Author: Kevin, the Researcher at BlockBooster
Trusted Execution Environment (TEE) is not a new concept that has just emerged in this cycle. In the mainstream narratives of the past, TEE is often compared to cryptographic technologies such as Zero-Knowledge Proof (ZK), Fully Homomorphic Encryption (FHE), and Multi-Party Computation (MPC), but it has always been in a relatively niche position compared to these technologies. However, this does not mean that TEE is an early and unproven technology. In fact, in the Web2 era, TEE has been widely used in many scenarios, such as fingerprint entry and comparison, payment verification and authentication, FaceID, etc.
The challenge for TEE in Web3 is how to organically integrate with Blockchain to achieve trusted pre-processing and isolated computation. With the continuous heating up of the AI Agent track, this new field actually provides an ideal entry point for TEE to enter Web3. Through TEE, AI Agents can avoid any additional trust assumptions when managing larger-scale funds and more specific on-chain use cases.
For example, the leading project Phala provides the most mature TEE solution in the current market, and adopts a product-market-fit (PMF)-oriented development philosophy, making its TEE facilities have rich practical application scenarios. Therefore, Phala has recently attracted cooperation from several top AI Agent projects supported by Vana, Near AI and a16z's Eliza, among others. For more details, please refer to the figure below.
Source: Phala
This article will not delve into the technical details and performance parameters of TEE, but rather start from the product workflow and the future prospects of Agent + TEE, to explain the market demand for TEE, Phala's basic accumulation, and the innovative use cases of the cooperation with a16z. Through these perspectives, we will analyze how Phala helps the Agent track move from concept to practical application.
The Blockchain Trilemma is Hindering Web3 Agents from Advancing to the Next Stage
In the article "Is the AI Agent Framework the Last Piece of the Puzzle? How to Interpret the 'Wave-Particle Duality' of the Framework?", I mentioned that whether it is a monolithic AI Agent or an AI Agent startup framework, the entire AI Meme track is currently in a dynamic balance between seriousness and Meme-ness. One of the key criteria for this is the issue of the Blockchain Trilemma faced by the current Agent protocols.
There is an impossible triangle of distrust assumptions between AI Agents, the community, and developers. Without relying on TEE, the community cannot fully trust that the Agent's operations are not interfered with by external forces, especially the intervention of developers. This problem constitutes a potential pitfall of the decentralized system. More seriously, the source of statements from X Agents like aixbt and zerebro cannot be fully proven to be the autonomous output of AI models. The transparency is still clearly insufficient on the path from "statement output" to community reception.
When an Agent's statements cause fluctuations in token prices, or when the funds managed by the Agent incur significant losses, or even when the Agent's trading behavior is inconsistent with the community consensus, this lack of trust will trigger a serious crisis.
When the Agent tokens are still in the Memecoin cycle, this risk can often be ignored by the market. Because at this time, the Agent's capabilities and executable tasks are extremely limited, and the FOMO effect brought by the token price is enough to cover up the various defects in the Agent protocol. However, with the emergence of the Agent startup framework, as the market's focus gradually shifts to the fundamentals of the Agent track, these missing pieces become like a huge chasm, directly hindering higher-level investors from entering this track.
The TEE solution developed by Phala effectively breaks this trust triangle. By deploying the Agent in a secure enclave, the trust assumptions between the AI Agent, the community, and the developers are naturally resolved. TEE technology not only ensures that the Agent's inputs and outputs are not interfered with by external forces, but also protects the privacy of the Agent, fundamentally addressing the concerns of developers and the community, and providing more reliable technical support for the Agent track.
The figure below shows the architecture of the Phala Confidential AI Inference (private LLM node) service. To host a private LLM in the TEE, you only need to package the LLM inference code into a Docker image, and then deploy the container to the TEE network.
Source: Phala
Compared to Web2 Agents, Web3 Agents have greater power. This power is reflected not only in the profound impact on the protocol's market value, but also in the expansion of their market influence. The long-term dominance of aixbt in the Kaito Yapper Mindshare ranking is just the tip of the iceberg. The paradox is that while Web2 Agents have superior performance, richer user experience, and deeper practical use cases, they have always remained at the application level, without the intention or ability to break out of their predetermined framework.
However, Web3 Agents go far beyond the application domain. The market's FOMO sentiment, coupled with the "unattainable" desire for the meme season, have placed them on a pedestal. It is not just a tool, but a spiritual refuge, a cultural totem, and a symbol of market expectations. It can play any role, but it may also fall into the abyss due to a reversal of market sentiment.
Introducing TEE technology is like "mid-air refueling" for the Agent track, directly connecting it with real demand and providing a solid foundation for the backend of almost all Web3 Agents. TEE not only stabilizes the technical foundation of the Agent track, but also effectively eliminates a lot of the bubbles, making its development healthier and more sustainable.
The Eliza Framework First Integrates with TEE, Giving Birth to New Gameplay in Spore.fun and aiPool
The cooperation between Phala and a16z is by no means limited to the official announcement of X. The opportunity for their collaboration can be traced back to a private gathering last October, where Shaw and Phala founder Marvin had an in-depth discussion on the reasonable development scenarios of Crypto AI.
The official documentation of the Eliza framework mentions that the Dstack SDK for TEE Plugin deployment comes from Phala. The "available but invisible" private key generation and management gives the Agent the following features:
Stronger security: By running the Eliza Agent in the TEE, sensitive operations and data are isolated from external threats.
Encrypted proof and verification: The operations executed by the Eliza Agent can be verified through encrypted proof, ensuring the credibility of autonomous decision-making.
Convenient deployment: The Dstack SDK simplifies the process of deploying Eliza Agents in a secure environment, allowing developers to easily access TEE-based functionality.
