I think the main problem is how to adjust costs. If you discovered a software bug, you just fix it, recompile it. With hardware design, if you have a problem, then you have to make a new PCB, manufacture a new PCB…

Their CEO is a really open guy, and their engineers are really, really, really good. I talked with those guys, and I love it when I have a conversation with someone… I walk out, and I feel really stupid, because he is so smart.

I noticed, really, really… except RCL Semiconductors, and they are a semiconductor factory in Hong Kong. Just recently, they built a new factory in Shenzhen. They now moved their equipment over there and go online. I think they are actually online already, or they go online soon with the new ...

The problem with LibreSilicon, as such, is the mindset in the industry.

He is a professor. He is the guy who runs the HKUST Lab. He is from Taiwan. He is very proud of Taiwan, let us just say…

Ah, Professor Lee. Here, I have his name, wait a second. That is…

Do you have like contacts with them? Maybe you could introduce us?

Yeah, “Take my love, take my land.”

…that was my mood when I saw that HKUST was closed.

Yeah. [laughs] That is the basic situation right now. I do not have a lab available. The entire R&D at the moment for LibreSilicon is on hold, because the universities are closed, including the university where the lab is we need, because of the protests.

Lanzendörfer, yeah, Lanceville.

Lanceville Technologies. The logo is by the way… my family name, my German family name, literally translated into Chinese. I am, that is also a problem why I cannot just run away…

Yes, it would be collaborative, anyway. My company actually… so, I have a company registered. You somehow have to write invoices and get taxed, so I have a company there in Shenzhen.

A), just a lab, or B) a lab plus funding, or C) you take it over. That is like those three things, which I would be interested if any of those three things would be possible.

That would make me even more mobile. I could take some holidays, go to Brazil, hang around there a little bit. It would be cool.

Yeah. For me, it would actually be a super jackpot if someone else would run around in the clean room in a bunny suit.

…walking around in a clean room an entire day…

Do they have a lab? What I would be interested is… from Taiwan… I have multiple options, which I would like to check what of those are possible. One option would be that we could use one of the labs here. Or maybe that some department in Taiwan basically picks ...

Yes, there are multiple folks rising.

Or Taiwan, exactly.

Hey, Duolingo sucks a little bit with recognizing my pronunciation.

Or have you noticed this little…?

The semiconductor manufacturing in Mainland China is super paranoid. The manufacturers… they will not even let you close to the fab. That is a problem. I found an alternative lab in Portugal. So now, I am kind of torn between an unlimited flow of “pastéis de nata”…

Can I ask for help? Our situation right now is: this lab is not available. I now am in a situation where I have the difficult choice to make. The lab is broken. I have three choices. Because we applied for funds, the donation fund from NLnet that should come ...

And now, the lab is closed. I am right now mostly running between Hong Kong… because I am too lazy to move back to Shenzhen… between Hong Kong and Shenzhen. In Hong Kong, I do…

OK, things went up in flames.

Yes.

…build things like that. Now, there is… kind of… well, things happened. Things happened. Things, well, initiated, accelerated oxidation…

Problem is right now, that is why I came – I flew here from Hong Kong. I went to Hong Kong for the R&D, because in Hong Kong, there was in HKUST – the only available lab you could rent in Greater China, excluding Taiwan, which would allow external users ...

They download the process spec. They adapt it to their machines. They make these mask sets from the online GDS2 files on git (Revision Control System), and then they can calibrate all the recipes until they match all the analog parameters.

It is really, you let it verify automatically – you have a needle bed for your PearlRiver wafer. You put it on there, you measure out all the analog parameters, and when everything is fine, then you start producing.

Yeah, so, actually, you can characterize every single manufacturing step of the process. Then something… you can actually automatize that with needle beds.

What you do when you manufacture for LibreSilicon, the first thing, first, you send through some wafers with the PearlRiver and measure out all the test devices over the wafer.

That is the process step layers. That is the mask set we specify. You are free in how you manufacture the masks, because exposures, separate exposures are different in any fab. But you have parameters: for that is PearlRiver. The PearlRiver test wafer basically allows you to basically check whether ...

Oh, yeah. Our process is actually antinode. There has been something called scalable CMOS in the open circuit design from Tim Edwards. There has been something like LibreSilicon, but only on digital level. It is called MOSIS design rules. We actually went one step further, and we said, “OK, now, ...

Oh, so I should bring details?

It would be already… for companies from a business perspective… it would be like an absolute improvement. Then the other thing is, of course, the security increase. That is from that thing, so we have improvement of security… of reliability of the chips. It is scalable – OK, that is ...

They also need to invest money in R&D to develop their library to have components to develop the chips. LibreSilicon would turn this around by 180 degrees. You suddenly would have a continuously organically growing IP catalog, and foundries actually get paid to silicon-verify part of these components going into ...

The entire design flow is open, so you can basically follow the design and can verify that these components are not compromised. For companies supporting LibreSilicon, the advantages, right now, they have to have money on the side, not only investment for the factory equipment, for building the factory.

Developers from many countries can collaborate in designing IP cores – components in the chip, like telecommunication parts.

Yes, it is a lottery in reverse, exactly. I like that thought. The situation we have – what is LibreSilicon first? First, what is it? In general, to make it precisely, it is meant to be something like a meta foundry, meta factory – so that we can standardize manufacturing, ...

“Maybe next time.”

I have seen a TED talk from you. Nice talk, too, thanks – especially the strategy on how to organize funding. [laughs] So, it is really laisser-faire. “So OK, it went wrong. Thanks for the investment.” [laughs]

You have watched the congress talk?

Yes. What, I do not know, have you seen my congress talk?

That is actually why I am here, but we can come to that later. I will, Hernâni…

Taiwan is perfectly capable of manufacturing their own telecommunication components, also, at semiconductor level.

Yes, it is the market leader.