Hacker News new | past | comments | ask | show | jobs | submit login

You can pontificate all you want about money alone being insufficient, but mine isn't a comment about printing money, nor about Intel.

Long term semiconductor capital expenditure won't give fruit for several years, and reliance on foreign production binds us into disadvantageous geopolitical power projection games. Companies are cautious in laying out the funds because they want to be certain of demand and efficient with their resources. They don't particularly care about the externalities. A government has power and scale that makes that job easier, and an incentive to establish better long term solutions that may cost more than the short-term greedy solution. In contrast, corporations have shorter time horizons, less risk tolerance, and a greater incentive to seek the greedy solution.

You get brains when you hire them. You get foundry tooling when you buy it. For the specific problem of building a manufacturing ecosystem around semiconductor fabrication, both are widely available.

Executive orders about China's influence upon the semiconductor supply chain are just lip service unless and until eleven or more zeroes are included somewhere in the package. It's not a PR problem. As you correctly point out, someone actually has to roll up their sleeves and build lots of shit. That doesn't happen for free, and it takes coordination and risk that almost no corporation can pull off unilaterally.

We have an incredibly sophisticated defense industrial apparatus for the same reason.

(BTW...Not all spending results in money creation. You can tax, issue bonds, or just recover the cost from the productivity of the project. The idea of "printing money" isn't necessarily relevant to capital projects that return revenue, although I suppose it is always a backstop.)




Intel has been spending about 14B/year on long term capex and another 14 B/year on R&D.

It doesn't seem to have gotten them a whole lot.

https://www.macrotrends.net/stocks/charts/INTC/intel/researc...

You may be able to buy an old chip fab, but you can't just buy organizational capacity and the ability effectively run an R&D department.

US spending on R&D has been increasing steadily, the private sector funds 600B/year https://fas.org/sgp/crs/misc/R44307.pdf

There has been no slowdown in either public or private R&D spending.

The world is literally filled with the carcasses of failed R&D projects when nations that tried to develop industries but failed to do so due to deficits in organizational skill, excess corruption, lack of human capital, and ineffective business practices. Meanwhile tiny Taiwan went from a nation that produced shoddy toys and bicycles to a world power. You can't buy solidarity, unity, or human capital.

As for our defense industrial complex, I would say that real innovation there, too, has substantially slowed down compared to the 50s-60s. Look at how quickly we obtain supersonic flight or nuclearized the navy, and how much slower progress has been since. Similarly innovation per real dollar spent has dramatically fallen in the medical field, too. Software is the main area where the US remains a leader, but outside of that it's slim pickings, even as we spend more money on R&D than ever before.


> It doesn't seem to have gotten them a whole lot

Intel has majority market share in datacenter and client computing. I learned in a professional setting that you can't ever judge the quality of a semiconductor company based on its price action because the vast majority of the money deployed on Wall Street lags the technology because technologists are rarely the ones expressing discretionary views on the stock.

> The world is literally filled with the carcasses of failed R&D projects when nations that tried to develop industries

This is a bad take. The largest pure play foundry makes up essentially half of its host economy, and there's a difference between building factories/ports/plants and doing research. Even failed research is a step forward but that's not what we're talking about here. There's no geographic reason why Taiwan should be the epicenter of semi fabrication, in fact there's a political and economic disincentive for it.

You have this assumption that there's somehow a shortage of human capital in the semiconductor space but that's just not true. The present bottleneck comes from the fact that demand has grown beyond what the industry can facilitate, and capex outlay is a risk that takes time to bear fruit.

I disagree with your view of defense innovation, but I suppose we can't know because all of the juiciest recent developments are probably a secret anyway. But think of your ten favorite technologies and try to determine whether they were affected by defense spending. Here's an example: there'd be no Uber without GPS, and no GPS without the defense industry.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: