🌌 Reshaped #38

How to fix the smart city, VCs and monopoly power, carbon neutrality, Big Tech earnings, generational wealth divide and much more

Oct 31, 2020

Welcome to a new issue of Reshaped, a newsletter on the social and economic factors that are driving the huge transformations of our time. Every Saturday, you will receive my best picks on global markets, Big Tech, finance, startups, government regulation, and economic policy.

Can we fix the smart city? In some contexts, it is even impossible to refer to it without assuming it is a very bad thing for cities and citizens. Many companies released their earnings reports for Q3 this week. You will find a brief recap of their performances in The markets. I also briefly commented on the impact of the antitrust case against Google for VCs.

Please, take a moment to share this newsletter with your network!

Share Reshaped

New to Reshaped? Sign up here!

Can we fix the smart city?

Disclosure: I am currently involved in a smart city project that partly reflects some of the ideas covered in this short essay.

Sidewalk Labs, the Alphabet-owned urban innovation company, has recently launched Delve, a generative design tool that “empowers urban development teams with cloud computation and machine learning to design better cities, faster, with less risk”. The tool uses ML algorithms to generate almost infinite design possibilities for new urban neighborhoods, including not only buildings but also streets, outdoor areas, amenities, and energy infrastructure. It also builds financial and energy models to measure the impact of each component on the whole system.

Applications like Delve raise many concerns among architects and urban planners, worried that such disruptive technologies will brutally replace activities that they consider as profoundly human-driven. This fear is diffused in the whole urban innovation sector, which is under attack for being too technology-oriented and focused on data extraction. A few months ago, Sidewalk Labs had to leave its core Quayside project in Toronto due to long-lasting backlash from local citizens — even if the official statement mentions the “unprecedented economic uncertainty” as the only reason for this choice.

There are many factors behind this sentiment of backlash. One is linked with broader privacy concerns over data extraction processes. In the West, this fear is amplified by the use of smart city technologies in China and other Asian countries, where these applications are functional to greater control of individuals. Another factor is the diffused sense that these technologies miss the critical link between the city and its dwellers, which translates into a cultural conflict between the creative (sometimes chaotic) evolution process of the city as a complex system and the attempt to simplify it through technology-driven urban planning and policymaking.

We can categorize smart city applications on a scale that goes from the enhancement of processes already in place to the full integration of each aspect of urban life into a comprehensive digital system. This apparently utopian scenario is not new: attempts to centralize governance through technology have always existed — think about the Chilean Cybersyn project in the 1970s. However, these technologies have evolved so much that centralized governance can be implemented at a much smaller scale (potentially up to individuals). This is basically where the smart city concept comes from.

This ascending technological complexity is paired with two other dimensions: the investment needed to put in place these processes and the ambition of technology-driven policymaking. Assessing the actual benefits for citizens, however, is harder. Ideally, these benefits should be positively correlated with a higher degree of all the other dimensions (technological complexity, initial investment, and political ambition). Nonetheless, the already mentioned complex nature of urban systems makes this correlation far from automatic.

Technically, the market for smart city solutions stems from the application in the urban landscape of technologies already in place elsewhere. This applies not only to software but also to hardware applications like environmental and motion sensors. Hence, the economics of smart city markets can be easily described as offer-driven, with tech giants constantly competing for key cities to scale their urban technologies divisions. Leading incumbents include corporations like Cisco, Huawei, Microsoft, and IBM. There also many startups in the field that either specialize in narrow segments of the smart city or target smaller (often domestic) towns. This means that tech giants face few threats from new entrants, which have to settle for smaller urban areas or side projects in cities where the major infrastructure is already provided.

This technology-driven idea of a smart city, however, raises a fundamental question: to what extent is it a neutral representation of the city? Smart city solutions use various inputs (sensors, satellites, human input, third-party sources like social media, etc.) to collect data and draw a comprehensive picture of the city in real-time. This means that what cannot be collected through specific algorithms is naturally left behind. If local administrators only rely on these technologies to do their job, relevant parts of the city can be excluded from public policy interventions. At best, we can say that technology here plays the role of the mediator between the real city and its actionable, data-derived counterpart.

Fixing the smart city means acting to make human-driven public decision-making processes be back at the core. In this scenario, urban technologies would still play a relevant role. However, mediation would remain a mainly human responsibility, made easier by an open and transparent dialogue with citizens and the possibility to rely on decision-support systems designed around the specific needs of the city. In practice, this goal can be reached by applying two fundamental changes to the dominant smart city framework.

First, smart city projects should start from the existing processes of local governments and Public Administrations. Going smart is much more than applying sensors and installing software. It means promoting open government standards and investing in the digitalization of institutions with a solid and ambitious agenda. An inefficient local government that invests in a smart city suite will very probably end up being an inefficient local government with a smart city suite. The path towards smartness in the urban context starts from ruthless self-analysis and deep knowledge of the city.

Adapting these processes also means making them open to a constant discussion with citizens and local organizations to imagine the future of the city and solve its most pressing problems. This is considered a (hard to achieve) priority in many European countries like Spain, which has launched two of the most renowned digital platform to promote citizens’ participation (Consul in Madrid and Decidim in Barcelona). Some startups followed this example to develop SaaS solutions based on civic technologies, which aim at building solid and transparent relationships between citizens and local authorities.

Bundling together these citizen-oriented solutions with data-driven applications can solve most of the perceived problems of smart cities. In Europe, where privacy and data governance issues are considered more important than anywhere else, a more holistic approach to urban innovation — an approach that accepts its complexities instead of making them seem simpler — can lead to renewed trust towards local governments and technology providers. The pandemic can accelerate the adoption of smart city applications, in one sense or the other. Hopefully, it will do so in the form that best respects the city and its multifaceted subjectivity.

The State

Industrial policy

After the EU, China, and Japan, South Korea announced that it will go carbon neutral by 2050 (The Guardian). It will be far from easy, as the country “relies on coal for about 40% of its electricity generation, with renewables making up less than 6%”. To reach carbon neutrality, South Korea will need to reform its energy policy and abandon coal-based projects already under construction. Similarly, China will build its upcoming five-year plan trying to put together technology dominance (Financial Times) and the ambitious goal of going carbon neutral by 2060. A recent article by Nature explains how it could be done in practice.

The plans differ in their details, but agree that China must first begin to generate most of its electricity from zero-emission sources, and then expand the use of this clean power wherever possible, for example switching from petrol-fuelled cars to electric ones. It will also need technologies that can capture CO2 released from burning fossil fuels or biomass and store it underground, known as carbon capture and storage (CCS).

This is consistent with the scenario drawn by analysts at Macro Polo, who, in a recent report (very recommended), forecast the evolution of China into a more sustainable competitor in the global economy.

A China that will be near-majority middle class for the first time, with increasing technological parity with Silicon Valley and a less carbon-intensive energy landscape, all under the aegis of a stronger Xi Jinping and his vision of governance. Achieving these outcomes will require trade-offs, in this case a China that will likely redouble on domestic priorities and moderate its appetite for global adventurism.

However, doubts persist about the fairness of Chinese competition. An FDD report released a couple of weeks ago says that China “seeks to ‘leapfrog’ the world’s developed countries”, with a special focus on Germany, which is targeted “first as a source of technology, second as a partner through which to export standards favorable to China, and third as a competitor for the lead in the current industrial revolution”.

On the potential effects of Chinese dominance in 5G technologies through Huawei, see the latest essay by Evgeny Morozov on Le Monde Diplomatique. This is my pick of the week and an extremely interesting long read for the weekend.

Antitrust and regulation

EU and German antitrust authorities are thinking about starting an investigation about the potential abuse of market dominance by SAP, the German ERP provider (Politico). Margrethe Vestager also explained her doubts about breaking up tech companies. Meanwhile, in France, advertising companies and publishers have filed an antitrust complaint against Apple due to its upcoming privacy changes that will require “apps to obtain permission from users for data collection” (The Wall Street Journal).

On Wednesday, in the US, the CEOs of Facebook, Google, and Twitter participated in a Senate hearing regarding the influence of their platforms on the political debate of the upcoming elections (The New York Times). However, the core topic (Section 230) was only marginally addressed: “the theatrics, which often devolved into shouting, meant that the topic of the hearing — the future of a legal shield for online platforms — was barely debated”.

The markets

Big tech

Big Tech corporations beat earnings expectations, registering some record results in Q3. In particular, Microsoft reported growth in its Azure (+48%) and Surface (+37%) businesses. Amazon benefited as well from the growth of its cloud division (+29%), which had a positive impact on profits, despite the postponement of Prime Day (CNBC). To sustain these revenues sources, Amazon is heavily investing to expand its data centers: its Capex expenditures more than doubled relative to 2019 and, for the first time, they surpassed that of Alphabet (The Information). For a deeper overview of Big Tech earnings, take a look at this nice recap by The New York Times.

Not-so-big tech

A few days ago, Advanced Micro Devices (AMD) announced the acquisition of Xilinx for $35 billion in an all-stock deal. AMD is motivated to foster its “rapidly growing data center business” and other key product lines such as 5G technologies. This is not only a historical acquisition in the semiconductor industry, but also a major step towards its consolidation after Nvidia’s Arm deal last month. After having closed the gap with Intel in terms of product performances, AMD seems ready to differentiate its offering and gain an edge in rapidly growing sectors.

In addition, I strongly recommend the piece on Palantir by Michael Steinberger in the last issue of The New York Times Magazine. A detailed, objective, and complete overview of the company and its most interesting characteristics.

[…] Palantir’s work on the coronavirus has also highlighted the mistrust that trails the company. In Europe, it is viewed with suspicion because of the C.I.A. connection. But the main source of apprehension is simply the nature of Palantir’s work. Although Palantir claims it does not store or sell client data and has incorporated into its software what it insists are robust privacy controls, those who worry about the sanctity of personal information see Palantir as a particularly malignant avatar of the Big Data revolution.

