Admit it - you aren’t going to bring the “next billion people” onchain by telling them about the latest Arbitrum perp DEX that allows for automatic cross-leveraging of Liquid Staking Tokens. Non-crypto people don’t care about the behind-the-scenes tech or the incremental improvements in capital efficiency we all seem to crave.
Do you think STEPN breached $2b in market cap because of its “novel tech” and “sound tokenomics?” NO. It exploded because your buddy (who wouldn’t know his seed phrase from a bowl of alphabet soup) could make money during his evening stroll.
So, how do we onboard new people into crypto? You make it applicable to their everyday lives. By connecting the dots from the offchain to the onchain (and doing it in a way that improves processes), everyday people might someday understand that our shared obsession is not just about speculating on dog coins.
Why Solana? Why DePIN? (Also… WTF is DePIN?)
As much as it pains the research analyst in me, as usual, price generally drives fundamentals in these markets (and not the other way around). Now that SOL has risen from the ashes and has put in impressive YoY returns, people are finally admitting to themselves that maybe Solana does have something to offer the crypto ecosystem.
Even when compared to Ethereum, Solana’s advancements have been quite impressive.
- Their parallel computing allows for simultaneous transactions to occur on the network if these transactions are wholly separate from one another.
- Also, Solana’s isolated fee markets mean that a hugely popular NFT mint on the network won’t interrupt DeFi activity. Have you ever tried to make a Uniswap trade on Ethereum when there is a hot mint or airdrop claim going on that day? Good luck.
Even taking a step back from these updates, if you were to personally onboard a friend into crypto these days, where would you start?
If you pick Ethereum - they aren’t going to use a network that costs $20 (if you’re lucky) every time you want to swap a token.
If you pick a Layer 2… which one do you pick? There seems to be new ones popping up each week. Liquidity fragmentation is still an issue, with solutions like cross-chain messaging protocols just ramping up. What if we want to onboard users today?
For me, Solana makes sense for crypto newbies due to its low fees, incredibly-fast speeds, and focus on User Experience (UX). And it is also for these reasons that retail-focused DePIN projects are flocking to the network.
Decentralized Physical Infrastructure Networks (DePIN) are organizational protocols that connect a group of consumers to a group of providers. Imagine them as the middlemen that connect users to the resources they need.
Solana’s first big step towards connecting the digital world to reality came with the Solana phone (“Saga”). Designed to be a seed vault with easy access to Web3 apps (that also can make phone calls), the Solana phone has now become a hot commodity given how many Solana-native projects have airdropped to buyers. Many of the protocols listed in this report build off the back of this connection Solana has with practical, real-world use cases - if Ethereum is science, Solana is engineering.
DePIN is a huge umbrella term, and in this report, we are going to dive into projects that attempt to build marketplaces for:
- Computing / AI
- Price feeds
- File storage
In all of these cases, the role of the protocol is to provide a network that simplifies the experience of users on either side of the exchange. The blockchain takes on the role of a mediator, connecting demand with supply of real-life infrastructure.
Distributed computing and AI
Do you remember when Ethereum’s Merge was right around the corner and people were excited that high-end GPUs would once again become obtainable for the masses? Well…
Demand for computing power is at an all-time high, and with computationally-intensive tech like AI on everyone’s minds, my bet is that this won’t change any time soon.
A major issue with the uneven distribution of these resources is that, often, users seeking computing power find it difficult to obtain access to these machines. Orders can take months to fill and data centers require intense planning and execution to stand up. DePIN offers a decentralized marketplace for this trade to occur - where hardware owners can lend out computational power to users looking to leverage it.
Centralized options exist, such as those from large cloud providers (think Amazon Web Services), but decentralized options could provide a better experience, security, and price point for both users and providers.
The elephant in the room is Akash - a decentralized computing provider built on Cosmos, which serves generalized computing power (meaning it doesn’t specify how you have to use their network of CPUs/GPUs). Akash is incumbent in this space, with ~60 active (generally, more professional) providers, capable of providing almost 7,000 CPUs and 200 GPUs all for substantially lower prices than their centralized competitors. Often, other DePIN protocols have opted to specialize their offerings, instead catering to AI applications, 3D rendering, or other specific use cases for decentralized compute.
Nosana - a GPU grid for AI applications
First on our list is Nosana, connecting a grid of user-sourced GPUs to consumers looking to build AI products. People or companies with idle consumer GPUs (such as NVIDIA 40x0 and 30x0 series) can become a node in the Nosana network. On the other side of the equation, builders looking to play with AI tech can use Nosana’s in-house AI Interface, fully kitted with Stable Diffusion and Llama 2 workloads.
