Category: Technology

  • Why Nvidia is rising?

    Nvidia’s market cap crossed $3Tn in June 2024 joining the league of Apple and Microsoft. It is 10 times the market cap from December 2020.

    nvidia's market cap have increased ten time since 2020

    What does Nvidia do?

    It is imperative to know what Nvidia does before delving into the reasons why its market cap has risen so quickly in the past few months.

    Fab vs fabless

    On a high level, the semiconductor industry can be divided into two types based on their work – fabs and fabless companies.

    Fabs are the companies that have the manufacturing facilities that actually produce the semiconductor chips. These manufacturing facilities are highly expensive to set up and maintain. With the demand for chips increasing they have to keep up with the latest advancements in manufacturing technology as well. They have to keep working at nearly full capacity always to make them sustainable.

    The fab companies generally depend upon the orders given by fabless companies to manufacture the products. An example of a fab company is Taiwan Semiconductor Manufacturing Company (TSMC). Fabs companies are essentially the backend for the actual products.

    Fabless companies are the ones who design the chips and they depend on the fab companies to manufacture them. They focus on creating intellectual properties and distribution rather than manufacturing it. These companies are the brains behind the chip. They decide the design of the chip.

    The fabless companies outsource the manufacturing to the fab companies and work on contracts with them. Nvidia is a fabless type of semiconductor company.

    Nvidia designs and supplies the graphics processing units (GPUs), that are required for high-end computing. They are used in a variety of devices that are used for varied applications like research, engineering, construction, gaming, and medicine to name a few.

    Reasons

    Rise in demand for Generative AI

    Generative AI has been taken across industries since 2023. More and more companies are being created based on generative AI. Diverse industries from entertainment to manufacturing are adopting generative AI to make their businesses more efficient. 

    Nvidia GPUs are used to train the models by companies like OpenAI. With the increase in the usage of generative AI, the demand for Nvidia is only rising.

    Tech giants Amazon, Google, Meta and Microsoft have all signaled they plan to spend $200bn this year on chips and data centers needed to train and operate their AI systems.

    Demand for data centers and supercomputers

    Tech companies are investing in supercomputers and data centers. The demand for data centers has risen post covid pandemic. Over the last few years, the demand for Nvidia’s most powerful and advanced computer chips necessary for artificial intelligence has boosted sales and profits. Google, Meta, Microsoft, Amazon, and OpenAI buy billions of dollars of Nvidia’s GPUs.

  • The global semiconductor industry scenario

    Semiconductor is, hands down, one of the world’s most crucial products. Right from mobile phones to automotive, semiconductor chips are important. The world would stop without semiconductors. The semiconductor supply chain in the world is a very complex one with the involvement of a lot of geopolitics.

    Huge dependency on Taiwan

    The world is unequally dependent on this island nation of 24 million population. It is the home to Taiwan Semiconductor Manufacturing Company (TSMC), which is the world’s largest chip manufacturer. It manufactures the chips for all the major designers like Apple, Nvidia, Samsung, Intel, and AMD. Though these companies provide the products for all the consumer products, the main manufacturer is TSMC.

    TSMC is the market leader with more than 60% of the market share with them. It is as good as a monopoly given the sheer market size and ever-increasing demand. It is the leader in manufacturing smaller, efficient, and performing chips. The most advanced chips in the world are 5 nm chips. TSMC is heavily investing to produce groundbreaking 3 nm chips, expected to be up to 15% faster and use far less power. They are expected to roll out in the coming years. This way, it is far ahead of the competition.

    Taiwan hosts a robust ecosystem of semiconductor supply chains. It has a strong network of design firms, equipment suppliers, and packaging and testing companies. This makes the local ecosystem of the industry stronger. This integrated supply chain allows for efficient production and innovation in the semiconductor industry.

    By nature, the semiconductor manufacturing industry is very capital-intensive. On top of this, because it is a crucial part of the global supply chain it has always been the case that leading economies of the world have taken the interest in it. The island has always been contested with China claiming it. This adds a layer of complexity. If ever tensions between China and Taiwan, it poses a huge threat to disrupt the supply chain.

    Any challenges in the semiconductor supply chain directly impact various industries like electronics, automotive, and defense, to name a few, and eventually the lives of normal people.

    Exploring other options

    This huge dependency on Taiwan for such a crucial product that affects the whole world makes it very vulnerable. To reduce this dependency nations are making efforts by heavily investing in research & development and establishing manufacturing facilities. This is easier said than done, as the ecosystem in Taiwan is times ahead of it’s competition. 

    India, in March 2024, has led the foundation of three manufacturing facilities. India is aiming to be a 10 trillion dollar economy in the next decade. It is projected that a significant portion of the economy will be contributed by semiconductor manufacturing.

    The uncomfortable relationship with China has made the United States to prioritize on semiconductor chip manufacturing. US companies have pledged 200 billion dollars for chip manufacturing. 

    The European Union has undertaken the initiative to invest $45 billion in research & development and manufacturing facilities. The aim is to get a market share of 20% in semiconductor chip manufacturing by 2030. 

  • Challenges of internet of things (IoT) implementation

    As someone who has worked in the Internet of Things (IoT) industry for more than five years, I have an unpopular opinion. The IoT industry has not grown at the pace at which it was expected to grow. I’ll focus this article completely on Industrial IoT. I feel this because of the number of challenges that businesses are concerned with this type of digital transformation, which are technical as well as financial. Let’s see what challenges businesses face while implementing an Industrial IoT system.

    Challenges of iot implementation faced by industries

    1. Hardware installation and working with legacy systems

    Factories generally have a wide variety of machines with different makes and models. Every machine has its own way of working and connecting with 3rd party devices. Industrial IoT implementation starts with connecting with the existing assets. With such a range of legacy devices, it is a tough task to connect IoT devices and sensors to them. There are multiple cases that arise from this.

    Does the machine support sharing data at all? If it does, what are the ways? If not, can the data be taken from raw circuitry? Different machines supporting data sharing by specified protocols might be locked and need to be paid to share the data.

    The Cherry on the cake, different machines will take different amounts of time and many times unpredictable amounts of time to install sensors and share the data.

    2. Data security

    IoT devices are distributed across different parts of a factory. In many cases, they are distributed across geographies. Different devices generate different types of data. The data has to be processed in real-time to provide quick insights to the clients. This is necessary because if there are any anomalies, actions have to be taken quickly. This data then has to be stored in different locations depending on the architecture. It is distributed across local edge devices and cloud storage.

    Such complicated architecture means a lot of points of failure. With new threats and vulnerabilities rising every day, there is always a threat of cyber attacks on IoT systems. Especially given the fact that we are working on physical and costly machines, and the confidentiality of the data generated.

    3. Connectivity losses

    Let’s say the IoT system has been implemented in a successful way. All the installation and security aspects are taken care of. But what if the connectivity itself of the place is gone? This is something that is not under the control of either the solution provider or the customer. Such external issues do mean a loss to the client in the end. What is the point of implementing an IoT system if one cannot get the analysis and reports they need, right?

    4. High implementation cost

    Let’s get straight to the point. Industrial IoT projects are very costly. Right from the initial surveys for feasibility to installation of the system to the configuration to the handoff, it takes a huge time investment as well. This huge investment will typically take months to years to get returns. The returns may not solely come from the system that is installed. Additional equipment may need to be purchased and installed to do so.

    Hence it is very difficult to predict from where the returns would come for organizations.