Author: Adeniyi Abiodun(CPO & Co-Founder @Mysten Labs) 

Translation & Comment: Steve Kim(CEO & Co-Founder @Four Pillars) 

We Wrapped Up 2024 With a Bang; Now the Real Party Starts in 2025

Folks, 2024 was the year we finally showed the world what “building for the long haul” really looks like.

Sui kicked off with a boom in DeFi, saw massive developer momentum, and showed the world what zero-friction onboarding really means.

By December, everyone was talking about how Sui wasn’t just another Layer 1 but something entirely different:

a truly user-friendly platform, capable of handling real-world workloads effortlessly.

link: https://x.com/EmanAbio/status/1873756691650728034

Now here we are, gearing up for 2025. And let me tell you, if you thought 2024 was crazy, wait till you see what’s coming next.

In 2025 we’re going way beyond “faster finality.”

We’re designing a future where Sui’s core engine becomes the backbone for finance, gaming, AI-driven agents, and everyday apps.

2025 will prove that Sui is the first blockchain that can handle every piece of that puzzle from start to finish.

So, let’s dive in.

Last Year We Delivered Sub-Second Finality; This Year We’re Taking It Even Further

With Mysticeti V1, we showed the world how a DAG-based (Directed Acyclic Graph) consensus could deliver sub-second finality under heavy load.

But since the moment we started designing Sui's core architecture, we knew our initial consensus design wasn’t the final word.

Our plan from day one was to build a chain architecture flexible enough to evolve once real builders came knocking with real needs.

That moment arrived in 2024 when developers, especially DeFi folks, pushed Sui’s shared-object model to the limit.

Sub-second finality was already a selling point, but DeFi demanded near-instantaneous finality for shared assets.

This meant cutting the last bit of latency, boosting throughput, and giving developers a simpler, more direct path to building lightning-fast apps.

That’s exactly what Mysticeti V2 is all about.

Sui's current architecture follows a "two-protocol" design.

We had a consistent broadcast layer and a consensus layer running in tandem, which meant your average transaction needed a flurry of signatures and message exchanges just to get validated.

By the time all the certificates are aggregated, the system piles up a whole bunch of cryptographic checks. This leaves only fewer CPU cycles available for efficiently executing user transactions.

Mysticeti V2 / FastPath basically bakes that broadcast logic right into the DAG itself.

Instead of pinging every validator for multiple signatures and gluing them back together, we’re paring things down to fewer certificates, fewer heavy crypto ops, and a lot less back-and-forth.

Practically, it will feel more like a classic transaction processing system: submit once to a validator, see it come out confirmed, and you’re good.

No more micromanaging a gazillion signatures or messing with complicated fallback flows. It’s all handled under the hood.

What’s the payoff?

For starters, validators are no longer swamped verifying big piles of proofs, so they can push more CPU power into actual transaction execution.

That means you will see more throughput and lower latency, plus simpler transaction flows.

What I’m personally excited about is how seamlessly this evolution fits into Sui’s architecture.

That let us seamlessly swap out Mysticeti V1 for Mysticeti V2, with no painful reboots for our validators and no massive disruptions for devs.

We’re aiming to roll this out thoughtfully in 2025 to ensure we keep that speed-stability balance we pride ourselves on.

And once we’ve tuned a single validator to the absolute max, the next move is to multiply that capacity by scaling beyond one machine.

The Speed Barrier is Already Broken; Now, Remora Will Keep Sui Scaling With Demand

Back when we first showed the world that Sui could finalize transactions in under a second, folks were already asking, “How long until you hit a bottleneck?”

The truth is, as any blockchain grows, there’s only so much you can push a single validator node before you’re hitting 100% CPU or rolling over memory limits.

We knew that if you throw hundreds of thousands of transactions per second at a single validator, no matter how well you tune the hardware, you’ll eventually max out.

But the promise of Sui is that it will never run out of capacity to process transactions when demand hits the network.

Not by magically claiming 'infinite bandwidth,' but by letting us scale resources so we can handle any traffic spike without breaking a sweat or ballooning your fees.

That’s why the next natural step in 2025 isn’t just about going faster on one big machine, it’s about spreading the work across many.

Yes, we’re turning what used to be a single-track process into a massively parallel system that can scale out as easily as you can add more hardware.

