“The removal of data, and as a consequence AI monopolies, by more decentralised blockchains is as much a cultural backlash against Googlenomics as Bitcoin was to the financial crisis” — Jamie Burke, Outlier Ventures.
Masayoshi Son, the charismatic leader of Softbank the $100bn venture fund and his 300 year vision of the future has been making headlines but for me it misses a fundamental paradigm shift happening in the Internet today. A change that could render many of his investments vulnerable. In short, the most transformational technology of tomorrow, from data collection to AI, will be increasingly open source, tokenised and highly decentralised.
Softbank’s model takes the logic of the economies of scale and network effects core to Web 2 venture investing to the extreme, focusing on ‘winner takes most’ sectors. Son often opens meetings with ‘’imagine capital were no longer a problem’’ enabling the team to prioritise aggressive growth over profitability.
If we look at Softbank’s portfolio they focused on a handful of synergistic categories such as: multi-sided marketplaces, real estate (offices, storage, robotics), e-commerce and logistics (last mile delivery). Often these investments are based on one primary premise: they are ‘AI optimised’ companies that at scale, and with a data advantage, can better optimise utilisation and capacity to outcompete on the fundamental economic unit costs of their market over the long-term.
To support this thesis they are also investing in hardware companies like ARM and Nvidia that are making machine learning easier, faster and cheaper as well as software startups like Improbable, which provide a virtual environment to support the development of complex digital worlds. Interestingly, with Improbable, they have recognised that the gamification of networked problem solving may be the best way to parallelise machine learning at scale. And it’s here there is a hint at what may await them.
This story starts with a shift of compute power away from cloud to open source internet protocols that enable the same benefits of scale but through distributed marketplaces. The most famous example being the Bitcoin blockchain: where billions in resources, including specialist hardware and energy have emerged ‘bottom-up’ coordinated by a simple algorithm and reward system. However, this was just the first example. Where Bitcoin mining was focused on mobilising hardware to solve a very particular mathematical game there are many other distributed systems emerging where parties commit resources to solve a variety of problems in exchange for digital rewards.
In our own portfolio we have Haja Networks (enabling interoperability between traditional database silos), Ocean Protocol (for that data to be sold through decentralised data marketplaces to help train algorithms) and Fetch.AI (where those algorithms can gain economic agency and sell their value to others). When combined they provide an emerging stack of technologies that level the playing field for startups by bringing economies of scale for a longtail of market participants. This means many of the old advantages that underpin Son’s vision such as data monopolies and centralised IT infrastructure will be available to all through these new ‘global public utilities’.
The advantages of proprietary datasets, their algorithms, and in-house talent will gradually diminish as open networks offer an attractive alternative. As governments push for sovereignty of data and users gradually respond to the incentivises of tokenised networks these new protocols it will challenge Son’s ‘winner takes most’ platforms.
If these trends are not responded to by Softbank’s portfolio then the dream of building sustainable companies with any meaningful AI advantage is under threat, and so too are their valuations. Uber will become a company nobody chooses to drive for, Doordash a company that cannot recruit couriers and WeWork a heavily indebted commercial real estate company.
In fact, the very economic model of Silicon Valley’s freewheeling ‘move fast and break stuff’ capitalism, at the heart of Son’s vision, is itself falling out of vogue as eloquently described in: ‘The Age of Surveillance Capitalism’. The idea of trusting platforms like Uber to develop benevolent AI to extract and exploit the ‘’digital exhaust’’ of our interactions is quickly becoming unacceptable and dated.
Well, given they have already deployed 40% of their $100bn the best they can do is hedge at least some capital into the counter-narrative. However, what’s unique about these new protocols is because they are open source they aren’t as vulnerable to the typical response by incumbents: buy, assimilate or kill.
Currently the ‘crypto’ market is 90% down from it’s 2017 all-time highs and furthermore, many next-generation protocols are only just coming to market, this time with bear market valuations. One thing that is in short supply right now in crypto is patient capital.
So, If Son repeated what he did so well after the Dot Com bubble, with investments into Yahoo and Alibaba, he could ‘buy the dip’. Theoretically, Softbank could lead the investment round of every threatening new protocol coming to market or, more importantly, those which compliment its portfolio.
Why? Because all these protocols are looking for adoption to effectively turn their technology into internet standards. If Softbank’s portfolio are the early adopters of these technologies they can adapt and protect their multi-sided marketplace cash cows secured by ubiquity of service, brand and network effects..
Bringing that much capital into the market combined with real-world adoption, would see the mother of all bull runs, with the potential for a trillion dollar crypto industry.