Nosana boasts cost savings to developers of up to 85% when compared to traditional cloud infrastructure providers. This claim is set to be tested with their initial testnet having just gone live in December. As part of their Genesis phase, node operators can put their idle GPUs to use to earn some early $NOS tokens and help test out the network.
Their main grid is scheduled to launch in April/May of 2024 with their Galactica upgrade, along with handy development toolkits to help encourage tinkerers to play around with Nosana’s network.
Their token, $NOS, was released in January of 2022 and is used as staking collateral for node operators. In the case of a node operator misbehaving, they risk their $NOS being slashed. The token is also used as a medium of exchange within the Nosana ecosystem, where consumers can pay for jobs with $NOS.
Our take on $NOS
The combination of Solana + AI sets the stage for $NOS to do very well in a bullish market. Paired with its relatively-low market cap, expect some attention to be pointed their way.
When investigating the token, it's worth mentioning the drawbacks. First off, only 38% of the initial supply was earmarked for users (if we are being generous and including the liquidity and public sale allocations). The large insider allocations should give retail investors pause.
Additionally, the token itself does not claim any governance power nor revenue claims, these both flow to the founding company. So… what exactly are you buying when you purchase some $NOS? Speculation. Speculation that someone else in the future will value the tokens higher than today. Which, in crypto, can actually get you quite far - especially when you can leverage the hype of a hot chain (Solana) and an even hotter narrative (AI). But for the fundamentally-minded investor, there isn’t a whole lot to chew on for token value accrual.
Render - the 3D rendering powerhouse
Let’s get this big boy out of the way - Render is generally one of the first projects people think of when they hear “DePIN.” They have been around since 2017, offering a network of GPUs for 3D artists to leverage when they hope to graphically render virtual scenes for games or movies.
Render began as a spinoff of a company called Otoy, well-known for their rendering product Octane. With Render, 3D artists can outsource their rendering jobs to the GPU network all within their software suite, leveraging Octane.
Going so far as to be used in major movies and TV shows, Render has become a point of pride within crypto, showing a real-world use case being adopted by the outside world.
Recently, Render has announced a few upgrades that have excited followers (bagholders?) - the migration from Ethereum to Solana to decrease friction, the broadening of scope to include AI/ML and spatial computing, and the usage of native Cinema4D files. Along with these upgrades, the network has been seeing some decent upticks year-over-year:
Render has voted to adopt Helium’s Burn-and-Mint Equilibrium model (more on Helium later), which allows renderers to pay for services in dollars and GPU providers to share a predictable amount of $RENDER inflationary tokens.
Since so many Solana DePIN projects leverage this model, let’s show how this works in practice.
Burn-and-Mint Equilibrium (BME)
The problem: consumers want to pay in dollars, but the protocol wants to distribute a predictable number of governance tokens per epoch.
From the front end, consumers pay in dollars, which are converted behind-the-scenes to a proportional amount of $RENDER. This complexity is abstracted away from the consumer, but means that $RENDER is still the medium of exchange.
For every given epoch, node operators (GPU providers) earn points equal to the work they accomplished. At the end of the epoch, they will earn a share of the $RENDER rewards, in proportion to their contribution to the total amount of work done by the network.
At the end of the epoch, consumer-spent $RENDER and node operators points are “burned,” new $RENDER is minted per the inflation schedule, and this newly-minted $RENDER is distributed to node operators. Note that in Render’s case, newly-minted $RENDER is also distributed to consumers and liquidity providers, to encourage participation.
For Render, the DAO has set aside a specific inflation schedule, set to decrease with time.
Our take on $RENDER
The over-meme’d adage “old coin bad” is over-meme’d for a reason. Older projects tend to come with tons of bagholders that bought into the hype, only to realize >70% drawdowns. On top of this, newer projects, with the benefit of hindsight, can shape their architecture in order to avoid the pitfalls of their predecessors. Finally, these older projects tend to sit at higher valuations compared to newer ones, making their risk/reward calculus fall short of the competition.
Despite all of this, if any “old” project can rise from the ashes, it may well be Render. They have caught the collective mindshare as the DePIN project (next to Helium), proving their product-market fit (PMF) with real-world adoption of their protocol and recently putting in a new all-time high in market cap. Whether they can maintain that adoption after incentives dry up is anybody’s guess, but if the charts have anything to say about it, the market believes there is a chance.
Still, if taking the trade, I would walk in with tempered expectations due to:
- The likelihood of selling from underwater bagholders and
- The lack of upside due to an already-high market cap
io.net - spinning up AI clusters in 90 seconds
In 2022, China dropped more cash on GPUs than on oil. Read that again. Forget fossil fuels; processing power is the new powerhouse driving zero-to-one innovations and reshaping the geopolitical game. We're talking about a shift from black gold to silicon supremacy. The only problem? The pipeline for GPUs moves at a snail’s pace.