So, we set out to crack the horizontal-scaling puzzle, starting with our Pilotfish research.

Pilotfish gave us a glimpse of how linear scaling should look: you add more machines to the validator cluster, and you get a straight-up multiplication in TPS without wrecking your sub-second finality.

That research evolved into Remora, a family of techniques that systematically splits validator workloads across multiple machines.

No matter the complexity (be it DeFi mania or a huge on-chain gaming spree), Sui can adapt.

Those initial results looked so good that not only did we keep pushing, but universities started contributing, too.

So far, the results have been mind-blowing.

In test environments, we’re already clocking throughput in the hundreds of thousands of transactions per second (tps), and we’re not done yet.

What’s cool is that we’re doing this in a way real validators can actually run. You won’t need a single supercomputer hidden in a data center; you can piece together a cluster of smaller machines.

We’re taking it slow and steady, of course.

We’re watching how DeFi, NFTs, gaming, and data-heavy apps start testing the limits of single-machine validators.

By the time real-world use cases actually demand five or ten times more throughput, Remora will be primed to deliver it without forcing a complete tear-down of your existing infrastructure.

We’re basically future-proofing Sui so that, as the network grows, we don’t lose our biggest advantage: unbeatable speed.

Bottom line: if you thought last year’s performance lifts were impressive, Remora will show you how deep the rabbit hole goes.

Of course, once you’ve got a chain that’s effectively unbounded in throughput, you can start doing some pretty wild off-chain interactions that eventually settle on Sui.

And that’s exactly where Programmable P2P Tunnels come into play.

On-Chain Is Blazing Fast; Let’s Supercharge Off-Chain Transactions Too

When people first hear about “off-chain tunnels,” they often think of Lightning Network for Bitcoin, just a simple, closed-ended channel for P2P payments.

On Sui, though, we’re dialing that concept up a notch with Programmable P2P Tunnels.

link:https://x.com/kostascrypto/status/1872518706837766299

These P2P tunnels aren’t just for one-off transfers. they’re built to handle anything from low-latency grocery payments to real-time game logic.

It can even enable multi-player interactions that reintroduce their final state on-chain when you’re done.

That means you can lock some SUI tokens (or any Sui-based asset) into a tunnel, collaborate or transact off-chain, then close it out and confirm everything on-chain in a single sweep. Best part: Sui’s programmable P2P tunnels can provide unlimited throughput with zero latency.

How’s that possible? The secret sauce is combining a few key elements you won’t easily find elsewhere.

Since we treat each asset or data chunk as its own object, you can create as many tunnels as you want, no global contention.

Need extra privacy for a casino game or a complex P2P arrangement? You can slip in zero-knowledge proofs so that neither side sees the other’s details until it’s all wrapped up.

For complex use cases (like storing partial game states),@WalrusProtocol provides a robust off-chain file storage layer that’s still cryptographically tied to Sui.

And because we’re hiding all the complexity, like specialized cryptography or transaction bundling, behind a developer-friendly interface, the overhead for devs is minimal.

On top of that, Mysten Labs is actively collaborating with researchers @CarnegieMellon(Carnegie Mellon University) to further develop and refine these programmable P2P tunnels.

Now, once you have all that raw capacity to play with, the conversation naturally turns to how transactions get ordered and prioritized in a fair manner. Especially under heavy load.

We’re Not Just Patching MEV; We’re Rewriting The Playbook

Raw speed doesn’t help if users get front-run left and right.

For most blockchains, “MEV” sounds like a dirty word, a hidden tax on users.

Bots front-run your trades, pocket huge gains, and the average user just gets burned.

SIP-45 is here to bring clarity.

https://github.com/sui-foundation/sips/pull/45

It not only raises the gas-price ceiling (so you can bid more if it truly matters), but it also changes the rules of how transactions get broadcasted and slotted into the chain.

The goal? Make it more predictable to pay for guaranteed inclusion when it’s actually worth it.

Sure, you’ll see more intense gas-bidding wars if a thousand people all want to mutate that same liquidity pool object.

You might be thinking: So is this just another pay-to-win scenario where big spenders always cut the line?

Actually, the point is to give everyone, devs, validators, even everyday users, a clearer picture of who’s paying what and how those fees get priority.

At the same time, this also paves the way for some truly creative MEV solutions.

Think redistributing part of that “rush to trade” profit back to users or staking participants, rather than letting a handful of malicious searchers siphon it all.

Of course, we’re not naïve. MEV can’t just be wished away.

But if you can structure it so that regular folks get a piece of the upside, everything changes. Protocols can defend themselves from front-runs and sandwich attacks.

We’re designing a future where MEV isn’t a zero-sum game but a public resource fueling the entire ecosystem. SIP-45 gets that ball rolling by setting a fair playing field for ordering logic.

link:https://x.com/w_riches/status/1873923352332800012

But SIP-45 is just the first step in taming MEV.

Deeper projects like DAG observability (giving the entire network real-time insight into incoming transactions) ensure no single party holds an “information advantage.”

We’re creating a platform that doesn’t just shrug at high contention. We want high volume, we want intense usage, and we want a system that handles it with fairness built in.

But we’re not stopping there.

Decentralized Storage Used to Be Clunky; Walrus Changes Everything

In Q1 2025, Walrus Is Hitting the Mainnet.

Walrus is our bold new approach to storing data at scale, built natively on Sui’s coordination layer.

It’s not just some IPFS-like file dump.

Walrus is our vision of a distributed storage layer that runs with the same ethos as Sui itself: no single point of failure, massive scalability, and a built-in sustainable economic model.

Most blockchain storage systems stop at basic file saving or referencing.

Walrus goes further by letting developers build rules, compose with smart contracts, and unlock new use cases for data.

With Walrus, you can build truly end-to-end decentralized Apps.

You can store front-end files, user-generated content, or anything else you need. All of it is governed by the same trustless principles that Sui applies to the rest of the stack.

Soon anyone can anchor an entire website front-end, user files, or even large AI model artifacts into Walrus.

By using Sui’s object model, each piece of data can be treated like a strongly typed object on-chain.

Developers can now tie storage events directly into Sui smart contracts, letting them spin up dynamic applications.

Suddenly, large-scale data that used to be locked away in someone’s AWS bucket can be globally accessible, programmatically managed, and transparent in terms of who’s storing what.

Yet raw decentralized storage is only half the story.

If everything is plain text by default, you can’t very well build a DocuSign alternative or a private Dropbox on top of it.

That’s where ‘SEAL’ steps in.

We’ve Got Storage Figured Out; Now SEAL Is Making Data Privacy & Access Control a Non-Issue

SEAL is an encryption and access-control framework designed to give your data the confidentiality it demands, complete with sophisticated, programmable policies.

Think of SEAL as the rulebook that says, “These addresses can decrypt this data, under these exact conditions.”

You might be streaming a high-definition video, or building a DocuSign-type workflow where only specific parties can unlock the legal documents.

With SEAL, you’re not forced to rely on naive “public or private” storage.

Instead, you define the cryptographic policies that say, “Grant read access to user X for a week, then revoke it. Or require a certain on-chain event, like a contract signature, before data is shared.

This synergy of Walrus and SEAL delivers the Holy Grail for decentralized apps: a place to store big data plus a built-in mechanism for robust encryption and policy enforcement.

Building “Dropbox on Sui” or “Netflix on Sui” doesn’t have to be a pipe dream anymore; you actually have the primitives to make it real.

And because Walrus is set to arrive in mainnet form soon, 2025 is shaping up to be the year we see the next generation of enterprise-grade Web3 apps.

Storing data on a chain is one thing; unlocking it seamlessly, conditionally, and securely is the game-changer.

All of this sets the stage for the next step in Sui’s evolution: ensuring that we can route data and transactions in a way that’s path-aware, stable, and extra resilient.

Sui’s Core Is Fortified; It’s Time Reinforce Sui’s Networking Layer With SCION

When we talk about building an internet-ready blockchain, there’s one massive elephant in the room: the internet itself is far from bulletproof.

You can have the most secure consensus and the best cryptography, but if your packets get hijacked halfway around the world, all that blockchain security goes out the window.

That’s exactly why we’re embedding SCION into Sui’s global validator network.

By integrating SCION, Sui can bypass the endless single points of failure we rely on in the conventional internet.

We’re talking robust, unstoppable connectivity that seamlessly survives DNS hacks, BGP hijacks, and massive DDoS floods, all while preserving the open nature that blockchains stand for.

link: https://x.com/SuiNetwork/status/1841968585578840299

In 2025, we’re aiming for Sui Mainnet to fully adopt SCION under the hood.

That means the network gets triple redundancy, multiple SCION paths, multiple providers, plus a fallback to standard IP if needed.

That’s a real game-changer.

But none of this matters if building on Sui still feels too complicated or if devs can’t easily leverage all this horsepower.

That’s why, in 2025, we’re doubling down on developer experience and all the tooling that makes shipping on Sui feel just as smooth.

The Tech Barriers Are Conquered; Now Let’s Make Building on Sui Effortless for Devs

Sui is already the blockchain with the best DevEx in the whole industry.

link: https://x.com/EmanAbio/status/1838597119411929458

But for 2025, we’re turning that up to eleven, streamlining everything from how you query on-chain data to how you verify your Move code.

Our guiding principle? Write less boilerplate, get more done.

Whether you’re spinning up a DeFi protocol, or tinkering with some AI-fueled dApp, Sui wants you coding features, not wrestling with infrastructure.

First Up, RPC 2.0; A Smarter Way to Query Sui

We’re ditching JSON-RPC and going all-in on GraphQL, which pairs perfectly with Sui’s object-centric model.

That means fewer endpoint hacks and more natural queries, especially when you’re dealing with complex relationships (like multiple child objects or dynamic NFTs).

Plus, the RPC service is splitting off from fullnodes.

Meaning, no more “read vs. write” choke points. Indexers can grab transaction data in bulk, do the heavy lifting, and feed it into a stateless RPC layer that stays snappy even under massive load.

Meanwhile, the Move Registry is about to turn building on Move into a frictionless, community-driven process.

Next Is Sui Move Prover

Security is front and center on everyone’s mind, so we’re finally rolling out a developer-friendly formal verification tool.

link:https://x.com/b1ackd0g/status/1874531793476497837

The Sui Move Prover checks the correctness of your Move code by letting you write specs in plain Move assertions (instead of learning a new spec language).

The upside?

You can systematically prove code properties without switching mental contexts every five minutes.

This is a big step forward for advanced devs who want to guarantee safety in their DeFi protocols or next-gen dApps.

Speaking of security, Sui isn’t stopping at static analysis.

Then Comes Bugdar; The AI-Powered Smart Contract Auditing Tool

Bugdar is an AI-driven scanning system that pounces on potential vulnerabilities before they ever see the light of mainnet.

Built using an ensemble of frontier LLMs (like o1, Gemini, and Claude), Bugdar runs in real-time, so you can catch issues in your code from day one.

It’s like having a security researcher embedded in your IDE, constantly checking for reentrancy traps or funky code patterns.

And because it’s trained on real exploit scenarios, it can spot subtle threats that might slip past conventional audits.

We’re Also Rolling Out Move Registry

Move Registry acts like a universal naming service for packages and modules.

Similar to how NPM or crates.io changed the game for JavaScript and Rust, the Move Registry ensures you don’t need to memorize cryptic addresses every time you want to import someone else’s code.

Just reference human-readable package names and let the registry resolve everything for you.

Whether you’re building PTBs, referencing another dev’s on-chain library, or handling cross-project dependencies, the Move Registry cuts the overhead so you can actually focus on shipping.

Think fewer headaches, fewer collisions, and a smoother path to code reuse.

We want Sui devs to borrow and share code as easily as they do on Web2 platforms.

Speaking of making this easy for developers: we’re rolling out new SDKs to ensure that building on Sui is a matter of writing a few lines of code, not wrestling with complicated zero-knowledge circuits.

Code Less, Ship Faster; Let the SDKs Do the Heavy Lifting

A lot of people ask, “How do I actually build a Web3 app without hiring a dozen cryptographers?”

The answer: SDKs that let you tap into Sui’s power in just a few lines of code, no advanced crypto wizardry required.

We’ve been working hand-in-hand with devs to figure out what exactly needs to be packaged up: from basic transaction calls to advanced zero-knowledge hooks.

And because Sui’s architecture treats each asset as an object, once you have a well-structured SDK, mixing and matching functionality is a breeze.

By bundling up core Sui functionality into straightforward APIs, be they JavaScript, Python, or Move-based DSLs, we’re reducing the friction that used to slow down project timelines.

