AI strikes back
And also: Big Tech regulation in 2021? Sustainable and not-so-sustainable winners of the pandemic. The Second Cold War: China gets stronger, while EU and US could be allies again. The chip revolution
As people try to look beyond the pandemic, AI becomes a hot topic (again)
DeepMind announced a new “major” breakthrough, with potential impact on drug development: Researchers at the company have trained an AI algorithm that is capable of very effectively predicting how a protein with a given composition will fold. For many analysts this opens the door to exciting developments in drug innovation (FT)
Impact is not yet clear, but even experts that question short term benefits in new medicines believe this is important: Of course, there are many skeptics, because this wouldn’t be the first time we have an overhype on the potential applications of AI to medicine, including at the start of the coronavirus crisis. But a nice post in Medium by a guy who seems to know about this stuff, mentions that even if there could be an overhype on expectations of short-term pharma developments, DeepMind’s breakthrough is still important, because of implications on (1) learning how proteins interact, (2) learning about how folding errors (which cause a large number of illnesses) happen, and (3) enabling “synthetic biology” (i.e. ability to design new proteins, adapted to specific needs) (VishalGulati)(FT)
To moderate excitement, let’s look at self driving cars, where the hype seems to have faded. Two articles at the WSJ this week point to how the big promise of autonomous vehicles, an enormous hype by 2016, has largely faded, with e.g. Apple having laid off more than 200 workers related to this project, and companies like Google (Waymo), Intel (Mobileye) and Nvidia still having very modest (in some cases even declining) revenues from this segment. Connected vehicles (not yet autonomous) seem a much more tangible reality for now. On the other hand, the field is becoming a hot topic among Chinese startups, so let’s see what happens in the coming months / years… (WSJ)(WSJ2)
Some people think that specific components, more than cars themselves, are the key opportunity This week we had news about Luminar, a startup specialized in LiDAR (the laser-based method to measure distances e.g. from an autonomous car), which has just gone public, through its acquisition by a SPAC (aka “blank-cheque company”). The massive valuation (close to $8bn after the public debut), apart from turning the 25-year old founder Austin Russell into a very rich guy, might be showing us something relevant about the self-driving car sector (as the attached blog says): there are far fewer LiDAR companies than car startups, so these few companies could be immensely valuable (TheLastFuturist)
In parallel to all this, a storm has started over Google on AI ethics: We learned this week that Jeff Dean, Google’s head of AI, fired one of the company’s two co-heads in AI ethics, after an ugly public discussion that is still going on in social networks. The person in question, Timnit Gebru, had written a paper on AI-bias that she intended to publish, but an internal review within the company recommended not to go ahead with the plan. As a result, she sent an email with a set of conditions to stay in the company, which were not accepted. That’s the official story. Apart from that, the MIT Tech Review has published an analysis of what they claim were the contents of the banned paper, including the analysis of three key risks created by AI: (1) the environmental effects of AI model training, impacting communities that typically don’t benefit from the models, (2) the risks of algorithmic bias linked to training the models with text taken from the internet, and (3) the opportunity costs from Big Tech companies focusing AI research on short-term opportunities, e.g. language “manipulation”, as opposed to “understanding”. Nothing nice for Google, of course, but also nothing that one would imagine as a cause for employment termination… (FT)(WSJ)(MITTechReview)
2021 could be the year of the big “Big Tech” regulation
The EU is on the way to regulate Big Tech. According to a leaked document, they would be working on an asymmetric system of digital regulations, with more obligations for the large platforms, and looking to facilitate growth for “emerging competitors”. But even this is not clear, because a couple of weeks ago we heard comments (by Margareth Verstager) about “going after fast-growing companies before they are able to achieve market dominance”. Which would point out to potential restrictions to some of the “alternatives” to current Big Tech (FT)(FT2)