In order to bridge the gap and distribute processing power more efficiently, io.net is building decentralized access to computing power for machine learning (ML) applications. By establishing a network of GPUs from data centers, crypto miners, and established crypto projects (they have already partnered with Filecoin and Render), io.net is positioning itself to be the best option for projects in need of computation.
As always, a project can be better, faster, or cheaper to stand out. Io.net aims to be all three.
- Already pushing 23,000 GPUs in their network (3x more than Akash), io.net has the goal of reaching one million processors.
- Unlike their competition in the decentralized space, they allow for up to thousands of GPUs to be clustered to a user’s specifications, acting as a single unit. This is necessary for ML workloads, but currently not available elsewhere. This alone sets them apart from the competition.
- Deploy or employ assets permissionlessly - no KYC required. #MakeCryptoCypherpunkAgain
- Spinning up an AWS instance can take days. Spinning up an io.net instance takes a couple minutes (check out their demo at Solana Breakpoint 2023 if you don’t believe me).
- Being built on Solana means that micropayments are possible - a huge step for removing friction for end users. Users can also opt for direct credit card payments if they don’t want to manage a crypto wallet.
- Built-in integrations of Kubernetes, Ludwig, Pytorch, HuggingFace, Unreal Engine 5, and Unity mean that io.net is as close to plug-and-play for AI/ML applications as you can get.
- Io.net aggregates from existing hardware - consumer GPUs (like from Render), crypto mining farms (whose profitability tanked after the Merge), and independent data centers can all plug their hardware into the network and start earning.
- 70%+ savings when compared to using traditional products like Azure or AWS for cloud computation.
- It’s a win-win: idle/underused machines make more money and consumers pay less money. Connecting these users via cloud architecture and removing the middlemen makes this possible.
Our take on $IO
To say we are excited about io.net would be an understatement. The future of AI/ML development necessitates easy access to customizable clusters, and io.net is checking all of the right boxes.
$IO looks to be utilized in a similar BME architecture as many of the other projects in this report, where customers pay in dollars and node operators can be paid in their share of dollars and/or $IO token rewards (more details here). Io.net is tweaking some supply- and demand-side details in order to encourage network usage, which makes a lot of sense to us.
While the token is not yet live (expected towards the tail end of Q1 2024), we expect it to launch at a very high valuation (>$1B). Truth is, crypto x AI projects are hardly ever much more than vaporware - usually 1% fundamentals and 99% pump-amentals. But when a project like this comes along that fits a real need by using a decentralized network, and can do it better than its centralized competition… it goes to show why we are all fascinated by this space.
Navigation networks are generally controlled by state-level actors, or companies that might as well be state-level - if it isn’t the US government, then it’s your Googles, Apples, Ubers, etc. calling the shots, collecting the data, and reaping all the benefits. And, no surprise - establishing these networks is a daunting task, hence why these behemoths are the only game in town. That was, until, some enterprising protocols stepped up to build their own networks.
No doubt that it’ll take a lot to dethrone the incumbents, but that doesn’t mean it’s not worth trying.
Hivemapper - a crowdsourced Google Maps/Street View
When you pull up Google Maps on your phone, Google is consuming vast amounts of data about your route, merging it into their databases, and charging companies huge sums to access the data.
By using their custom dashcams, Hivemapper drivers are mapping the streets of the world. And when companies need mapping data, they can go to Hivemapper instead of their centralized competitors - passing on the profits to the community.
Consumers of this map/street view data could include:
- Surveying businesses
- Home insurance
- Navigation apps
- Training data for self-driving cars
- Up-to-date traffic/construction data for robo-taxis
- People/businesses looking to move to a new area
GoogleMaps data is notorious for skipping over large swaths of “unimportant” locations - not mapping entire countries, skipping small towns, and only updating high-impact urban areas. Hivemapper changes the calculus, attempting to map the entire globe. And they are well on their way.
>26,000 contributors have mapped 100 million km (over 60 million miles), and roughly 10% of the world’s roads after just under a year of operation. At this rate, Hivemapper is outpacing Google Street View (and its billions in funding). They have been able to achieve this rapid coverage by paying more to drivers that map new areas (more on this later).
Hivemapper is able to offer its mapping data at a fraction of the cost of Google - roughly $1/km compared to Google’s $4/km. Businesses can also incentivize new areas to be mapped or more up-to-date data via Hivemapper’s Burst feature. By paying Hivemapper, businesses can specify cities or neighborhoods to be mapped with higher priority, paying drivers extra for their services.
