visitor activity monitoringweb statistics

NVIDIA Is Skynet

3 Reasons Why NVIDIA Stock (NVDA) Is Undervalued.

3 Reasons Why NVIDIA Stock (NVDA) Is Undervalued

Analysts are ill equipped to value the company and yet it underpins some of the fastest growing industry verticals and sits on more than one cash cow. Here are three reasons why I believe NVIDIA is well positioned to keep growing into a technology juggernaut the likes of which we’ve never seen before.

Firstly, I am not an analyst, nor a paid evangelist of NVIDIA Corporation. I’m just a guy that spends too much time tinkering with new technologies and new media, specifically what happens when you put them together. My 360 digital agency, TWOMC, produces boutique video and has been doing a lot of VR / 360 work recently. When we invested in new PCs for this work with GeForce GTX 980 Ti GPUs I was re-introduced to CUDA rendering. If you’ve worked with video at all, especially VR or 4K and up, you know the pain of rendering as it ties up your CPUs. I’ve seen renders take days. CUDA, invented by NVIDIA, allows you to use muscular GPU(s) to do most of that heavy lifting. It can mean the difference between a 4-minute and 8-hour render.

Roughly, this one innovation, CUDA rendering, has saved our inhouse video team 30 hours of downtime a week – per PC. That got my attention. I’d heard of their GPU business and their domination of the hardware layer of the multi-billion dollar gaming industry. What else was NVIDIA working on? It turns out they are into a lot more.


1. Stock analysts do not have the tools or perspective to properly value this technology.

Take a look at NVIDIA’s Q1F19 earnings. Forget the record revenue of $3.21 billion, up 66% from last year. Forget the data center revenue of $701 million, 71% higher than the previous year. Remember that HPC (high performance computing) is shortly going to be a $10 billion dollar market because of something called server side AI. While many voices are debating how to define AI and whether we should proceed with AI, it’s quickly becoming a prerequisite for modern cloud data centers, among a host of other applications.

Server providers like Lenovo, Supermicro, as well as ODMs like Foxconn and Quanta are already building configurations based on NVIDIA’s new HGX-2 platform for some of the world’s largest cloud data centers. The HGX-2 platform is special for a long list of reasons, not the least of which is the speed it can train AI. More on that a bit further in. The point here is that understanding NVIDIA’s long list of revolutionary breakthroughs like the TensorRT AI inference accelerator software, NVIDIA NVSwitch GPU interconnect fabric and GPU acceleration for Kubernetes for enterprise, requires first deeply understanding these things, understanding why they’re revolutionary – and then being able to conceptualise what the market potential is for them. It’s like explaining the advantages of CUDA rendering to non-video folks: If you’re not already seeing direct benefits in your own backyard from these innovations it’s forgivable to be nonplussed about them. But if you start to think about what kind of competitive advantage they afford a technology company – especially the one producing them – you start realising that traditional market indicators might not apply in this case.

The current P/E ratio of 42.25 reflects this outlook as most investors focus on quarterly growth in the gaming industry and believe the value is already baked into such a relatively high share price ($256.95 as of August 7th, 2018). This is what many analysts claim. Remember these are the same people that love to short Apple after each iPhone release. They’re frequently wrong on that front and Apple retails consumer technology, which is a much more familiar beast to market “experts”. Much like the fictional Skynet, NVIDIA has fingers in many, many more technology pies than Apple.


2. It’s not always apparent but NVIDIA’s hardware and software quietly underpins and enables some of today’s most breathtaking technologies: self driving cars, virtual reality, high end gaming, cryptocurrencies and we’ve already touched on AI.

Last week Elon Musk gave NVIDIA the axe saying Tesla’s home grown replacement GPU solution was 10 times faster than NVIDIA’s at the same cost. However, NVIDIA announced its own next-gen chip for autonomous driving called Pegasus, back in May, which is ten times faster. Losing Tesla might seem like a blow until you realise they still supply Toyota, Mercedes-Benz, Audi, and VW. Wall Street shrugged off Musk’s announcement and the stock has gone up steadily since. They’ve had time to process this so NVIDIA’s earnings report August 16th should be interesting. Expect growth in everything except GPUs for cryptocurrency mining, which had spiked due to speculatory crypto investment – which is tapering off now. Their lead in the gaming hardware business provides NVIDIA the perfect perch to cash in on virtual reality and the human VR training boom set to take off. But the bigger deal for NVIDIA is AI for robots.

Given the moral debate on AI and confusion around what it really is, most people can be forgiven for believing that real AI is years away. But deep learning is already here. You don’t have to be a total nerd to appreciate what is being made possible by NVIDIA’s deep learning work: remember those funny / scary DARPA videos of robots navigating obstacles in the lab? An artificially intelligent robot is a lot like a human child, it needs to be taught all about the world and painstakingly trained how to deal with its environment. Traditionally, this has meant thousands of hours of designing, building and trialing both the robot’s physical form as well as practicing real-time navigation skills on physical obstacle courses – a lot of them. Now imagine how that can be accelerated when you combine photo-realistic VR environments, real-world physics engines and AI deep learning. It’s like VR training for AI robots – but in fast forward. New robot AI can train thousands of hours in a few seconds across multiple obstacle courses without building a single physical course. Boot camp happens in a few hours instead of a few years. Even more impressive, the monitoring AI can quickly surmise which obstacle courses are the most efficient and optimise the process further. If NVIDIA is producing AI-enabled advancements like this for other companies, imagine how it is applying it to their own business processes.


3. Strong, effective leadership.

This is probably the most important factor. While some hardware OEM giants have legions of executives NVIDIA is still run by a founder, CEO Jensen Huang, and a relatively small team of officers. Jensen is a bit of a rock star, having been ranked No. 1 on the Harvard Business Review’s 2017 list of the world’s 100 best-performing CEOs, Fortune’s 2017 CEO of the year, among many other business accolades. And he’s still young with many years left in him. These awards are mostly based on profitability and at around 50% annual profit growth Jensen deserves them. But to me the award most indicative of Jensen’s leadership prowess was Glassdoor’s 2017 survey of the highest employee-rated CEOs, where he had the best showing in Silicon Valley, a 99% approval rating from anonymous NVIDIA employees (No. 6 on the list). He’s winning without “leaving bodies in his wake”.

You would have to look really hard for a leader more dedicated than Huang. How many CEOs have a tattoo of the company logo on their arm? Maybe even more indicative of his calibre of leadership is how he communicates with customers about failures. When the company incorrectly stated specs in their GeForce GTX 970 GPU back in 2015, Huang wrote an open letter apologising and explaining the controversy that followed. It was clearly a communications issue but the CEO stepped up to take the hit. When was the last time Tim Cook apologized publicly for an Apple product fail?

To be clear, NVIDIA is not a neural net-based conscious group mind and artificial general intelligence like Skynet from the film, The Terminator. It is, however, uniquely positioned to continue pulling away from the competition by leveraging the very AI technology it produces. Hopefully, what it does with this very unique competitive advantage is a net positive for humanity, and more interesting to watch than the film.

See original article: Linked

Other Blogs

VR / AR Training Is Here

This Is What Digital Disruption For The Learning

When Worlds Collide

It’s time to redefine the term Digital Convergence.

Thoughts On Thingking Fast And Slow (Part III)

How do you sum up a masterwork like Thinking Fast and Slow?

Leave a Comment

Secured By miniOrange