Disclaimer: After researching and writing this article, I am now holding a position in AIOZ. Please take any bias with this in mind.
An introduction to AIOZ Network
Unless you have been living under a rock these past 12 months, then you would have noticed that there have been significant developments in the advancement of artificial intelligence (AI).
We have all played around on ChatGPT for plagiarising articles, Midjourney for turning your dog into a Disney character, or even began a burgeoning relationship with your new AI-generated Waifu, unbeknownst to your actual living humanoid wife. Whatever the case, it has taken the world by storm of late, and this is only the beginning.
With such acceleration comes some very real concerns. Not just on how quickly this thing is evolving but more so on who is at the wheel feeding it the Rare Candies to seek world domination. That was a Pokemon reference; by the way, if you didn’t get it, I kinda feel bad for you.
The issue is as old as time itself, the powers that be, aka the Magnificent 7: Alphabet (GOOGL; GOOG), Amazon (AMZN), Apple (AAPL), Meta Platforms (META), Microsoft (MSFT), NVIDIA (NVDA), and Tesla (TSLA) have such a vast power law advantage that it is slightly daunting given their past record of ethics and monopolies.
All of the above are working directly on their own implementations of AI, Cloud Computing, or hardware to enable the next generation of this wave, which is to be expected but, as I say, kinda gives me the willies…
I, for one, am not going to trust the future of civilization in the hands of Zuck, despite his recent PR revamp.
Not only do these guys have the infinite money glitch, but they also have the infinite data glitch, too. It just so happens that they all (maybe barring Nvidia for different reasons) have an ever-increasing data set as well. This is fuelled by their users, who can’t help but insistently post their whole lives, warts and all, on social media/the internet.
Note: Tesla has been developing using AI for a while now, but Elon’s adjacent company x.AI, announced in November 2023 that they will be working closely with the self-driving car company.
Speaking of Elon (and bear with me, this is a very longwinded segue)… OpenAI, the developer of ChatGPT, actually started out as an open-source project funded by Musk himself.
xAI looks to be spun out of frustration with the way that the company has been heading since his departure, i.e., “maximum-profit company effectively controlled by Microsoft.” I’m not even going to get into why Bill Gates having something as powerful as OpenAI in his grubby little mits is a bad thing; I’ll leave that to you.
An adjacent issue is the ever-increasing demand for GPUs (hardware that enables the processing power that enables AI models to learn and grow). They simply cannot make them fast enough. The next-gen GPUs that have been promised to purchasers are severely backlogged due to the chip shortage and explosive demand.
This is why we have seen a ridiculous rise in the NVDA price over the past 12 months, with Nvidia being the market leader in GPU production.
No, that is not a meme coin; that is a $1.83 Trillion market cap company, Nvidia.
What’s even more frustrating for start-ups looking to build in this space is that Nvidia has entered a partnership with almost all of the Magnificent 7, which are not only getting priority over their existing GPU inventory but are now also working to develop bespoke GPUs for their specific internal use cases.
So yes, huge demand playing into a crazy power law dynamic of huge centralizing forces that already control nearly every aspect of your waking life… we need a solution.
Well, without sounding like that really annoying friend at the party, crypto could be the answer…
The Rise of Crypto, AI, and DePIN
If you have been following our research recently, you will have seen a DePIN primer and also a more in-depth report on the current state of Solana DePIN.
Why are you talking about DePIN and AI in the same breath?
Well, my good friend, we can enable one with the other, and crypto underpins the full system. Sit back and let me explain.
DePIN stands for Decentralized Physical Infrastructure Networks. That is a bit of a mouthful, so think of it like this: Centralized infrastructure, such as cloud computing through Amazon Web Services (AWS)/ Amazon S3, enables you to access processing power, storage, and databases without having to go out and create your own server farm or go through the hassle of maintaining it.
Come now, ye trusty start-up founder, and let old Bezos handle that for you. We are extremely secure (lmao), have your best interest at heart, and certainly wouldn’t turn off access whenever we feel like it.
AWS revenue was around $90.8bn last year alone, servicing over one million users and scooping up over 41.5% of the cloud computing market share, which is more than all its major competitors combined (Azure, Google Cloud, IBM).
If you are running a decentralized application that is focusing on AI or DeFi, does it really seem sensible to be operating on a single point of failure through a massively centralized company/service like this? It all seems a little silly.
Decentralized applications need to be built using decentralized Infrastructure, or if not, what is the point?
Now, back to DePIN, if only there were a solution that democratized access to a global, permissionless, and performant cloud computing competitor… you can see where I am going with this.
