AI Sector

AI has been the hottest narrative and mindshare in the Web3 space for a while. It is the Metaverse of this cycle, and its growth in both technical development and price action has validated its significance in the market.

We believe there is still much room for growth for AI in Web3. As shown above, the top 10 marketcaps of AI are still lingering at around 1/3 the valuation of MEMEs and still valued lower compared to DeFi. The internal consensus is that there is a good chance AI marketcap will catch up to these top 2 sectors.

However, funds and token holders have different tastes and thesis behind buying and holding a token. Despite being bullish in the AI sector as a whole, rotations are inevitable, and in the short-mid-term timeframe, we believe mindshare will flow from ASI (FET + AGIX + OCEAN) to other major AI projects like RNDR and AR.

Thesis

The merger of FET, AGIX, and OCEAN was the main topic of the first week of April. Its spotlight also marked the local top of both AI and the general crypto market. Although the merger seems like a bullish news a whole, we internally believe this will be somewhat bearish for ASI as a whole post-merge.

1. Valuation

It seems not many have dug in-depth into the official ASI Alliance Vision Paper. Here is a brief summary.

They aim to build Artificial Super Intelligence and they are merging to make it happen. The snapshot valuation (dictated by the team) is simply a sum of the max supply FDV of all three projects - valuing ASI at ~$6.9 B (FDV) (a snapshot of TWAP of past 15 days from the snapshot date.

The conversion of the tokens is 1 ASI: 1 FET: 2.309 AGIX: 2.308 OCEAN. You can already see a few tokens trading at a discount that you can trade towards the final conversion date—however, this probably deserves a separate article.

The key point is that ASI's valuation will come in second in terms of FDV (1st being TAO) and the largest coin in terms of MC in the AI sector (categorised by Coingecko) (this chart is with the ASI total supply implemented).

Based on the back-calculation of the % supply of MC of FET, AGIX, OCEAN combined - the MC of ASI is estimated to be around $4~4.1 B (@$6.9 B FDV). For the fundamental value of ASI to be priced as the top MC AI token seems unreasonable relative to other projects, such as RNDR or AR, being valued at much lower valuations.

2. Tokenomics

We have analysed the tokenomics and vestings of the three tokens. The FET token is close to being fully distributed and doesn’t seem to have any issues.

However, AGIX and OCEAN have direct quotes on distributing potentially additional supply that may be sold in the market towards the ASI conversion.

AGIX

In the ASI Visionary Paper, they propose the release of 100 M AGIX (roughly ~$100 M) to fund the purchase of more GPUs and CPUs to provide computing resources for the SNET network. The funding will obviously have to come from selling AGIX (ASI) in the open market. Unfortunately, the conversion of ASI is a very good event to offload some tokens while the spotlight is on them.

Although $100 M relative to ASI FDV may look relatively small, $100 M of TWAP selling when there is less interest from new buyers of the newly launched ASI token can eventually bring ASI into a much lower valuation.

OCEAN

OCEAN also had ~$800 M worth (currently around $635) of OCEAN ready for community incentives. Although this does not have the concept to the AGIX distribution, mature projects giving out community incentives often lead to selling pressure - because there is less upside relative to a fresh project diluting incentives to grow their ecosystem. An example is the dYdX DAO incentives - the effectiveness of the dYdX incentives wasn’t worth the dilution looking back.

This will have to be monitored more closely post-merger. However, it is something that we can keep in mind for the mid-run.

April 15 Snapshot

May 8 Snapshot

Because the three entities will govern and run as independent legal entities, there is a small chance that the interests of the three token holders may not coincide in the future post-merge

For example, FET has been 90%+ vested for a reasonable amount of time. This is a strong valid enough thesis for some whales to hold and stake a big stack. However, as the merger realises, the ASI tokens will have potential vesting dilution coming from the AGIX and OCEAN which are only 40-60% fully diluted relative to MAX Supply. Also the $800 M worth of OCEAN incentives may also not be in-line with the consensus of, for instance, FET and AGIX holders.

There is a lot of uncertainty that will need to be re-evaluated post-merger since the proposal was initiated by the foundation, not the community.

Bullish on AR, RNDR

Easy calculation is the sum MC / FDV of RNDR and AR. The combined MC of RNDR and AR is ~$6.7 B (@$8.3 B FDV). Although the MC is larger than ASI - this is mainly because RNDR is 72% and AR is 100% diluted and AR is fully vested. FDV is around similar value.

Although ASI and AR, RNDR (more often categorised as DePINs) are not the same project per se, market treats them as both an AI sector beta token and their fundamental value in the AI sector is different in an impactful manner.

RNDR / AR - the backbone of AI

NVDIA’s value in the market comes from AI’s demand for more GPU power. The same flows within crypto - most crypto AI projects require GPU power and a storage layer for efficient AI computing power. Ultimately, the foundational layer often deserves a higher valuation relative to applications built on top, and this is the case in both Web3 and Web2.

https://companiesmarketcap.com/

https://www.coingecko.com/

The majority of top MC projects are L1s, and rarely does a project built on top of an L1 surpass the original L1’s MC / FDV. Although the concept is not directly applicable here, the point is that RNDR and AR are built as a general infrastructure for AI in Web3 to thrive—and without these Lego blocks, the chances of AI blooming in the Web3 scene are less likely.

