Superintelligence Alliance (ASI): Complete 2026 Guide

The Artificial Superintelligence Alliance (ASI Alliance) was the most ambitious consolidation crypto had seen in the decentralized AI sector. It was announced on 27 March 2024 by three of the oldest AI-focused blockchain projects: Fetch.ai (FET), SingularityNET (AGIX) and Ocean Protocol (OCEAN). The plan was to merge three separate token economies into one, build a single counterweight to the centralized AI labs, and pool research, compute and developer communities under one governance structure.

Two and a half years later, the story has changed. The token merger went ahead in mid-2024. CUDOS, a decentralized compute network, joined later that year. Then in October 2025 Ocean Protocol left the Alliance after a public dispute. Anyone holding FET, or looking at the AI-crypto sector in 2026, needs to understand both the original design and how the Alliance actually played out.

What you will learn

• The mechanics of the merger: the fixed conversion rates (0.433350 FET per AGIX and 0.433226 FET per OCEAN), why FET was the base token, and why the planned ticker change to ASI was repeatedly postponed.
• The technology stack: Fetch.ai's autonomous agent framework (uAgents, Agentverse, ASI:One), SingularityNET's AGI research (OpenCog Hyperon, the MeTTa language), the decentralized compute layer CUDOS brought in, and the data-sharing tools Ocean contributed before it left.
• The market picture: how FET traded from its early-2024 peak through the merger and the Ocean split, and where it sits among competing AI tokens such as Bittensor (TAO), Render (RENDER), NEAR and Virtuals Protocol (VIRTUAL).
• Governance and supply: the planned supply of about 2.63 billion tokens, how much of it was actually converted, and what the Ocean dispute showed about the Alliance's governance.
• Investment considerations: a balanced look at what could drive the ecosystem and what could hold it back, including unconverted supply, foundation treasuries and concentrated selling.
• Outlook to 2030: scenarios for the ecosystem, with clear caveats about how uncertain they are.

Key takeaways at a glance

• The merger was technically successful but strategically mixed. Major exchanges handled the token conversion smoothly. Merging the organizations themselves went much less smoothly, as Ocean's exit showed.
• FET remains the live ticker. The Alliance's token still trades as FET on most venues. AGIX and OCEAN holders were moved into FET through exchange-side and on-chain migration tools.
• The Alliance is a federation, not a single company. Each member kept its own foundation, team, treasury and roadmap. That arrangement is the root of both the Alliance's flexibility and its governance weakness.
• Agents are the core thesis now. In 2024 the pitch was about pooling AI infrastructure broadly. By 2025-2026 the focus had narrowed to autonomous AI agents, agent-to-agent payments and a decentralized LLM layer (the ASI-1 family of models).
• The sector has many competitors. Bittensor's subnet economy, AI agent launchpads on Base and Solana, and compute networks like Render and Akash all compete for the same capital and developers.

How to use this guide

Each section stands on its own. If you hold FET or are thinking about it, read Investment Considerations and Competitive Landscape closely. If you are a developer, start with Technology Deep Dive. If you held AGIX or OCEAN and are unsure about migration status, see the FAQ, which covers conversion paths and the current status of OCEAN after the split.

A note on data: the AI-token sector is very volatile and live prices change by the minute. This guide uses fixed protocol facts (conversion rates, dates, supply design) and well-documented historical price points. For live price, market cap and circulating supply, check a real-time data source before acting on anything here.

The Artificial Superintelligence Alliance is a collective of independent organizations that agreed to share one native token and coordinate on building decentralized artificial intelligence. Its stated mission is to give an open, community-governed alternative to the closed research labs (OpenAI, Google DeepMind, Anthropic, Meta) that dominate frontier AI. It aims to do this by combining agent infrastructure, AGI research, data markets and decentralized compute into one stack.

The founding members

• Fetch.ai, founded in 2017 in Cambridge, UK, by Humayun Sheikh (an early investor in DeepMind), Toby Simpson and Thomas Hain. Fetch.ai builds infrastructure for autonomous economic agents: software that can discover services, negotiate, transact and coordinate for users or businesses. It runs its own Cosmos SDK-based mainnet, and FET also exists as ERC-20 and BEP-20 tokens. Sheikh became chairman of the Alliance.
• SingularityNET, founded in 2017 by Dr. Ben Goertzel, a long-time AGI researcher best known for the OpenCog project and for being chief scientist behind Hanson Robotics' Sophia robot. SingularityNET runs a decentralized marketplace for AI services and funds research toward beneficial AGI, mainly through the OpenCog Hyperon framework. Goertzel became the Alliance's CEO.
• Ocean Protocol, founded in 2017 by Bruce Pon and Trent McConaghy. Ocean built tools for tokenized data sharing: data NFTs, datatokens, Compute-to-Data for privacy-preserving analysis, and the Predictoor prediction-feed product. Ocean was a founding member but left the Alliance in October 2025.

Later additions

In 2024 the CUDOS community voted to join the Alliance. CUDOS is a decentralized cloud and GPU compute network built on Cosmos. Its token was folded into FET through a governance-approved migration, giving the Alliance its own compute layer. This fit the broader strategy: Fetch.ai had already announced a roughly $100 million commitment to GPU infrastructure (often called Fetch Compute) for training and serving models for the ecosystem.

