Disclaimer: This is not financial advice. Anything stated in this article is for informational purposes only and should not be relied upon as a basis for investment decisions. Triton may maintain positions in any of the assets or projects discussed on this website.
Summary:
One of the most common hurdles that investors with deep experience in ‘traditional’ assets encounter is getting comfortable with how digital assets are valued and the associated return potentials. You’ve likely heard the go-to refrain that Bitcoin should be avoided as an investment and has zero intangible value because “there are no cash flows” or “Bitcoin is not a productive asset” or “there is no place for Bitcoin in a buy-and-hold portfolio”. While this stance may have been (arguably) valid in the early days of development 15 years ago, it is sorely outdated today and worse, has prevented investors who adopted that stance from having exposure to the best-performing asset class of the past decade. Shown as cumulative returns over that time, the opportunity cost of that outdated advice really hits home. Even had an investor allocated immediately before BTC’s massive drawdowns in 2014 and 2018 (e.g. invested December 2013 and December 2017), they’d still have outperformed equities by year end 2023, considerably.
We’ll touch on these points, and more, as we present a refreshed view on how to approach and evaluate digital assets with a more current perspective.
How much is a message worth?
In 1792, it cost roughly $0.20 to send a letter 300 miles in the US, or $6.47 adjusted for inflation. By 1855, this had come down to $0.02 ($1.07) and today costs just $0.68. The time for the letters to be delivered, measured in weeks then, now down to days. The Pony Express tried to improve upon this service in 1860 and offered trans-continental mail delivery within 10 days, at a cost 250x higher than that of the traditional mail service, before ultimately going bankrupt due to the establishment of the cross-country telegram network.
In 1850, the cost of sending a telegram from NYC to Chicago was $1.55, or $61.67 adjusted for inflation. By 1890 that had dropped to $0.40 ($13.64). Though state of the art, this was still highly inefficient and burdensome, giving way to the telephone. The cost of a 3 minute phone call in 1902? $5.45 ($196.66), down to $1.45 ($15.20) in 1960 and essentially free today (if you ignore the $40-per-month network access plan primarily for data). Phone networks relied on manual operators until the mid-1900s, and in the US, it was not until the 1950s that over 50% of households had a telephone. Less than 25% of Americans have landlines today as mobile phones have become the norm.
And now with email, the marginal cost of sending a message is free, assuming access to the internet and a willingness to be subjected to ads.
The fact that technological improvements and the migration to digital services drive the price of goods and services down while increasing efficiency is no secret. But yet, even though the marginal revenue per digital message today is $0, the companies that offer these services are some of the most valuable in the world, such as Google, Meta, TikTok and soon to join the club, OpenAI. At their core, what do each of these companies actually do? Meta curates internet connections and communities; TikTok curates attention; Google curates information; and OpenAI curates information, but in a more useful way.
Where is the true value accrual in these? Is it in the network, usage, and flows? Or is the value in the computers, kombucha bars, and pricey California real estate? Of course, there might be residual value in these ‘real’ assets in case the company goes bankrupt, but otherwise, they are largely major cost sinks and overhead expenses required to maintain the network and services these companies provide. Internet-native companies are fundamentally different from traditional widget-production companies. Former Y-Combinator president and OpenAI’s CEO Sam Altman predicts we’ll soon see the first $1B company run by just 1 person.
Financial networks are no different. Visa and Mastercard are brick-and-mortar OpCos built to facilitate their payment networks. Importantly, they do not actually move value or settle transactions themselves, but rather communicate money flows between financial institutions, ultimately executed in batches by banks at some later point in time. Put another way, Visa and Mastercard curate financial transaction information, coordinating transactions, settlement and clearing. Clearly this is extremely valuable: Visa and Mastercard are together worth $1 trillion.
Quick thought experiment: assuming the value of Visa’s network is accurately reflected by how much entities are willing to pay them to use it (e.g. revenues), how much more valuable would Visa be if it maintained that network and functionality but reduced its 2024 OpEx of $10.9B to just the costs associated with maintaining the network, or $736M? At 32 P/E, roughly $350B more valuable, or $900B.