Hivemapper’s token, $HONEY, also uses a Burn-and-Mint Equilibrium (BME) model for distribution. This means that incoming dollars to the protocol are used to burn $HONEY, and newly-minted $HONEY inflationary rewards are then doled out to drivers. The only difference between Render and Hivemapper’s BME models is that instead of being released on a time-based schedule, $HONEY is released when progress is made in mapping the world. On top of this, if additional profit is realized, more $HONEY will be bought back from the market and distributed to drivers.
This aligns the objectives of the drivers with the protocol - the quicker the globe is mapped, the quicker rewards are distributed. Drivers are rewarded based on a combination of Coverage (unique surface coverage submitted in a region), Activity (the total amount of surface coverage submitted in a region), and Resilience (how well a region’s activity is distributed); with additional rewards coming from Burst incentives.
Our take on $HONEY
In general, $HONEY is mostly a reward token for drivers, with about 40% of the supply earmarked as distribution rewards. It doesn’t carry any governance power nor revenue share opportunities, and likely is traded in for cash by the drivers. Additionally, a 60% share is set aside for insiders (20% to investors, 20% to the team, 15% to Hivemapper Inc, and 5% to Hivemapper Foundation) - with vesting for both the team and investors just starting now.
For all of these reasons, we don’t love the token from an investment standpoint. What we do love is the project’s ambition and vision. Crowdsourcing at this scale is truly impressive and it would benefit all of us to bring this decentralized option to a sector dominated by basically-nation-state actors. If Hivemapper can leverage their BME model to scale their platform and accomplish their goal of mapping the world, it’ll prove that crypto incentive models can drive meaningful change in the “real world.”
Teleport - Uber, on the blockchain
Ridesharing companies like Uber market themselves as platforms, only to act as middlemen, connecting a web of drivers to customers looking for a ride. For all of the hype these companies get, their workers are not exactly thrilled with their business practices. Why? Because even after all of the VC subsedes to lower fees, Uber still takes a 30-40% cut of the money paid, lowering the bottom lines of their drivers in a race to zero.
In Teleport’s design, they would replace this middleman architecture with an open and transparent marketplace, where drivers and riders can be matched with minimal friction. Just like VC money, Teleport can bootstrap early riders/drivers with their native $TRIP token, helping to jumpstart the flywheel.
Establishing a complete rideshare network is complicated, and involves a lot of moving parts. For this, Teleport has proposed several different actors participating (and being rewarded for their help):
- Balancers - invite Drivers/Riders in order to overcome imbalances in the network
- Verifiers - inspect driver licenses and drivers’ cars
- Compliance Auditors - legal and operational readiness for Verifiers and Operators
- Operators - handle the regulatory and operational requirements of running a rideshare service
- Client Companies - mobile app developers
While not-yet fully deployed, Teleport turned some heads when it announced a $9m raise from VCs back in October of 2022. And now with deployment imminent, we’re definitely paying attention.
Our take on $TRIP
As of today, not a ton is known about Teleport’s token $TRIP. We do know:
- It’ll be used to incentivize network participation
- It’ll have governance power
For us, we will be watching their progress closely. Uber is valued at over $120b (with a ‘b’) and just entered the S&P500, with Lyft nibbling at their heels at $4b. Uber is a monopoly when it comes to rideshare platforms - and Teleport looks to disrupt this. If they can, better hold onto your hats.
Whether the purpose is measurement or connectivity, establishing a web of nodes throughout the globe takes immense coordination. You have to:
- Get the hardware nodes into the hands of all participants
- Provide easy onboarding in order to plug them into the network
- Incentivize participants to position their hardware where it is useful to the network
- Incentivize participants to stay online for long periods of time to ensure liveness
Blockchain tech is uniquely capable of tackling this problem set. Its built-in reward structure (tokens) allows designers to properly apply pressure where it is most needed. On top of this, its distributed nature removes worries of centralization and rent-seeking behavior of middlemen collecting disproportionate value.
Onocoy - giving satellites a helping hand
You can’t put an aerospace engineer in charge of covering a Solana DePIN report and not expect him to love the idea behind Onocoy. Here’s the deal - GPS satellites are really good for a lot of positioning use cases. They can give you a good idea about your location, but for any kind of operation requiring a lot of accuracy, GPS positioning alone usually doesn’t cut it.
Additional sensors give satellites a helping hand when measurements require high accuracy. One of these sensor types is known as RTK (or Real-Time Kinematics), which use on-the-ground receivers to help correct GPS’ meter-level accuracy to millimeter-level. Now, these receivers have a decent range (about as wide as a small city), but are generally only placed near major urban centers.