How the merger was structured

The core mechanism was a token merger, not a corporate merger. The details matter:

• FET became the base token. Fetch.ai had the largest market cap of the three and the most mature chain infrastructure, so AGIX and OCEAN were converted into FET instead of all three being swapped into a brand-new token.
• Fixed conversion rates were set using price ratios at the time of announcement: 1 AGIX = 0.433350 FET and 1 OCEAN = 0.433226 FET. Because the rates were fixed, arbitrage kept AGIX and OCEAN prices tied closely to FET between the announcement and the conversion.
• Planned total supply of about 2.63 billion tokens. FET's existing supply of about 1.15 billion was to be expanded by about 1.48 billion newly minted tokens to cover the converted AGIX (about 867 million) and OCEAN (about 611 million).
• A planned rename to ASI. The original roadmap had FET rebranded to ASI. That rename was postponed during the 2024 rollout and has not been fully put into effect as a market ticker. Most venues still list the token as FET, sometimes with ASI branding next to it.

Timeline

• 27 March 2024: Alliance announced. The combined value of the three tokens was widely reported at around $7.5 billion at the time.
• April 2024: Each community voted to approve the merger through governance.
• June to July 2024: The token merger, originally scheduled for 13 June, began on 1 July 2024 after a short delay. Binance, Coinbase, OKX, Kraken, Upbit and others converted user balances automatically. Self-custodied holders could use the SingularityDAO-hosted migration portal and Ocean's migration contracts.
• Second half of 2024: CUDOS joined, and work began on ASI-branded products and a shared compute effort.
• 2025: Fetch.ai released ASI-1 Mini (February 2025), described as a Web3-native large language model tuned for agentic workflows, and launched ASI:One as a consumer and developer interface for agents.
• October 2025: Ocean Protocol Foundation announced it was leaving the Alliance. Fetch.ai leadership publicly alleged that Ocean-linked entities had converted large amounts of OCEAN into FET and moved them to exchanges. Ocean disputed how this was described. The episode is covered in detail in later sections.

What the Alliance is not

The Alliance is not a single legal entity that owns all the code. Each member keeps its own foundation, staff, treasury and roadmap, and coordination happens through an Alliance governing council. This federated design let projects keep their own identity, but it also meant a member could leave without the others being able to stop it, which is exactly what happened in 2025.

The Alliance was created at the height of the first big AI-token cycle, which followed ChatGPT's launch and NVIDIA's surge in 2023-2024. To see why it still matters in 2026, you need to look at the wider AI economy and at how decentralized AI has developed since then.

The centralization problem has gotten bigger

When the Alliance launched, its central argument was that frontier AI was concentrating in a handful of companies with access to huge compute budgets, proprietary data and favorable regulation. That trend has only grown stronger. Training runs for frontier models now cost billions of dollars. Hyperscalers have committed hundreds of billions to data center buildouts. Access to advanced GPUs is shaped by export controls and long-term supply contracts. The ASI Alliance argument, that open, permissionless infrastructure is a needed counterweight, is more relevant than in 2024. The open question is whether a token-coordinated federation can actually compete at that scale.

Agents became the dominant theme

The big change between 2024 and 2026 was the move from models to agents. Across mainstream AI, the focus moved from chatbots to software that takes actions: booking, buying, coding, negotiating and calling other services. This lines up closely with what Fetch.ai had been building since 2017. Several developments raised the value of agent infrastructure:

• Agent communication standards. Protocols such as Anthropic's Model Context Protocol (MCP) and Google's Agent2Agent (A2A) made it normal for agents to discover and call tools and other agents. Fetch.ai positioned Agentverse and its uAgents framework to work with these standards instead of competing with them.
• Agent payments. Stablecoin rails, account abstraction and new payment protocols for machine-to-machine commerce made it practical for agents to pay each other. Crypto-native agent networks have a structural advantage here, because wallets, micropayments and programmable settlement are native to them.
• Agent token speculation. The 2024-2025 wave of agent tokens on Base and Solana (Virtuals Protocol, ai16z/ElizaOS and many others) proved there was retail demand for the theme. It also split the sector's attention and capital.

Decentralized compute matured

GPU shortages in 2023-2024 created real demand for alternative compute markets. Render, Akash, io.net, Aethir and the CUDOS network (now inside the ASI ecosystem) all grew, but their utilization and revenue have been uneven. The lesson for 2026 is that decentralized compute works best for inference, fine-tuning and burst workloads. It is less suited to frontier-scale pretraining, which needs tightly connected clusters. The Alliance's compute strategy reflects this: it focuses on serving agents and mid-sized models, not on competing with hyperscaler training runs.

Consolidation as a sector-wide test

The ASI merger was the first large-scale attempt at token-level consolidation among independent crypto projects. Its outcome is a reference point for the whole industry:

• It showed that a multi-token conversion can be carried out at scale with exchanges cooperating. Billions of dollars in tokens moved with relatively little disruption.
• It showed that merging tokens does not merge incentives. Each foundation still had its own treasury, contributors and strategy, and those differences surfaced over time.
• The Ocean exit raised new questions: who controls converted treasury tokens, how a departing member is handled, and what protections token holders have when foundations disagree.