No more fumbling with half-documented endpoints or rewriting the same transaction logic from scratch.

You just import, configure, and start building.

Put it all together, and you’ve got a dev workflow that’s faster and more secure than ever.

But as we keep streamlining life for developers, we also want to make sure everyday users enjoy the same level of frictionless experience.

So, Devs Are Happy; Now Let’s Remove the Last Bit of Friction for the Users

One of our biggest pushes in 2025 is making Sui feel invisible to the average person.

That’s why we’re bringing together all the major breakthroughs under a single banner: frictionless user experience.

Next-Gen Passkeys

Passkeys are shaping up to be a game-changer for day-to-day usage, bridging biometrics and ephemeral session keys so you don’t need to sign over and over.

Imagine unlocking your DeFi swap with Face ID, then playing a Sui-based game without reauthorizing every action.

We’re also experimenting with ephemeral keys that automatically disappear once you’re done, meaning you never leave behind any “digital residue.”

It’s about giving you the speed of Web2 apps, plus the trust-minimized security of Web3, all in one flow.

But, even with Passkeys, people lose devices, forget passwords, or want a fallback plan for disaster scenarios. That’s where KELP comes in.

KELP: The “Safety Net” That Feels Like Magic

Think of KELP as your personal safety net for Sui accounts: you can set up flexible recoveries, attach 2FA logic, or even pair it with zkLogin as a backup login method.

KELP offers a smart layer of recovery logic, combining 2FA, fallback mechanisms, and even zkLogin for extra assurance.

Want to set up a scenario where Google sign-in stands in as your backup if you forget your mnemonic? Done. Want to tie your recovery to your Facebook or an additional passkey device? KELP has you covered.

link: https://x.com/kostascrypto/status/1859694527738446191

No more horror stories of lost seed phrases. KELP means you always have a path back to your account, minus the tears.

For power users and institutions, we’re taking things a notch further with multi-factor wallets. By weaving in hardware enclaves, multi-sig logic, and advanced key-loss protection from KELP, you get a flexible safety net. It’s enough to satisfy both self-sovereignty purists and those who prefer a guardian to bail them out in worst-case scenarios.

You can pick your level of custody: pure self-ownership, or a “shared guardian” approach where a trusted third party can step in if something goes south.

Put simply, it’s a best-of-both-worlds model that ensures you won’t lose sleep (or your assets) if you misplace a key. Now, with all the infrastructure and user-friendly features we have built for Sui, it has reached a point where interacting with Sui doesn’t feel “crypto-like” at all.

It’s just fast, safe, and user-friendly.

This sets the stage for the next big wave: gaming.

We’ve Nailed UX; This Year Gaming Will Bring Sui to the Masses

We used to hear that Web3 was too confusing, too slow, or too niche. But in 2025, we’re flipping the script with the @SuiPlay device and @PLAYTR0N OS.

Gaming is Sui's trojan horse for mass adoption.

Instead of forcing players into obscure wallets or complex sign-ins, it layers blockchain right into the gaming experience at the OS level.

No friction, no weird fees or toggles, just pure gameplay that taps into Sui’s real-time asset trading behind the scenes.

In 2025, you will be able to get your hands on the @SuiPlay devices that you have ordered. Behind the scenes, we’ve lined up big-name studios and major titles: think @panzerdogs, @playDARKTIMES, @xocietyofficial ,@samuraishodown, plus more.

We’re not just talking about a handful of indie experiments.

Over 65 studios are already building for Sui, with around 70 titles queued up for 2025 releases.

That’s how you spark a user explosion, thousands of gamers jumping into frictionless experiences without the usual wallet nightmares.

You want to sell a legendary sword mid-raid, buy an armor upgrade instantly, or swap NFTs with someone halfway across the world? Done in under a second

And that’s exactly how we bridge the gap between “on-chain assets” and the frantic reality of in-game action.

Then they start discovering the broader Sui ecosystem: maybe they convert their in-game gold into stablecoins, or stake it in a DeFi protocol.

Suddenly, you’ve got millions of new entrants in DeFi who never thought of themselves as “blockchain users.”

A single top-tier title can bring millions of active users who log in daily, and all those microtransactions flow into liquidity protocols, token markets, stablecoin swaps, you name it.