These regulatory actions also aim to control online disinformation and promote media plurality, and the obvious way to do this (also under discussion in the US) is to make “Big Tech” platforms responsible for some of the content they carry. Some regulatory “hawks” are already asking form a legislation with “teeth”, which we assume that wouldn’t look good at all for the platforms (FT)
In the US, rumors continue about potential antitrust actions against Google. Facebook, Amazon and Apple: It looks like the Biden victory won’t stop the trend, and everyone is talking about a “bipartisan” agreement on the need to act. Initiatives are proliferating from different areas of the administration, and linked to different potentially anticompetitive behaviors (e.g. Search, Advertising, and Android / Apps for Google) (WSJ)
In China, the case of Ant Group has started the country’s race to regulate its own Big Tech companies. With the massive “invasion” of Chinese people’s personal finances by the app in just 1 year, including loans to 500m people, the government has finally made up its mind and is now on the way to regulate this and other emerging digital industries. The WSJ published a deep dive on the Ant Group case this week (WSJ)
In the short term, Big Tech could just get bigger
Meanwhile, Google and Facebook are (with Amazon) are getting even bigger during the COVID crisis, as ad spending further shifts to digital: More than 50% of the US advertising spending is expected to go to digital platforms this year. As the digital ad market is dominated by just 3 competitors (almost 70% of the market among Google, Facebook and Amazon), it is clear that this is helping Big Tech getting bigger (and creating more regulatory concerns) (WSJ)
And it is not just advertising. As with capital in the past, companies controlling consumers’ attention have plenty of (additional) opportunities to grow. Facebook, looking to create a non-ad-based monetization engine for WhatsApp is an example of this. The central element for the new model will be the app placed as the window through which businesses will interact with their customers (NYTimes)
Facebook just acquired a company to build WhatsApp’s non-ad monetization engine: They just announced the acquisition of Kustomer, a startup specialized in customer-service platforms and chatbots. All these skills could be very useful for what Facebook wants to do with WhatsApp, or at least that seems to be the way they see it, because they’re paying $1bn (not bad…) (WSJ)
Other pandemic winners face the future
Zoom posted another quarter of record sales: The company’s 3Q20 results were (again) pretty impressive, with revenues now almost 5x vs. last year, and beyond analysts’ expectations. The company also raised its guidance, and now expects to reach more than $2.5bn revenues in 2020. They have 433,700 “enterprise” customers (companies with more than 10 employees), 6x vs. last year, and the number of customers paying more than $100,000/year has doubled (WSJ)
But investors are starting to feel “Zoom-fatigue”: Yes, the results looked fantastic, but the valuation was already very high (+600% this year), and investors are worried by what the coming vaccines could do to Zoom, as people could start going back to their offices and their schools, so usage could decrease. So the stock reacted negatively to the results, falling -5% (WSJ)
Also this week, Salesforce confirmed the acquisition of Slack, targeting Microsoft’s space and its position in remote work. But the general perception is that this will be enough to beat them, as a number of features from Teams would still be missing. An example: Teams has emerged as a leading app videoconferencing in the pandemic times, and Slack (or Salesforce) don’t have an alternative. And on top of this there is the fact that Teams is bundled with Office365, making it much more invulnerable (WSJ)(WSJ2)
COVID is accelerating mobile banking in Africa, from just payments to many more things. The continent had already been a global leader in terms of mobile payments, with sub-Saharan Africa concentrating more than half of the total 1.04bn registered mobile money accounts in the World. Now, with the pandemic, mobile phones are also being used to receive loans and government-sponsored social grants. Telecom operators are benefitting from this, and progressively turning into financial institutions (WSJ)
And this is supporting a bright view for the future of the telecom industry in the region. In Nigeria, the government has a plan to expand broadband, as a tool to get 100m citizens out of poverty in the next 10 years. Obviously, with mobile broadband becoming the infrastructure also for financial transactions, the expansion of broadband networks is one of the key priorities, and many consider this government-sponsored plan as a sign that the local telecom industry is a good investment opportunity (FT)
Second Cold War: China gets stronger. EU to ally with the US?
