-A_N_K-

Kite has always believed in a future where AI agents can handle everyday tasks smoothly and safely. That’s why the arrival of Google’s Agent Payments Protocol (AP2) feels like a perfect match for our vision. AP2 sets the standard for how AI agents make payments, while Kite provides the blockchain layer that actually processes and settles those transactions. Together, they create a complete system—AP2 defines what needs to happen, and Kite delivers how it happens.
We’re embracing AP2 because it aligns beautifully with our agent-first philosophy. By making Kite fully “AP2 compatible,” we aim to support easier adoption, encourage innovation, and help the entire ecosystem grow. Our goal is to build smooth integrations, reference examples, and developer-friendly tools that make AP2 more accessible for everyone.
What Exactly Is AP2?
AP2 is an open standard designed to let AI agents make payments securely on behalf of users. It is being developed with help from more than 60 global leaders—including Mastercard, PayPal, Coinbase, Adyen, and American Express.
Think of AP2 as a universal language for payments. Whether the transaction uses credit cards, stablecoins, bank transfers, or something else, AP2 ensures the same core principles:

clear authorization, strong verification, and transparent accountability.
For Kite, AP2 fits directly into our broader roadmap, especially the upcoming Kite AIR app store, where agent-powered applications can operate and monetize safely. AP2 provides the trust layer for agent payments, while Kite’s AI-native blockchain ensures those payments can be executed and settled quickly, securely, and at scale.
Why AP2 Matters for the Future
Today’s payment systems still assume that a human clicks “buy” or confirms a screen. But in the next phase of the internet, AI agents will take over many of those tasks—finding flights, renewing subscriptions, buying groceries, and more.
Without a shared standard, this shift could lead to confusion, fraud, or incompatible systems. AP2 solves this by giving AI agents a trusted, consistent way to interact.
AP2 also connects with Web3 through extensions like A2A x402, opening doors for crypto payments and stablecoin transactions. This is essential for the coming “agentic internet,” where AI agents operate across traditional finance and blockchain with the same level of trust.
How AP2 Works in Simple Terms
At the core of AP2 are Mandates—secure digital instructions signed by the user.
For instant purchases, you give an Intent Mandate (“Find me running shoes”), and your agent later sends a Cart Mandate for final approval.For ongoing tasks, like auto-buying concert tickets, the Intent Mandate sets the rules. Your agent then generates Cart Mandates only when the conditions match.
This creates a transparent trail that links your intention to the final payment, making each step safe and verifiable. Google has also released open-source documentation and developer kits, making it easier for builders to start using AP2 in real applications.
Where Kite Comes In: The Settlement Backbone
AP2 defines how agent payments should work. Kite makes those payments real.
Much like ERC-20 tokens rely on Ethereum for transaction settlement, AP2 relies on execution layers like Kite to bring transactions to life.
AP2 covers the “what” — authorization, mandates, complianceKite delivers the “how” — execution, settlement, interoperability
This harmony makes Kite and AP2 a powerful combination. We’re committed to strengthening this connection through integrations, reference implementations, and advanced features that make it easier for developers to build AP2-ready applications.
Together, Kite and AP2 can unlock new experiences—from automated shopping to smarter B2B purchasing—pushing the world closer to reliable, agent-driven commerce.
Looking Ahead
We believe AP2 is a major leap toward a future where AI agents can transact with full trust and independence. At Kite, we’re excited to help build that future by becoming a key settlement layer for secure agent payments.
If you’re exploring AI-powered payments or building agent-first applications, we invite you to join us in shaping this new digital economy—one where AI agents can act safely, confidently, and autonomously.
#KITE @KITE AI $KITE

We’ve entered a new moment for tokenized stocks. They are no longer just digital versions of company shares. Today, they unlock real on-chain liquidity and allow users to earn yield while still keeping full exposure to companies like Tesla or Nvidia. This shift is changing how people move capital, how collateral works, and how users interact with both traditional and blockchain-based markets.
Recently, Falcon Finance hosted a live X Space to explore this transformation. Two experts joined the conversation:

Leo, Marketing Manager at xStocks & Backed, and Artem, Chief RWA Officer at Falcon Finance.

Together, they broke down how tokenized stocks actually work, why investors are adopting them, and what the future might look like when most major assets live directly onchain.
A New Chapter for Tokenized Equities
Leo explained how quickly things have changed. A few years ago, mixing stocks with blockchain felt strange for many people. But today, institutions understand blockchain better, regulations are developing, and users are comfortable handling their own digital assets through secure, self-custodial tools.
In this environment, tokenized stocks like those from xStocks are not synthetic products or CFDs. These tokens are backed one-to-one by real shares held with regulated custodians. This means an AAPLx or TSLAx token always represents an actual share stored safely in a segregated account. These aren’t just digital numbers—they are real assets that now live onchain.
Why Falcon Finance Added Tokenized Stocks
Artem shared that Falcon’s thinking is simple:

DeFi should not be limited to crypto tokens alone.
If users can hold traditional equities onchain in a legal and transparent way, those assets should also plug into the same DeFi systems used by ETH, BTC, and stablecoins.
With tokenized stocks as collateral, Falcon users can create new strategies. For example:
A user can hold tokenized Tesla or S&P exposure,Mint USDf against that collateral,And earn delta-neutral yield through Falcon’s diversified strategies.
The collateral—whether BTC, stablecoins, or tokenized stocks—remains in a protected reserve. Falcon’s USDf yield does not depend on the asset’s direction. Instead, it comes from market-neutral activities like arbitrage, funding farming, options-based portfolios, and other transparent strategies shared on Falcon’s transparency page.
Where Users Can Access Tokenized Stocks
Leo explained that Backed is responsible for issuing xStocks, while many partners help distribute them. These include:
Global exchanges like Kraken,Wallets such as Phantom and Solflare,Telegram-based apps and onchain vaults.
Minting and redeeming is available to verified institutions and qualified retail users, depending on local regulations. Every country’s rules differ, so availability varies case by case.
How Tokenized Stocks Work Inside Falcon
Once a Falcon user completes KYC, using tokenized equities feels very similar to using ETH or BTC.
Falcon’s Classic Mint requires a minimum of 10,000 USDf. Artem gave an example:

Holding around 30 TSLAx, valued at roughly 13,000 USD, is enough to mint about 10,000 USDf after Falcon’s 20% overcollateralization requirement.
After minting:
The user still keeps their Tesla exposure,Gains fully liquid onchain USD,And can earn yield or use that liquidity across DeFi.
You stay invested, stay liquid, and stay onchain.
Turning Tesla Into Working Liquidity
Artem described a common scenario:

Many people own valuable Tesla stock but don’t want to sell it. Tokenized stocks provide a way to unlock liquidity without giving up their position.
The user can:
Use TSLAx as collateral,Mint USDf,And deploy that liquidity in lending, liquidity pools, or Falcon’s yield strategies.
Instead of choosing between “sell” or “hold,” tokenized stocks offer a third path: stay invested and stay liquid.
Who Uses Tokenized Stocks Today
Leo highlighted three major user groups:
Crypto-native users, mainly in Asia, who want exposure to large stocks without leaving blockchain rails.Centralized exchange users, who access tokenized stocks through platforms like Kraken or Bybit.Traditional investors, who are discovering crypto because it offers lower fees, real ownership, and fast settlement.
Demand is naturally highest for familiar names: Tesla, Nvidia, Circle, Coinbase, AMD, and emerging crypto-focused companies like BitMine and SharpLink.
How xStocks Are Used Outside Falcon
The xStocks ecosystem is already growing fast. According to Leo:
Users can borrow and lend on Kamino.They can provide liquidity on Raydium pools and earn fees.Market makers keep token prices aligned with real-world shares through RFQ systems that connect to Jupiter and Pyth Express Relay.
There is also work underway to create onchain ETF-style products using xStocks, bringing traditional market structures into the DeFi world.
Falcon’s Future Collateral Plans
Artem shared that Falcon’s roadmap is straightforward:

If an asset can safely live onchain and has good liquidity, Falcon aims to support it.
Today that includes tokenized treasuries, crypto assets, and tokenized stocks.

Next could be:
Short-duration notes,Tokenized credit products,CLO tranches,And non-USD treasury assets.
Each new asset type expands the strategies users can build using USDf as the core liquidity layer.
What 2030 Could Look Like
Looking ahead, Artem imagines a world where the boundary between traditional finance and onchain markets disappears. Users won’t ask “Is this onchain or offchain?”—they’ll simply choose the fastest, most transparent system.
Onchain rails are more efficient for liquid global assets like equities, bonds, gold, and major currencies. Falcon aims to make these assets usable in DeFi as collateral, liquidity, and programmable components.
Leo added that this shift is happening faster than many realize. xStocks were launched only a few months ago and have already crossed billions in trading volume. Interest is coming not only from crypto users but also from major traditional players like NASDAQ and BlackRock, who are exploring their own tokenized equity systems.
The big question now is whether the industry will keep these systems open and accessible, or restrict them behind gatekeeping.
Final Thoughts
By the end of the session, one message stood out clearly:
Falcon and Backed are not just putting equities onchain—they are transforming what those equities can do.
A Tesla share no longer needs to sit unused in a brokerage account.

In tokenized form, it can:
Stay fully collateralized and protected,Be used to mint liquidity,Earn yield through DeFi strategies,And still give the user full exposure to the underlying stock.
Capital that used to be “locked” can now move freely without losing investment exposure.

This is the true power of tokenized stocks on Falcon: real assets, real ownership, real liquidity—fully onchain.
#FalconFinance @Falcon Finance $FF

The global payments industry is growing fast. According to research, it may rise from $3.16 trillion in 2025 to more than $5.3 trillion by 2030. In traditional finance, people already have access to many types of ETFs that follow different trading styles—like covered calls, volatility strategies, and managed equity approaches. These products make it easy for anyone to earn structured yield without needing deep financial knowledge.
For example:
The JEPI ETF by JPMorgan has delivered around 8–9% yearly, using a covered call strategy.The SVOL ETF earns 10–12% annually by shorting volatility with hedging.The SYLD ETF has generated over 11% yearly by choosing companies that return capital to shareholders.
But in Web3, most of these structured yield strategies still do not exist in tokenized form. A few projects, like Ethena’s USDe, have shown that on-chain versions of complex financial strategies can work. Still, the wider market is missing tokenized versions of other advanced strategies such as risk-parity models, volatility plays, hedged equity products, or trend-following futures.
This gap opens a large opportunity:

bringing real, institutional-level trading strategies on-chain and making them accessible to everyone.
Introducing the Financial Abstraction Layer (FAL)

To solve this, Lorenzo created the Financial Abstraction Layer (FAL) — the backbone that turns complicated financial strategies into simple, usable, on-chain products.
FAL acts like a layer that:
Tokenizes strategiesExecutes tradesManages fund operationsDistributes yieldConnects DeFi and CeFi in a single seamless system
It is the engine behind On-chain Traded Funds (OTFs), handling everything from capital routing to NAV accounting to yield payouts.

How FAL Works — A Simple Three-Step Model

FAL is built around a clear and transparent cycle:
1. On-Chain Fundraising
Users deposit funds into smart contracts.In return, they receive tokenized shares that represent their portion of the fund.
2. Off-Chain Trading Execution
The collected capital is used to run real financial strategies, such as:
Arbitrage between exchangesDelta-neutral tradingVolatility harvestingOther active or automated methods
These strategies are managed by approved trading firms or automated systems with clear rules.
3. On-Chain Settlement & Distribution
After trades are completed:
Profits and losses are pushed back on-chainFAL updates performance, NAV, and payoutsUsers receive yield through tokens, claimable rewards, or fixed-maturity assets
This brings professional-grade financial operations into a fully transparent blockchain environment.
What Are On-Chain Traded Funds (OTFs)?
OTFs are blockchain-based versions of ETFs. They are fully tokenized funds that live on-chain and are managed by issuers who use Lorenzo’s infrastructure.
Each OTF represents exposure to a trading strategy or a combination of strategies.
OTFs are different from traditional ETFs because they:
Use smart contracts for real-time NAV updatesIssue and redeem tokens directly on-chainIntegrate instantly with wallets, dApps, and DeFi protocolsAllow both retail and institutional users to choose very specific strategy exposures

Strategies That OTFs Can Support

OTFs can represent a wide range of trading styles, such as:
Delta-neutral arbitrage across CEXs and DEXsCovered call income strategiesVolatility harvesting (short VIX, hedged positions)Risk-parity portfoliosTrend-following or managed futuresPerpetual funding rate optimizationTokenized CeFi lending or RWA-backed income
This brings traditional financial diversity into Web3 in a modular, customizable format.
FAL + OTF = Scalable, Modular, On-Chain Finance
Together, FAL and OTFs create a powerful new building block for the crypto space:
“Tokenized access to actively managed trading strategies, with modular yield mechanics, fully governed and settled on-chain.”
This combination unlocks:
Institutional-quality yield for everyday DeFi usersTransparent access to active trading strategiesCustom yield profiles based on personal risk preferencesA scalable framework that any protocol can build on
With $BANK and Lorenzo’s infrastructure, a new generation of on-chain investment products becomes possible — simple, secure, and accessible to all.
#LorenzoProtocol @Lorenzo Protocol $BANK

The BANK token is the main utility and governance token for the Lorenzo ecosystem. It is designed as a digital tool that helps the platform run smoothly, while giving active users a fair voice and rewards for their participation. You can think of BANK as a token that represents your contribution and level of involvement in Lorenzo — not ownership in a company.
What BANK Meant For
BANK a multi-purpose token used inside the Lorenzo platform. Its goal is to encourage people to take part in the ecosystem, support the protocol, and help it grow. Users who are active, contribute value, or complete transactions on Lorenzo can earn BANK rewards.