If we tie everything together, we get a decentralized network that has the power to harness DePIN to enable further development and democratize access to AI. This is where the protocol we will be focussing on today comes in. Introducing AIOZ Network.
What is AIOZ?
“AIOZ Network is DePIN for Web3 AI, storage, and streaming and empowers a fast, secure, and decentralized future”
I know what you’re thinking; there are a lot of big buzzwords in there, but once we are done with the article, hopefully, it’ll all make sense.
AIOZ Network is a distributed infrastructure network that allows anyone, anywhere, anytime, access to decentralized services such as storage, AI computation, content distribution, VOD/live streaming, and even IPFS/IPFS pinning.
The network itself is a sovereign blockchain built using the Cosmos SDK but is also EVM compatible to allow seamless integration of any decentralized applications (DApps) written in the most adopted smart contract programming language there is: Solidity.
The chain has extremely low fees (as shown below), instant finality, and up to 1,400 transactions per second. Given its roots in the Cosmos ecosystem, the interoperability between chains is secured by IBC, whilst heading over into EVM-land is handled by the Gravity Bridge.
Info: https://explorer.aioz.network/
Currently, there are around 50 validators in the AIOZ Network that maintain the operations and security of the chain, leveraging the trusty Tendermint for consensus. These 50 Validators are chosen by the community, whereby the top 50 with the highest AIOZ stake weight will be included in the validator set. Validators and delegators stake AIOZ and earn AIOZ block rewards plus transaction fees, as you might expect.
Info: https://explorer.aioz.network/nodes
On top of that, anyone can run an AIOZ Edge Node, which contributes to the security, storage capacity, bandwidth, and computing power of the network. And, as with the standard of dPOS systems, they will get rewarded in AIOZ tokens for doing so. They even have a cool little GUI for smooth brains like you and me, who have no technical know-how whatsoever, to be able to spin up a node quickly and seamlessly.
The idea of an Edge Node is really important to the very purpose of AIOZ. Edge computing refers to a network that processes the data/transactions close to its source. Which, if you think about it, is paramount when handling large data sets for AI model training or streaming video and doing it quickly!
There is no point in shifting to a decentralized system if the performance is lacking. We need better, faster, cheaper.
So, what are you actually doing when you are running a node?
The three main contributions you are making to the network with your hardware and infrastructure are:
- Storage: Large language models and big data need A LOT of storage.
- Computing: Both your CPUs and GPUs help with the transaction processing for things like video transcoding, software ops, and AI-related tasks.
- Bandwidth - Utilizing Edge Computing through your local bandwidth can drastically increase the performance of the network. This helps massively with latency and reduces costs for tasks being handled on the AIOZ Network.
Oh, and you’re helping with the decentralization of the network!
A brief history lesson before we get into the next advantages of having such a powerful network.
AIOZ Network has been working on DePIN products before every thread-boy on Twitter started engagement farming this narrative. They initially rolled out a proof of concept called AIOZ Tube (a decentralized YouTube competitor) to show off how the AIOZ Network, and specifically their dCDN (decentralized content delivery network), could support and handle products such as streaming.
So, what is actually going on under the hood to allow this to happen on a decentralized network?
Well, this is all possible due to the AIOZ dCDN or decentralized Content Delivery Network. If you think about the steps that are needed to allow you as the end user to watch your favorite episode of the blocmates podcast, is the following…
- We upload our video to YouTube, which is then stored on YouTube servers and distributed across the platform, and then hopefully, people see the thumbnail and click on it.
- Next, YouTube has to pull that data from wherever it is stored locally and stream it to you whilst you eat your food in your underwear.
Now, the whole process of storage, sourcing, and the redistribution of content would be made a hell of a lot easier with a centralized setup. The beauty of the dCDN design is what makes AIOZ Network interesting with regard to efficiency, delivery, and scalability.
This is how the dCDN works (in a nutshell).
Info: https://docs.aioz.network/aioz-dcdn/overview
At YouTube, for example, the media files are stored on their own servers. On the AIOZ network, they are stored on a decentralized network, which rewards those nodes in the network for maintaining, securing, fetching, and storing the content. This is one of my favorite aspects of tokenization. It aligns incentives in a way incumbent Web2 systems can’t.
And what about scalability, I hear you ask?
Well, usually, a blockchain's throughput and performance tend to scale inversely with validators on the network. This is because if you have 10 validators in a network, they can come to a consensus much easier than 1,000 and even easier than 10,000. The key here is the Edge Nodes. With each node added to the network, the whole system becomes more scalable and performant due to the increase in hardware and localization of storage. Again, this makes no sense if the system isn’t better, faster, or cheaper, but with a decentralized peer-to-peer system like AIOZ through its dCDN, this can be achieved and scale linearly with the network.