This is exactly what blockchain networks such as Bitcoin do. They decouple these payment networks from the brick-and-mortar OpCos weighing them down.
In fact, they even expand upon what existing payment networks are capable of by enshrining the clearing, settlement, fee payments and fraud prevention directly into a true value transfer network. Whereas Visa relies on multiple banks and other service providers to facilitate a single transaction between 2 entities, Bitcoin requires only those 2 entities and the open source Bitcoin codebase.(* - see footnote). That’s it. Even today, Bitcoin is considered a dinosaur with its 10-minute block times, but compared to the T+3 settlement of Visa, it is absolute hyper speed. With other major blockchains such as Solana and Ethereum L2s, transaction times are now easily sub-1-second, which is why Visa is building on Solana in an attempt to stave off disintermediation. As for fees? As we highlighted in our recent post, median fees to execute a transaction are now just a fraction of a cent, regardless of the value moved, and far cheaper than the 3% credit card network charges.
Further, blockchains turn value transfer networks into platforms. Ethereum is itself a value transfer network, but also enables entire applications to be built on top of it, natively incorporating payment functionality directly into any program utilizing the Ethereum Virtual Machine (EVM). Currently, there is $120B in value among the applications on Ethereum alone, with roughly another $60B on other blockchains. Equivalently, imagine if Apple’s App Store was built entirely on top of Visa’s network. How much more value would that add to Visa? But more importantly, what else could that enable that does not exist today?
In this new model, every company shares a value transfer network allowing for complete interoperability with each other. Put another way, Venmo users in the US could send value directly to Google Pay users in the UAE or AliPay wallets in China, instantly, who from the same wallet can use those funds in their new AAA video game to purchase a skin. Expanding even further, a Canadian could send $20,000 back to their family in Africa instantly and for near-zero cost. Today, that would take multiple days and cost hundreds of dollars in platform and exchange fees through traditional remittance rails. Expanding even further, banks and other commercial entities can conduct wholesale settlement via blockchain rails, in effect doing what SWIFT has been set up to do. In fact, SWIFT is actively exploring how to use blockchains to facilitate the movement of tokenized assets, leveraging the Chainlink oracle protocol.
So, do we think the entire world’s financial infrastructure will move on-chain? Definitely not, and definitely not overnight. Technological progress takes time, and financial institutions are famously slow and risk-adverse when adopting new technologies (rightfully so). But we are also adherents to Amara’s Law: “One usually overestimates what can be achieved in a year but underestimates what can be achieved in 10”, especially because technological penetration to critical mass has happened far more frequently on a 20+-year time horizon.
Bitcoin itself is barely 15 years old, and Ethereum less than 10. Back in 2015 advocates were loud about how these new blockchains were going to change the world, but by 2016, did things really look that much different? No. But fast forward to today and millions of active users on Ethereum have engaged in over $4 trillion in trading of just ETH in the past year, with billions of dollars worth of projects built on top of it in the forms of layer 2 chains and applications. And yes, they generate staggering levels of cash flows for such small teams: individual projects with a handful of developers facilitate billions-of-dollars per day in activity and generate over $200M per year in protocol cash flows with high operating margins and essentially zero PPE. At launch in 2015, Ethereum had a market cap of just $80M, and today is worth $400B. Where will we be in another 10 years?
*A few arguments typically come up here applying this analogy to Bitcoin, primarily 1) miners’ costs are real and should be included in the cost to upkeep the network, and b) new token issuance paid to miners for their work is a cost to the network. Re: 1) correct-ish, but those are resources spent by 3rd party service providers. Much like you don’t include Bank of America’s interest expense in Visa’s OpEx, you don’t include a miner’s costs in the cost to run the Bitcoin network, and if you do, that brings you to 2) the network’s costs are the block rewards that go to miners. Currently, this inflation is 1.6%, thus, BTC network’s operating margin is effectively 98.4% - this is the one single cost the network incurs, in the form of payments to its service providers, the miners, and covers the entirety of the service they provide to the network.