Those who don’t participate will not receive incentives — rewards are based only on real activity.
The token is essential for the platform’s health. Without $BANK , users would have no reason to spend time, effort, or resources helping the system function.
What BANK Token Meant
Holding BANK mean you own shares in Lorenzo or its partner companies.

It does not promise profits, dividends, or returns.

It is also not considered a security in major jurisdictions like Singapore or the British Virgin Islands.
BANK only to be used inside Lorenzo’s ecosystem — it has no external rights or guarantees. The price of BANK markets is not managed or controlled by the Lorenzo team.
Lorenzo’s Journey So Far
Lorenzo started as one of the earliest BTCFi staking platforms. Over time, it evolved into a full-scale asset management system built for institutional-level yield strategies.
Today, Lorenzo has:
Integrated with 20+ different blockchainsConnected to 30+ DeFi protocolsHelped manage over $600M in BTC through products like stBTC and enzoBTC
Now, BANK is the token powering governance, community incentives, and the long-term sustainability of this growing ecosystem.
Token Supply & Allocation
Total supply: 2,100,000,000 BANKInitial circulating supply: 20.25%
The entire supply will be divided across ecosystem needs such as governance, rewards, community initiatives, and long-term development.
Unlock Schedule
All BANK tokens will unlock gradually over 60 months.

To protect the community and ensure long-term alignment, no tokens for the team, early buyers, advisors, or the treasury will unlock during the first year.
What You Can Do With $BANK
BANK main functions inside Lorenzo:
1. Staking
Users can stake BANK for special benefits like:
Access to governanceAbility to voteExclusive featuresInfluence over incentive gauges and reward flows
2. Governance
BANK vote on key decisions, including:
Product and protocol changesFee adjustmentsAllocation of ecosystem growth fundsFuture emission updates
This makes governance community-driven and transparent.
3. Rewards for Active Users
A portion of the protocol’s revenue will be used to reward users who:
Use the platform regularlyVote on proposalsJoin campaignsTake part in community activities
Active participation = More BANK is Introducing veBANK (Vote-Escrowed BANK)
To activate all BANK tokens can lock their tokens to receive veBANK — a special non-transferable token used for governance.
How veBANK works:
The longer you lock $BANK , the more voting power you receive.veBANK holders can vote on incentive gauges.Long-term lockers earn boosted rewards and more influence in governance.
This ensures the protocol is guided by people truly committed to its future.
#LorenzoProtocol @Lorenzo Protocol $BANK

Kite AI has revealed its new Kite AI Ecosystem Map, showing how a wide network of leading tech and blockchain partners is helping autonomous AI agents work smoothly across both Web2 and Web3.

This ecosystem brings together cloud platforms, AI models, payment systems, and blockchain tools—creating a strong foundation for agents that can think, act, transact, and coordinate on their own.
Connecting Two Digital Worlds
AI agents today must move freely between Web2 and Web3. They need to store data, run models, process payments, and complete on-chain actions without waiting for human input.

Kite AI provides the trust, identity, and payment systems required to make this possible—acting as the “bridge” that connects traditional tech with decentralized infrastructure.
Web2 Partners: Scale, Intelligence, and Global Reach
Kite AI collaborates with major Web2 players that power the modern internet. Each partner gives agents new abilities:
Google – Secure cloud storage and fast data processing.

Amazon – Global e-commerce access, price comparisons, and automated order fulfillment.

Meta (Llama models) – Strong language understanding and generation.

DeepSeek – AI models focused on advanced, domain-specific reasoning.

Alibaba – Cross-border shopping and supply-chain visibility.

Shopify – Payments and commerce tools across millions of stores.

Claude (Anthropic) – High-level reasoning and problem-solving.
These integrations give AI agents the same tools humans use every day—only now, agents can use them automatically.
Web3 Partners: Trustless Logic and Programmable Money
On the blockchain side, Kite AI works with leading decentralized infrastructure projects:
Solayer + EigenLayer – Restaking systems on Solana and Ethereum to boost capital efficiency.

Huma Finance – Real-world asset credit access for autonomous agent operations.

LayerZero – Cross-chain messaging so agents can act across multiple chains.

Chainlink – Reliable oracles and real-world data feeds.

Privy – Embedded wallets and smooth user authentication.

Different Wallets – Secure transaction execution for agents.
Kite AI also works closely with next-generation AI and data teams such as 0G, Sentient, Virtuals, Codatta, Vana, Story, and Flock, strengthening the network’s intelligence layer.
A Complete Ecosystem for Agent Payments
Many platforms focus on either Web2 or Web3.

Kite AI is unique because it unites both—giving agents the scale of Web2 and the trustless programmability of Web3.
The ecosystem shows one clear fact:

Autonomous agents must operate everywhere—not in isolated systems.
What This Means for Builders
Developers building AI agents now have access to powerful new possibilities:
Store data on Google Cloud and settle payments on EthereumLet agents shop on Amazon with smart contract spending rulesUse Claude for reasoning while executing payments through Coinbase WalletCoordinate across chains using LayerZeroAccess onchain credit through HumaBoost capital efficiency using restaking
All of this is powered by Kite AI’s trust layer, Agent Passport identity system, governance tools, and payment rails—making agent operations safe, scalable, and transparent.
Join the Ecosystem
The companies in this map form the early foundation of the agentic internet.

But the network is growing fast. As more agents launch and new industries adopt automation, the Kite AI ecosystem will continue expanding.
If you’re building the future of autonomous AI, this is your chance to join a network designed for real, practical use cases across Web2 and Web3.
The agent-driven future isn’t a prediction—it’s happening right now, and Kite AI is building the core rails that make it possible.
About Kite AI
Kite AI is creating the base layer for the agentic internet—a decentralized system that lets autonomous agents authenticate, transact, and coordinate without middlemen.

Using an open network design and a protocol-to-protocol token model, Kite AI supports identity, payments, governance, and interoperability while enabling long-term sustainability for the entire ecosystem.
#KITE @KITE AI $KITE

Falcon Finance has taken another big step by adding Centrifuge’s real-world credit token JAAA as approved collateral to mint USDf. This is one of the rare examples where a diversified, AAA-rated corporate credit portfolio can be used directly inside DeFi. Falcon is also adding JTRSY, a short-term tokenized Treasury product, expanding its list of high-quality collateral options.