So, hopefully that explains at a high level, the underlying infrastructure of the AIOZ network. Now, let’s take a look at the products that are being built on top of it, which is the exciting part.
What is AIOZ W3S?
WEB3S is a decentralized cloud/object storage solution for crypto and Web3. I mentioned earlier in the article the issues with traditional cloud service providers and the need for decentralized applications to be built on decentralized rails. If we think back to the underlying AIOZ P2P architecture, we can quickly see how this is a really interesting solution to the centralized incumbents.
Here are a few key points about WEB3S that a user might consider when compared to traditional providers like Amazon S3:
- No single point of failure - This is a big one; you don’t want to upload your files only for there to be a data breach, hack, or misplacement of the files, and they are gone forever. The AIOZ Network creates multiple copies, which are distributed globally across the network, which also aids in quick retrieval via the dCDN network.
- Security through cryptography - This one is a given, and everyone should understand the benefits of being situated on crypto-rails, but I think we sometimes forget the beauty of how these systems are designed. You and only you have access to your files.
- Fees - Pay-as-you-go fees for storage come in at around $17/T, whilst delivery bandwidth comes in at $5TB at an average download speed of 15MB/s, and yes, this is all paid for in AIOZ tokens. This is a flat rate with no hidden costs. Here is what Amazon S3 pricing looks like…
Info: https://aws.amazon.com/s3/pricing/
Good luck working that out, and oh yeah, it changes per location, too…
“BuT MuH CoMPanY AlReAdY RuNs oN AmAZoN s3!”
Cool, they have an adapter for that. Users can switch or integrate AIOZW3S into their current Amazon S3 set-up; it is fully compatible with other cloud services.
What is W3IPFS?
It doesn’t take a genius to work out that this is an IPFS-like product. For the non-technical folks who are not up to speed with IPFS, here is a terrible explanation of why it is important and how it works.
InterPlanetary File System (IPFS) is a peer-to-peer network that enables file storage and sharing across its network. Kinda like a blockchain network? Yeah, kinda, but not really! IPFS is a decentralized system that spans the globe and even uses cryptography to help locate and securely store files.
So, why do we need W3IPFS?
Well, with all good things, particularly when they are (for the most part) decentralized, we usually see some trade-offs. There is no way a smooth brain like me is using IPFS. I have no technical capabilities whatsoever, so there is absolutely no way I am figuring that one out.
There are also well-documented issues with things like latency and general UI/UX when it comes to search capabilities and retrieval. By all accounts, the documentation around IPFS is a little all over the place.
There is also this process of “garbage collection” whereby nodes reaching storage capacity will free up space by deleting older files to make way for new ones… even I know that is a bad thing.
As a workaround, the idea of IPFS “pinning” was developed, which enables data to be stored on a couple of nodes in the network and be flagged, basically saying, “plz don’t delete this during garbage collection.”
On top of that, pinning makes the UI/UX of storing and retrieving data/files much easier for a wider user base.
Again, why is this relevant to crypto and web3? Well, think of your beloved over-priced jpegs, aka NFTs. The underlying token may be stored on-chain, but the file associated with said token will usually be stored off-chain on IPFS. What you see when you view an NFT collection on OpenSea or Blur is a reference to their related files located off-chain.
With WEB3IPFS from the AIOZ Network, vital NFT data from collections, NFT marketplaces, and games can be stored in a decentralized way, on-chain. This comes with the added benefit of native pinning to ensure lifelong access to your favorite jpeg data as long as the AIOZ Network is running.
Usually, the issue with pure-on-chain storage is the storage limitations. A lot of on-chain NFT collections currently are pixel art as the chain can only handle small data. This is not the case with NFTs stored natively on the AIOZ Network. As we now know, AIOZ is a network that can store, handle, and process BIG DATA due to its Edge Nodes and dCDN.
What is W3Stream?
Video streaming services have overtaken television. Who would have thought it when, back in the early 00s, you were watching Charlie Bit My Finger? Now, there are full-on shows with incredible productions that rake in views and advertising spending that the big dinosaur broadcast agencies could only dream of, and it isn’t stopping any time soon.
Streaming services globally are a $500bn a year industry, accounting for over 38.1% of all TV usage worldwide.
That is all very impressive, but over recent years, we have seen an incredible amount of familiar faces, including Google, Apple, Amazon, and Netflix, owning the lion's share. Once again, we have slowly begun to see the perils of this monopoly, with content creators being censored, shadowbanned, and even completely platformed with no comment on the reasons why.
