Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?threads/let%E2%80%99s-rebuild-the-us-microchip-industry-%E2%80%93-not-give-it-a-50bn-plus-check.16382/page-3
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021370
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

Let’s rebuild the US microchip industry – not give it a $50bn-plus check

Look up the definition of a monopsony.

Warning, the last time I made a claim about market structures in this forum it was to call the EDA industry a cartel, and folks jumped down my throat saying "it's not a cartel it's an oligopoly and that's fine".. a bit like saying "there's no fire but it's so smoky I can't see any startups".

My opinion derives from direct experience with the utterly pointless burden imposed by EDA licenses on my small semiconductor company, the fact that Cadence EDA cost us just as much as the masks and a minimum wafer run on XFAB 180nm. Their opinion is presumably informed by loyalty to whichever of the big 3 signs their paychecks.
 
Intel should stop waiting on the US Congress to pass this bill. They should put their money where their mouth is and begin building the fab complex in Ohio. They have plenty of money to do this.
 
Will this heavily Intel oriented Chips Act and the subsidies strategy actually punish innovation and discriminate against those well managed companies?

I think it is misleading to say that it is Intel oriented b/c the fabless guys aren't making money. For starters these guys aren't in any trouble/can be replaced in a short period of time (all of the startups are proof of that). What cannot be build up in under a decade is a more robust American semiconductor manufacturing environment. It is also wrong to say intel is going to stop the Ohio plant if they don't get their handout. From day one Intel has said funding would decide fi they built a smaller fab like the old plans for the German fab, or if it would be Intel's biggest. The CHIPS act doesn't discriminate against non Intel semi manufacturers, with other US companies getting money, as well as TSMC Samsung and a plethora of semi suppliers.

If memory serves the plan for an even bigger TSMC Arizona fab and a Texas wafer factory were threatening not be built at all unless the extra costs of US operations were paid for by the US government.

As for the CHIPS act itself, as others have mentioned I am skeptical how much of an impact we shall see. Will falling PRC demand/the loss of trailing end PRC capacity be offset by higher demand for accelerators, datacenters, and AI? Or will most of this new capacity sit at 50% utilization, let's hope this is money well spent, because while it isn't that much cash (by US stnds) the risk is concerning.
 
Intel should stop waiting on the US Congress to pass this bill. They should put their money where their mouth is and begin building the fab complex in Ohio. They have plenty of money to do this.
They have. If memory serves they already broke ground. But they skipped the fancy ceremony until they get "their"money.
 
To @benb and @Fred Stein,

If you were the Congressman, Senator, or US President, what would you like to do with the US semiconductor manufacturing and the industry as a whole?
Here’s what I would do. Starting with some history.
1945: Vannevar Bush, “Science the Endless Frontier”. This Is the eloquent, very influential document for why the US government supports scientific research. https://www.nsf.gov/od/lpa/nsf50/vbush1945.htm#ch3.8

After 1945, Universities receive government support, while industry does not, in the US.

Bush set the US on a path with a balance between government support, for basic research, and free market competition.

1980s: Antitrust law changes under Reagan, allowing a merger boom. This sets the US in the direction of merging businesses rather than innovating.

1994: NAFTA opens US markets up to Mexico and Canada manufacturing.

2001: China joins WTO and more outsourcing of manufacturing occurs.

2024: The hypothetical benb presidency. To address monopolization, the safe harbors that permit M&A discussions would be treated as collusion and result in jail time. This would end all mergers. Higher interest rates would address financialization. Innovation rot is harder to address, in the short term; mergers drying up helps as it will create more competition, which is ultimately where innovation comes from.

In the short term, I wouldn’t play the quick fix game, with CHIPs act type bills. That is reinforcing the monopolies, not breaking them up.

In the mid-term, I would create a sovereign wealth fund to invest in small companies with a long time horizon. Smaller investments, in small businesses, since I want smaller businesses, not bigger. And it would need to be blue ocean in the sense it wouldn’t fund a Foundry 300mm fab but the US first 8inch SiC fab would be OK. The pool of money would be operated by ex-venture capital. They would do annual reports and seek to grow the fund by 20% a year, with higher risk/higher return investments, favoring job creation over exits. Exits of course would be impossible since you’d go to jail, if I was president.
 
Back
Top