Together, these additions help Falcon move into its next chapter of real-world asset (RWA) integration.
Falcon Finance has always positioned itself as a universal platform for onchain collateral and delta-neutral yield. With JAAA joining the ecosystem, Falcon is bringing in a new category of structured, institution-grade credit — shifting RWAs from being “static assets” to becoming active tools for liquidity and yield generation onchain.
Bhaji Illuminati, CEO and Co-Founder of Centrifuge Labs, perfectly captured this shift, saying that tokenization is only the beginning. The real breakthrough happens when these assets can actually do work onchain — especially as collateral. Falcon’s support for JAAA and JTRSY is a big milestone in that direction.
What Makes JAAA Special?
JAAA, currently holding more than $1B in total value, is managed by Janus Henderson. It represents a carefully selected portfolio of short-duration, investment-grade corporate credit. Instead of being locked away in traditional markets, JAAA packages these structured credit products into a simple onchain format that DeFi users can actually interact with.
With Falcon accepting JAAA as collateral, holders no longer need to sell their high-quality credit positions to unlock liquidity. They can mint USDf against it and use that liquidity across different DeFi strategies — all while keeping exposure to the underlying RWA.
This transforms corporate credit from a passive holding into a productive onchain asset.
A New Phase for Falcon’s RWA Engine
Artem Tolkachev, Falcon’s Chief RWA Officer, explains that the RWA world is expanding rapidly — moving beyond tokenized Treasuries and into more complex, higher-yield credit products. Falcon’s goal is simple:

support any well-structured, compliant tokenized asset and make it usable as reliable collateral.
Importantly, Falcon does not rely on the yield from RWAs to power USDf. RWA tokens sit safely in segregated reserves, while USDf yield comes from Falcon’s own market-neutral strategy layer. This keeps risks separate and maintains consistent performance across all collateral types.
Building Toward a Cross-Asset Collateral Future
Adding JAAA aligns with Falcon’s broader vision:

a future where many types of real-world assets can be posted as onchain collateral.
Falcon already accepts:
Tokenized equitiesGoldU.S. TreasuriesOther institutional-grade instruments
Now, with structured corporate credit joining the mix, Falcon is paving the way for even more diverse and professionally built collateral portfolios. Over time, users will be able to unlock liquidity from a wide mix of RWAs without having to liquidate their positions.
As Artem shared, the industry is heading toward a world where nearly all major assets — credit, equities, commodities, treasuries — become programmable collateral by 2030. Falcon wants to make sure these assets can be put to work, not just represented onchain.
How Users Can Get Started
After completing Falcon’s simple KYC process, users can:
Deposit JAAA or JTRSY into Falcon as collateralMint USDf against their collateralKeep full exposure to the underlying RWADeploy USDf into staking, liquidity pools, restaking, or market-neutral strategies
This gives users the best of both worlds:

the stability of high-quality real-world credit and the flexibility of onchain liquidity and yield.
#FalconFinance @Falcon Finance $FF

Falcon Finance has introduced a new earning feature called Staking Vaults, giving users another way to grow their assets. Below is a simple guide on what these vaults are and how you can use the FF Vault to earn USDf.
What Are Staking Vaults in Falcon Finance?
Staking Vaults are a fresh addition to Falcon’s yield options. They work alongside the platform’s Classic and Boosted Yield models, which normally require staking Falcon’s synthetic dollar, USDf.
These vaults are built for users who prefer holding their tokens long term. Instead of selling your assets, you can stake them, stay fully exposed to their price growth, and still collect yields paid in USDf.
The first token supported at launch is FF, Falcon Finance’s governance and utility token. Inside the FF Vault, you deposit FF tokens and receive USDf rewards in return. To keep the system efficient, staking FF comes with:
a 180-day lockup perioda 3-day cooldown before withdrawal
This FF Vault is only the beginning—Falcon plans to introduce more vaults for additional tokens in the future.
Why Use Staking Vaults?
Staking Vaults give users a way to earn extra rewards without selling their altcoins.
Here’s what you get:
Weekly yield distributions during the 180-day lock-upRewards paid in USDf, a liquid synthetic dollarFull exposure to your original FF tokens—no selling requiredAn expected 12% APR, delivered daily in USDf
Once the lock-up period ends, you are free to unstake and withdraw your original FF.
How to Stake FF in the Vault
Falcon has made the staking process very user-friendly.
Connect a crypto wallet that holds your FF tokens.Go to the Staking Vaults section on the Falcon platform.Select the FF Vault.Enter the amount of FF you want to stake.Click “Stake” and confirm the transaction in your wallet.
You will need to pay a network fee. After the blockchain confirms your transaction, you will see updated information inside the vault interface, including:
Your current staked amountHow much USDf you’ve earned so farRemaining lock-up timeYour full vault transaction historyAny USDf rewards you can claim
Staking rewards are not automatically sent to your wallet—you must manually claim them from the vault page whenever you like.
Final Thoughts
The introduction of Staking Vaults opens up a powerful earning path for FF holders. It allows users to make their governance tokens more productive while still keeping ownership. With an attractive APR and rewards paid in USDf, these vaults strike a strong balance between long-term value and everyday utility.
And thanks to Falcon’s clean, simple interface, earning steady yields has never been easier or more accessible.
#FalconFinance @Falcon Finance $FF

In the past few years, one of the most interesting things happening in crypto is the rise of the Move programming language. It was created to fix some of the biggest security issues we’ve seen in older blockchain languages. Today, Move is powering new networks like Sui and Aptos, and now, it's slowly making its way toward both the Ethereum and Bitcoin ecosystems.
Ethereum has a long history of adopting new tech quickly. But this time, the Move ecosystem is growing at the same moment that new Bitcoin liquidity layers are emerging. Because of this, Bitcoin Finance (BTCFi) and Move may end up working together in ways that were impossible just a few years ago.
Let’s explore who’s building this ecosystem and how Bitcoin can fit into this new wave of DeFi.
What Exactly Is Move?
Move was first built by Meta for its discontinued Diem (Libra) project. The idea was simple:

create a programming language that treats digital assets as real, protected objects instead of just numbers in code.
Move was inspired by Rust and designed so that every asset has:
one ownerno duplicatesstrong protection from common hacks
This solves many problems found in Solidity—the language behind Ethereum—which has suffered from well-known issues like reentrancy attacks.
Even though Diem never launched, Move lived on. Today it powers networks like Sui and Aptos. Move also comes with its own fast virtual machine (MoveVM), which is built for:
parallel transaction processingstrong security verificationefficient memory handlingdeveloper-friendly modular design
Another huge advantage is that Move can work side-by-side with Solidity. Developers can deploy both contract types without needing to rewrite their apps.
Important Projects in the Move Ecosystem
The Move ecosystem is still young, but several major projects are already active:
1. Sui
Sui is a fast Layer 1 blockchain built for high-speed transactions. Many of its creators worked on Meta’s original Diem team.
Highlights:
Sub-second finalityVery low feesParallel transaction processingStrong focus on digital assets and gamingExtended version of Move called Sui Move
Its complex consensus design uses DPoS, DAG, and a BFT model called Mysticeti. Together, they help the network achieve high throughput without sacrificing reliability.
2. NAVI
NAVI is Sui’s main liquidity protocol, similar to Aave. Users can lend, borrow, or earn yield—but NAVI adds several advanced features:
automated leverage strategiessafer "Isolated Markets"flexible collateral rules
It is one of the core DeFi pillars on Sui.
3. Aptos
Aptos is another Layer 1 network built around Move. It launched in 2022 and quickly became known for its speed—up to 160,000 TPS with less than one-second finality.
Key traits:
Parallel execution engine (Block-STM)BFT + PoS consensusHighly scalable designBacked by major investors like a16z and Coinbase Ventures
Aptos has one of the strongest communities among Move-based chains.
4. Cetus
Cetus is the leading DEX in the Move ecosystem. It offers:
deep liquidityconcentrated liquidity poolspermissionless access for developerssmooth trading and low-slippage swaps
Cetus is often compared to Uniswap v3 but optimized for Move.
5. Movement Labs
Movement Labs is bringing Move to Ethereum through an L2 called Movement.
Its MEVM engine allows:
Move-based and EVM-based apps to run togetherhigher security for smart contractsfaster transactionscustom rollups for developers
The goal is to merge the large Ethereum user base with Move’s stronger security model.
Where Does Bitcoin Fit In? — Enter Lorenzo
This is where things get exciting.
Lorenzo Protocol is working to connect Bitcoin directly into the Move ecosystem. It is the first omnichain Bitcoin liquidity layer built around MoveVM.
What Lorenzo enables:
Bitcoin liquidity can move through Move-based ecosystemsBitcoin holders can earn more through liquid stakingBTC can finally be used inside high-performance DeFi apps
This solves one of Bitcoin’s biggest limitations:

its base layer is extremely secure but not flexible.
Lorenzo adds the flexibility without touching Bitcoin’s core design.
With Lorenzo, Bitcoin can be:
stakedused in lending protocolsadded to liquidity poolsintegrated into Move-based chains like Sui, Aptos, and Movement
This opens the door for a true BTC-powered DeFi economy.
Why This Integration Matters
Move brings stronger contract safety and faster execution.

Bitcoin brings unmatched trust, liquidity, and decentralization.

Lorenzo acts as the bridge between them.
Together, they create:
a more secure DeFi environmentdeeper liquidity across ecosystemseasier access to BTC inside modern applicationsa new financial landscape where different chains work together instead of competing
This growing connection between Bitcoin and Move-powered platforms might be the beginning of a more open, more interoperable DeFi future.
And as Bitcoin’s liquidity enters these high-performance networks, it could also accelerate mainstream adoption across the entire blockchain ecosystem.
#LorenzoProtocol @Lorenzo Protocol $BANK

Introduction
AI models and multi-agent systems are getting better at an unbelievable speed. They’re learning to handle tougher and more complicated tasks — things like cleaning up an entire codebase, booking flights, or arranging an appointment. But as we give them more responsibility, the chances of making mistakes also grow. And the bigger the task, the bigger the damage those mistakes can cause.
This creates a major problem: trust.
If people still feel responsible for every action taken by an AI system, they simply won’t use it unless they trust it completely. And trust is not something that grows from one number or one performance score — everyone has their own threshold.
That’s why understanding the “trust gap” is important. To build systems people rely on, we must map out where trust breaks, what causes it, and how to strengthen it. Only then can AI agents slowly move toward the level where people feel confident enough to adopt them.
A recent research study, “A Survey on Trustworthy LLM Agents: Threats and Countermeasures,” introduces a framework called TrustAgent. It helps explain what makes an agent truly trustworthy and what pieces need to be improved. Below is a simplified walkthrough of this framework.
The Trust Agent Framework
The Trust Agent framework divides trust into two groups:
Intrinsic Trustworthiness — everything happening inside the agentExtrinsic Trustworthiness — everything happening outside the agent
Let’s explore both.
Intrinsic Trustworthiness (Inside the Agent)
1. Reasoning and Decision-Making
This is the “mind” of the agent — where decisions, planning, and task execution happen.
If something goes wrong here, the whole system becomes unreliable.
Trust usually breaks due to:
Jailbreaking attemptsPrompt injections that override instructionsBackdoor attacks through bad training data
Stopping these issues isn’t easy. Tools like alignment training, screening prompts, and using multi-agent “shields” can help, but the risk never disappears entirely.
2. Memory and Retrieval Systems
Agents often store extra information outside their main context window through retrieval systems (like RAG). This helps personalize tasks or provide missing context.
But it also opens several risks:
Fake or harmful data can be addedMemory can be leakedStored data can be used to bypass safety rules
Many of these issues are hard to detect from inside the model. So the best defense involves monitoring changes, limiting direct access to memory, and putting guardrails on outputs.
3. Tools and External Actions
Tools extend what an agent can do — like querying databases or writing code.
But tools are powerful, and misuse can be dangerous. Problems include:
Manipulating how tools are usedMisusing tools to affect external systems
Because tool-calling is a fairly new capability, defensive research is still limited.
Extrinsic Trustworthiness (Outside the Agent)
1. Agent-to-Agent Interaction
In multi-agent setups, several agents work together. This adds complexity — and more room for things to go wrong.
Agents can:
Spread “infections” through promptsWork together maliciouslyPass harmful information around
Defensive strategies include collective monitoring and analyzing network behavior to stop harmful patterns before they spread.
2. Agent-to-Environment Interaction
Agents that operate in real-world environments — robotics, industry, healthcare, finance — need extremely high trust. A small mistake can become a major disaster.
Even digital environments, like financial systems, require careful security and reliability checks.
But research often overlooks this part, focusing more on performance than real-world grounding.
3. Agent-to-User Interaction
This is where human expectations matter most.
Users adjust their trust based on how an agent behaves — and this dynamic is not well studied yet. Understanding how trust grows or breaks from the user’s perspective is a huge research gap.
Where Does This Leave Us?
We see trust failures everywhere — from AI generating fake legal citations to models executing unintended tool actions. The truth is:
There’s no universal rule for when an agent becomes trustworthy.Every industry has different expectations.Trust changes over time as people interact with the system.
At the heart of the problem is responsibility. People trust other humans because someone is accountable when things go wrong. But with AI systems, this accountability is unclear.
Some believe better alignment training is the solution. Others think large language models will never reach true trustworthiness. Another approach is to build infrastructure that gives AI agents identity, verifiable actions, and transparent governance.
This is exactly what Kite AI aims to build — a foundation where autonomous agents can operate responsibly with clear rules, roles, and verification.
If you want to explore this new world of agentic AI, visit Kite AI’s platform and see how you can start building.
#KITE @KITE AI $KITE