Other sectors, including gaming, DeFi and Layer 2 ecosystems, have watched these lessons closely. In that sense the ASI Alliance's importance goes well beyond FET's price.

Macro and regulatory context

The 2025-2026 market has been shaped by clearer US crypto regulation, bigger institutional participation through spot ETFs for Bitcoin and Ether (and, increasingly, other major assets), and a liquidity cycle that has favored large-cap assets over long-tail tokens. Current sentiment readings across majors such as BTC, ETH, SOL, NEAR and BNB are strongly bullish, but sector-specific tokens like FET have historically moved with higher beta in both directions. In AI specifically, regulators in the EU (the AI Act), the UK and the US are paying closer attention to model accountability, which could eventually favor transparent, auditable, open infrastructure. That is a long-term tailwind for the Alliance's thesis, but it is not a near-term driver of demand.

Why 2026 is a decisive year

By 2026, decentralized AI has to show traction, not just a story. The relevant measures are agent activity, fee revenue, model usage, developer retention and compute utilization. The Alliance has a long history, deep research credibility through Goertzel's team, a working agent platform through Fetch.ai, and its own compute layer through CUDOS. Whether those assets turn into lasting demand for the token is the central question the rest of this guide looks at.

The ASI ecosystem has one main token (FET), several legacy tokens at different migration stages, and a group of affiliated projects. This section covers each component, then the closest related tokens that investors usually compare with ASI.

1. FET (Fetch.ai / ASI Alliance token)

• Role: The unified Alliance token, used for staking, governance, network fees on the Fetch.ai chain, paying for agent services in Agentverse, and access to ASI products.
• Chains: Native on the Fetch.ai Cosmos SDK mainnet, with ERC-20 and BEP-20 versions bridged to Ethereum and BNB Chain.
• Supply: The planned post-merger total was about 2.63 billion, but the real circulating supply depends on how much AGIX and OCEAN was actually converted, plus foundation and treasury holdings. Check live supply data. Do not assume the planned figure.
• Price history: FET hit an all-time high near $3.45 in March 2024, around the time of the Alliance announcement. It fell sharply after the merger, in line with the wider correction in AI tokens, and came under extra pressure during the October 2025 Ocean dispute, when large conversions and exchange transfers were reported.

2. AGIX (SingularityNET, legacy)

• Status: Converted to FET at 0.433350 FET per AGIX. Most exchanges delisted AGIX after automatic conversion. Self-custody holders were directed to official migration tools.
• Ecosystem that lives on: The SingularityNET AI marketplace, the OpenCog Hyperon AGI framework, the MeTTa programming language, DeepFunding grants, and spin-offs like Rejuve (longevity) and SingularityDAO. Some of these ecosystem tokens still exist separately.

3. OCEAN (Ocean Protocol, former member)

• Status: Converted to FET at 0.433226 FET per OCEAN during the merger window. After Ocean left the Alliance in October 2025, the status of unconverted OCEAN and of the migration contracts became contested. Some exchanges and community channels discussed renewed support for OCEAN as a standalone asset. Anyone holding unconverted OCEAN should check official Ocean Protocol and exchange announcements for the current position.
• Technology: Data NFTs, datatokens, Compute-to-Data, Ocean Nodes and the Predictoor product for AI-driven price prediction feeds.
• The dispute: Fetch.ai leadership publicly alleged that Ocean-associated entities had converted hundreds of millions of OCEAN into FET and sent a large share to exchanges, putting selling pressure on the token. Ocean rejected that framing. The full facts may come out through further disclosures or legal proceedings. Investors should treat specific allegation figures as claims, not established facts.

4. CUDOS (compute layer)

• Status: Joined in 2024 through community governance, and CUDOS tokens migrated to FET.
• Role: Decentralized cloud and GPU infrastructure for training and inference across the Alliance. It sits alongside Fetch.ai's announced investment of about $100 million in compute.

5. ASI products: ASI-1 and ASI:One

• ASI-1 Mini (released February 2025) is a large language model built for agentic reasoning and Web3 tasks, and it is part of the Alliance's attempt to own a model layer.
• ASI:One is a chat-style interface where users can interact with agents registered on Agentverse. It connects the model layer to the agent marketplace.

Related tokens investors compare with ASI

• Bittensor (TAO): A decentralized network of machine learning subnets where miners compete to produce useful intelligence and validators score them. It has a Bitcoin-like fixed supply of 21 million and halving schedule, and its dynamic TAO (dTAO) upgrade in 2025 introduced subnet-specific tokens. TAO is the main rival for the title of AI-crypto blue chip.
• Render (RENDER): A decentralized GPU rendering and compute network. It migrated from Ethereum to Solana and has a burn-and-mint equilibrium token model. It competes with the CUDOS-based compute layer.
• NEAR Protocol (NEAR): A Layer 1 that repositioned around "user-owned AI", AI agents and chain abstraction, led by co-founder Illia Polosukhin, a co-author of the original Transformer paper. Its sentiment reading is currently 90/100.
• Virtuals Protocol (VIRTUAL): A launchpad and co-ownership protocol for tokenized AI agents, mainly on Base. It is the leading consumer-facing agent platform and competes with Agentverse for agent creators.
• Akash Network (AKT): A Cosmos-based decentralized compute marketplace with growing GPU capacity. It competes directly with CUDOS-based compute.
• Internet Computer (ICP): Promotes on-chain AI inference and fully on-chain applications. It is a different architecture but part of the same narrative.
• The Graph (GRT): Indexing infrastructure that has repositioned toward AI data access. It overlaps with Ocean's original data thesis.

