Skip to content
 

Even more on Nvidia. Goldman raises NVDA’s target to $1,000. It’s already edged up above $914. Jensen is on CNBC tonight at 8 PM EST

Goldman raised Nvidia to $1000. It closed today at $914.35.

Here’s the last five days:

Jensen Huang, Nvidia CEO, will be on a show called “Leaders” tonight at 8PM EST on CNBC.

My latest theory on picking stocks. Luck is best. Our Nvidia engineer was lucky to be in the right place at the right time. He’s the one who turned $20,000 into $18 million. Click here.

If If not luck, then love.

Love is when you know enough about the company to be in love with it and its future.

Love comes from extensive and intensive knowledge. Learning. Studying. Reading.

That’s why I keep writing about Nvidia. I keep learning more and more about it.

My take on Nvidia and speed

More than anything else, Nvidia is about speed, and taking advantage of it. So here goes:

NVIDIA’s recent advancements in GPU technologies showcase an enormous leap in computing, networking, and AI:

+ Speed. Think of auto-driving. Your car has many cameras. Output from them needs to be processed very quickly. if not sooner. Otherwise the consequences can be severe. NVIDIA’s latest GPUs ensure that autonomous vehicles can react in microseconds.

+ Parallel Computing. The laptop I’m using to write this blog processes one calculation, then the next. One after another. Today’s software is written to handle that sequential process. NVIDIA’s GPU-based computers process millions of calculations simultaneously. This speeds things up enormously. Tasks you couldn’t do before because you’d crash the car or the robot or the surgery can now be done. Speed is the elixir or the gods.

+ Advanced Networking. Nvidia chips talks to other chips. They talk to memory. And they talk to nearby and distant computers. Today’s Internet speeds are far too slow. NVIDIA’s networking technologies such as NVLink and InfiniBand address these communication bottleneck problems. Nvidia’s networks are millions of bits per second faster than anything you and I commonly use today. This not only enhances performance in high-performance computing (HPC) environments but also supports the demands of large-scale AI applications and data centers, where speedy transfer of data is critical.

+ Data centers. a.k.a. The Cloud. Racks and racks of computers holding zillions of files from zillions of applications of zillions of companies. They need all the above, plus an ultra-fast, ultra-reliable network to get bits and bytes around and between them all. NVIDIA’s technology significantly increases efficiency, reduces power consumption, and supports the enormous computational and networking demands of modern cloud services. By integrating their GPUs into data centers, NVIDIA enables faster processing of vast datasets, supports complex AI and machine learning workloads, and supports quicker data retrieval and analysis. This is terrific for businesses relying on cloud services for real-time analytics, AI services, and scalable computing resources. Companies want the power to analyze my online purchases and suggest my next magic tennis racquet, which will allow me to beat Mark, my partner.

+ Digital Twin Modeling/Simulations. NVIDIA loves the idea of a digital twin – a virtual representation that mirrors a real-world physical asset or system. It’s literally a true-to-reality simulation of physics and materials, capturing the behavior and characteristics of a physical object or process. It’s continuously updated to stay synchronized with its physical counterpart.

Applications? Besides representing inanimate objects and processes, such as machinery, buildings, the weather, or vehicles, they can also model human interactions, behaviors, and processes. In terms of cybersecurity, digital twins can serve as “sandboxes” for simulating cyberattacks on computer networks. And they can be used for simulations involving optimization, testing, and product improvements based on real-time data. Indeed, digital twins enable extensive testing and training in a safe, virtual environment. They prepare autonomous robots for real-world scenarios without risk. Since they interact and can share data bidirectionally with connected sensors and devices, digital twins are closely tied to the Internet of Things (IoT).

By 2028, NVIDIA expects the global market for digital twin platforms to reach $86 billion.

+ AI and Machine Learning. NVIDIA doesn’t just provide the hardware for AI; it also offers comprehensive software frameworks, such as CUDA for parallel computing and the TensorRT open-source library for high-performance deep learning inference. This ecosystem enables rapid development and deployment of AI models across all kinds of industries, from healthcare diagnostics to financial modeling.

At a higher level there’s NVIDIA Inference Microservices, (NIMs), pre-trained AI models that you can download and run on a workstation or cloud or corporate data center, as well as their AI foundry service launched in collaboration with Microsoft Azure. Enterprises and start-ups can use this end-to-end solution to create their own custom generative AI models. These models can power applications such as intelligent search, summarization, and content generation.

Soon chatbots will be intelligent and know what I want, what to sell me and what the solution to my problem is. (Other than rebooting.)

