complex computations required in AI and machine learning. 

What is a GPU?

A GPU is a type of microchip designed to perform many calculations simultaneously, known as parallel processing. GPUs can handle thousands of tasks simultaneously. They are specifically designed for tasks that require heavy mathematical computations, such as rendering graphics and performing complex calculations for AI and machine learning.

Origins of NVIDIA

NVIDIA was founded in 1993 by Jensen Huang, Chris Malachowsky, and Curtis Priem. The company’s stated mission at the time was to revolutionize the world of computing graphics. To that end, NVIDIA initially focused on developing GPUs for gaming and professional markets. 

However, the company has since expanded its focus to include highly advanced computing technologies, including artificial intelligence and data science.

What is data science?

Data science involves extracting meaningful insights from vast amounts of data, which is critical for advancing AI and machine learning technologies. Here’s how data science fits into NVIDIA’s strategy and offerings:

• NVIDIA’s GPUs are designed to handle massive parallel-processing tasks, making them ideal for data science applications. They accelerate complex computations and enable faster data analysis and machine learning model training.
  
• The company’s robust ecosystem of AI software and development tools, such as CUDA and RAPIDS, further increases the appeal of NVIDIA’s products to researchers and developers. 
  
• NVIDIA’s CUDA (Compute Unified Device Architecture) is a technology that allows programmers to use NVIDIA graphics processing units for general-purpose computing tasks, such as scientific simulations, machine learning, and data processing.
  
• RAPIDS is an open-source suite of software libraries and APIs that enables data scientists to execute end-to-end data science and analytics pipelines entirely on GPUs, which significantly speeds up workflows. (An API, short for “application programming interface,” is a set of rules and protocols that allows different software applications to communicate with each other.) 

In summary, NVIDIA’s GPUs are the backbone of many AI research and data center operations. The company’s dominance in this space is largely due to the outstanding performance of its GPUs in deep learning and AI workloads. 

NVIDIA GPU pricing and profit margins

The prices for NVIDIA’s professional and data center GPUs, such as the Quadro series and the A100 Tensor Core GPUs, often exceed several thousand dollars per unit, depending on the configuration and capabilities​.

NVIDIA’s profit margins on their GPUs are impressive. The company’s net profit margins on their GPUs have been reported to be 50% and higher. These margins reflect their strong market position and the popular value-added features of their GPUs.​

[Image source]

How big is the AI microchip market?

The answer is, very big and growing by leaps and bounds.

Driven by the increasing demand for AI capabilities across a wide range of sectors such as data centers, automotive, and consumer electronics, the global revenue from AI chips is expected to top $70 billion in 2024, which is a massive 33% increase over 2023 revenues​​.

NVIDIA’s data science solutions

Here’s a sampling of the markets NVIDIA targets with its data science solutions:

Academia and research

• AI research: Universities and research institutions use NVIDIA GPUs to advance AI research, which drive breakthroughs in areas like natural language processing, computer vision, and robotics.
• Educational tools: NVIDIA provides tools and resources for academic institutions to teach data science and AI.

Enterprise and industry

• Data centers: NVIDIA’s data center GPUs, such as the A100 Tensor Core GPU, are optimized for AI and data analytics workloads. A variety of enterprises use these for high-performance computing and big data analytics.

 

• Healthcare: NVIDIA Clara is an AI-powered healthcare platform that is used to improve medical imaging, genomics, and drug discovery.

Autonomous machines

• Self-driving cars: The NVIDIA DRIVE platform uses data science and AI to develop and train autonomous vehicle systems. The platform processes and learns from vast amounts of driving data to improve vehicle safety and autonomy.
• Robotics: Data science enables robots to interpret sensor data, learn from their environment, and make intelligent decisions.

How data science fits into NVIDIA’s business strategy

Data science is a fundamental component of NVIDIA’s strategy to push the boundaries of AI and machine learning. By providing GPU-empowered tools and platforms, NVIDIA can be more than just a supplier of products to industries.

Instead, it is becoming a partner of an impressive range of organizations. The intent of these partnerships is to mutually harness the power of data for innovation and efficiency. 

NVIDIA’s newest toy: Blackwell architecture 

NVIDIA’s Blackwell architecture is poised to strengthen the company’s dominance in the AI field. Scheduled to be released in stages beginning in mid-2024, Blackwell architecture features dramatic performance improvements over the company’s already powerful suite of products. 

What does the term “architecture” mean?  

In the context of computer hardware, the term “architecture” refers to the design of a system’s components and how they are organized to interact with each other. 

Calling GPU designs like NVIDIA’s Blackwell an “architecture” emphasizes the systematic approach that was taken in the design of the GPU. That design process took into account the integration, functionality, and performance of all of the system’s components. 

The term underscores the complex strategic planning involved in creating advanced hardware that meets the needs of modern computing tasks.

Manufacturing Blackwell’s mega-powerful GPUs

NVIDIA’s Blackwell architecture GPUs contain 208 billion tiny switches called transistors. These GPUs are made using a state-of-the-art process by Taiwan Semiconductor Manufacturing Company (TSMC), the world’s leader in advanced microchip manufacturing. 