How the pieces fit together

Taken together, the ASI ecosystem covers more of the decentralized AI stack than almost any rival: models (ASI-1), agents (Fetch.ai), AGI research (SingularityNET), and compute (CUDOS). With Ocean gone, it has a gap in data-sharing infrastructure. Bittensor competes mostly on incentivized model production, Render and Akash on compute, and Virtuals on agent distribution. ASI's selling point is breadth. Its weakness is that breadth without enough depth in any single layer can leave it beaten by focused competitors in each one.

The ASI Alliance's technology is best seen as four layers: the settlement layer (the Fetch.ai chain and FET token), the agent layer (uAgents, Agentverse, Almanac), the intelligence layer (ASI-1 models, SingularityNET's AI services and Hyperon), and the compute layer (CUDOS and Fetch Compute). This section explains how each works and how they connect.

Settlement layer: the Fetch.ai chain

The Fetch.ai mainnet is built on the Cosmos SDK with CometBFT (formerly Tendermint) consensus, a Proof-of-Stake design with fast finality. Validators stake FET, and delegators can stake through validators to earn rewards. Because it uses Cosmos, the chain supports IBC (Inter-Blockchain Communication), so FET can move across the Cosmos ecosystem. Bridges maintain ERC-20 FET on Ethereum and BEP-20 FET on BNB Chain, and that is where most DEX liquidity and most of the merger conversions happened.

Key properties:

• Staking and governance: FET stakers secure the chain and vote on governance proposals. The Alliance's merger and the CUDOS integration were both approved through governance on the respective networks.
• CosmWasm smart contracts: These support contracts for agent registration, payments and service agreements.
• Low fees: These are important for agent micropayments, which would be too expensive on many Layer 1s.

Agent layer: uAgents, Almanac and Agentverse

This is the most distinctive part of the stack.

• uAgents is an open-source Python framework for building autonomous agents. A developer defines an agent with an identity (a cryptographic address), message handlers, scheduled tasks and protocols, which are typed message schemas the agent understands. Agents can hold wallets and send or receive FET.
• The Almanac is an on-chain registry contract. Agents register their address, endpoint and the protocols they support, which makes them discoverable. Registration needs a small FET fee and periodic renewal, which helps discourage spam.
• Agentverse is the hosted platform and marketplace where developers deploy agents without running servers, publish them for discovery, and connect them to large language models. It works as an agent app store with search, analytics and hosting.
• DeltaV and ASI:One are user-facing interfaces where people describe a goal in natural language and the system routes it to suitable agents. DeltaV was the earlier interface. ASI:One is the newer chat-style product tied to the ASI-1 model.
• Interoperability: In 2025 Fetch.ai added compatibility with MCP-style tool calling and other agent frameworks (such as LangChain-based agents) so that agents built elsewhere can be wrapped and registered in Agentverse.

The result is a full lifecycle: build an agent, register it on-chain for discovery, host it, connect it to a model for reasoning, and get paid in FET or stablecoins when other agents or users call it.

Intelligence layer

ASI-1 models. ASI-1 Mini was introduced as a large language model tuned for agentic workflows: multi-step planning, tool use and coordinating other agents. Its main role is to orchestrate Agentverse agents, not to compete head-on with frontier general-purpose models. Later versions and fine-tunes have extended this role.

SingularityNET AI Marketplace. AI services such as vision, language and specialized models are published with standard APIs (gRPC) and paid for per call through a payment-channel system originally built for AGIX and now working with FET. It was one of the first working examples of a decentralized AI service marketplace.

OpenCog Hyperon and MeTTa. This is the most research-heavy part of the Alliance. Hyperon is an AGI framework built around the Atomspace, a distributed knowledge metagraph. MeTTa (Meta Type Talk) is a programming language designed for self-modifying, neural-symbolic AI. The goal is to combine symbolic reasoning, probabilistic logic and neural networks so that systems can generalize beyond pattern matching. It is long-horizon research, not a near-term revenue product, but it is the scientific case behind the Alliance's "superintelligence" branding.

Compute layer: CUDOS and Fetch Compute

CUDOS brought a Cosmos-native network for providing cloud and GPU resources. Hardware providers offer capacity, and users rent it for model training, fine-tuning and inference. Alongside this, Fetch.ai's roughly $100 million compute commitment aimed to deploy high-end NVIDIA hardware (H200/H100-class GPUs and similar) for ecosystem use. Typical workloads include:

• Serving ASI-1 and other models used by Agentverse agents.
• Fine-tuning models for specialized agent tasks.
• Providing compute credits to ecosystem builders and researchers.

