America, 5G, and Industrial Policy

ARTICLES - 25 June 2019

By François Godement

The global debate about 5G – Huawei’s marketing lead and the security issue, Europe’s fragmented landscape– has usually taken for granted that America is almost absent from the scene. Yet America’s pullback is more of an appearance than a reality. Yes, it is no longer the world’s maker of handheld devices, or of much hardware gear. Once a leader for standards with CDMA (after Europe’s GSM norm, now forgotten), it has not been half as present as Chinese entities – companies, public organizations and lobbying – in defining new norms for 5G inside the international standards body 3GPP. "The Valley" is not focused on complete hardware products but on software for which it obtains huge global returns, chip design and subsets that go inside others’ brands. The three telcos – AT&T, Verizon, and Sprint (instead of the 23 European companies, not counting virtual operators) - are also very profitable. But they do not build, of course.

The simmering Huawei security issue, brought to a climax by Donald Trump’s decision on May 15 to ban Huawei from U.S. telecommunications, may seem to have been a self-inflicted wound. Where Europe has two underfinanced suppliers (Ericsson and Nokia), the U.S. seemed to have none.

That perception is wrong. American software, chip design and subsets run inside almost every existing IT system – including nascent 5G networks. Rather than take risky bets over brand dominance (with the exception of Apple), the US digital industry makes itself indispensable to competitors.

Yet, it is on that perception of America’s absence that Europe’s choice for 5G, inasmuch as it values a quick deployment, appears to be only between two options. One option is a dependence on Huawei, eventually mixed with some other suppliers. The other option is a trendy but difficult move to "digital sovereignty": it requires enough funding for infrastructure and financial support for the two European manufacturers, so that they can match the volume and price of the Chinese supplier. In short, digital sovereignty is only as good as its toolbox. As attractive as the magic words "industrial policy" and "digital sovereignty" might be, this is a high-risk road. Large pan-European projects have cumbersome processes for decision-making in a very fast-moving sector. There is no independent European authority that can craft an industrial policy beyond support for innovation – NASA, DARPA, the Office of Net Assessment are institutional midwives that are hard to match. A subsidized European industry, perhaps legitimized by the precedent of Chinese government support for its own digital sector, is not legal in terms of WTO rules. It might also suffer from all the defects of state-run industries – as was in fact the case for many French telecommunications and computer companies until the 1980s.

The U.S. policy community has awakened to the issue of a growing gap in telco development and in the standards of the future.

But what if there were solutions in relation to new American developments? The U.S. policy community has awakened to the issue of a growing gap in telco development and in the standards of the future. Part of it is domestic politics: rural communities are not well served, and the concentration of telcos brings high prices for consumers. The Trump camp, in keeping with its populist approach that is neither Republican nor Democrat, has in fact talked about a nationalization of networks that would place the entire sector under government authority. Congress and its Republican members, egged on by the three telcos, are making sure they defeat the idea within a larger bill proposal for "secure telecommunications."

Yet several developments catch the eye. One is a largely private venture, associating Cisco, Rakuten (Japan’s platform and mobile company) and Red Hat (world leader in open source software). It has begun testing a software-based 5G solution, expanding the virtual network systems that manage these new networks and their edge architecture, routing data through a private cloud. The solution cuts through a lot of the hardware needed for 5G, relies on open source, and therefore, verifiable software that many say is the key to enhancing security for the data networks. Multi-layer networks or "slicing" are in any case needed for the data with very different requirements that will travel through the 5G networks of the future. The Cisco & Rakuten solution deals with this issue at the root, instead of grafting new software to hardware networks that were not designed for this. Rakuten is already installing base stations in association with NEC, at prices said to be considerably lower than other 5G suppliers.