High-quality positioning data is useful for applications including:
- Deformation monitoring
- Natural disaster warning (tsunami/earthquake)
- Drone/robotics positioning
- Autonomous vehicles
And with so much of the world’s surface not currently being monitored, there is definitely a market need for this effort.
Onocoy proposes incentivizing the placement of these RTK receivers in order to cover large percentages of land. And even though they are only in beta testing, their growth has been impressive, with almost 2,000 participants plugging into the network with their receivers.
New users are encouraged to map out uncovered areas, with rewards decreasing if your area becomes oversaturated (3 overlapping signals). Like other projects mentioned, Onocoy uses a BME model, which allows customers to pay for services in cash, and network participants to be rewarded in $ONO tokens, in proportion to their contributions.
While these receivers aren’t necessarily cheap (they will set you back 1-2 grand), hobbyists interested in electronics and radios might be willing to learn a thing or two about blockchain tech if it means earning a bit of money.
Onocoy is still in beta testing, but interested enthusiasts can try out their network by requesting a whitelist spot.
Our take on $ONO
The token itself is used as a means to incentivize network participants and for governance. Here are the allocations:
- 30% to reward pool
- 20% to DAO
- 20% to investors (10% to both rounds)
- 15% to founding team
- 15% to future operations
Overall, not the worst allocations we’ve seen. The real struggle that Onocoy will probably face is the initial hurdle of getting receivers into the hands of network participants and making the onboarding process as smooth as humanly possible. Hell, if the participants are being drowned in cash, you better believe that outsiders will be interested in learning some new tech to join up.
We’re not saying that this is a trillion-dollar idea, but if it is a billion dollar idea that is being valued as a million dollar idea, you know we will be interested. We will be watching the transition from beta into mainnet closely to see if $ONO becomes an interesting opportunity.
Helium - a hivemind of hotspots
Another one of the OG DePIN protocols now calls Solana home after beginning life on their own blockchain. Helium is a distributed network of IoT and 5G mobile hotspots, powered by network participants. They’ve seen incredible adoption, with 12,000 nodes in place around the globe. These nodes provide LoRaWAN (Low Range Wide Area Network) connections for IoT-enabled (Internet of Things) devices as well as 5G hotspots for mobile coverage.
As with many of the previous projects, node operators are compensated in tokens for their participation in the network, which in turn is used by a huge ecosystem, including:
- Location tracking
- Mining operations
- Air quality/weather monitoring
- Food distribution
- Hotspot providers
- Agriculture/food safety
- Autonomous vehicle tracking
Just to name a few.
With the demand in bandwidth being up-only, Helium is positioning itself to be the decentralized option when it comes to mobile and IoT coverage. By providing a marketplace where node operators are rewarded based on the usefulness that they bring to the network, Helium hopes to become the go-to solution for connectivity.
A spin-off use case is Helium Mobile, a mobile provider that leverages Helium nodes when available (and T-Mobile when out-of-range of nodes)... all for $20/month. Pretty great when the average American spends much more for a phone plan.
The giant benefit is that people can use this network without even knowing that blockchain tech is supporting it. If we ever want decentralized technology to become the standard, this is the type of Trojan horse we need.
And now with Solana phone owners able to use their phone as a hotspot and earn rewards, the flywheel can really start spinning.
Helium also uses a BME model to dish out token rewards to node operators. Participants operating a 5G node earn $MOBILE, while LoRaWAN operators earn $IOT tokens. Users that tap into these networks pay dollars, which in turn pair up to whichever nodes they accessed during that epoch. At the end of the epoch, earned $MOBILE and $IOT can be burned to mint $HNT, which is distributed to node operators in proportion to how heavily their nodes were used.
$HNT emissions are on a two-year halving cycle, meaning that annual emissions are halved every two years, teetering down to <1,000 $HNT per year in the 2050s.
Our take on $HNT
As with $RENDER, $HNT has to overcome the hurdle of “old coin bad” to regain its former glories of a $5b market cap and beyond. The limited use cases of $HNT outside of protocol governance provide additional headwinds for price, as most network participants generally view their earned $HNT as income, likely dumping it for cash on a regular basis.
Network activity has seen a great uptick in the past few weeks, however, the network generally does not generate significant income in its current state (on the order of hundreds or 1-2 thousand dollars per day). Definitely heading in the right direction, and we will keep an eye on their progress for sure. These are still the early days.
At this point, today’s 9-10 figure market cap makes an $HNT allocation tough to justify from a risk/reward perspective if purely looking at fundamentals. However, it’s completely possible that $HNT (and likely $MOBILE) could see strong price action in the short/medium term with the uptick in Helium Mobile activity and hype surrounding Solana DePIN.