Data layer: the gap left by Ocean

While Ocean was a member, the planned data layer used Ocean's data NFTs (ownership of a dataset), datatokens (access rights), and Compute-to-Data, which lets an algorithm run on private data without the data leaving its owner's environment. That privacy-preserving model was a strong match for AI training on sensitive data. Ocean's open-source code remains usable, but its exit means the Alliance no longer has a committed data-layer partner. Expect the ecosystem to fill the gap with in-house tooling, new partnerships or integrations with other data protocols.

How a request moves through the stack

1. A user asks ASI:One to find and book a hotel within a budget.
2. ASI-1 interprets the goal and searches the Almanac/Agentverse for agents that support travel-search protocols.
3. Several agents return offers. The orchestrator compares them, possibly calling SingularityNET services for things like review sentiment analysis.
4. The chosen agent completes the booking, and payment settles in FET or a stablecoin on the Fetch.ai chain.
5. Model inference runs on Alliance compute (CUDOS or Fetch Compute), paid for by the service operator.

Each step is a possible source of token demand: registration fees, service payments, staking to secure settlement, and compute rentals. Whether those flows become large enough to matter compared with speculative trading volume is the main economic question for ASI.

FET has been one of the most traded AI-sector tokens since 2023. Its market behavior reflects four forces: the general crypto cycle, AI-sector speculation, the supply effects of the merger, and the governance shocks that followed.

Phase 1: the run-up to the announcement (2023 to March 2024)

FET was one of the biggest winners of the 2023-2024 AI rally. The token climbed from under $0.10 in late 2022 to an all-time high near $3.45 in March 2024. The rally was driven by the ChatGPT boom, NVIDIA's earnings and a general rotation into crypto tokens seen as proxies for AI. The Alliance was announced close to the peak. The combined value of FET, AGIX and OCEAN at that point was reported at roughly $7.5 billion, which made the new entity one of the 20 to 30 largest crypto assets at announcement.

Phase 2: the merger window (April to July 2024)

Because the conversion rates were fixed, AGIX and OCEAN traded as proxies for FET during the governance and migration period. Any gap from the 0.4333-0.4332 ratios was quickly closed by arbitrage. The merger began on 1 July 2024. Major exchanges converted balances, delisted AGIX and OCEAN, and kept trading FET. In the same period, the wider market corrected after the March 2024 high, and FET gave back a large share of its gains even though the conversion itself went smoothly.

Phase 3: post-merger consolidation (late 2024 to 2025)

FET took part in the late-2024 rally, but agent tokens and Bittensor increasingly won the attention of AI-focused traders. Through 2025, FET traded far below its 2024 peak. Several structural factors weighed on it:

• Bigger float: The merger roughly doubled potential supply. Even with the same total market value, per-token price targets from 2024 no longer applied directly.
• Treasury overhang: Foundations and early stakeholders from three projects now held the same token, which created several possible sources of selling.
• Competition for attention: AI agent launchpads, subnet tokens and new Layer 1 AI stories split speculative flows.

Phase 4: the Ocean dispute (October 2025)

The sharpest shock came when Ocean Protocol Foundation said it was leaving the Alliance. Public statements from Fetch.ai leadership alleged that large amounts of OCEAN had been converted into FET and moved to exchanges. On-chain analysts tracked major flows to centralized exchanges. FET fell steeply during the episode as the market priced in both the selling and the reputational damage. The episode is important for market structure analysis because it showed that converted treasury supply was not locked. A departing member with a large token holding could sell it into the market.

Supply analysis

• Original FET supply: about 1.15 billion.
• Minted for AGIX conversion: about 867 million FET (enough to cover the full AGIX supply at 0.433350).
• Minted for OCEAN conversion: about 611 million FET (enough to cover the full OCEAN supply at 0.433226).
• Planned total: about 2.63 billion FET/ASI.
• CUDOS: A small additional allocation tied to the CUDOS migration.

Investors should pay attention to how planned supply differs from actually converted supply. Tokens reserved for conversion that were never claimed are an unresolved supply question. The Ocean exit made that question bigger for the OCEAN tranche. Always use live data that separates circulating supply, total supply and holdings controlled by foundations.

Liquidity and venues

FET is listed on almost every major centralized exchange, including Binance, Coinbase, OKX, Kraken, Bybit and Upbit. Korean exchanges have historically driven a large share of AI-token volume, which adds sensitivity to regional retail sentiment. Perpetual futures markets for FET are deep, so funding rates and open interest are useful sentiment indicators. During sharp moves, leverage-driven liquidations have tended to make volatility worse.

Correlations

FET has historically shown:

• High beta to BTC and ETH: It typically moves several times more than majors in both directions. With majors currently showing strong sentiment (BTC 90/100, ETH 87/100, SOL 88/100), the usual pattern is that AI tokens lag at first and then accelerate if risk appetite spreads into the long tail.
• Strong correlation with other AI tokens: FET, TAO, RENDER, NEAR and agent tokens often move as a group, reacting to NVIDIA earnings, major AI model releases and AI-sector headlines.
• Idiosyncratic shocks: Alliance-specific events (the merger, Ocean's exit, product launches) have produced large divergences from the sector.

Metrics to track

• Active agents registered on the Almanac and Agentverse, and the transaction volume between them.
• Staking ratio on the Fetch.ai chain (how much supply is locked compared with liquid).
• Net flows of foundation and treasury wallets to exchanges.
• ASI:One usage and the number of developers building on uAgents.
• Compute utilization and revenue from the CUDOS-based layer.
• FET's relative strength compared with TAO and the overall AI-sector index.

