Re:Staking Weekly #10
The Now and Future of Restaking and AI, Eigen Games, and Aethir's New Edge Product Joining Restaking
Welcome to issue 10 of Re:Staking Weekly!
This week, I spent most of my time pondering why crypto would become not just a nice-to-have, but a must-have for an agentic future. More than that, I'm quite certain that restaking could play a big part in filling the gap between the current AI landscape and a fully autonomous future.
Also covered this week are the big EigenLayer x College DAO hackathon happening during ETH Denver, numerous EigenLayer product updates, and Aethir, a revenue-generating GPU cloud, launching its restaked edge GPU product.
Let's dive in!
AI x Restaking pt.2: The Now and Future of Agents and Restaking
This week, I read Satyaki's summary thread (read here, it’s great!) about the 12 AI AVS built on top of the EigenLayer ecosystem.
While the accelerated ecosystem onboarding deserves accolades, I couldn't help but notice that most AI projects, beyond their fancy marketing language, still follow a traditional model: providing a network of resources (e.g., GPUs) where users rent space to deploy AI models or agents.
Of course, starting here makes sense, but this model's fundamental problem is its reliance on a centralized or monopolistic entity controlling resource expenditure. If the resources vanish or the centralized entity becomes defunct, everything running on those resources — whether models, agents, or anything in between — disappears.
When we envision the future of AI, we imagine a landscape where agents and models break free from single-entity control. These agents would have autonomy over not only their actions (and therefore control their outcomes and revenue) but also the resources they consume (both physical and economic) and their ability to collaborate with each other.
A significant gap still exists between today's reality and the future we aspire to create.
The Keyword is Alignment
To achieve full autonomy, the key is actually alignment.
AI models and agents need to align with external actors — namely other agents, you, me, and the world. They need to align in terms of their autonomous actions but also in their existence and resource consumption. Not all agents are created equal — resources should flow toward higher-output and aligned AI.
But how? There are two ways to go from here.
The first way, similar to how actions are being hard-coded as pluggable modules, is giving agents the authority to decide how many resources they want to consume. This is a pretty mid idea, since alignment is hard to maintain. The worst case? Think SkyNet from Terminator — not exactly the future we're aiming for.
The second way, which requires quite a bit more ambition, relies on market incentives. You probably see where I'm going at this point — a market-driven, coordinated approach that directs both physical resources and economic security toward aligned agents. Yes, I'm talking about restaking.
Restaking addresses the alignment issue by dynamically allocating resources on-demand, ensuring that agents have the computational power they need when they need it (remember the past narrative of restaking being the verifiable SaaS/developer platform?). The more aligned an agent is with its broader audience, the more resources — both in terms of economic security and physical compute resources — will be allocated to it. Agents that don't perform will naturally be limited by the resources they can pull, therefore limiting their growth potential.
Meanwhile, this mechanism is completely market-driven and doesn't rely on agents' decision-making. It's good old capitalism applied to our non-human friends. Of course, it will be hard to take away the ownership, but I believe this shift from centralized/monopolistically controlled agents to fully autonomous ones not only creates more aligned and useful AI agents but also encourages agent-to-agent and agent-to-human collaboration in an efficient AI ecosystem.
Building Our Way Towards the Future
“The best way to predict the future is to create it”. — Peter Drucker
What products, then, can we build to fill the gap? Restaking is still pretty barebone, where a ton of new features can be added to quickly get us 80% there.
First, we need a platform to not only launch an agent but also allow people to restake, operate, rank, and use the agent. The current example would be EigenLayer's dashboard or Autonome by AltLayer (we are also working in a similar direction at EigenExplorer), but many features are still missing.
Second, we need an API or gateway that creates a unified interface for agentic communication. This allows for communication and collaboration among agents, ensuring they work together rather than in isolation. An example of this use case would be Hyperlane (see recent bounty request).
Lastly, there's restaking MEV. There's potential for a market where economic security can be speculated and arbitraged upon, allowing for the trading of resources and the creation of derivative markets on top of agents. This could add a new layer of market-directed efficiency to the AI ecosystem, further bridging the gap to a future where AI agents operate with greater autonomy and adaptability.
If you're interested in building to fill the gap of the agentic future or want to discuss any of these ideas further, send us a DM!
Eigen Games
EigenLayer and College DAO are giving out $50,000+ in prizes with the first Eigen Games, bringing together teams from 30+ universities to compete for glory and push the boundaries of innovation.
Hosted at the Curtis Hotel in Denver from Feb 22 to Mar 1, Eigen Games will create a focused environment for discovering and developing the next generation of builders in the Eigen Ecosystem, combining technical development, expert mentorship, and investor exposure.
For protocols and DAOs watching the AVS space, this represents a unique opportunity to identify early technical talent deeply trained in restaking mechanics.
If you're attending ETH Denver and interested in supporting university teams or connecting with emerging talent in the restaking ecosystem, reach out to Mustafa (mustafa@eigenlabs.org).
News Bites
Rewards feature will be upgraded to v2 (work payments) in 10 days.
EigenLayer’s first token-holder town hall was conducted on Jan 16th.
EigenLayer launches Sidecar v2.0, its real-time reward indexer.
Kairo’s research shows that Binance's bnSOL has recorded the largest weekly LST inflow in Solana’s history during the past week.
Aevo launches its $500k $EIGEN trading rewards campaign.
Aethir launches Aethir Edge, an edge-based GPU device, and announces its integration with EigenLayer.
That's it for this week's newsletter! As always, feel free to send us a DM or comment directly below with your thoughts or questions.
If you want to catch up with the latest news in the restaking world, give our curated X list a follow!
See you next week and thanks for reading,