Look at this from our fellow crypto content creators and their experience with YouTube recently…
I rest my case.
There are “free-speech” competitors that have recently gained a lot of traction; even Twitter (I refuse to call it X) has made a massive push into video hosting and streaming. The issue is that anything that goes against Musk’s view of the world gets strangled on distribution by the algorithm. God forbid you to post a link that pushes you off of Twitter… It’s all very tiresome.
Anyway, it is glaringly obvious that decentralized streaming platforms that have great web2-like UX are very much needed.
W3Stream is another great proof-of-concept product on AIOZ Network that shows the performance of the chain whilst allowing users to retain their sovereignty and access to their own content.
This product enables video streaming, hosting, encoding, management, and analytics. It would be really interesting to see some crypto-native advertising or even micro-payments for content creators and media companies, but who knows if that is in the works… I suppose, being a permissionless system, anyone could build it.
What is WEB3AI?
Right, the big one you have all been waiting for… WEB3AI.
We covered the ridiculous situation happening with GPUs across the globe in the opening paragraphs of the article. We have also covered the much-needed democratization of such technologies, whether that be access, processing power, or even a permissionless AI model marketplace.
In one way or another, the magnificent seven are all trying to push towards a monopoly because, remember, competition is for losers. Some would even suggest that Sam Altman’s recent testimony in front of Congress calling for regulation in the development of AI and large language models was a bid to stunt competitor growth through regulatory capture… Who knows…
In a nutshell, it is a cesspit and will only become more cutthroat in the race to be number one. Democratizing access is a must, not a nice to have.
So, how can AIOZ Network help?
By now, you should be up to speed with the way the network works. The multi-faceted approach to distributing computer power across a decentralized network enables powerful execution and handling of large data sets, which are needed for training things like AI models.
Let's say you were creating a text-to-video model like the recent Sora AI product from OpenAI. You need to train that model on Video content and also interpret the input data, i.e., the text.
Where do you get the humongous video data set to learn from? You can’t use Google or other streaming platforms; that data is gatekept. Where do you get the input data for the text prompt to teach the model?
This is very difficult to obtain and certainly gives the big dogs a huge head start with their different social media verticals where people are willing to submit their data for free and are none the wiser to carefully crafted terms and conditions that opt them into all sorts of repurposing of their own content.
This is obvious, but what if there was a way for users of a network to opt into selling their data, which is already stored on the network (WEB3S3), to those wanting to train their AI models? Better yet, what about those who wish to sell their pre-trained models to those wishing to build on the back of the data?
Well, a lot of this is primed and in the works with WEB3AI from AIOZ Network.
The key is the dual-sided marketplace coupled with the execution environment powered by the nodes on the network.
Now, tying this back to DeFi apps… If you are building, say, an AI yield optimizer that has been trained to seek out the best yield-bearing strategies over a specific time period. What does that look like?
Maybe the model is expecting a large amount of volatility over the coming weeks based on IV, so the optimal strategy over this period is for some form of LP strategy. Or maybe the markets are pretty flat, but there is a new liquid staking project paying high APRs on ETH deposits.
With the amount of new L2s popping up and the inefficiencies across their new naive DEXs, there could be opportunities for cross-chain arbitrage. Whatever the strategy, this is something that a DeFi AI model could be trained on. Blockchains are data-rich and inherently accessible to anyone.
The possibilities are truly endless, and we haven’t even scratched the surface of what is yet to come, but with all the tools open and accessible to anyone and everyone on the AIOZ Network, plus the added composability across the EVM and Commos, this thing becomes something to be aware of.
AI Airdrop farming, anyone?
Meet the AIOZ team
I thought it best to include this section because we are entering the beginning of a hockey stick growth curve at the intersection between crypto and AI. What this will result in are a lot of people with absolutely no idea what they are talking about beginning to raise a shit tonne of money for “AI” projects which will be nothing more than a GPT-Wrapper.
The background of AIOZ Network and when they started building in this space is very important. The team has been focussing on AI since 2013, with an ever-increasing overlap in crypto. So, I thought it best to point out that if you have just heard of this protocol, it isn’t just another flash-in-the-pan hype product that came in for a quick buck.
They have researched and developed many products in the DePIN and AI space, including ideas like smart advertising, IoT, AI-powered healthcare assistance, and smart parking products. So, yeah, it's not a copy-paste idea with a nice shiny website.
When AIOZ Network was officially formed in 2017, it joined forces with its sister company, AIOZ AI, to tackle the world of DePIN.
The team has worked alongside some of the world's most prestigious universities in publishing research in the fields of IoT, AI, and Robotics.