And which protocol could run most of this trading activity? Think @DeepBookonSui.

When a player trades, they're actually using DeepBook, Sui's fully on-chain order book capable of delivering a trading experience similar to that of a CEX.

The beauty is that gamers won’t even realize they’re doing DeFi until they see the results: cheaper fees, real asset ownership, unstoppable markets.

That’s when you get mainstream adoption, when the underlying tech stops being a barrier and becomes a gateway to more vibrant digital economies.

That exact same, near-invisible blockchain experience that can win over gamers also speaks volumes to traditional institutions.

With Sui, they see a platform that can scale globally, handle real-money transactions, and keep everything simple for the end user.

Next Up Is a RealFi Revolution; And We’re Just Getting Started

In 2024, the Sui ecosystem made waves by pushing its Total Value Locked in DeFi well past $1.75 billion.

That was a wake-up call for more than just crypto traders.

What’s coming in 2025, however, goes far beyond crypto speculation and yield farming.

Partnerships like the one with Ant Digital will result in the inflow of billions in fixed-income assets directly onto Sui.

That means an investor in Hong Kong can own a slice of a bond originated in Europe without cutting through layers of middlemen.

We’re already seeing stablecoin liquidity (USDC, FDUSD, and more) feed into this, along with cross-chain BTC integration from Lombard Finance.

On the retail side, think “wallet as a bank account” for people who’ve been shut out of legacy finance.

Instead of fighting cross-border wires or currency instability, they can log in with zkLogin, deposit (or receive remittances) instantly, and hold assets in digital form without local bank hoops.

That’s a complete game-changer for countries where opening a checking account is more painful than pulling teeth.

The result is a robust, interconnected network that feels less like a crypto playground and more like the global financial system 2.0, only cheaper, faster, and transparent by design. Once you open the door to real assets you're going to need trading infrastructure that can handle the rush without glitching.

That’s where DeepBook becomes a force multiplier.

Here, DeepBook becomes a global central limit order book for literally any asset you can tokenize, providing sub-second swaps and full transparency.

No gatekeepers, no clunky order handling, just a smooth, unstoppable engine that ties into DeFi’s massive liquidity.

Institutions see that and realize “on-chain” no longer means “slow or risky”; it’s an upgrade from siloed, aging financial pipelines.

And because Sui hammered out user friction last year, 2025 will see Sui welcoming an influx of institutional volume that was previously waiting on a reliable, low-latency L1.

The logical next phase?

Extending that same powerhouse infrastructure to become a global coordination layer for every corner of the digital economy.

And Finally, This Is the Year Where Sui's Arc as a Global Coordination Layer Really Begins

The core idea of Sui has always been to be more than just a “faster, cheaper blockchain.”

From day one, Sui was designed to solve a fundamental challenge: coordinating human activity across distance and time, without handing all power to a centralized middleman.

By 2025, we’re finally seeing that vision play out at scale.

Instead of a loose collection of siloed dApps, Sui is evolving into a global coordination layer for digital assets, connecting everything from cross-chain liquidity to 3D printing supply chains.

Coordinating Cross-Chain Assets with IKA

For Sui, “Global Coordination Layer” isn’t a tagline, it’s the logical endgame.

Once you realize Sui can manage complex digital assets, off-chain infrastructure, and cross-chain liquidity all in one place. You see how Sui can become the natural underlying fabric.

For anything that requires people or machines to converge on a single source of truth.

The proof is in the projects already leading the charge.

Imagine controlling your Ethereum or Solana assets straight from a Sui-based contract, but with zero reliance on the usual bridging headaches.

That’s what IKA brings to the table.

It’s a system that allows Sui-based smart contracts to sign transactions on other chains, thanks to distributed MPC signers and two-party computations (2PC).

There’s no single “bridge” holding all the keys.

Instead, each user remains an essential participant, so even if half the network turned rogue, your tokens wouldn’t magically vanish.

This changes cross-chain interactions from a fragile patchwork of solutions into a secure, user-centric model of true interoperability.

Suddenly, Sui becomes the universal ledger coordinating different networks, turning multi-chain complexity into something you can manage with a few lines of Move code.

3DOS: Rebooting Manufacturing Logistics

Then there’s 3DOS, a platform that coordinates global manufacturing via 3D printing networks.