After Biden’s win, EU is offering an alliance to the US, to face the “Chinese challenge” together. After Trump’s efforts to cut the links between the US and Western Europe, the new US administration is seen as an opportunity by the European Union, that now has a plan to start cooperating again, with an agenda that includes a wide spectrum of topics, from the virus fight to the digital “Cold War” against China. Given the latest news from China, this deal would probably be beneficial for both sides of the Atlantic (FT)
China keeps showing evidence that they’re making fast progress in advanced strategic technologies:
Their satellite-based geolocation system Beidou already eclipses GPS in 165 countries: The Chinese satellite navigation system, Beidou, is growing fast out of China, and the push from the Chinese government to expand it is perceived as a sign to create an international “zone of influence” for Chinese technologies, including those with (obvious) military applications. According to an American satellite company, capital cities in 165 out of 195 countries are observed more frequently by Beidou satellites than by GPS (FT)
Chinese scientists just announced a they’ve achieved quantum computing “supremacy”, not much later than Google, and with (apparently) a more powerful system. According to official Chinese media, the new system performs some computations up to 100trn faster than conventional computers, and that it is able to process 10bn times faster than the Google machine that first achieved “supremacy”, last year (Bloomberg)
China is doubling down on its plans to catch up with the US in chip design tools, often leveraging Chinese investments in US chip companies, that could be providing them with valuable information. Many startups being founded in China these days are focused on developing chip design tools, aiming to cover a gap revealed by the US bans, that threatens to slow down the Chinese tech innovation machinery. What is interesting is that in many cases these new startups are being created by executives or are hiring engineers that were previously working for American companies. In a parallel story, Chinese-State backed funds are (still) investing in American chip companies, acquiring stakes that often may give them right to access internal information that could be valuable for China. This could be the next target for the American authorities (FT)(FT2)
The UK is getting more aggressive in its position against Huawei in 5G: Political pressure against Huawei keeps raising in the UK, fostered by a sector of the Conservative party that claims for more aggressive measures against the presence of Huawei in the British 5G networks. There was already a ban against the Chinese vendor’s equipment after 2027, but some operators were expecting to install and use some elements until then, as this would allow them to deploy 5G faster or at lower costs. Now the government is forbidding the installation of Huawei network kits starting Sept 2021, and operators say that this could increase their costs by approx. £2bn (FT)
And in India, China seems to have lost the battle:
Tata is close to acquiring an Alibaba’s Indian partner After the clear message by the Indian government that they’re against Chinese technology companies, they keep leaving the market. The latest example is BigBasket, an online grocery app that was backed by Alibaba, and now is being sold to the local Tata Group, which is expected to acquire a 80% stake (for $1.3bn!) (Bloomberg)
Chinese tech investors turn from India to Indonesia: Another sign of the shift is the way Chinese VCs are moving out of India and increasingly setting their focus in Indonesia, another large market, where investment in the tech sector have grown +55% in 1H2020, to a total of $2.8bn, in part because of this. Indonesia already hosts the largest number of “unicorns” in South East Asia (FT)
Still, globalization might be here to stay: Even if all this landscape looks disappointing, maybe some of the good things of globalization are not going to disappear. A recent report by DHL and NYU’s Stern Business School confirms that the trend towards globalization has turned around, with the index they use (based on a combination of metrics including movements of people, money, goods, and ideas) slipping to the 111-121 range this year, after peaking at 126 in 2017. But at the same time, the report predicts that there will be a rebound next year. The reason? Although people and money flows have slowed, international trade has already started to grow, and ideas (the fourth component) are more global than ever, thanks to digital technologies (FT)
A revolution under way in chips?
Amazon Web Services announced Macs as a Service this week: Amazon Web Services is holding their annual re:Invent event these weeks, in an online format forced by the current circumstances. And one of the most interesting announcements has been this move to make Apple computers available through the cloud, a move clearly targeting the mass of Apple developers that can be interested in building and testing their apps remotely, rather than having to buy an Apple machine for that. The deal also shows how Apple is prepared to renounce to some hardware sales in exchange for strengthening their ecosystem (WSJ)
This can have implications on Intel: Amazon may be doing this just because the Apple developer market is very significant, and they need to keep them with AWS. But the collateral effect is that, as Apple moves away from Intel chips and into Arm, this could affect also the mix within Amazon’s data centers, and have a significant impact on what today is one of the most important markets for intel (hyperscalers’ facilities)(NYTimes)
At the same time, AWS is trying to run fast in quantum computing, where rivals Microsoft and Google had moved ahead. The company has been hiring for a Quantum Hardware Team for some time now, and has even onboarded the mythic John Preskill from CalTech, and they launched their first quantum computing service (Braket) in August. Now Amazon seems to be prepared to build an actual machine, based on a superconductor model, similar to the ones that IBM and Google have been working on (Bloomberg)