TSMC uses a specialized 4NP process to manufacture Blackwell GPUs. This involves crafting intricate, high-tech circuits at a microscopic level.

Advanced lithography machines developed by the Dutch company ASML are used in this process. These lithography systems use extreme ultraviolet (EUV) light to etch incredibly fine details onto the silicon wafers. The microscopic components needed for the Blackwell GPUs are created with this astounding technology.

Blackwell: designed to solve the most challenging AI problems

Each Blackwell product consists of two massively powerful microchips that are connected to each other. This connectivity allows the chips to communicate at an extremely fast speed of 10 terabytes per second. This design creates a single, unified superchip that significantly improves AI performance.

This architecture is specifically built to solve AI-related problems and the handling of AI tasks that were previously too difficult or time-consuming.

NVIDIA’s Blackwell architecture incorporates advanced features such as multi-chip module designs, enhanced tensor cores, and improved ray tracing capabilities. Blackwell GPUs deliver up to 30 times faster AI inference performance and significantly higher energy efficiency compared to its predecessors. 

These advancements allow Blackwell GPUs to handle the most demanding AI workloads with ease.

NVIDIA as an American company

NVIDIA being an American company is important from a cybersecurity perspective and for broader geopolitical, economic, and technological reasons. 

Here are some key points relevant to NVIDIA’s status as a US corporation:

Data sovereignty

• Protection of sensitive data: As an American company, NVIDIA is subject to US laws and regulations regarding data protection and cybersecurity. Because NVIDIA is a US corporation, sensitive data processed by NVIDIA’s technology is governed by a strong legal framework designed to protect against unauthorized access and breaches.
  
• Compliance with US regulations: NVIDIA adheres to stringent US cybersecurity standards and regulations, such as the Cybersecurity Information Sharing Act (CISA)  and the National Institute of Standards and Technology (NIST) guidelines. The goal of this compliance is to create a high level of security for critical infrastructures. 

Though a discussion of CISA is beyond the scope of this article, it is worth noting here that its supporters believe it strengthens US cybersecurity through improved information sharing and collaboration, while its critics allege that the law opens doors to privacy infringements, data protection issues, and the potential for government abuse.

Supply chain security

• Trusted supply chain: As a US-based company, NVIDIA’s supply chain can be more easily monitored and controlled by US authorities. This reduces the risk of infiltration by malicious actors.
  
• Preventing espionage: Having a critical tech company like NVIDIA within US jurisdiction helps mitigate risks of industrial espionage and cyber threats from foreign adversaries, protecting national security interests.

Collaboration with government agencies

• Public-private partnerships: NVIDIA can collaborate closely with US government agencies, such as the Departments of Defense and Homeland Security, on cybersecurity initiatives and defense projects.
  
• Access to classified information: As an American company, NVIDIA enjoys a certain level of trust with classified information. The company participates in confidential projects that are critical to national security.

Geopolitical and economic reasons

• Technological leadership: Having NVIDIA in the US helps the country maintain its technological leadership in advanced computing and AI, which are critical areas for economic and military superiority.
  
• Innovation hub: The presence of a leading company like NVIDIA in the US drives innovation and attracts top talent, which in turn spurs advancements in AI, machine learning, and high-performance computing.

Economic impact 

• Job creation: NVIDIA contributes to the US economy by creating high-skilled jobs in engineering, research, and manufacturing. This boosts local economies and supports the tech industry ecosystem.
  
• Economic security: A strong domestic tech industry, anchored by companies like NVIDIA, improves economic security and reduces dependency on foreign technologies, which can be subject to geopolitical tensions and trade restrictions.

Strategic autonomy

• Independence from foreign influence: By having key tech companies like NVIDIA based in the US, the country can maintain strategic autonomy and reduce reliance on foreign entities for critical technologies. This not only improves US national security but also contributes to global security by providing a stable and secure supply of critical technologies while reducing the risk of geopolitical disruptions and safeguarding against cyberattacks by rogue governments and other bad actors.
  
• Controlling critical infrastructure: Because NVIDIA is an American company, it must work with the US government to make sure the critical infrastructure technologies it develops support American national interests and are not subject to foreign control or influence.

Intellectual property (IP) protection

• IP laws: The US has strong intellectual property laws that protect the innovations and technologies developed by NVIDIA. This protective legal climate encourages continuous investment in R&D and helps safeguard the company’s competitive edge.
  
• Patent enforcement: Being a US company allows NVIDIA to more effectively enforce its patents and intellectual property rights.

Global Influence

• Setting standards: As an American company, NVIDIA plays a leading role in setting global technology standards and influencing international policies related to AI and high-performance computing.
  
• Strategic alliances: NVIDIA’s position as a US company allows it to form strategic alliances with other American and allied companies, such as Taiwan-based TSMC. This creates a collaborative ecosystem that can collectively address global challenges in technology and cybersecurity.

In summary, NVIDIA’s status as an American company is essential for cybersecurity, economic security, technological leadership, and maintaining strategic autonomy.

It guarantees adherence to stringent cybersecurity standards, protects sensitive data, supports national interests, and contributes significantly to the US economy and global technological influence.