Evaluating FET as an asset means balancing a real, long-running technology ecosystem against structural supply and governance issues that the merger brought to the surface. This section sets out both sides.

Opportunities

• Breadth across the stack. Few decentralized AI projects cover agents, models, research and compute together. If the agent economy grows quickly, ASI is one of the few ecosystems that can capture value at several layers at once.
• A long track record. Fetch.ai, SingularityNET and CUDOS have been building since 2017-2019. Their code, research publications and partnerships have survived several market cycles. That longevity is rare in a sector full of short-lived launches.
• Research credibility. Ben Goertzel's AGI research program, including Hyperon, MeTTa and the annual AGI conferences, gives the Alliance intellectual depth that agent-launchpad competitors do not have.
• Agents are what the market wants now. Agent interoperability standards (MCP, A2A) and agent payments are mainstream topics in 2026. Fetch.ai's lead in agent registries, hosting and machine-to-machine payments lines up with the direction of the wider industry.
• Liquidity and access. FET is listed widely with deep spot and derivatives markets, which makes it one of the main tools institutions and funds use for AI-sector exposure in crypto.
• Possible relief once the overhang clears. If converted treasury supply is distributed and absorbed, and the market sees less forced selling, sentiment can recover. Supply shocks that are already priced in sometimes come before recoveries.
• Strong majors. With BTC, ETH, SOL and many large caps at strong sentiment readings, history suggests capital can eventually rotate into higher-beta sector plays like AI tokens.

Risks

• Governance fragility. Ocean's exit showed that the federation has no strong legal or economic mechanism to stop a member from leaving or selling its converted tokens. Other members could, in principle, also leave or change direction.
• Supply overhang. The merger roughly doubled potential supply. Foundation treasuries, unclaimed conversion reserves and ecosystem funds are all possible sources of selling. Their transparency is uneven.
• Token value capture. Much ecosystem activity (developer hosting, model use, research) does not strictly require holding FET. Agents can increasingly settle in stablecoins. Without strong mechanisms that create demand for the token, such as staking requirements, fee burns or required registration fees, usage growth may not turn into token value.
• Competition. Bittensor has a clearer incentive design and a stronger story about scarcity. Virtuals and similar platforms have stronger consumer distribution. Render and Akash have more focused compute businesses. Centralized AI companies are building their own agent platforms with far more resources.
• Execution risk. Coordinating several organizations with different cultures, priorities and treasuries is hard. Roadmap delays, such as the long-postponed ticker change to ASI, suggest operations can be slow.
• Narrative dependence. AI tokens trade heavily on mainstream AI headlines. A slowdown in AI investment, disappointing earnings from AI hardware makers or a broad risk-off market would hit the sector hard.
• Legal and regulatory risk. The Ocean dispute could lead to litigation, and the outcome is uncertain. Regulators may also look more closely at token mergers, treasury conversions and foundation-controlled supply.
• Volatility. FET has had drawdowns of more than 80% from cycle peaks. Position size and time horizon matter more than entry timing.

An analytical framework for FET

A disciplined way to judge FET is to track three questions over time:

1. Is real usage growing? Look at registered and active agents, transactions between agents, ASI:One users and compute revenue, and whether these are rising quarter over quarter, not just during hype periods.
2. Is supply becoming cleaner? Check for clear reporting of foundation holdings, falling net inflows to exchanges from treasury wallets, and a rising staking ratio.
3. Is governance becoming stronger? Watch for clearer Alliance rules, any formal legal structure, enforceable commitments from members, and resolution of the Ocean dispute.

If the answer to all three is improving, the long-term picture is constructive. If usage stays flat while supply keeps growing, rallies are more likely to be driven by sector rotation than by fundamentals.

Practical considerations for holders

• Migration status: Leftover AGIX or legacy tokens in self-custody may still need official migration tools. Use only official links announced by the project and be careful of phishing sites pretending to be migration portals, which have been common.
• Staking: Native FET staking on the Fetch.ai chain earns rewards but involves an unbonding period. That matters for liquidity during volatile periods.
• Chain choice: ERC-20, BEP-20 and native FET are different representations. Make sure you use the correct network when depositing to exchanges or bridges.

The decentralized AI sector in 2026 has several distinct segments: incentivized model networks, agent platforms, compute markets, data networks and AI-focused Layer 1s. The ASI Alliance spans several of them, so it competes in each. This section compares ASI with its main rivals.

ASI vs. Bittensor (TAO)

• Model: Bittensor runs subnets, each a competitive market for a specific AI task, where miners are rewarded in TAO based on validator scoring. ASI is a federation of products and research groups and does not have a single competitive incentive mechanism.
• Tokenomics: TAO has a fixed 21 million supply with Bitcoin-like halvings, which gives a strong scarcity story. FET's supply was expanded by the merger, which weakens comparisons based on scarcity.
• Strengths of ASI: More developed agent tooling for end users, mature research, and a working consumer product (ASI:One).
• Strengths of Bittensor: Clear incentive alignment, a fast-growing number of subnets after the dTAO upgrade, and strong mindshare among crypto-native AI developers and funds.
• Verdict: TAO has generally overtaken FET as the market's preferred AI-crypto blue chip. ASI's best counter-argument is usability and the integration of agents with applications.

