How Nvidia Got Lucky

Brit Ryle

Posted August 23, 2023

When I started in this biz 25 years ago, internet stocks were starting to heat up. There was one category of stocks called “internet incubators” that were pretty popular. The incubators were basically venture funds that bought shares of internet companies, both private and public. Owning one of these “incubators” was like owning some swath of the burgeoning internet economy.

There were probably 4 of 5 popular incubators, but I can only remember the names of two: Internet Capital Group and Softbank, from Japan. Internet Capital Group was actually the most popular one, because it focused on B2B companies – business to business was considered a bigger market than business to consumer (B2C). 

I don’t remember exactly what actually happened to Internet Capital Group. But I do remember when some of its portfolio companies began to fail. 

But Softbank had made some pretty good early investments in Yahoo! and Alibaba. It survived the bursting of the internet bubble just fine. There was a time when its founder – Masayoshi Son – was considered one of the best investors in the world. 

Investing can be hard. And Softbank’s debacle with shared-space company WeWork kinda undermined Masayoshi Son’s reputation as a great investor.Softbank first invested $4 billion in WeWork in 2017. Another $2 billion investment in 2019 valued WeWork at $47 billion…  

Long story short, WeWork imploded before it could go public and stick regular investors with the stock. Softbank doubled down, eventually putting $18.5 billion in WeWork, even paying the founder $1 billion to walk away. 

By 2020, Softbank had written down the value of WeWork to $2.5 billion. Today WeWork shares trade for 11 cents and the company is valued at $94 million. But even that is fanciful: the company will be declaring bankruptcy very soon. 

“I’ve been admitting that several times I was foolish…” – Masayoshi Son

Self-awareness is among the greatest – and most under-appreciated – traits any investor can have. Because it’s just not possible to be right all the time. Like it or not, we all carry our own biases and preferences that influence the decisions we make. 

A little humility can go a long way. And nothing will do more to undermine humility than success. Make a few good investing decisions, and it’s pretty easy to start thinking “I got this.”

Hence the phrase “bull market genius.”

Now I’m not here to pick on Masayoshi Son. I don’t even know the guy…

I want to talk about Softbank’s $32 billion purchase of ARM Holdings. Because it’s a great example of why investing – especially tech investing – can be very difficult. 

When Softbank bought out ARM Holdings in 2016, it was pretty much a no-brainer that ARM was among the most important companies in the world. Because ARM owned the architecture that allowed semiconductors to get smaller and smaller, and more and more energy efficient exactly the conditions needed to make smart phones and laptops smaller and batteries last longer. 

For a while there, just about every chip company out there was paying royalties to ARM Holdings to use their patented architecture. Apple (NASDAQ: AAPL), Qualcomm (NASDAQ: QCOM), Intel (NASDAQ: INTC) – all of them.

At the time of Softbank’s buyout of ARM Holdings, Masayoshi Son made a reference to tech investing as if it were a game of chess: “It’s easy to look at where your pieces are now and place the next one. This one is 10, 20, 50 moves ahead.”

Because Son thought that ARM Holdings’ architecture would be the building block for both the Internet of Things (IoT) and artificial Intelligence (AI). 

Well, yes and no…

All Hail Nvidia

We all know Nvidia is the king of AI chips. But some of that success is a bit accidental. Back in 2020, Nvidia actually had a similar vision for the future of chips as Masayoshi Son – that’s why Nvidia agreed to buy ARM Holdings from Softbank for $40 billion. 

The deal was nixed in February 2022 because of regulatory issues. But if you think about it, Nvidia got lucky…

Even in early 2022, nobody really knew what form AI computing would take. Turns out, instead of using a lot of small chips embedded in various gadgets, the ChatGPT model called for (relatively) massive chips in massive servers in massive data centers processing massive amounts of data. 

For Nvidia, it was better to be lucky than good, I guess. And for Masayoshi, well, I kinda wish him a little luck. In 2019, he sold the 5% stake he had accumulated in Nvidia for $3.6 billion. Today, after another blowout earnings report from Nvidia, that 5% stake would’ve been worth about $65 billion (thanks to Bloomberg for that little nugget).

That’s it for me today. Take care and I’ll talk to you on Friday…

Briton Ryle
Chief Investment Strategist
Pro Trader Today