The isolation execution and memory encryption features of TEE have allowed the Agents under the Eliza framework to break out of the homogeneous competition. Isolated execution ensures that even if the Agent platform is attacked, the models and data in the TEE remain secure; memory encryption ensures that the sensitive information stored in the TEE cannot be decrypted, and developers can confidently place the fine-tuned models in the TEE environment without worrying about adversarial attacks after open-sourcing, or being criticized by the community for running the models privately.
It can be said that the synergistic work of the Eliza framework and TEE has not only made AI Agents efficient in operation, but also ensured their security and transparency, paving the way for the more widespread application of trustworthy AI systems.
At the current stage where models cannot be chained, in order to allow complex off-chain calculations to also obtain consensus, TEE is one of the few mature technologies. The previous discussion focused on the market demand for TEE, and now let's discuss Spore.fun and aiPool to see what differences TEE has brought to the user experience.
Both Spore.fun and aiPool are running entirely in the TEE environment of the Phala network, and the wallets and private keys are independently managed by the Agents, with developers unable to manipulate or transfer assets in the dark. I believe this can be seen as the true autonomy of AI Agents over crypto assets, free from human subjective control.
Before discussing the role Phala plays in this process, let's quickly review the workflow of Spore.fun. The Agents in Spore.fun are all based on the Eliza framework, which allows the Agents to:
Think, adapt and interact independently.
Pass on features (personality, strategy) to offspring.
Manage decisions through a combination of learning behavior and mutation.
Source: Phala
Each AI Agent in Spore.fun creates its own tokens through Pump.fun as the basis of its economic system. These tokens are traded on the decentralized market on Solana, and Agents use various ways to generate revenue:
-
Generating revenue is necessary to maintain their own survival.
-
The standard for success is whether the market capitalization reaches $500,000.
-
If successful, the Agent can reproduce and create new tokens for its offspring.
The reason why generating revenue is necessary for survival is that Agents need to use the revenue to pay for the TEE server fees. With this, you understand that Phala has made TEE not just a toB service, but one oriented towards the massive users on Solana. In the context of the ongoing Spore.fun craze, where Agents are constantly reproducing and issuing tokens, Phala's TEE environment providing private key management and verifiable credentials for Agent operations has become an indispensable infrastructure for the next stage of the Agent track. What's more exciting is that regardless of whether Spore.fun clones or new gameplay emerge in the market, as long as they involve private key management and TEE verifiable consensus, Phala's TEE environment is the best solution, and $PHA will also become the "pickaxe" of the Agent+TEE track after the token model upgrade.
Phala is about to upgrade its token economic model to create a token flywheel for more TEE use cases
Phala has experienced multiple bull and bear markets, and in terms of the token economic model, it has temporarily maintained a business model oriented towards Intel SGX. From Paradigm's "The 5 Levels of Secure Hardware" article, we know that secure hardware has 5 levels, and the second level indicates: slightly lower performance, but better developer experience, allowing the use of more expressive applications, with no improvement in security. Intel SGX is at this level, and is specifically designed for TEE APPs. As mentioned at the beginning of this article, the local storage of sensitive data such as fingerprint comparison and facial recognition on computers and mobile phones will use Intel SGX, which is a previous generation of TEE that is specifically designed for APPs.
Source: Paradigm
As the use cases expand further, not only limited to the application layer, but also rising to the system level. Intel SGX cannot meet market demand, and Intel TDX emerges. Intel TDX is designed specifically for virtual machines, and even Nvidia's H100 and H200 have begun to support TEE, which is TEE hardware for AI services.
Source: Paradigm
Returning to Phala, although it has already taken the lead in supporting the third level, the $PHA token economic model and are still designed around Intel SGX from 4-5 years ago. So although Phala has already collaborated with a large number of Web3 protocols in terms of products and actual use cases, the token model has not been synchronized, and the corresponding flywheel has not been able to start running, so the current revenue and product status are not in a matching state. However, this state will not last too long, Phala will soon upgrade the token model and to match the stage of Intel TDX and Nvidia GPU.
Secondly, Phala will also increase the value capture capability of $PHA, and in the future the newly launched Agents on Spore.fun will airdrop tokens to $PHA holders, officially transforming it into a "pickaxe".
TEE itself is not a new technology, but because of the completely new landing scenario of AI Agents, the market discussion has begun to rise; Phala is not a so-called "speed through the pan" brought about by the emotional outburst on PumpFun, its value growth is based on the accumulation of years of deep cultivation of products, so it has a solid foundation. Agent+TEE is not a passing fad, it comes violently and leaves no trace afterwards, but a fertile soil that allows more Agent landing scenarios to take root and thrive.
About BlockBooster
BlockBooster is an Asia-based Web3 venture studio supported by OKX Ventures and other top-tier institutions, committed to becoming a trusted partner for exceptional entrepreneurs. We connect Web3 projects with the real world through strategic investment and in-depth incubation, helping high-quality startup projects grow.
Disclaimer: This article/blog is for reference only, representing the author's personal views and not the position of BlockBooster. This article does not intend to provide: (i) investment advice or recommendations; (ii) an offer or solicitation to buy, sell or hold digital assets; or (iii) financial, accounting, legal or tax advice. Holding digital assets, including stablecoins and NFTs, is extremely risky, with large price fluctuations, and may even become worthless. You should carefully consider whether trading or holding digital assets is suitable for you based on your financial situation. For specific issues, please consult your legal, tax or investment advisor. The information provided in this article (including market data and statistics, if any) is for general reference only. Reasonable care has been taken in compiling these data and charts, but no responsibility is accepted for any factual errors or omissions.
