When I think of GE and the old school industrials that it competed with, I can’t help but wonder if today’s hyper scaling tech companies are unwittingly waltzing into the same trap. They’re spending billions in capex on data centers to power their ai dreams—the very same ai whose large language models increasingly look like commodities: these are in some sense latter day dishwashers and plastics from ge’s era. I know that Zuck et al see trillion dollar markets in the future, and Zuck et al are savvier operators than Welch ever was, but history is littered with examples of companies that invested heavily in capex, only to not have revenues sufficient to generate a return on that capex.
It’s possible, but it’s strange to think of intelligence as being commoditized but also not valuable. I suppose it all comes down to the application, and at least in Meta’s case, the capex is already driving results. The GPUs he’s buying are paying for themselves via return on advertising spend, the LLMs are an embedded call option on AGI that also undercuts OpenAI and Google trying to monetize their LLMs.
I think this is a good reason to be long China- they deliberately squashed their finance sector to try and discourage this, and their general strategy of becoming the indispensable industrial nation has been very effective, with the hits (killing off the non-Chinese solar industry) still rolling in.
It's worth reflecting on the fact that China has never been able to make a good jet engine (or leading-edge semis). It's a testament to the difficulty of progress that GE has faced. China will continue to be a force for industrial goods, but they will face similar difficulties. They are struggling to attract foreign direct investment because investors are worried about low return on assets (and the trustworthiness of the government and partners). When the country subsidizes industries with lower returns, the result is almost always slower growth. If they keep their currency low to keep their industrial products cheap, it's effectively a tax on the workers in the factory, who will have lower purchasing power; and it aggravates trade partners who think they're dumping and enact barriers. If that happens enough, they're stuck with expensive factories that are severely underutilized, an angry, poorer population, and decades of low, no, or negative growth like Japan. There ain't no such thing as a free lunch.
The trophy tech is a lagging indicator- they make all the ballpoint pens now. SMIC has been impressive, it seems to have pushed the previous generation of tech much farther than it was ever supposed to go.
Hard to predict the geopolitics, but I suspect there's ultimately more appetite for cheap stuff than fighting the Colossus.
When I think of GE and the old school industrials that it competed with, I can’t help but wonder if today’s hyper scaling tech companies are unwittingly waltzing into the same trap. They’re spending billions in capex on data centers to power their ai dreams—the very same ai whose large language models increasingly look like commodities: these are in some sense latter day dishwashers and plastics from ge’s era. I know that Zuck et al see trillion dollar markets in the future, and Zuck et al are savvier operators than Welch ever was, but history is littered with examples of companies that invested heavily in capex, only to not have revenues sufficient to generate a return on that capex.
It’s possible, but it’s strange to think of intelligence as being commoditized but also not valuable. I suppose it all comes down to the application, and at least in Meta’s case, the capex is already driving results. The GPUs he’s buying are paying for themselves via return on advertising spend, the LLMs are an embedded call option on AGI that also undercuts OpenAI and Google trying to monetize their LLMs.
Great points.
I think this is a good reason to be long China- they deliberately squashed their finance sector to try and discourage this, and their general strategy of becoming the indispensable industrial nation has been very effective, with the hits (killing off the non-Chinese solar industry) still rolling in.
It's worth reflecting on the fact that China has never been able to make a good jet engine (or leading-edge semis). It's a testament to the difficulty of progress that GE has faced. China will continue to be a force for industrial goods, but they will face similar difficulties. They are struggling to attract foreign direct investment because investors are worried about low return on assets (and the trustworthiness of the government and partners). When the country subsidizes industries with lower returns, the result is almost always slower growth. If they keep their currency low to keep their industrial products cheap, it's effectively a tax on the workers in the factory, who will have lower purchasing power; and it aggravates trade partners who think they're dumping and enact barriers. If that happens enough, they're stuck with expensive factories that are severely underutilized, an angry, poorer population, and decades of low, no, or negative growth like Japan. There ain't no such thing as a free lunch.
The trophy tech is a lagging indicator- they make all the ballpoint pens now. SMIC has been impressive, it seems to have pushed the previous generation of tech much farther than it was ever supposed to go.
Hard to predict the geopolitics, but I suspect there's ultimately more appetite for cheap stuff than fighting the Colossus.