UnifAI is stepping into the Kite ecosystem as the first module built for “Agentic Finance,” often called AgentFi. Together, UnifAI and Kite AI are creating the base layer for a future where AI agents can manage money, make decisions, and run financial activities on-chain with full transparency.
Born from Kite AI’s vision of an “Agentic Internet,” UnifAI gives users a smarter way to handle their crypto finances. It helps people run strategies, manage funds, and act inside decentralized systems in real time. The goal is simple: make AI your personal guide — an advisor that can trade, analyze, and detect new opportunities across DeFi with safety and intelligence. At the same time, UnifAI gives developers the tools to build and scale advanced, AI-driven financial applications.
When UnifAI’s intelligence framework connects with Kite AI’s infrastructure, builders can create agents that discover profitable opportunities, perform secure transactions, and participate as independent players in the growing AI economy.
Chi Zhang, Co-founder and CEO of Kite AI, explained it clearly:

“UnifAI marks the next phase of agentic finance. AI agents won’t just serve as tools — they will become real participants in financial markets.”

The partnership between Kite AI and UnifAI reinforces their mission to build an open and transparent AgentFi ecosystem.
A Shared Vision
Solving the Core Problems of AgentFi
UnifAI is designed to tackle two major challenges:

autonomous coordination and capital efficiency.
Its architecture allows developers to build agents that can launch strategies, earn yield, and manage their own funds — all recorded directly on-chain.
Kite AI supports this journey by offering early access to its infrastructure, partner ecosystem, and protocol integrations. This positions UnifAI to fast-track its goal of creating AI agents that are productive, independent, and financially sovereign.
Why UnifAI Is Built Inside Kite AI
UnifAI sits naturally inside Kite AI because both follow an agent-native design.

Here’s how they complement each other:
Agent-Native Infrastructure: Kite AI provides identity, payments, governance, and verification layers that help agents become independent actors.Programmable Finance: Agents execute secure, verified actions through Kite AI’s smart contract layer.Agent Passport: Each agent gains a verifiable identity that works across the ecosystem.Cross-Ecosystem Support: UnifAI connects with data, compute, and validation layers within the Kite network.Autonomous Capital Use: Agents can find, judge, and execute financial opportunities without human input.Transparent Attribution: Every action, return, and decision is recorded on-chain, building trust and accountability.
What Makes UnifAI Technically Strong
UnifAI delivers an advanced on-chain intelligence framework that includes:
Dynamic tool discoveryAutomated strategy execution (trading, lending, LP, etc.)A single API that unlocks the entire DeFi landscapeAutonomous task executionStrong security where sensitive data stays client-side
How It Works — A Simple View
UnifAI acts as the execution backbone of the agentic finance world. It connects AI agents to DeFi protocols and liquidity layers, enabling them to work independently.
These agents monitor market conditions, scan for opportunities, and execute strategies in areas like trading, lending, and liquidity management.

For users, this means automated yield optimization and smarter portfolio oversight.

For developers, it means they can launch and scale intelligent agents easily.
Every action that UnifAI-powered agents take is recorded on-chain, creating a secure and transparent system for the next wave of autonomous finance.
Why This Matters
The UnifAI + Kite integration unlocks:
Simpler Decisions: Agents quickly spot the best opportunities.Automated Strategies: Complex DeFi moves happen on their own.Real-Time Advantage: Agents adapt to the market instantly.Developer Freedom: Build custom financial agents with modular tools.
Sunny, Co-founder of UnifAI, shared the vision:

“We want intelligent agents to act with clarity, responsibility, and real economic impact. This partnership with Kite AI helps us build that open and aligned ecosystem.”
What Comes Next
This is only the beginning.

As UnifAI expands alongside Kite AI, the ecosystem moves beyond basic infrastructure and enters a new era of agent-native finance. Both teams are laying the groundwork for a future economy driven by autonomous intelligence.
About Kite AI
Kite AI is building the first blockchain made for agentic payments. It gives AI agents verifiable identities, programmable governance, and native access to stablecoin payments. The company is founded by seasoned builders from Databricks, Uber, and UC Berkeley, and has raised $33M from major investors like PayPal and General Catalyst.
About UnifAI
UnifAI builds infrastructure for autonomous AI agents that simplify DeFi for everyone.

For users, it executes strategies automatically so they can benefit from DeFi without being online all the time.

For developers and projects, it offers a modular, secure, and interoperable system to build and scale AI-powered agents across the entire DeFi ecosystem.
#KITE @KITE AI $KITE

Falcon Finance has taken another major step forward by adding Centrifuge’s real-world credit token JAAA as eligible collateral to mint USDf. This marks one of the rare moments in DeFi where a diversified, AAA-rated credit portfolio can be used directly as onchain collateral. Falcon is also listing JTRSY, a short-duration tokenized Treasury product, further strengthening its collection of high-quality assets. With these additions, Falcon is preparing for its next stage of RWA integration, now including investment-grade corporate credit.
As a universal collateral platform for onchain liquidity and delta-neutral yield, Falcon continues to move beyond crypto, Treasuries, and tokenized equities. Bringing JAAA into the ecosystem unlocks a new class of structured, real-world credit that can now be used inside DeFi instead of sitting idle. This turns RWAs from passive holdings into active assets that generate liquidity and utility across decentralized markets.
Bhaji Illuminati, CEO & Co-Founder of Centrifuge Labs, shared an important perspective:

“Tokenizing real assets is only the beginning. The real impact comes when those assets can be used as collateral directly onchain. By bringing JAAA and JTRSY into Falcon, we’re giving holders more utility and moving closer to a fully connected onchain financial system.”
JAAA, which now holds over $1B in TVL, represents a carefully selected portfolio of short-duration, investment-grade corporate credit managed by Janus Henderson. Wrapped into an onchain format, it provides access to structured credit yields while still being suitable as collateral inside DeFi.
With JAAA accepted as collateral, Falcon users can keep exposure to high-quality real-world credit while minting USDf against it. This lets them stay invested in top-tier credit products while also unlocking onchain liquidity for staking, liquidity pools, and other strategies. Instead of being a static asset, structured credit becomes active collateral powering DeFi activity.
Artem Tolkachev, Falcon Finance’s Chief RWA Officer, explained the shift happening across the market:

“We’re expanding Falcon’s RWA engine and working with leaders in the space, and Centrifuge is clearly one of them. The market is moving beyond tokenized Treasuries toward higher-yield, more complex credit. Our mission is to support this evolution by enabling liquidity for any well-designed tokenized asset. JAAA is a perfect example of what’s possible when real-world credit becomes usable collateral.”
Falcon keeps all RWA tokens strictly as collateral, stored in separate reserve accounts. The yield from USDf does not depend on the underlying asset; returns continue to come from Falcon’s market-neutral strategy stack. Because of this structure, collateral risk, strategy performance, and user returns stay clearly separated, keeping USDf predictable across all supported collateral types.
A Step Toward Falcon’s Cross-Asset Collateral Vision
Adding JAAA aligns with Falcon’s long-term plan to support a wide mix of tokenized real-world assets. The platform already accepts tokenized stocks, gold, U.S. Treasuries, and other high-quality assets. Now, with structured corporate credit included, Falcon is building toward a future where users can deposit diversified RWA portfolios and unlock liquidity from many different onchain asset classes.
Artem summed up this future direction:

“By 2030, most major liquid assets will exist as programmable collateral. Whether it’s credit, equities, commodities, or Treasuries, Falcon’s role is to make sure tokenized assets can actually be used—not just displayed.”
Once users complete Falcon’s KYC process, they can deposit JAAA and JTRSY as collateral, mint USDf, and continue benefiting from their underlying RWA exposure. That USDf can then be used across Falcon’s ecosystem — staking, liquidity pools, restaking, or the platform’s delta-neutral yield strategies. In short, users can stay fully invested in real-world credit while gaining onchain flexibility and liquidity.
#FalconFinance @Falcon Finance $FF

Bitcoin has been the number one cryptocurrency ever since it launched in 2009. But even though it’s the oldest and most trusted coin, it doesn’t control the entire market anymore. Today, Bitcoin holds a little under 55% of the total crypto market share — far from the 95% it owned back in 2013.
Over the years, many new blockchains have grown because they offer things Bitcoin didn’t originally provide, like advanced smart contracts, cheaper transactions, or better privacy. But a new wave of Bitcoin Layer-2 networks is starting to change this balance again. These L2s are bringing all the innovation that happened outside Bitcoin back home to the Bitcoin ecosystem.
For almost a decade, people discussed the idea of bringing DeFi, smart contracts, NFTs, and more to Bitcoin. Now, thanks to new Bitcoin Layer-2 technologies, it’s finally becoming real. And this shift could unlock a massive value opportunity worth $600 billion or more.
Let’s break down how this return to Bitcoin could happen — and what it means for the future.
Bringing DeFi Back to Bitcoin
Most alternative cryptocurrencies exist because they offer smart contracts. Ethereum is the biggest example, with a market cap of around $408B. Other blockchains like BNB Chain, Avalanche, Tron, and Solana borrowed from Ethereum and built their own ecosystems.
These networks power:
Decentralized exchangesLending platformsLiquid stakingNFTs and meme coinsAnd billions of dollars in stable coin activity
Because these apps require flexible smart contracts, users moved to these chains. That’s why Ethereum and other L1s captured almost all DeFi activity.
DeFiLlama reports nearly $100B locked in DeFi, but less than 2% of it exists on Bitcoin. Even when people use BTC in DeFi, it’s often through wrapped bitcoin (WBTC) on Ethereum — which depends on centralized custodians.
Bitcoin Layer-2s are changing that.
Bitcoin L2s: Unlocking Real DeFi for BTC
Since Bitcoin’s base layer doesn’t support advanced smart contracts, most DeFi use cases need to live on L2s. Many new L2s — including the Lorenzo App Chain — bring full EVM compatibility, meaning developers can deploy Ethereum-style apps directly on Bitcoin’s L2 ecosystem.
This means:
LendingDEXsLiquid stakingNFT marketplacesComplex smart contracts
can all run on Bitcoin-backed Layer-2 networks.
A major breakthrough came from new two-way peg designs powered by BitVM, which improved how funds move between Bitcoin and L2 environments. This makes Bitcoin L2s safer and more reliable than past attempts.
This shift also challenges the need for other base-layer blockchains. If Bitcoin can stay decentralized and secure while L2s handle the innovation, then many alternative chains may lose their purpose.
Why Other Cryptocurrencies Exist — and Why Bitcoin Can Replace Them
Most altcoins fall into two categories:
1. Coins for Privacy
Examples:
MoneroZcash
Their technologies were actually proposed for Bitcoin years ago, but weren’t added due to trade-offs. Now, Bitcoin L2s are beginning to bring strong privacy options through systems like:
Ark (private off-chain payments)Mercury Layer (private swaps)Fedimint (anonymous digital cash)
In the future, even Tornado-Cash-style systems or Zcash-like L2s could launch on top of Bitcoin.
2. Coins for Cheaper Payments
Examples:
XRPDogecoinBitcoin CashLitecoin
Bitcoin’s answer is the Lightning Network — but Lightning can be difficult for new users. Privacy-first L2s like Ark may become easier alternatives that remove the need for opening channels on-chain.
In short: every major altcoin use case is slowly being absorbed back into the Bitcoin ecosystem.
The $600B Opportunity
The global crypto market is valued at around $2.4 trillion. But much of that value comes from:
EthereumSolanaAlt L1sDeFi tokensPayment coinsWrapped assetsChains that Bitcoin L2s can replace
Based on current valuations, Bitcoin has at least a $600 billion+ opening to absorb:
Ethereum-like DeFiHigh-speed payment chainsWrapped BTC usagePrivacy networksStablecoin ecosystems (eventually via L2 minting)
Even if Bitcoin doesn’t replace meme coins or NFTs completely, it can dominate almost every other major use case.
This also helps DeFi itself — because using Bitcoin as the base asset makes the entire ecosystem:
More liquidMore secureMore predictableEasier for everyday users
Instead of juggling multiple tokens, users can simply rely on one strong digital money: BTC.
Bitcoin’s Path to Becoming the Internet’s Native Currency
Jack Dorsey famously said that Bitcoin is the “native currency of the internet.” But before Bitcoin can rival gold, the dollar, or global financial systems, it first needs to beat:
ETHXRPDOGESOLAnd other major altcoins
By shifting all major crypto functionality to Bitcoin L2s — such as the Lightning Network for payments and Lorenzo App Chain for DeFi — Bitcoin gains both the liquidity and the network effects needed to dominate the entire crypto economy.
Bit by bit, liquidity from other chains will migrate back to Bitcoin, strengthening its position as:
✔ A store of value