ASI vs. Virtuals Protocol (VIRTUAL) and agent launchpads

• Model: Virtuals lets anyone launch a tokenized AI agent with co-ownership through agent-specific tokens, mainly on Base. It is consumer-first and driven by speculation.
• Strengths of ASI: Infrastructure depth (registry, hosting, protocols, payments), an orientation toward enterprises and developers, and agents built for practical tasks rather than social-media personas.
• Strengths of Virtuals: Distribution, viral growth and a direct link between agent popularity and token demand.
• Verdict: They serve different markets. Virtuals captures speculative agent demand, while ASI targets useful agent infrastructure. If agent utility grows, ASI's positioning may age better. If agent speculation keeps dominating, Virtuals-style platforms will capture more of the flows.

ASI vs. Render (RENDER) and Akash (AKT)

• Model: Render focuses on GPU rendering and AI compute with a burn-and-mint equilibrium. Akash offers an open Cosmos marketplace for general cloud and GPU compute.
• Comparison: ASI's compute layer (CUDOS plus Fetch Compute) serves its own ecosystem first. Render and Akash are standalone compute businesses with clearer metrics on utilization and revenue.
• Verdict: For pure compute exposure, specialists are easier to evaluate. ASI compute is a supporting layer, not the main reason to own FET.

ASI vs. NEAR Protocol (NEAR)

• Model: NEAR is a general-purpose Layer 1 that has rebranded around "user-owned AI", chain abstraction and agent infrastructure, with privacy-focused inference and an AI research program.
• Comparison: NEAR has a broader DeFi and application ecosystem and strong technical credibility (Illia Polosukhin co-authored the Transformer paper). ASI has a longer, more specialized history in agents and AGI research.
• Verdict: NEAR is a platform bet that includes AI. FET is an AI bet that includes a platform. Current sentiment on NEAR is strong (90/100), which reflects its wider ecosystem appeal.

ASI vs. data networks (Ocean, The Graph)

Ocean, now independent, competes as a decentralized data and prediction-feed protocol. The Graph has moved toward AI-friendly data access. Without Ocean, ASI has a weaker data story, and data access is an increasingly important part of AI value chains.

ASI vs. centralized agent platforms

The most serious competition comes from outside crypto. OpenAI, Google, Anthropic, Microsoft and Amazon are all building agent ecosystems with tool use, agent marketplaces and enterprise integrations. Their distribution and model quality are far beyond any decentralized competitor. ASI's realistic role is not to replace them but to provide neutral, permissionless infrastructure for agent discovery, identity and payments that works across providers. That is a niche, but potentially an important one.

Comparative summary

• Breadth of stack: ASI is the leader, followed by NEAR, then Bittensor.
• Incentive design clarity: Bittensor leads, Render is strong, and ASI is moderate.
• Consumer distribution: Virtuals leads, while ASI:One is building.
• Supply discipline: TAO is strongest. FET is weakened by the merger expansion and treasury overhang.
• Research depth: ASI (SingularityNET) and NEAR stand out.
• Governance stability: Single-team projects are generally steadier than the federated Alliance, as Ocean's exit showed.

ASI's competitive position is best described as a diversified incumbent under pressure. It has deep assets and a long history but faces sharper, more focused competitors in each segment. Its long-term relevance depends on proving that integrating several layers produces something specialists cannot easily copy.

Forecasting a sector as volatile and fast-moving as decentralized AI is uncertain by nature. The following scenarios are frameworks for thinking, not predictions, and each depends on observable milestones. They are meant to help readers judge new information as it comes in.

Key drivers to 2030

• Growth of the agent economy: Whether autonomous agents become a major part of digital commerce, with widespread agent-to-agent transactions and machine payments.
• Open vs. closed AI: Whether open-weight models stay competitive with frontier closed models, which affects demand for decentralized inference and model coordination.
• Alliance governance: Whether the Alliance develops a stronger legal and economic structure, settles the Ocean dispute, and keeps (or adds) members.
• Token value capture: Whether FET becomes structurally needed for agent registration, staking, settlement or compute access, or whether stablecoins take over settlement.
• The macro and crypto cycle: Liquidity conditions, regulation and institutional appetite for sector-specific tokens beyond the majors.

Bull scenario: ASI becomes the neutral agent layer

In this scenario, agent standards come together, and the Fetch.ai registry, identity and payment rails become a widely used neutral layer for cross-platform agents. ASI:One and the ASI-1 models get meaningful usage. CUDOS-based compute serves a growing share of agent inference. The Alliance tightens governance, publishes clear treasury reports, and introduces stronger token sinks such as fee burns or staking requirements for agents. In this case FET could regain a top-tier position among AI tokens, and its large potential supply would be absorbed by real demand. Signals to watch include sustained growth in active agents, published fee revenue, institutional products that include FET, and partnerships with mainstream AI or enterprise platforms.