Artificial Intelligence

A new process called “less than one”-shot (or LO-shot) learning could make an AI model “able to accurately recognize more objects than the number of examples it was trained on” (MIT Technology Review). In other words, this technique would allow researchers to train their algorithms with less data, which could prove fundamental to advance AI applications in areas where data collection is difficult, or particularly expensive. However, reducing the complexity of AI models goes beyond that and includes techniques to compress neural networks through knowledge distillation (The Next Web).

The speculators

Last week, the US Department of Justice accused Google of monopoly power in its search and online advertising businesses. As promised, starting from this issue, I will dedicate some space to the analysis of the downstream effects of a potential end of such a monopoly. In the short term — which means years before any final outcome of the trial — it is easy to forecast lower spending in M&A deals by tech corporations. The total number of Big Tech acquisitions is expected to decline in 2020 (see picture below from CB Insights) to a five year low.

Tech acquisitions were at the core of the House report on competition in digital markets and will play a relevant role during the Google trial. Fearing the consequences of antitrust scrutiny, tech giants might reduce acquisitions even more, which will have huge effects on the whole innovation economy. Four months ago I wrote about the relevance of this intense acquisition activity for new ventures.

The starting point is that Big Tech has more than $450 billion in cash, which makes it extremely resilient in front of exogenous shocks and always capable of making strategic acquisitions of new entrants. This secondary market for tech startups has come to be a fundamental source of exits for many actors in the innovation economy […]. This cash pile makes Big Tech a sort of exit monopoly in times of narrow IPO windows, when digital ventures find it hard to go public.

The reduction of exit opportunities will profoundly reshape the reward system of the innovation economy, making public equity markets even more relevant for founders and investors. Even if the current trend of IPOs and SPAC acquisitions might alleviate the consequences for startups with either relevant growth history or industry-driven attractiveness (like EV startups), in the medium term this will impact venture building, startup valuations, and venture capital investments.

To understand the effects on the latter, it is interesting to analyze the insights of a recent paper by Bushra Samimi. The basic theory of the papes is that VCs push their backed startups to pursue growth paths that naturally lead to acquisition. The excerpt below is taken from the abstract of the paper.

Although most startups intend to eventually go public through an initial public offering (“IPO”), the liquidity pressures from venture capital firms positions startups for an acquisition. Under these circumstances, venture capital policies create economic opportunities for tech giants to swoop in and acquire startups. This escalation in M&A activity in the technology space increases the market power of tech giants such as Google, Facebook, and Amazon. The result is a highly concentrated technology market which stunts innovation, heightens barriers to entry into the market, and reduces consumer welfare.

In turn, VCs face similar pressures as they “fear falling valuations stemming from the tech giants potentially duplicating the startup’s idea or from a fear that the startup will no longer be a hot topic, and thus their goal is to prematurely sell at the highest valuation they can”. To overcome this dependency, I suppose VCs will adapt in three ways:

They will specialize in the most IPO-prone industries to exploit any chance to profit from startups going public. Expect more vertical funds focusing on green and climate technologies, electric vehicles and energy storage, and consumer applications for the low-touch economy.
They will build stricter relationships with non-tech corporations to foster open innovation programs and create corporate VC funds. Yes, it sounds like VC-as-a-service and is not so innovative, but it will make exits easier.
They will differentiate their portfolios to include more mature companies that generate more stable revenues. This is a trend that is already in place among VCs like Andreessen Horowitz, but I expect wider adoption of similar investment strategies.

Meanwhile, Apple is developing an alternative search engine for its own devices, which would make it independent from Google (Financial Times). Both Google and Apple largely benefited from the deal as it “cemented their status at the top of the tech industry’s pecking order” (The New York Times). However, the antitrust threat puts in danger the pact and forces Apple to consider valid alternatives.

The big picture

A recent article on Bloomberg analyzes the evolution of the generational wealth in the US from 1989 up to nowadays. The chart below represents the share of total wealth retained by each generation. The Millennial generation curve is simply shocking.

It’s not unusual for younger age groups to be significantly poorer than their elders. Even so, Millennials remain far behind where previous generations were at the same age. In 1989, when the median Boomer was 34, the generation controlled more than 21% of U.S. wealth. To match that, Millennials, with a median age of 32 now, will need to quadruple their wealth share over the next couple of years.

To prevent a climate catastrophe, research institutions and government agencies are reconsidering a strategy to artificially cool the atmosphere (The New York Times).

That strategy, called solar climate intervention or solar geoengineering, entails reflecting more of the sun’s energy back into space — abruptly reducing global temperatures in a way that mimics the effects of ash clouds spewed by volcanic eruptions. The idea has been derided as a dangerous and illusory fix, one that would encourage people to keep burning fossil fuels while exposing the planet to unexpected and potentially menacing side effects.

Do you study or teach Economics? Then you might like this report by Rethinking Economics.

Thanks for reading.

As always, I am waiting for your opinion on the topics covered in this issue. If you enjoyed reading it, please leave a like (heart button above) and share Reshaped with potentially interested people.