Policy decisions can be acrimonious because those with funds face “difficult choices” in the face of intense lobbying. The big question is: Do corporations need public funds support, especially as they grow and declare billions of dollars in the dividends for their shareholders? The answer lies in decreasing R&D spends and gradual shift towards maximising quarterly profits. For example, General Electric under Welch had shown substantial profits, but despite his lionisation, people now realise it was mostly paper wealth valuation without too much to show. I measure the value of a company if they create tangible assets (manufacturing) rather than services, because you are then scraping the bottom of the pile for “contracts” and not moving up the value chain. Likewise, Intel had spent billions of dollars in buybacks or doling out dividends, but shut down its fabs, moving towards “designs”, leading to its sorry state and being knocked off from the podium.
Chipmakers battle for slice of US government support | Financial Times
The House of Representatives last week followed the Senate in passing a broad law to counter China’s rise as a technology power, including $52bn in grants to support advanced chip manufacturing and research and development in the US. The law, which has yet to be signed, unlocks an estimated $24bn more in investment tax credits for chipmakers by letting them write off 25 per cent of the cost of new factories, or fabs, against their profits in the first year.
Here, look at the competence of the US in the chip “manufacturing”:
The problem is investments and ecosystems. As again, I can’t comment on the prevailing geopolitics, but the rise of China is attributed to its special economic zones and tightly knit supply chains. The current ongoing supply chain disruptions due to pandemic (and zero tolerance policies) have knocked off the values from manufacturing.
A little bit from TSMC:
TSMC is also angling for extensive support to justify its most significant attempt so far to put down roots in the US, where costs are higher than its home base. Chair Mark Liu said in June that the US plant was turning out to be “more costly” than TSMC had expected and that a shortage of chip manufacturing talent was causing the company problems.
The reason why I post this is to peer into supply chains and better understand how hardware (and consequently servers) will impact availability. It might be instructive to wait for new GPU’s, for example, which might offer a lower cost to run algorithms for your shiny AI project. All these are usually impacted by upstream policy decisions (including availability of local talent).
RadOncNotes 