The State of Blockchain Infrastructure
I believe that the biggest barrier to blockchain adoption is poor infrastructure. There are other huge challenges — lack of regulatory clarity and poor UX among them. But to get buy-in from policy makers, the industry needs to demonstrate that the technology can deliver positive use cases. And for applications of the technology to delight users, the underlying infrastructure needs enable developers to more easily create great products. It all starts with the infrastructure.
Web3 infrastructure (by which I mean all the software and tools that enable the development and delivery of end user applications) is lightyears behind Web2 cloud computing infrastructure. In Web2, AWS, Microsoft Azure, and GoogleCloud deliver powerful and resilient compute, storage, and networking to enterprises all over the world. By contrast, blockchain networks are often slow, low throughput, and unstable. Furthermore, the leading cloud providers and many other Platform-as-a-Service companies also provide a vast array of sophisticated, battle-tested development tools and middleware that help developers access the underlying infrastructure. Blockchain application developers often need to build complex, non-core solutions that would be available off-the-shelf in the traditional cloud world.
All this means that only the most blockchain-obsessed developers have the expertise to build well-functioning blockchain applications, and only those deeply committed to the mission of decentralization are motivated to try. The key to blockchain’s next phase of growth is reducing cognitive and technical roadblocks for new blockchain developers — and therein lies a big opportunity for industry startups.
Blockchain Networks
The first category of blockchain infrastructure that needs improvement is the base-layer crypto networks themselves. Today’s biggest blockchains, Bitcoin and Ethereum, prioritize data integrity and broad consensus, which is extremely expensive in a highly distributed network — thus, these networks have high latency and low throughput. Many projects are trying to overcome these limitations with new consensus algorithms that offer at least some of the security and decentralization of Bitcoin and Ethereum. Some are trying to solve the scalability problem without sacrificing security, while others are willing to trade-off some security for scalability.
There are many horses in this race, including the next generation Ethereum, Algorand, Blockstack, Cosmos, Hedera, Polkadot, and more. They face many technical and mathematical challenges, but even those that architect theoretically flawless networks face the greater challenges of going to market.
Because there is so much competition among general smart contract platforms, I am particularly excited about more domain-specific chains. Dapper Labs’ new Flow protocol, which is built for gaming and social use cases, was born out of the company’s direct experience crashing the Ethereum network with a popular app. Flow favors a distribution of labor amongst different types of nodes, rather than sharding, because the latter technique reduces the real-time interactivity amongst otherwise interoperable apps. It is also a prime example of an app developer solving its own infrastructure problems, and I believe their firsthand encounter with Ethereum’s constraints gives them a big advantage in architecting the right network for their use case. Figure’s Provenance, a private, permissioned distributed ledger network with a native cryptocurrency called Hash, is seeking to digitize trillions of dollars in financial assets. It is less open and more controlled than other crypto networks, which will help major financial institutions get comfortable issuing securities on the platform. These use-case specific projects know what to optimize for, and also benefit from more focused go-to-markets.
It may be that trade-offs at the base layer cannot be fully reconciled, however, and there is a growing recognition that not every state change needs to be recorded as an on-chain transaction. Thus, companies providing second-layer scaling solutions are running important experiments. The most prominent layer-two technology is Bitcoin’s Lightning Network, which is still quite nascent but has strong momentum. DCG portfolio company Lightning Labs is the driving force behind this, while DCG portfolio company Radar is also making it vastly easier to interact with the network, with an App Store that allows a user to connect to Radar’s Lightning node just once, instead of re-connecting to every app. Meanwhile, projects like Loom Network are offering SDK’s to spin up more sidechains (which offer greater speed and throughput) on smart contract platforms like Ethereum. It is early days for these layer-two solutions, but they may play a critical role in scaling blockchain networks.
Additionally, many application developers are tackling scalability challenges themselves, and then open sourcing their code for others to use. For example, Livepeer developed a sophisticated probabalistic micropayment protocol, and gave the code to the community. Filecoin is committed to enabling a network that can process millions (or even billions) of transactions per second, through a combination of innovations in their consensus protocol, off-chain solutions, sharding, and hierarchy. These innovations can be re-used and stitched together to build other protocols and applications.
Blockchain Middleware and Dev Tools
The second category of critical blockchain infrastructure includes any software that makes it easier for developers to build on top of the networks. This is akin to Platform-as-a-Service in today’s cloud market. The protocols themselves play a critical role in the developer experience — each one is responsible for cultivating its own developer community. Those that prioritize great documentation, tooling, and APIs will have an edge, and it is perhaps more important that protocols use paradigms familiar to developers. Parity, the company behind Polkadot, is aiming to do all of the above with its Substrate framework.
And third-party service providers are springing up to provide critical solutions, particularly around node infrastructure — it needs to be easier for developers to access the underlying ledgers. Startups such Alchemy, Blockdaemon, Infura, are providing enhanced API’s so developers can easily deploy smart contracts and read/write to the blockchain. Two DCG companies just entered this market with exciting new products. The aforementioned Radar is launching Deploy, an infrastructure service provider born directly from the company’s experience building dApps. And Terminal recently announced a testing, logging, deployment, and monitoring service for dApp developers. The goal of these companies is to provide tools that most developers do not have the time, energy, or expertise to build in-house.
Furthermore, once a project is in the wild, there’s a critical feedback loop that is often missing — protocol and dApp developers rarely understand the impact of off-chain activity on on-chain activity. Several blockchain analytics startups, including Flipside and Scout, are building solutions that allow developers to translate abstruse on-chain data into human-readable information. By understanding the impact of their actions, projects can more effectively test hypotheses, adjust their strategies, and drive adoption.
These crypto PaaS companies have not gotten much attention, but they will dramatically accelerate the adoption of blockchain tech by making it vastly easier for developers to experiment. And experimentation is the key — more shots will lead to more hits, and even the failures will provide key learnings.
Conclusion
The blockchain industry is in the very early days of building new financial infrastructure and an entirely new web stack on top of open source, permissionless networks, from the ground up — this is what makes the space so exciting. This is a massive undertaking, and progress in the space might look slow to casual observers. But the infrastructure is maturing, and as it comes of age, many of revolutionary blockchain use cases that people have talked about (and many more that nobody has yet thought of) will come to fruition.
Thank you to Avichal Garg of Electric Capital, Alex Lines of Notation Capital, James Hutchins of North Island, Alan Curtis of Radar, Doug Petkanics of Livepeer, and Jim Myers of Flipside Crypto for your extremely helpful thoughts and feedback.