Data gathering and handling
Since the early days of blockchain tech, a distributed ledger and data storage/management have been an obvious fit. Filecoin is probably the most recognizable player in file-storage space, boasting the most reach and wide scale adoption of any decentralized file storage protocol. Likewise, Chainlink has cemented itself as the oracle provider, providing a direct link from onchain and offchain data and messaging.
Naturally, all newcomers to this broad “data handling” ecosystem will find themselves being compared to these behemoth incumbents. But because Filecoin and Chainlink are aligned with Ethereum and the Ethereum Virtual Machine (EVM), speed and cost have persisted as nagging obstacles towards growth. Perhaps the “lightspeed” nature of Solana will give these new protocols an edge over the kings.
With the rise of Machine Learning, data has become a hot commodity. AI-training data sets take immense resources to develop, but good AI needs good training data. Turning off the data faucet is a surefire way to handicap your model’s progress.
GenesysGo - file storage that you’ll actually use
GenesysGo are positioning themselves to be the go-to file storage application within the Solana ecosystem with their main product ShdwDrive (which uses their D.A.G.G.E.R. consensus mechanism). In their mind, Solana is the perfect bedrock on which to build an enterprise-scale storage and compute network on crypto rails.
Filecoin’s slower network speeds mean that a 1 MiB file takes about 5 minutes to store and receive in a best case scenario (and about 4 hours in the worst case, source). That same file takes about… 3-7 seconds for ShdwDrive. Multiply that out across hundreds of users and millions of files and you have yourself the setup to a hostile takeover.
ShdwDrive is able to accomplish this by running their D.A.G.G.E.R. consensus atop the Solana blockchain, both purpose-built to deliver fast and cheap interactions. This type of application can be beneficial to any project/person hoping to leverage cloud services:
- Web hosting/content management - where users can store any type of file
- Social media - an immutable backlog
- Archival - erasure coding embeds implicit redundancies in the system
- Datasets - an onchain library, accessible by all
- Personal & editable storage space - with optional immutability, ShdwDrive could become your own personal Google Drive
ShdwDrive is currently in Phase 1 of their Testnet, where individual users can run “Wield Nodes” to support D.A.G.G.E.R. consensus. According to their roadmap, they are set to run testnet phases through 2024, to bring ShdwDrive and D.A.G.G.E.R. to the big leagues.
Our take on $SHDW
The token itself was initially distributed back in 2022, via a public IDO and to holders of their SSC (Shadowy Super Coder) NFTs. It’s worth noting that the IDO price was about $1.73. While the team wasn’t given direct allocations, they were of course allowed to participate in the IDO/NFT sale. Additionally, no insiders were given allocations - quite unprecedented to see.
The $SHDW token itself is used to pay for services within the ecosystem, and additional future utility is hinted at within the D.A.G.G.E.R. documentation. While not groundbreaking use cases, and certainly isn’t as great as governance and/or revenue share, it is better than nothing from the demand side.
With $FIL sitting at a substantially-higher valuation compared to $SHDW, the comparisons become as left-curve as you could imagine. Anyone can open up MarketCapOf, see a big number, and ape. It also helps that $SHDW is almost fully distributed (unlike $FIL) and the Strategic Reserve Fund was burned.
When you combine that with a mixture of good tech and a potential product-market fit that actually makes sense, you have yourself a potential winner within this small corner of crypto.
Pyth - not precisely “Solana’s Chainlink”
After a very successful airdrop, all eyes are on the “Chainlink for Solana” (which, in all honesty, isn’t a supremely accurate descriptor). Pyth is the largest first-party oracle provider - where “first party” means that the data is provided from the data origins themselves and not routed through a third party.
Pyth is able to route data between the on- and offchain worlds by leveraging Wormhole’s cross-chain messaging infrastructure, connecting offchain price feeds to onchain applications (and vice versa). Being native to Solana means that Pyth enjoys a low latency environment - critical for oracles, which can be vulnerable to minute price fluctuations. Note that users (Consumers) of these data feeds generally center around Solana, but do operate elsewhere too.
You can think of Pyth market participants as falling into one of three categories:
- Publishers, which, as you may guess, publish price feeds onto the network. For their troubles, they are granted a share of the data fees from the Consumers. Publishers are rewarded in proportion to the quality and quantity of data they provide.
- Consumers ingest price feed data and leverage it for their (on- or offchain) applications, and sometimes pay data fees (in $PYTH or stables) if they want priority access.
- Delegators can stake tokens with Publishers to earn a portion of data fees, but also stand to suffer slashing if the oracle is proven to be inaccurate.
Pyth is positioning itself as the low-latency choice for crypto price feeds - a key building block for any trading app that necessitates quick feedback.