RNDR - What would big funds buy?

First, RNDR is the leading decentralised GPU powerhouse in the industry. There are a few competitors, such as IO.net (which hasn’t launched a token yet)—however, RNDR has been around the longest and has the biggest connection in both Web3 and Web2.

Funds have the fiduciary duty of protecting the fund's capital - their top priority is to incrementally increase the fund's size without blowing up the fund. Therefore, they are forced to trade conservatively - and this is one of the biggest reasons why funds only trade top MC projects. High FDV projects are backed by top-tier VCs and have a very low chance of “rugging” token holders.

Therefore, funds often trade the top 1-5 MC projects in a sector - explaining explosive moves behind tokens such as NEAR even if its FDV was over $5 B. Vaneck also has a very good report noting multiple projects in the AI sector and you can easily notice they are mostly talking about $1 B+ FDV projects

https://www.vaneck.com/us/en/blogs/digital-assets/matthew-sigel-vanecks-crypto-ai-revenue-predictions-by-2030/

Therefore, RNDR, as the market leader, is a very safe and easy selection for portfolio inclusion for big funds such as Vaneck. Additionally, RNDR is one of the few Web3 projects that spoke on the panel during the NVIDIA Conference in March. Additionally, one can imagine the amount of exposure RNDR has globally in the Web2 scene relative to all the other Web3 isolated AI projects.

Additionally, the uncertainties behind the synergy behind the merger of ASI are not favourable risks for funds to gamble on by choosing the highest FDV token in the sector. Funds would rather prefer a safer and less uncertain roadmap-based project such as RNDR.

AR - AO

https://www.communitylabs.com/blog/a-quick-guide-to-ao-the-hyper-parallel-computer

On February 13, 2024, Sam reveled AO - a parallel computer built on top of Arweave. This is a revolutionary development for AR as its main functionality as a storage network wasn’t getting much of the attention it needed from the AI community. As an independent network, the leverage they have in seamless interoperability between AO and AR is a significant advantage AO has over other projects, such as building something similar and integrating with other independent platforms such as Filecoin or ICP.

https://x.com/samecwilliams/status/1757161860028150159

We internally believe the market is mispricing AR as simply a storage layer (it is valued at $2.5 B MC and FDV) and holds an opportunity for a lot more upside when it starts valuing AR as an AI project itself. This is also because not many understand AO should be treated as an efficient computing Layer 1, benefiting from the direct storage accessibility provided by AR.

There are also ongoing speculative (which we think is relatively reasonable) rumours on a potential AO airdrop to AR holders as the two protocols are directly intertwined. However, this was not the case with Fluence (FLT) and Filecoin (FIL) at the launch of FLT (a similar relationship between AO <> AR).

In essence, an L1 with extremely fast computing capacity built on top of Arweave should give the valuation of AO to be minimum $1 B+. The expectation from the community regarding AO is extremely bullish - and the only way of front-running the launch seems like a holding spot for AR.

https://stats.dataos.so/arweave

https://www.ao.link/

As we approach the launch of AO, Arweave's fundamentals have been thriving. While AO’s launch hasn't garnered widespread attention yet, the team's dedicated efforts to grow and expand the AO ecosystem are evident. For instance, AO Ventures has been incubating projects, and it's now in the third week of the incubation program. Qualified projects will soon be pitched to top-tier VCs, which could significantly alter the landscape of the AO ecosystem.

https://x.com/CommunityLabs/status/1778410876799795491

Additionally, there are already a few protocols that have successfully launched (or expecting to launch) on AO such as PADO Labs (zkFHE), AR Swap, Permaswap, etc… The ecosystem is already receiving some traction with some MEMEs launching on top too.

Conclusion

In the crypto world, projects with upcoming narratives and a degree of uncertainty often attract more attention than those that have already revealed 90% of their plans. For example, the outlook for ASI and its token conversion may seem less exciting when considering the potential of AO or the possible allocation of funds into RNDR.

While AI is a highly promising sector in this cycle, there will inevitably be winners and losers. Large funds and traders can't trade the intraday moves of micro-cap launches occurring daily. We see a potential opportunity as money could flow from ASI basket tokens into fundamentally stronger projects like RNDR and AR.

Internally, we're optimistic about RNDR and particularly AR, due to their attractive FDVs, which are relatively cheaper compared to other protocols. We believe AO is still in its early stages and could easily lead the market. While there are DeFi, NFT, and Gaming-focused L1s, AI-focused L1 / L2s are still under development. Given this, we believe AO, backed by Arweave (an ideal infrastructure for AI-focused layers), could get a significant head start in the AI L1 race.

Disclaimers

The content of this article should not be construed as financial advice.”