Another development is a public initiative with private partners, and it originates, like much of America’s innovation and industrial policies, with the Department of Defense. DARPA has long entertained non-conventional contract and funding practices designed to achieve leapfrog innovation or so-called "moonshots". The possibility now extends to a number of agencies within the Departments of Defense (DoD), Energy (DOE), Health and Human Services (HHS), Homeland Security (DHS), and Transportation (DOT), as well as NASA. OTAs (for Other Transaction Authority) are consortia (company of companies) sought by a government agency among private firms – in DoD’s case, both usual suppliers and "non-traditional" entities that do not customarily deal with the DoD. They submit studies leading to contracts for a "prototype." The consortium’s role stops where infrastructure construction and maintenance begin. Although the first phases are not subject to tender rules, the studies are anonymous and judged by a qualified jury. The goal is to cut red tape and accelerate innovation – doing away with many public procurement rules including the Buy American Act. One source from the consortium explains: "this is to cut through our traditional acquisition system, seeing that China is unencumbered by rules." The process has indeed mushroomed – by the end of 2017, there were already 27 OTAs, directly under the DoD’s supervision, with many more associated with one or another Defense arm.

One such OTA specifically deals with the issue of frequency bandwidth or spectrum sharing and efficient use, with 5G in mind. According to participants, the DoD has started from the security issue of 5G networks dominated by Chinese suppliers. But the concerns extend further: the availability and interoperability of 5G networks for military deployments abroad and with allies, for example. Dual use is also a key concern of the DoD. A National Spectrum Consortium (NSC) has been formed with 227 private participants – including large Defense contractors, "Mom and Pop" research outfits or companies, university labs – together with DoD agencies, including major intelligence agencies. A 1.25-billion-dollar fund has been allocated in 2015 – from the 44 billion proceeds of the mammoth 2014 1775-1780 Mhz bandwidth sale. There is a view to include other government departments, and a question as to foreign participants: so far, only American subsidiaries (and for instance, Ericsson and Nokia) have been allowed to join, but others can be suppliers (as is Thales in at least one contract with the NSC).

The process has innovative traits: a mix of large and small firms, from inside and outside the defense circle; fast-track processes at the research stage, increasingly moving towards usual public tender rules as the results go past the development stage into production. And management is private, although the consortia have to follow public accounting rules. The use of some of the windfall profits from frequency bandwidth sale should be a lesson to Europe, where the large proceeds from these sales fall into general public budgets.

The OTA consortium’s role stops where infrastructure construction and maintenance begin.

Unsurprisingly, Congress is not totally enthused by a spreading way of doing business that dispenses with some of the rules of Federal Acquisition Regulation (FAR). The whole process locates public financing at the R&D and early innovation stages with collectively shared results. The next development is open to competition from all members of the consortium, with production going back to the usual rules. OTAs could therefore escape the accusation of undercutting markets with subsidies.

There could be further developments. Such an R&D and innovation effort into spectrum use is likely to yield results. The Cisco-Rakuten-Red Hat-NEC virtualization is perhaps not the last major innovation we shall see.

It is still unlikely that large software, chips and subset firms would go directly for the entire 5G hardware business: flexibility is actually profitable. But participation, or acquisition, or cross-acquisition of one of the two European manufacturers could still make sense if there is a strong development perspective, which the obstacles placed in front of Huawei suggest. In fact, a situation could also arise where Europe would be deprived of one of these two industry leaders through an M&A.

At any rate, it would be irresponsible for Europe to count solely on its regulatory process, or on a fuzzy compromise over Huawei, and on an underhanded American response to the Chinese challenge. We are likely to see new action from the other side of the Atlantic, and in tandem with the Japanese industry. New alternatives to the hugely costly 5G infrastructures could be emerging, and it may be in Europe’s interest to be on the inside of these initiatives rather than on a separate path. Before we see a European DARPA – or an OTA process that cuts through much of the red tape in the interest of innovation – we may have to consider the innovations coming across the Atlantic. In a broad perspective, Europeans should take a cue from some of the innovative solutions for public-private partnership that are now emerging in the United States. After the Space X and Blue Origin competition to Ariane, this may be the next high-tech challenge from America.

America, 5G, and Industrial Policy

By François Godement

Add new comment

Commentaire

Envoyer cette page par email