They’re swapping out old-school supply chains for real-time, on-demand production, all governed by Sui’s unstoppable finality and minimal transaction friction.

You push a button, the system routes your request to the nearest idle 3D printer, and the item materializes, no warehouses, no half-baked logistics nightmares.

Meanwhile, Walrus ties it all together by offering a way to store massive data on-chain in a decentralized, composable manner.

It lets you store and retrieve data in a decentralized fashion, enabling you to build an entire off-chain solution, like a big 3D-printing pipeline or a supply-chain aggregator.

At the same time, you can rely on Sui’s consensus to coordinate who stores what, for how long, and at what price.

Now, imagine those same protocols powered by AI agents that can autonomously transact, verify, and reason about the data they process.

That’s where 2025 will really get exciting.

We Broke Down Every Barrier; AI Agents Are the Last Piece of the Puzzle

IMO, 2025 will be the year where we finally cross over from AI living in back-end servers.

AI will start living in front of our eyes, right in the thick of human interactions, social feeds, and on-chain transactions.

What makes these agents truly revolutionary is their ability to act with minimal human oversight while tapping into real money flows and services around the globe.

In practice, that means an agent you spin up to handle project management could instantly spin off a second agent specialized in web design. It could pay for its deliverables autonomously.

Finally, it could merge the results into a final package with no human micromanagement needed.

The challenge is making sure these agents can trust each other’s identities, run on verified code, and interact with real data without leaking private information.

That’s where Sui comes in.

By easily giving each agent a crypto wallet (zkLogin), Sui can remove the hassle of credit cards and bank accounts, letting them transact directly in digital currency. And because Sui is cheap and fast, microtransactions or frequent data updates don’t choke the system or blow up the budget. Meanwhile, features like attestation ensure an agent you hired for, say, code review is actually running the correct LLM model, not some hacked or backdoored version.

On top of that, you can store large data sets or entire AI models privately in Walrus, letting agents share and consume data without risking unauthorized leaks.

Add in Sui’s emerging DRM-like capabilities, which ensure your data or content is delivered exactly as you intend. It remains encrypted at rest and is protected by the chain of trust.

The result is a perfect sandbox for building next-level AI Agent-to-Agent coordination.

This is how we’ll see entire agent ecosystems emerge in 2025.

We’re already seeing early AI dev platforms like @Atoma_Network building on Sui.

IMO, we will soon see AI agents that can read, write, execute, and pay for services at scale with sub-second finality.

That’s not a pipe dream anymore; it’s the next logical step.

It’s the start of a whole new era, and we’ve barely scratched the surface of where these autonomous agents can go.

We’ve Built the Perfect Foundation; Now Let’s Take Sui Mainstream

Sui started out by blowing past old limits on speed and throughput, but in 2025, you’ll see that was just the opening act.

Looking back at how far Sui has come, it’s obvious Sui has outgrown the old narrative of “just another L1.”

Sui’s technology stack is finally hitting a point where even the biggest skeptics can’t ignore what’s been built.

Now, the only step left is to make the entire process of building on Sui as straightforward and inviting as possible. Our aim is simple: if you’ve got a bold idea for a Web3 app, Sui should be your first choice. Not because of hype or marketing, but because it’s undeniably the smoothest path to success.

We want Sui to be the bedrock of everything digital in 2025, and that means making Sui the easiest platform developers have ever touched.

Comments by Steve Kim

Even the achievements Sui made in 2023-2024 were things that existing Layer 1s could never accomplish (zk-Login, Mysticeti, Stashed, Dynamic NFT, Deepbook, suiNS, SuiPlay0X1), and the initiatives lined up for 2025 are truly remarkable to the point of being astounding. What's particularly bullish is that protocols like Sui-Walrus-Deepbook-SCION aren't operating independently, but are deeply intertwined throughout the 2025 roadmap, showing very tight relationships. This explains why Deepbook has been receiving constant attention. And naturally, this leads to high expectations for Walrus.

This isn't just about token sales - you can sense their ambition to build a truly complete Web3 stack. If this isn't just for show but genuine intent, I think Sui might show us another "Solana moment" of this cycle. And I believe that moment will likely come in 2025.

As someone who has been consistently following Sui for the past two years, I am incredibly excited about Sui's future trajectory.