Friday, February 5, 2021

What nonsense is happening around Silterra?

We know that Silterra is a problem child for Khazanah and ultimately our country. It has been bleeding money since inception probably around 2000. Although the idea came noble, the execution was not. It probably has lost us money to the tune of more than RM10 billion (I do not think anyone keep track as it is already a sunk costs). While on the financial side it was never a win, on the technology acquisition side it did some justifications. Malaysia now has some knowledge of how to handle a foundry albeit a very old one. We also have now people working in several foreign owned foundries most of them in Kulim and Penang.

While we have thrown substantial sum of money, for a long period we have tried to keep the fab afloat running on positive EBITDA - more or less cashflow positive. But we know that running foundries on positive EBITDA is not the right measurement as foundries capital costs are high - latest technology can go up to $20 billion (a 5nm or 3nm fab). There is no point for us to catch up.

While it has been a sore thumb, today, foundries have become an important asset in return thanks to two things, firstly IoT, IR4.0 autonomous vehicle etc. - as some businesses are not chasing for Moore's Law where the theory is that semiconductor is reducing half in size every 18 months. So old fabs have value. Another thing is China. They are hungry (all the while) and especially today where US is blocking them off from foundry technologies. So an old fab in Silterra is becoming an asset back albeit lower value than the newer fabs.

However, what is there to do with DNEX and Green Packet? Those two companies do not know how to run a fab! They probably have not stepped into a fab before 5 years ago, until now. Why is there a need to sell to a Malaysian company. If one notice both deals are with a China partner. The Chinese would rather work with one which do not need to know how to run a fab - hence DNEX and Green Packet. It is a flipping exercise if I may call it.

We have people and companies in Malaysia which may be closer to knowing the intricacies of running a foundry - even that is tough. I can probably see that the deal would have Silterra providing support for China's demand for chips but why not sell direct? DNEX and those other Malaysian companies do not know much. One is in some software support and Oil and Gas - another gotten some China's AI technology and try to deploy them in Malaysia.

Get on with real stuff and be direct. We have made mistakes on this for 20 years and it still has not ended.   

No comments: