And here is the long form update on all the themes we track and the lessons we've learned this year.
Digital Wealth Management (Roboadvice)
- Angel and small Series-A funding into B2C roboadvisors will be pretty much $0. Series B funding into digital wealth management companies more broadly will be a little more than half that of 2016.
- One digital wealth company from the second generation, like Acorns, Stash or Digit, will get bought for a ridiculous amount by an unlikely and surprising bidder.
- Almost all roboadvice assets will be (by %) at Vanguard, Schwab, and Bank of America.
- Everyone will start saying "artificial intelligence" instead of "roboadvice" but none of the business fundamental will have changed
The higher granularity than in the top 3 predictions makes us more wrong. Indeed most roboadvisor funding and interest shifted to micro-investing services (e.g., Stash, Acorns) and financial chatbots (e.g., Earnest, Trim, Cleo). Instead of AI for roboadvice, people do say micro-investing or chatbots. The underlying business change, however, is that those services have now customers/users numbering in the millions. That implies very small accounts (about $1,000) at very large scale. And nobody has been bought just yet at a massive price.
Blockchain and Digital Ledgers
- Bitcoin, the technology and not the $ value of the digital currency, will be relevant again, and will see at least one multi-billion dollar company leverage its public blockchain
- Big banks will try to patent troll their vertical solutions on top of the open source projects (Hyperledger, etc.) and will feel a sense of false confidence. It will be the year of lawsuits between the incumbent and startups communities, and open source will win.
- A blockhain-based production ready system,in either trading or money movement, will replace the legacy tech stack of a multi-billion dollar firm. It will actually increase costs in that year, and observers will draw the wrong conclusions even though the firm is more competitive in the long run.
Still feeling pretty good here. While the Bitcoin $ value did become extremely important (climbing towards $20,000 at one point as futures on CBOE were about to launch), the technology of the blockchain has catalyzed over $4 billion in Ethereum-based token launches and thousands of related startups. Banks certainly are trying to patent their enterprise blockchain solutions, but that is unlikely to matter in the long run we think as open source code bases grow exponentially.
- Intel and AMD will keep making hadware chips optimizing for neural networks, and will drive the execution speed of image recognition and similar tasks to be 10x faster than what we have today
- Real time video face recognition and editing will become a consumer toy. The first instance of someone releasing a viral video impersonating a politician using face-overlay and voice-replacement technology will get a massive negative backlash from conservative thinkers
- A political movement against unemployment caused by self-driving cars, chatbots, and other AI products will rise, but it won't be able to articulate its concerns in a way our political system understands or can address
It wasn't Intel and AMD, but NVIDIA that experienced the most growth resulting from their neural network hardware and developer resources. There is an arms race in determining who will power all our self-driving cars, trucks and augmented-reality devices. That race is not yet over. In terms of image editing and forgery of people's visages, check outFaceApp, which allows you to morph gender, age and other variables in real time. And as this Verge story describes, such algorithms ("deepfakes") are now being applied to adult content. This dangerous technology has not yet hit the mainstream and crossed wires with fake news and propaganda bots, but we have a better idea than ever just how dangerous AI-based communication has become, and the political effects it has had.
Neobanks & Challenger Banks
- Incumbent banks will wake up and come down extremely hard on challengers and neobanks. They will fight them on mobile apps, chatbots, instant payments, and user experience. You won't be able to tell apart the website designs of new and old firms.
- Startups will try to do everything to become financial supermarkets. Expect to see lending, payments, banking, data aggregation and insurance combined in an attempt to grab at least some consumer attention. Regardless, neobanks will struggle getting to any scale beyond 10,000 early adopters.
We were right that all established banks will trend towards becoming neobanks, with native mobile apps, chatbots and digital capability. One example of this is payments app Zelle vs PayPal owned Venmo. We also got the trend of unicorn Fintech startups becoming financial markets generally right (though it was easy) -- see SoFi, N26, Paypal/Acorns, or across to China as Ant, Baidu and Tencent build the Fintech of the future. We were wrong on the magnitudes. Incumbents aren't particularly scared of Atom or Monzo, as the tech banks struggle to get licensed or funded. The real story in licensing is the ability to hold deposits to lend out -- you know, to have a business model. The other part we missed is the scale of early adopters. Revolut has a million customers, as does Transferwise. You can say that those are international payments apps, to which we say, yes, you mean global banks?
Financial APIs & Banks as a Service
- In a watershed moment, every single large retail and corporate bank will have a Developer portal with open API keys. However, most of those APIs will do nothing other than pull information. Actionable APIs will be too expensive to use, but Venture Capital firms will overfund companies that tell a story about how to tap into them.
- BBVA and Santander will retain their position as marketing leaders of a BaaS offering, and we will start to see apps using their infrastructure.
- Someone will write a malicious app that crashes one of the API stores, generating tons of press and cyber security premiums.
APIs are coming not with a bang, but with a whimper. 2018 will be the year where APIs really get test and implemented by incumbents, so we were off here. BBVA and Santander indeed do maintain their marketing positions, reinforced in a quantitative study we've done on bank innovation. But JP Morgan and Barclays also matter. And the biggest security breach of the year wasn't from these APIs, but from horrible security procedures at Equifax, exposing over 140 million Americans.
Chatbots, Conversational Interfaces
- Virtual assistants will start to appear in business locations, from banking branches to shopping experiences. People in customer service from a major brand will be laid off.
- Amazon Echo, Google Siri and Google Home will start a nuclear war over the connected smart home. Like with its other products (Kindle, Fire), Amazon will lose its lead, especially as self-driving cars develop their own AI assistants.
- Facebook will get into the hardware business through an acquisition, and Messenger/Whatsapp will gain a physical form
Hmm, maybe too much science fiction juice in this one. 2017 did see millions of hardware sales of virtual assistants -- Google Home vsAmazon Alexa has Amazon in the lead; Apple HomePod is nowhere. But anthropomorphic AIs are still in early development. As an individual data point, the incredible human renders from Soul Machines are starting to emerge from the uncanny valley. Not so much for the robotic Sophia, who despite being granted citizenship in Saudi Arabia is still a pretty creepy robot. And we're not aware of any hardware acquisitions yet to give the brand-name AIs tangible form. Amazon did buy Body Labswhich does 3D scans of human bodies, but that's probably to sell clothes in Augmented Reality.
Democratization, Regulation and Crowds
- As consumer protection is rolled back, equity and real estate crowdfunding experience major scams that lead to a public backlash.
- Crowdfunding technology ends up creating asymmetric benefits for those already in power (think about who can really use Syndicate investing on Angellist, stock earning estimates from Estimize, or invest in Numerai's crowdfunded AI hedge fund). Income inequality becomes sharper despite equality of access to cheap investments.
This happened in a surprising but big way with the crypto economy. Consumer protection did not exist at all as Initial Coin Offerings raised increasingly more capital in 2017, fulfilling on the original promise of crowdfunding. The year ended with over $4 billion in ICO funding. Scams and hacks proliferated, from the $170 million Parity failure, to fraudulent ICOs being shut down in China, to the hundreds of millions of crypto locked up in the Tezos litigation. And of course, the story would not be complete without the huge concentration of mining power in Bitcoin, the 170 crypto funds trying to profit from the opportunity, the billions "created" in the Bitcoin Cash fork, and the general distribution of wealth in favor of crypto whales. Black swan outcome.
Bitcoin & Cryptocurrency
- 10 countries will have issued central-bank backed cryptocurrencies, and Americans will still not understand why that's important
- Another massive Initial Coin Offering like that of the DAO, over $50 million, will get people's heads shaking again. This one won't get hacked. Traditional Venture Capital firms will have invested 10% of the amount.
Again on the ball here, especially around ICO funding. Think about Filecoin, the CoinList powered SAFT launch where venture investors put in $50 million at $0.70 per coin and regular investors bought $200 million worth at $2.50 per coin (give or take). That's a 20% allocation of the overall raise to private investors, and a massive uptake by the public. More generally, ICOs are a 10x increase on all annual venture investment into blockchain and Bitcoin companies.
- Venture investment will continue to pour in, increasing by over 50% globally.
- There will be a major exit at a price point comfortably above $100 million to a legacy insurance company
- A scandal, either on the underwriting or regulatory side, like the licensing issue with Zenefits, will hit an insurtech startup and lead to renewed finger-wagging
According to Coverager, venture investment in Insurtech sat at $4 billion in both 2016 and 2017, so the theme is not seeing increased YoY activity. But it's not falling off either. In terms of acquisitions, there is healthy appetite in the sector to modernize and own software assets, with over$10 billion being spent on M&A. One example that fits the bill isTravelers buying Simply Business (online broker for small business insurance) for $490 million. CVS acquiring Aetna is certainly a big deal, but that has more to do with the state of healthcare in the US than artificially intelligent drones powered by machine vision. There is also something to be said for China's Ping An Insurance going public at a $200 billion valuation, but it is a stretch to tie that to the predictions above.
And on the scandal point we did get it directionally right. SoFi, which does have an insurance offering via partnership, got into the news on account of the CEO's sexual harassment allegations, which led to his resignation amid the broader transformation of our society with the #MeToo movement.
- Generation Z will become a bigger buzzword than Millennials. Banks will wonder whether then Snapchat generation even knows if banks exist.
- Valuations in the Gig economy (AirBnB, Uber, Task Rabbit) will crater, due to poor exit opportunities in the public markets and a churning contractor force dissatisfied with the lack of stability and benefits
So Gen Z is definitely on the radar, and includes everyone born after around the year 2000. But Millennials still reign. The brightest example is Olaf Carlson-Wee, the 27 years crypto-king of Polychain Capital. And the gig economy is still around. Despite setback for companies likeUber in London, most private unicorns are continuing their journey, and the crypto economy in particular is adding more steam to pinballing valuations. The perceived failure of Blue Apron's IPO (falling from a $2 billion offering to $800 million), however, does create a cautionary example.
- Snapchat holds on to its lead with younger demographics by understanding its customers better than anyone else, resulting in a flashy public exit that leads to multiple new businesses being started in the Los Angeles ecosystem, many of them in VR film
- One of Google, Apple, Facebook, Microsoft and Amazon will get into financial services in a way that makes financial incumbents extremely anxious, and will create public outcry. Financials will fail to understand why the tech firms are pursuing what seems like a dead end, but is really a way to engage with users and get more data
Snapchat certainly did go public, at $24 billion, but has since fallen to $18 billion. But unlike the GAFA tech giants, it has not been perceived as an artificial intelligence leader, and its bet on Spectacles has not gone the way it wanted. But the big tech firms have indeed gone further into finance -- from Amazon's SME lending, to Apple and Facebook messaging payments, to Google's machine learning and financial cloud. We would highlight Amazon's digital lending moves as particularly threatening to incumbents, given their higher quality data for automated underwriting.
Virtual and Augmented Reality
- Millions of virtual reality headsets will hit homes across the United States, leading to an explosion of VR apps powered by Valve's Steam store, Playstation, and Google.
- A multi-billion dollar investment firm will open a virtual reality branch with live customer service, potentially on Google Earth or another VR world
- Payments in VR will go live in Asia first, and will set the default behavior for the rest of us to follow. Credit card networks will open VR incubators or invest in VR companies.
- Augmented Reality will still be mostly unknown to consumers, but will see multi-million dollar contracts between enterprise clients
This is one is almost right, helped by Apple's latest hardware. Millions of VR headsets have indeed been sold to American consumers, but the general sense is that VR is not ready for mass adoption. Augmented Reality, however, has absolutely landed in the hands of consumers in the form of iPhone X, which comes with a ARkit, a development kit of software engineers. Ikea, Amazon and other retail apps are thinking about how to implement AR commerce, which we continue to think will drive payments activity. And of course, don't forget Alibaba's $20 billion revenue "singles day" shopping spree, in part powered by this technology.
Internet of Things & Wearables (IoT)
- Unsexy businesses in the "old economy" of manufacturing physical objects will position themselves as technology companies that generate terabytes of proprietary data. Hedge funds and other investors gladly pay for that data, repeating the experience of expert networks.
- Self-driving cars continue to beat performance expectations and end up in production-mode on roads all across the world, leading to unemployment and outrage. Governments, trying to stem the bleeding, file lawsuits against tech companies for breaking regulations. It doesn't work.
Broadly speaking, Alternative Data is a big growth area with many companies looking for ways to monetize their internal data exhaust. But the role of old manufacturers is perhaps not as drastically helpful as we had hoped (yet). As for self driving cars, we got it backwards. For some reason, the US House of Representatives is in favor of looser regulations for self-driving cars, and had passed a bill as such. Perhaps the auto industry as a whole, and not just Tesla, is seeing this as a way to boost its fortunes. And the cars aren't quite in production mode at scale, but may hit the roads by 2020.
- Human addiction to technology will reach a new height. People will spend more than 12 hours a day on their screens.
- The cost of sequencing the genome for an anonymous consumer falls to $25. Genome data can be made available between services via API.
- 3D printed organs will be implanted into human patients successfully.
- Scientists will claim they have simulated a full rat brain, which will become easier due to newly available hardware. There will be no implication in the physical world from this discovery.
Yep. People now spend 12 hours per day on media. The cost of sequencing a genome is somewhere between a $1,000 and $70, depending on granularity. Organs are not routinely 3D printed yet, but a Wisconsin company offered to implant NFC chips into its employees. And no, there are no stimulated rat brains yet -- but here's a worm brain in a robot body and optical neural nets on a silicon chip.