The Rise of Giant

BEYOND THE HEADLINES - AI
The Rise of Giant

We’re all familiar with the buzz around AI right now, and companies like OpenAI and Tesla are getting all the attention. It’s understandable as they’re the ones building the products that we actually use - they’re DTC brands (or at least have DTC products).

However, like any industry, there is a backbone of infrastructure that enables these companies to thrive and AI is no different.

The Rise of Nvidia

As we wrote about in our AI deep dive, to build an AI model you need a shed load of data. Depending on what the model output is, dictates what the ‘training data’ input is.

ChatGPT for example requires a huge amount of text to learn from - in fact, they pay Reddit $60mn a year just to be able to use data on the site - and Tesla needs millions of gigabytes of footage of roads, intersections, pedestrians etc. 

This is where Nvidia comes in. They make Graphical Processing Units (GPUs), which are one of the key parts of the puzzle of training a model.

And wow are they in the right place at the right time. Take a look at their stock price performance:

They are now the third most valuable company in the world, overtaking Amazon, Google, Meta and Tesla amongst others. They are now worth $2.2 trillion - more than the GDP of Canada….

Take a look here to see the world’s most valuable companies

A huge portion of the money raised by all these new shiny AI startups goes directly into the pockets of Nvidia. Just look below to see how their quarterly revenue has increased over the last year.

GPUs are the new gold. 

What does this say about the future:

Nvidia right now is a bit like the being only shop selling shovels in California during the gold rush. They’re not the ones running around with all the gold (AI), but they’re currently the ones making the real money. 

Having enough computational power to build and train these models is the limiting factor right now, and it doesn’t look like it’s changing anytime soon. 

A researcher at OpenAI recently tweeted that he could change the models he was training pretty fundamentally, but if the training data never changed the output would stay the same. In other words, the value is where the data is. And the money is where the value is.

What is so great for Nvidia shareholders is that they have an enormous moat. Unlike companies whose competitive advantage is built on software - like Yahoo or MSN…..- Nvidia produces hardware, with extreme precision and technical expertise learnt over decades. Not something easily copied.

A giant has risen and doesn’t show any sign of going back to sleep!

Although when Jensen Huang was asked would he do it all over again this was his response:

BEYOND THE HEADLINES - AI
The Rise of the Robots

What happened?

A company called Figure raised $650 million from organizations like Nvidia, OpenAI, and Microsoft to bring forward an era of robotics.

How did we get here?
When science fiction novelists and scriptwriters of the 20th century wrote about the future, the world saw visions of human-like robots running everything. Films like Star Wars have characters like C-3PO and R2D2 completing tasks for their human companions.

What these fictional stories didn’t depict was the internet, mathematical supercomputers and, AI image and text generation. Maybe that’s because it was less exciting but why has it taken us almost 50 years to build C3PO? What was so hard about creating a future of humanoid robot helpers?

As it turns out, our brains are very complex organs that make very complex tasks look simple. Computers are very good at doing things humans find hard and bad at things humans find easy. Maintaining balance, walking, lifting, and holding are tasks we do without thinking but bring huge engineering challenges for robotics teams.

With all that said, the time of C-3PO is coming. Check out the video:

Who is FigureAI and why does it matter?

Humans still do a lot of mundane tasks that could be automated. There are over 10 million undesirable and unsafe jobs in the US. Manual labor accounts 50% of goods and service prices and a lot of those jobs could be done by AI-powered robots.

The distant future is one where humanoid robots can help care for the elderly, build the first structures on new planets, or just help humans with their day-to-day (C-3PO!). The near term for a company like future is automating warehousing, shipping, and logistics jobs.

This kind of engineering takes years and billions of dollars. The iteration cycles are slower than software and you are limited by the standard and quality of the components. Not only that but Figure has massive competition from Tesla, who are building a humanoid robot called Optimus.

A lot has gone right for a company like Figure AI to be successful.

There is also a broader theme here that we are writing about regularly at the moment.

People and companies are exploring the edges of what’s possible. Some will be right. Some will be wrong. Less than 2 years ago, ChatGPT didn’t exist. Thinking about where the next big technological wave is coming from is an important part of securing successful futures in a world where change is the only guarantee.

CHART OF THE WEEK - EDUCATION
Chart of the week

This graph shows the percentage change in where kids are in school in the United States. Obviously, there was a huge spike in COVID but most interestingly, the change has sustained. See source.