0x09 Thick and Fast
UK's Semiconductor Strategy, India's OSAT Puzzle, US-China Confrontation Heats Up.
Given the focus on semiconductors in geopolitical conversations nowadays, every week brings new stories and revelations. Even so, the past couple of weeks saw some crucial developments prompting this special edition of Siliconpolitik.
Curveball: The UK’s National Semiconductor Strategy
The UK announced its National Semiconductor Strategy on May 19. A lot more people should read it. Here's why.
Many news items ridicule how the UK's £1 billion spending commitment over the next decade is minuscule compared to that of the US (~$52 billion) or EU (~$46 billion). But I think the UK's semiconductor strategy is qualitatively and markedly different from the others. It only focuses on the UK's comparative strengths rather than trying to build the entire ecosystem locally.
The policy vision is clear. The UK has three strengths: "research and development (R&D), design and IP, and compound semiconductors". And the policy is about converting this potential into a definitive advantage. There is no mention of reducing "import dependence" at all. For commercial off-the-shelf silicon chips, the strategy relies on international cooperation rather than indigenisation.
Instead, the focus is on compound semiconductors. It's not about manufacturing silicon chips within the UK but about gaining an asymmetric advantage by getting good at something few others in the world are thinking about.
So, it presents a radical departure from national semiconductor strategies announced by most countries in the recent past. Analysts should observe closely how the UK fares. It will have broader lessons for the tech geopolitics community.
Finally, do read the strategy document. Having read the semiconductor strategy documents of some other countries, I found this one to be the best-written. The vision, objectives, and pathways are all tightly coupled and well-explained.
PS: The previous edition of this newsletter covered a related development — the UK government getting Nexperia to sell 86% of its stake in Newport Wafer Fab.
A Hard Nut to Crack: The Curious Case of a No-show by Taiwanese Outsourced Assembly and Test (OSAT) in India
On the face of it, India and OSAT should be a perfect match. OSAT plants test the manufactured chips for defects and ensure protective packaging for all finished chips. This stage requires high capital investment, though not of the same order as fabs. Further, this stage requires large numbers of relatively low-skilled labour, whereas the manufacturing stage requires a sizeable high-skilled workforce. With low-skilled labour better available in India, Taiwanese firms like ASE Technology and Powertech Technology can benefit by offshoring these operations to India. The Indian government has also announced several attractive financial incentives in this area that Taiwanese companies can avail of. And yet, there has been no breakthrough. Taiwanese players have shown more interest in Malaysia, Vietnam and the Philippines. Why?
The not-so-glamourous reasons
The lack of policy consistency and high import tariffs are the bottlenecks that can explain why Taiwanese companies haven't moved ahead.
Trade policy misconceptions
Taiwanese officials have asked for tariff reductions on products used to make semiconductors. It makes sense because OSAT would require a lot of imports. Further, Taiwan had also filed a case against India for violation of the agreement on tariff reductions in WTO - also known as the Information Technology Agreement. We should reduce import tariffs soon. There's not even an "infant industry" argument to be made because there isn't any of it at all.
Lack of policy consistency
Frequent changes in tax policies and restrictive labour laws have proven to be a significant deterrent. Surajeet Das Gupta in the Business Standard has a couple of reports covering this problem. He writes:
"Nor do they want coercive tax regimes followed by long legal battles against the revenue department in highly import-intensive export businesses... Investors are asking for deeper consultation between the nodal ministry and the revenue department so that they are in sync. And they are seeking more liberal labour laws (especially longer and more flexible work hours) as an essential condition to investing." (Source).
So, fixing trade and tax policies is mandatory if we need industrial policy measures such as production-linked incentives to deliver. These pro-market policies can't be substituted with financial incentives.
Hot Spot: The US-China Confrontation Over Semiconductors
There was no shortage of action on this front.
In what can be construed as a first direct response to the export controls by the US, China banned “operators of its critical infrastructure” from using Micron memory chips. The stated reason is that “Micron's products have relatively serious potential network security issues”. A commercial off-the-shelf memory chip is unlikely to have specific network security issues targeted at China.
This response, then, is a signal. Moreover, Micron is chosen as a recipient of the signal because these chips have other international substitutes. The US will likely respond by adding more companies to the export control list. Keep watching this space.
Moving on, the Nvidia CEO struck a discordant note in an FT interview, arguing that decoupling in the semiconductor domain is unsustainable. He gave two reasons: one; there is no substitute for the Chinese market, and two; China will build the chips themselves if the American companies don’t.
He’s right on both counts, of course. But the Overton Window concerning China has shifted in the US. For at least a couple of years, the US would be willing to accept some self-harm, provided that it causes more damage to China. And hence arguments about the indispensability of the Chinese market are unlikely to make headway. That apart, it’s also true that China will be able to make high-end chips only after considerable effort and with a significant delay due to the export controls.