Base scenario: a durable but niche ecosystem

In this scenario, the Alliance stays a respected but secondary player. Agent infrastructure gets some real adoption, mostly among crypto-native and research users, while mainstream agents run mainly on centralized platforms. FET trades as a high-beta AI-sector token, rallying in risk-on phases and underperforming in downturns, without a sustained break from the sector. Governance stabilizes but stays federated. Bittensor and focused specialists keep leading their segments. This looks like the most likely path given current evidence, but it can change quickly in either direction.

Bear scenario: fragmentation and dilution

In this scenario, more members drift away or reduce their commitment, treasury selling continues, and agent usage moves to centralized platforms and stablecoin settlement without needing FET. Legal disputes drag on, and new competitors attract developers. The token's relative performance keeps declining, and the brand becomes mainly associated with the 2024 merger rather than lasting products. Signals to watch include falling developer activity, continuing exchange inflows from foundation wallets, roadmap slippage and departures of key leaders.

Milestones to watch, 2026-2028

• Resolution of the Ocean split: Clarity on the status of unconverted OCEAN, any legal outcome, and how the Alliance changes its membership rules.
• The ticker change to ASI: Whether the long-planned rebrand finally happens across exchanges, signaling operational maturity.
• Token economics upgrades: Governance proposals introducing burns, agent staking requirements or fee-sharing that tie usage directly to token demand.
• New ASI-1 releases: Whether the Alliance's models stay useful for agent orchestration as open-weight models from major labs improve.
• Hyperon milestones: Production-grade releases of the OpenCog Hyperon framework and MeTTa tooling, which would support the Alliance's AGI research credibility.
• Compute metrics: Published utilization and revenue for the CUDOS-based layer and Fetch Compute.

Longer horizon: 2028-2030

By the end of the decade, the core question is whether open, token-coordinated AI networks can capture a meaningful share of an AI economy dominated by a few companies. If AI agents carry out a significant share of digital transactions by 2030, neutral infrastructure for identity, discovery and payment could be very valuable, and crypto networks are naturally suited to provide it. The ASI Alliance has a strong historical claim to this space. Whether it holds that position depends on governance discipline and token design, not just technology.

Important caveats

• Price targets for AI tokens over several years have historically been unreliable. Supply changes, sector rotation and macro shocks dominate outcomes.
• The Alliance's federated structure means the organization's makeup in 2030 may look very different from today's.
• Breakthroughs in centralized AI, such as dramatically cheaper inference or agent platforms with native payments, could either grow the market for neutral infrastructure or remove the need for it.
• Regulatory treatment of token mergers, foundation treasuries and AI systems is still changing and could significantly affect the outlook.

Apart from FET's price, the ASI Alliance is a case study in how crypto projects consolidate. Several lessons apply to investors, builders and future merger attempts.

Lesson 1: Token mergers are technically feasible

The conversion worked. Billions of dollars in tokens across several chains and dozens of exchanges were converted with fixed ratios, coordinated delistings and migration portals. Fixed ratios plus exchange cooperation turned a complicated operation into one that was easy for most retail holders, who simply saw their balances change. Future consolidations now have a tested template to follow.

Lesson 2: Merging tokens does not merge organizations

The Alliance unified the token but left the foundations, teams and treasuries separate. That kept each project's culture and avoided legal complexity, but it also meant incentives could drift apart. When Ocean's priorities split from the others, nothing structural kept it in. Future alliances will probably need enforceable agreements, treasury lockups or vesting on converted tokens, and clear exit terms.

Lesson 3: Treasury conversions should come with commitments

The most damaging part of the Ocean episode was the large converted holding that could reach the market. Lockups, vesting schedules or public treasury dashboards for converted foundation holdings would have reduced uncertainty. Investors evaluating any token merger should ask how much converted supply belongs to insiders and what limits apply to it.

Lesson 4: Ticker and brand changes matter for execution

The long-delayed rename from FET to ASI created ongoing confusion about what the token is called and which assets are current. Clear, finished branding helps liquidity, onboarding and institutional acceptance. Partial rebrands show operational friction.

Lesson 5: The story has to become usage

The Alliance was announced at peak AI enthusiasm, and the story did much of the early work. Over a longer horizon, markets want usage numbers. Projects that publish clear metrics on agents, transactions, compute and revenue build credibility. Those that rely only on narrative lose it.

Lesson 6: Federations need conflict-resolution mechanisms

Multi-party crypto governance often assumes members will stay aligned. The ASI experience suggests alliances need formal dispute processes, defined steps for a member leaving (including the status of migration contracts and unconverted tokens), and transparent communication to holders when conflicts arise.

What it means for FET holders today

• The remaining Alliance is smaller but more aligned around agents, models, research and compute.
• Supply questions linked to the OCEAN tranche are an important variable to watch.
• Governance changes after the split, such as stronger membership rules or treasury transparency, would be a positive signal.
• The long-term case depends on Fetch.ai's agent platform and the Alliance's model and compute layers producing activity that truly needs FET.

What it means for the wider sector

Other AI projects, and crypto sectors more broadly, will study the ASI precedent before trying similar consolidations. The likely result is more careful structures: phased mergers, conditional conversions tied to milestones, legal entity integration, or looser partnerships without full token unification. In that sense, the ASI Alliance's biggest legacy may be what it taught the industry about consolidating, more than anything it built for FET itself.