+ Edge Computing. Beyond the points above, it’s important to highlight NVIDIA’s impact on edge computing. By bringing GPU-accelerated computing closer to the source of data, NVIDIA enables faster and more efficient processing at the edge. This is particularly relevant for Internet of Things (IoT) applications, smart cities, and industry 4.0, where immediate data processing is needed for real-time decision-making. Shoe sales are booming in Brazil. We need to get them more inventory. Where the heck are our shoes? Which ships? Which warehouses? Which stores?

+ Environmental Impact and Efficiency. NVIDIA is also big on addressing the environmental impact of high-performance computing. Their latest GPUs are designed to deliver maximum performance per watt (indeed they keep developing chips that do more with less power), thus reducing the carbon footprint of data centers and computing clusters, even as they explode in computing power.

Things Jensen Huang Didn’t Talk About That Much (or at all):

+ NVIDIA Mosaic. If you saw Huang’s presentation, you were probably impressed by the enormous multi-panel screen behind him. NVIDIA Mosaic, a technology for seamless multi-display setups, enables users to span work across multiple monitors (up to 16) as if they were working on a single, large display. This is great for graphic design, financial trading, and control rooms, as well as for immersive gaming experiences. And, of course, for CEOs like Huang wanting to impress people with big, spectacular presentations.

+ NVIDIA Warp. For those programmers not yet out of a job because of AI, NVIDIA Warp is a software development tool that lets Python programmers create GPU-accelerated programs for simulation, robotics, and machine learning. You can write code similar to native CUDA code, but with the friendly Python language –that everyone and their uncle has learned in the last several years. Important features of Warp include kernel-based programming, differentiable programming, and built-in functionality for spatial computing.

If you don’t understand everything above, neither do I. But I’m learning more every day. It’s incredibly exciting. Neither Intel, nor Microsoft, nor Dell, nor Apple, nor Meta (Facebook)  has turned me on like this — ever.

Instance: I love Instagram. I love elephants and lions and whales and penguins and tennis and ping pong. And all the stuff I see on Instagram.

Instagram thrives by selling stuff. Last night I bought a pair of shoes from Amazon on Instagram. It was my easiest purchase ever on Instagram. I’m convinced Amazon uses NVIDIA AI to move me quickly and painlessly through the buying process. It was actually fun.

Susan, my wife of 48 happy years (this Sunday), says I need more shoes like a hole in the head.

But, heck, they were only $29.99 (including free shipping).

What’s not to like about these beautiful shoes?

My take on Israel/Gaza:

+ Hamas wants only one thing – to kill Jews and/or remove all of them from Israel. That’s in its charter.

+ Hamas cares nothing for civilian Palestinians. It has never built one bomb shelter or one hospital. Nor has it provided access to its tunnels for Palestinian civilians to use as bomb shelter.

+ All the money it has received from its supporters — Qatar, Iran and others — it has been spent on weaponry and building tunnels – which are immense. They are longer than the New York City Subway system.

+ Every time the idea of a two-state solution has been proposed, the Palestinians have rejected it. The most “famous” rejection was in 2000 by Yasser Arafat. He wanted all the land for the Palestinians and nothing for Israel. The first formal two-state idea was in  1947. The United Nations General Assembly passed Resolution 181, which proposed partitioning Palestine into two states: one Arab and one Jewish. The resolution also called for Jerusalem to be governed by a special international regime. That started a war — Arabs versus Israel. The Arabs lost.

+ Hamas has been holding 125 or so hostages for over five months. Likely many are dead, but we don’t know the numbers.

+ Hamas knows Israel will never find those hostages in all those underground tunnels. Nor is Hamas motivated to release the hostages since that would sign its own death warrant.

+ Meantime, Israel, under huge pressure from the hostage families, searches for the hostages and kills and injuries many Palestinian civilians. But Hamas doesn’t care and hasn’t agreed to a ceasefire, which would stop the killing.

+ A solution is to stop funding Hamas. But that’s a really, really long-term solution. I bet those Hamas tunnels have enough food, fuel and military supplies for ten years.

+ Many Palestinians have left Palestine. Three big communities are in Jordan, Chile and The United States. Many Arab countries fear Palestinians and won’t let them in, e.g. Egypt. Kuwait expelled 95% of its Palestinians.

Meantime,

“The Biden administration and a small group of Middle East partners are rushing to complete a detailed, comprehensive plan for long-term peace between Israel and Palestinians, including a firm timeline for the establishment of a Palestinian state”.

and

Netanyahu rejects the two state solution. He says: “Israel will continue to oppose the unilateral recognition of a Palestinian state. Such recognition in the wake of the October 7 massacre would give a huge reward to unprecedented terrorism and prevent any future peace settlement”.

I have easy solutions to every one of the world’s problems. Just not this one.

Favorite recent New Yorker cartoon

See you soon. — Harry Newton