Tokenomics
This next section is what a lot of you will have been waiting for. So, what is the AIOZ token, and should we care at all?
The AIOZ token is a utility token in the sense it is needed for staking and delegation to secure the AIOZ Network. It is also used to pay for actions on the network whether that be for storage or for purchasing AI data sets and models.
As of December 2023, the AIOZ updated inflation schedule is now set to decrease by 1% a year, starting from 8% on 25 December 2023 to 5% on December 25, 2026.
The inflation distribution is split 50:50 into rewards for validators and delegators, whereas the remaining 50% is sent to the treasury.
There is a token burn implementation, too, which is programmatic based on network usage as follows:
- 50% of ALL AIOZ Network Blockchain Transactions
- 5% of ALL Node Rewards
- 5% of ALL Infrastructure Revenues
- 5% of ALL AIOZ Native dApp Revenues
This is the same as any network adoption playbook, in my opinion. The more usage for the applications on top of the network, the further the demand and subsequent burn, creating an increased demand with a diminishing supply.
There is a predictable inflation schedule, and by the looks of it, no supply overhang is in the hands of those looking to offload.
Note: This is all purely speculative.
Public and private sales occurred in March 2021 at the following prices
- Public Sales : $255,000 @ $0.015 per $AIOZ
- Private Sales : $1,095,000 @ $0.015 per $AIOZ
Despite the current roughly 10x price, probably due to inflation of the token, looking at the vesting schedule, this all appears to be fully in circulation.
Vesting Schedule:
- Public Sales: 100% Fully Unlock
- Private Sales: 25% Unlock upon Exchange Listing, 25% per month thereafter
- Team: 6 months Cliff, 8% Unlock every month thereafter
- Advisor : 3 months Cliff, 8% Unlock per month thereafter
- Marketing: 10% Unlock upon Exchange Listing, 5% per month thereafter
- Exchange Liquidity Provision: Reserved for DEX and CEX liquidity provision
- Ecosystem Growth : 3 months cliff, Reserved for Partnerships, User Acquisitions, Network Fees / Validators & Future Development of AIOZ
Info: https://medium.com/aioz-network/aioz-network-token-metrics-explained-791b4e0854af
So, make of that what you will.
Roadmap
Looking at the year ahead, there appears to be significant fine-tuning of existing products whilst also pushing out some new ideas, including AIOZ Drive App, AIaaS, native AIOZ token standards, and a native DEX.
Q1 2024
- AIOZ W3S
- AIOZ DTransfer dApp
- AIOZ Multi-Chain Wallet Browser Extension
- AIOZ Node V3 - AI Computing Integration
- AIOZ W3IPFS
Q2 2024
- AIOZ W3AI V1 - Decentralized AI as a Service
- AIOZ Drive dApp
- AIOZ Node V4 - Transcoding Integration
Q3 2024
- AIOZ W3Stream
- AIOZ W3AI V2 - AI Marketplace v1 for decentralized dataset monetization, open-source & user-customized AI model storage.
- Deployment of AIOZRC-20, AIOZRC-1155 and AIOZ RC-721 token standards
- AIOZ Gravity Bridge
Q4 2024
- AIOZ DEX
- AIOZ W3AI V3 - AI Training with Decentralized Federated Learning, Spaces for building AI dApps.
What we want to see here is stuff being delivered in a timely manner with the amount of competition and attention being focussed on this category; this stuff needs to be on the market ASAP.
Conclusion
As I’ve mentioned a number of times in this article, we are at the very beginning of the crypto x AI boom. The crossover between these two emerging industries is absolutely massive, and again, decentralized applications need to be built on decentralized rails.
I have no idea how to value anything in the AI space, and I don’t believe that anyone currently does. We have yet to scratch the surface of what is possible.
DePIN, AI, and crypto will all play a pivotal part in the democratization of such powerful technologies, and we can’t just allow the consolidation of power to continue in the hands of the established top tech companies. This technology is and should be a public good for anyone and everyone to use in whatever way they like. At the very least, we need an escape hatch and an alternative.
AIOZ Network is certainly something I will be keeping close tabs on throughout the next few years. I am looking for great execution on their roadmap and for them to be able to edge their way into the conversation when people think of AI and DePIN.
What I like from a speculative perspective is the fact that there is no ridiculous VC overhang with regard to token allocation, plus the fact the team has been building in this field for over ten years.
It is certainly something to add to the watchlist.
Please don’t take anything you read as financial advice; I literally don’t know what I am talking about. This article is purely for informational purposes, and everything read here is public knowledge.