Our take on $PYTH
The natural price-point comparison to make (and hence, we will make it too) is to $LINK, the king when it comes to oracle networks. Still, this isn’t an apples-to-apples comparison. Pyth is laser-focused on price feeds, while Chainlink is so much more.
In addition to price feeds, Chainlink has:
- Data Streams - onchain data that isn’t solely tied to price
- Proof of Reserves - to verify cross-chain and offchain backings for wrapped assets, pegged assets, and custodians
- Functions/Automation/VRF - compute smart contracts without the need to connect to another blockchain, inject verifiably random data from a trusted source
- CCIP - interact with a decentralized set of oracles for many functions, including cross-chain USDC non-custodial transfers
The all-in-one platform of Chainlink goes beyond simple price feeds and is the reason why the LINK Marines get all up in arms when Solana manlets start calling for an equal FDV for Pyth. For reference, at the time of writing, $PYTH FDV is roughly $2.5B while $LINK’s is about $14B.
After Pyth’s initial launch, we have about 15% of the total tokens released, with ecosystem growth and insider $PYTH set to be distributed over a 4-year timeframe. So while $LINK’s MC/FDV of 55% isn’t spectacular, $PYTH buyers should be aware of the upcoming token inflation (and potential insider dumping) over the coming years.
The Pyth network itself has seen some steady use since token launch, but nothing too crazy to suggest a market mispricing.
When we consider the current $2.5B FDV, it’s difficult to see massive upside on $PYTH - at least in the medium-term. The “Solana vs Ethereum” camps will no doubt pump the “Pyth vs Chainlink” narrative ‘til no end, but, in our opinion, there are better risk/reward opportunities in this space.
Synesis One - playing games to train AI
The lifeblood of any Artificial Intelligence platform is the data it uses to train its AI models. Just as a student needs a good teacher to be successful, without good training data, you have no hope of training a talented AI.
The problem? While the global training set market is expected to grow to over $10b, there haven’t been serious efforts to decentralize this ecosystem beyond the huge frontrunners.
Synesis One proposes an architecture of data training that actively works with AI builders, serving them useful data sets in a real-time feedback loop. Teaming up with Mind AI as their first consumer (which shares some co-founders to Synesis), the Synesis One ecosystem consists of “miners,” each with a crucial role in keeping this data-generation flywheel spinning.
- Architects - are the sophisticated of the bunch, in charge of ingesting the requests of the training set consumers (AI companies) and creating gamified jobs for the Builders to complete, which should satisfy the requests of the consumers.
- Builders - accept jobs from the architects. Generally, this means playing games and answering questions in a way that could help train AI (example).
- Validators - judge the responses by Builders to check their quality before their responses are ingested into the training data.
By posting a stake of $SNS tokens, these miners have skin-in-the-game, which is slashed if they are found to be underperforming or misbehaving. Synesis One can incentivize groups of specialists (Architects) and generalists (Builders/Validators) to participate in their ecosystem in order to build useful data sets for consumers, in a completely decentralized fashion.
On top of this basic architecture, Synesis One plans to lean heavily on the gamification of their process in order to build community. They’ve mentioned that they plan on bringing on MMORPG consultants to make the Builder/Validator roles as fun as possible (as this could be seen as a mundane process).
On top of permitting access to their Builder play-to-earn games, Synesis’ Kanon NFT collection allows users to “buy” keywords, with NFT artwork being produced by small artists. Since the first use case of this training data is for Large Language Models (LLMs) which work by vocalizing several different ways of conveying the same information, Kanon holders will be paid royalties each time their keyword is used. This is mostly a community-building exercise meant to reward members with a share of tokens.
Our take on $SNS
Synesis is taking a novel approach to a yet-untapped niche in crypto - attempting to gamify the process of data gathering from the community. If you’ve been online for a while, you’ve probably heard about pay-per-click work that is scattered around the web, where folks can earn a couple bucks by completing mundane tasks like completing surveys. Synesis looks to harness that same workforce and focus them towards building useful data sets for AI training. Honestly, if the games turn out to be fun and the onboarding process is simple, this concept could do really well.
Adoption has been impressive so far, with over 175,000 data points passing through the system in these early stages. Combine that with 3k Discord members and over 67k Twitter followers, and we could be seeing the beginning of wide scale onboarding.
The token itself, while it has governance power, had an initial distribution that favored insiders, with the team (current, future, foundation, and marketing) accounting for 40% and investors (incl. Private, seed, advisors, and shareholders) getting a little over 23%. Not great.
But, along with governance power, token holders may enjoy deflation in the future. A proposed mechanic would buy-back-and-burn $SNS from the open market whenever a profit is realized in a quarter. A nice little bonus if and when the project hits higher scales.
