Tech spending race
December 28, 2024
*********************
Just one of those charts from the dotcom bubble. There are a lot like this. This one went from 10 in late 1998 to 110 in Q3 2000. And then, poof.
The fortunes being spent today on data centers for AI are jaw-dropping, but tech leaders are actually worrying about spending too little. ‘When you go through a curve like this, the risk of underinvesting is dramatically greater than the risk of overinvesting for us here,’ Google CEO Sundar Pichai told analysts… Tech CEOs view their investments in data centers as all-purpose bets on the future. If the AI bubble pops, a data center can easily be put to work fueling whatever the next big wave in tech turns out to be.”
I clipped the above from Credit Bubble Bulletin, but the source Axios piece is linked at bottom. The Axios piece has a chart which shows aggregate Capex spending for MSFT, META, GOOGL and AMZN having gone from around $70b in 2019 to $218b in 2024.
When I read the Sundar Pichai quote, it immediately made me think of the dotcom bubble and fiber optic cable. Fiber optic was the backbone of the internet, and the lead company was Corning (GLW), stock pictured above. Of course, to check my memory, I searched for ‘Corning optic fiber’ and was rewarded with this spectacular piece by Kevin Maney from March of this year.
https://kevinmaney.substack.com/p/is-nvidia-remaking-the-movie-corning
He compares Corning with NVDA. It’s not the direct comparison to NVDA which I find interesting, it’s the general theme of a spending race brought on by transformational technology. The winners likely know there will be a shakeout at some point, but they will be left standing (in the rubble).
From the article: “For a while, the strategy looked brilliant. Demand from the Enrons and WorldComs and Level 3s went crazy. They had blank checks from investors to spend, and so they did.”…then…” What once looked like precious and scarce fiber optic bandwidth turned into a glut.”
The promise of AI is efficiency, and the replacement of many professional jobs. Like many transformations, there will be winners and losers.
Here’s a clip quoting Wendell Weeks, who was at the core of Corning’s fiber optic strategy and is now CEO (emphasis added).
“Wall Street had given (these telecom companies) around over half a trillion dollars, and so when people show up with hundreds of millions of dollars to buy your product, you tend to believe them,” Wendell said. “Our big lesson learned was, for one, it’s not enough just to be the best at what you do. You also have to understand your customers’ business model and your customers’ customers’ business model. We were risking more than money. We were risking significant dislocation of people’s lives.
The problem is that dislocations can’t be avoided. Perhaps in recent history they’ve been pushed a bit further forward. The volatility in stocks post-FOMC is a reminder to manage downside risks.
There’s a sobering line in the Big Short by Ben Rickert, in response to Charlie and Jamie celebrating their bet against CMO’s.
“If we’re right, people lose homes. People lose jobs. People lose retirement savings, people lose pensions. You know what I hate about f-cking banking? It reduces people to numbers. Here’s a number – every 1% unemployment goes up, 40,000 people die, did you know that?”
The great part about the Corning (and American) story is this: the company didn’t go out of business. It continued to innovate. Weeks: “We’re never about the products that we make because the products we make will always change. What won’t change is the core of the company being about values and innovation. …We took the opportunity to decide what type of company we were going to be for the next decade. A lot of companies, when they get in that spot, change fundamentally what they are. We instead really embraced the core of what we do, which is to create and make keystone components and to grow through innovation, and acknowledge at the same time there are inevitable downsides.”
The AI boom has created a lot of tech investment geniuses. The other part to remember is that at least part of the background is supported by liquidity. The last FOMC casts shade on the probability of unlimited liquidity.
https://www.axios.com/2024/12/20/big-tech-capex-ai