✔ A medium of exchange

✔ A settlement layer

✔ A DeFi and smart contract foundation via L2s
In short:

Bitcoin can absorb most of the crypto market — and L2s like Lorenzo may be the bridge that makes it possible.
#LorenzoProtocol @Lorenzo Protocol $BANK

Bitcoin is the biggest player in the crypto world. It holds more than $1.3 trillion in market value and represents almost half of the entire crypto market. Yet, even with all this strength, Bitcoin is barely connected to DeFi. Today, only a tiny amount—less than 0.3% of all BTC—is actually being used inside DeFi protocols.
Even Wrapped Bitcoin (wBTC), the most common version used in DeFi, holds under 160,000 BTC. That’s less than 0.8% of Bitcoin’s circulating supply. This huge gap shows a clear problem: most BTC is simply sitting idle, missing the chance to earn yield or interact with the fast-growing DeFi ecosystem.
Lorenzo Protocol is designed to change this story. Its Bitcoin Liquidity Layer creates a smooth bridge between BTC and DeFi. By offering the tools to mint several types of BTC-based derivative tokens—such as wrapped BTC, staked BTC, and yield-generating structured BTC—Lorenzo makes it easy for Bitcoin to become active inside decentralized markets.
With Lorenzo, passive BTC transforms into working capital. It can power lending markets, yield strategies, structured products, and a wide range of financial utilities across DeFi.
In simple terms, Lorenzo’s mission is clear:

Turn Bitcoin from a silent store of value into a powerful, productive asset for the decentralized economy.
#LorenzoProtocol @Lorenzo Protocol $BANK

Introduction
In earlier parts of the Kite AInsights series, we explored the problems that appear in multi-agent systems, how different agents can talk to each other, and how we can trust their actions.

Today, we look at something even more fundamental: planning.
There’s a famous saying, “When you don’t plan, you’re planning to fail.”

This fits perfectly with multi-agent systems, where several agents work together on one goal.

The article we’re summarizing discusses a paper called Agent-Oriented Planning in Multi-Agent Systems, which introduces a new way of helping agents break down and complete tasks more effectively.
We’ll keep this explanation high-level and easy to understand, but you can always check the original paper for deeper technical details.
Why Planning Matters in Multi-Agent Systems
Planning helps us handle complicated goals by dividing them into smaller steps.

For example, finishing university requires completing courses… which requires passing tests… which requires studying for specific chapters. Each level breaks down into smaller and clearer actions
For multi-agent systems, a similar process happens. Usually, there is a meta-agent—a kind of “team leader”—that:
Understands the user’s requestBreaks the request into smaller tasksAssigns each small task to the right agent
But the authors noticed a problem:

Meta-agents often struggle to assign tasks correctly just by reading the agent descriptions.
To fix this, they created three principles for better task planning:
1. Solvability
Each task should be something one agent can handle by itself.
2. Completeness
All parts of the user’s request must be covered.
3. Non-Redundancy
No extra or repeated tasks should be created.
Using these principles, they built the Agent-Oriented Planning (AOP) framework.
The Agent-Oriented Planning (AOP) Framework

At the start, AOP includes three main elements:
The Query: What the user wantsThe Meta-agent: The plannerThe Plan: A first draft of the broken-down tasks
After the first plan is created, AOP uses two components to refine it:
1. The Detector
It checks the plan for:
Duplicate tasksMissing stepsWrong dependencies
Then it suggests ways to improve the plan.
2. The Reward Model
It predicts whether each sub-task can be solved, without actually having to call the real agent yet.

It also helps detect when a sub-task is already completed.
Once refined, the plan is sent to the agents.

Their responses act as feedback, helping the system repeat the cycle until the main user request is fully resolved.
How Well Does AOP Work?
In the paper’s tests, the authors used GPT-4o as both the meta-agent and the specialized task agents (math, search, code, reasoning).

The Reward Model was built on all-MiniLM-L6-v2.
The results showed:
Around 10% better accuracy than single-agent setupsAround 4% better accuracy than traditional multi-agent setups
The main downside?

AOP consumes more time and token usage, which increases the overall cost.
What Challenges Still Remain?
Even with solid results, the AOP framework still has room for improvement. Here are three interesting paths the authors suggest:
1. Better Agent Collaboration
Agents shouldn’t work alone.

They could call each other—just like coworkers.

For example, a search agent may ask a code agent to help structure the data.
2. Human-in-the-Loop
In some cases, the system might assign a sub-task to a real human when automation isn’t enough.

This would mix human abilities with agent skills.
3. Stronger Trust and Verification
For real-world actions—like booking flights or paying for something—users must trust the system.

They need clear proof that tasks are done correctly, safely, and transparently.
This last point is directly linked to Kite AI’s vision:

building secure, verifiable infrastructure that lets agents operate with identity, payments, governance, and trust built into the system.
#KITE @KITE AI $KITE

Falcon Finance — the platform building a universal collateral layer for onchain liquidity and yield — has now added Tether Gold (XAUt) as a new collateral option for minting USDf. XAUt is the world’s largest and most trusted tokenized gold asset, and its integration opens the door for users to hold digital gold while earning stable, DeFi-native yield at the same time.
This update marks an important step for real-world assets (RWAs). By bringing tokenized gold into Falcon’s collateral system, the project continues to bridge traditional finance with onchain innovation. Users get access to liquidity backed by one of the most reliable stores of value ever known — physical gold.
Gold has an estimated global market cap of nearly $27 trillion, and more than $3 billion worth of gold is already tokenized on blockchain networks. Falcon’s move expands how tokenized gold can be used in DeFi: improving portfolio diversity, strengthening collateral quality, and allowing gold to serve as backing for USDf.
Andrei Grachev, Founding Partner at Falcon Finance, shared:

“Adding Tether Gold as collateral is a major step toward growing USDf adoption and connecting traditional value with onchain liquidity. Gold has always played a central role in global finance, and bringing it onchain through XAUt supports our mission to build a universal, yield-focused infrastructure.”
Tether Gold turns physical gold into a digital asset that offers 24/7 access, deep liquidity, fractional ownership, and secure custody. This gives users the ability to hold verified gold in token form — and now, through Falcon, use it as a yield-bearing asset.
Falcon’s synthetic dollar, USDf, has already grown to more than $2.1B in supply, backed by over $2.3B in reserves according to the latest attestation. Integrating XAUt will strengthen total value locked (TVL) and allow more users to earn stable, sustainable returns through sUSDf.
By making gold productive as onchain collateral, Falcon Finance continues its mission to build the foundation for universal asset collateralization in DeFi.
#FalconFinance @Falcon Finance $FF