A quick note on the team - one of the founders of Mind AI and Synesis One also founded Formation FI (which has since rebranded to Every Finance), back in early 2021.
Formation FI, which intends to bring risk parity to DeFi, has been plagued by slow development and poor team communication. And despite launching a token ($FORM) back in 2021 via an IDO, hacks/thefts, token delistings, and team shake ups have pushed protocol launch way back - with an initial test version just going live now. To say that early Formation FI supporters are disappointed would be an understatement, and the token price chart paints a clear picture of their frustration.
Now, Synesis One is not the same as Formation FI/Every Finance - they just share a founder. There’s a new chain, a clear use case, and an early working product (albeit with a close relationship to their one key partner, Mind AI). Synesis One was caught in the backlash of Formation FI’s hack in early 2022, but from all accounts, Synesis One has consistently delivered on their roadmaps without any major issues.
So with the narrative, fundamentals, and potential stickiness of gamification and community building, we could see $SNS fairing well from a risk/reward perspective - especially when the low market cap of <$50m lives in a frothy market.
Grass - selling your idle bandwidth to train AI
Know how we’ve mentioned that AI training involves LARGE data sets? While one way to provide these sets is by manual data entry (e.g. Synesis One), most AI-training labs opt to scrape the internet in order to speed up the process. The only issue is that the internet that I see is different from the one you see, depending on: where you live, what the web browser knows about you, what links you’ve clicked, etc.
These web scraping companies would sure love to get a complete picture of the internet from all of these perspectives - and Grass is giving them the chance.
Individual users can download the Grass browser extension to opt in being a node in the network. Companies will then pay to use idle bandwidth from your node, essentially paying you to borrow your internet access and gather data.
Grass is managed by its parent company, Wynd Network. And after being coached by some big brains, Wynd went on to secure a $1M pre-seed from top names, and Grass had their own $3.5M seed. Obviously, people are excited about it.
In December, Grass kicked off their mobile support with the Solana phone - a nod to the chain’s close DePIN connection. Mid-curved the Solana phone? Don’t worry, anon, Android and iOS apps are launching in Q1 of 2024.
Our take on $GRASS
While the token isn’t yet live, users can participate in their points program if they have access to a referral link (feel free to use ours here, or don’t, whatever). Upon release of their governance token, points will be convertible into $GRASS (or whatever they end up naming the token). Details are limited, but there are some hints in their FAQs. $GRASS will continue to be emitted via farming incentives to bolster dollar-denominated payments from early bandwidth purchasers - similar to bootstrapping methods already mentioned for other projects. Getting paid for doing nothing? Talk about a millennial’s dream.
We can see $GRASS performing well if the tokenomics and go-to-market are handled correctly. And with the tutelage of some top funds/individuals (shout out to the DePIN Daddy @Old_Samster), we’re quite excited for development and token launch details.
Why we’re keen on the Solana DePIN narrative
How often have you had this conversation with your friends or family -
“OK, but what can you actually do with crypto?”
And you just sit there, dumbfounded (with your mouth agape), trying to form a coherent sentence about why your parents should be excited about a trading bot that comes with a built-in smart contract wallet.
If we want crypto rails to power the world, it is these DePIN real-world use cases that prove to normies that we aren’t just peddling snake oil. Crypto is exceedingly good at powering networks as a thin infrastructure layer and win-win incentive structures. If there is a scattered group that has a common set of needs, and another group that has the means to provide them, crypto can give them the network to connect the two.
So, now, when you are sitting down for Christmas dinner and your annoying aunt asks what crypto is even good for besides making Richard Heart rich, you can at least muster the word “networks.”
Any one of the projects mentioned here can be a Trojan horse that wedges decentralized technology into an industry dominated by rent-seeking middlemen. When Amazon monopolizes a sector, do you think they pass cost savings down to the consumer? Hell no. Decentralized networks can be fashioned to align the incentives of all parties - a killer use case for hard-to-organize networks everywhere.
So even though we aren’t uber-bullish on every token mentioned in this report, we still want to see these projects succeed. Helium succeeding means we finally don’t have to bend over backwards for our ISPs. Hivemapper succeeding means that maybe Google Street View is finally getting a run for its money. Nosana or io.net succeeding means that Amazon Web Services isn’t the only game in town for distributed cloud access. Any of these is a win for the industry. It shows that crypto is not simply reserved for tech enthusiasts/gamblers, but is useful to everyday people.
DISCLOSURES: After conducting research for this report, the author, 563, began adding a small amount of $SHDW to his portfolio. Be aware of this bias while reviewing this report.