Thread by @jsuedekum

After reading this article, one can get the impression that things are going terribly wrong in the current 5G rollout.

Key questions: do we really want to allow Huawei to be part of the German network? And do we need the current auction format? /thread

washingtonpost.com/world/national…

First there's the issue of espionage: Do we really want to trust Huawei, a firm with close connections to the Chinese communist party, to participate in one of the most essential and sensitive infrastructures? It involves many risks, up to the possibility of an entire shutdown /2

So far, there’s no proof that Huawei ever engaged in such activities.

Still, many countries (e.g. Australia) are highly concerned about it, and have thus banned Huawei from their national 5G networks. But most European countries, incl Germany, haven’t decided yet. /3

But apart from security, there’s another issue: industrial policy.

The article claims that Huawei is flush with subsidies by the Chinese state. The European competitors which could in principle also build a 5G network – Ericsson and Nokia – don’t get such subsidies. /4

This has allowed Huawei to undercut Ericsson by 60 percent in price in the 5G rollout in the Netherlands.

From the currently ongoing German 5G auction, I am hearing that all bidders plan to work with Huawei since they regard the European alternatives as “too expensive”. /5

But do we really want to do this?

Let a subsidy-fed Chinese firm do the job, and give the (unsubsidized) European providers a friendly “no thank you”? Just to save on costs (and maybe time) for the network today? For a network that will be essential for many decades. /6

Notice this also has some geostrategic implications: Trump has banned Huawei from the American 5G and expects Europe to do the same.

Of course, we don’t have to obey. But if we don’t, this adds another scratch to the already fragile transatlantic relationship. /7

Thus, as a matter of industrial policy, isn’t there a case to deliberately decide to let Ericsson&Nokia build the German (or better: the European) 5G network?

Because the Chinese firm is not competing on fair terms, and it is too risky to work with from various perspectives. /8

And further on 5G, there’s another problem of the German rollout: the licenses are auctioned off to private firms and the German government insist to extract money out of it.

But this will probably lead - again - to a patchy network with lots of dead spots across the country. /9

Why? Because the private bidders who build the infrastructure (Vodafone, Telekom etc.) only care about the market in dense cities, but not about the countryside.

Thus, they have little incentive to build and to maintain a functioning network in the periphery. /10

But if autonomous driving, drones, industry 4.0, and other key technologies rely on 5G, they better be available and work properly everywhere in the country (even at the “letzte Milchkanne”) and not just in city centers. /11

Regulators can try to make companies care. But let’s fact it: this didn’t work well in the past.

Everyone who travels around Germany experiences this every day. Go to Austria and you have excellent connections even in the most remote places. In Germany, not so much…/12

Summary: @peteraltmaier - Why don’t we have a public infrastructure agency build the 🇩🇪 5G network, using European technology rather than Huawei?

Private providers can then access the infrastructure for a price differentiated by time, location, etc.? Very market-friendly /13

I’ve heard rumors that Telekom, Vodafone and the other players would even welcome this approach of vertical separation.

And the building of the infrastructure doesn’t even need tax money. It can be financed by state-guaranteed bonds of the infrastructure agency /14

To me, this sounds like a type of industrial policy (and infrastructure policy) that would actually work.

Would it be technologically feasible in reasonable time? And at what cost?
What am I missing?

@peteraltmaier @AchimWambach @haucap @armin_steinbach @pm_steinberg /END