Starknet q2 cAP
Introduction
Scaling never sleeps.
Between performance upgrades, STRK staking, new DeFi projects, and massive user onboarding, Starknet had a busy spring.
This Q2 recap walks you through what shipped, what’s coming next, and why Starknet is still the most ambitious L2 out there.
 |
 |
I. Bitcoin: real use cases, real integrations, real momentum
While most L2s flirt with Bitcoin narratives, Starknet is quietly building the stack to become Bitcoin’s execution layer, alongside Ethereum.
In Q2, this vision shifted from thesis to reality. Bridges are live, BTC is productive, Runes are usable, staking is coming, and both StarkWare & the Starknet Foundation are shipping on all fronts.
Here’s what happened.
a. The rails are live
The first layer of Bitcoin integration is simple: moving BTC into Starknet. And this quarter, the ecosystem delivered.
Garden Finance, an intent-based bridge, now supports BTC transfers to Starknet from Bitcoin, Ethereum, Arbitrum, Base, Berachain, and Hyperliquid.
Atomiq Labs enables direct BTC swaps between Bitcoin and Starknet.
Layerswap now supports BTC <> Starknet bridging, with Starknet being one of only two chains supported (alongside Ethereum).
Runes Bridge also went live, enabling users to bridge Runes from Bitcoin to Starknet. Currently, only the five largest Runes are supported, but more can be added based on community demand.
These rails are removing major friction for onboarding Bitcoiners to Starknet, and there's much more coming.
b. From static to productive: Bitcoiners in DeFi
While the BTCfi campaign hasn't launched yet, BTC on Starknet can already be put to work:
wBTC is now incentivized through DeFi Spring, with ~4% APR on lending markets like Vesu and Nostra.
Braavos launched a One-Click Earn flow: bridge, wrap, lend, done, all in one-click.
DOG, the top Rune on Bitcoin, is now fully integrated into Starknet DeFi:
Tradable on AVNU, Remus, Layer Akira and Fibrous
You can even use DOG to pay for gas via AVNU’s Paymaster.
Oh, and tBTC is now mintable directly on Starknet, with full integration into Starknet’s DeFi ecosystem currently underway.
To push things even further, long-time Bitcoiner and degen Jack Liu has joined the mission. He now leads a $2M fund focused on building consumer-facing Bitcoin apps on Starknet, accelerating the BTCfi movement from the front lines.
 |
c. Asset Runes: new primitive, new possibilities
Q2 also marked the launch of Asset Runes, Bitcoin-native tokens that offer 1:1 exposure to Ethereum assets while staying minted on Bitcoin.
The first one is USDC: minted on Bitcoin, redeemable anytime on Starknet.
On Bitcoin, you can trade or provide liquidity on DotSwap, or trade them on Magic Eden.
USDC Runes can be bridged back to Starknet via StarkGate at anytime.
More Asset Runes are coming: ETH, STRK, RWAs, NFTs…
d. Research & infra: ColliderVM, Broly, and staking
On the research side, StarkWare introduced ColliderVM, a major improvement over BitVM:
No 7-day fraud window
No fully collateralized withdrawals
Faster, more efficient, trust-minimized computation on Bitcoin
This is still in the research phase and far from production-ready. But it reflects StarkWare’s long-term commitment to scaling Bitcoin and turning Starknet into a Bitcoin execution layer, following the most trust-minimized path possible.
In addition to ColliderVM, StarkWare also released Broly, a proof of concept that lets users inscribe Bitcoin directly from a Starknet wallet, using a mechanism similar to storage proofs. It’s open-source and ready for developers to build on.
Finally, Bitcoin staking is expected to arrive on Starknet in Q3 2025. It will unlock three key benefits:
Sustainable rewards for BTC holders
Increased security for the Starknet network
A powerful new primitive for developers to build on
You can learn more in the official SNIP. Note that while BTC will become a stakable asset on Starknet, at least 75% of the staking power will still be driven by the STRK token, and a BTC/STRK alignment mechanism will balance incentives.
All of these Bitcoin integrations are laying the groundwork for the upcoming BTCfi campaign, which is set to launch in the coming months. Bitcoin on Starknet is still early, but it’s definitely real.
|
II. Network & Infra: faster UX, cheaper proofs, and a path to Stage 2
The infra vision behind Starknet has always been ambitious: lower fees, faster UX, and full decentralization, all without shortcuts or compromise.
In Q2, the ecosystem kept executing toward that vision, one upgrade at a time.
a. The engine under the hood: performance & network upgrades
This quarter marked a turning point for Starknet, with protocol upgrades that reinforce the foundation for everything that’s coming next.
v0.13.6 went live on Mainnet in early July. On paper, it’s a minor release. In reality, it lays the groundwork for S‑two integration, StarkWare’s new prover that will soon supercharge proving performance across the entire network.
v0.14.0, already live on testnet and scheduled for Mainnet in August, is the real milestone. It brings:
- A new sequencer architecture: 3 independent sequencers using Tendermint-style consensus, a major step toward decentralized sequencing.
- Pre-confirmations: UX finality drops from ~2s to just 500ms for most actions (swaps, LPs, lending, etc.)
- Block time: reduced from 30s to just 4–6s
- Fee market + mempool: transactions are now prioritized by tip (like Ethereum), providing a more predictable UX and preparing the ground for the upcoming STRK fee burn mechanism
- STRK becomes the only native gas token, but Paymasters like AVNU and Cartridge let users pay gas in other tokens
- For devs, it introduces some breaking changes, make sure to review the details.
 |
b. S‑two: the fastest prover in the world
StarkWare’s new prover, S‑Two, is now available to everyone, and it’s a monster. Across all zk benchmarks, it outperforms the competition.
 |
In fact, S‑Two is the first prover to make client-side proving viable in real-world apps. And some projects are already putting it to work:
Paradex is integrating it to enable privacy-preserving trading
Nexus is leveraging it for provable compute in Rust, C, and RISC‑V
Kakarot aims to bring real-time proving to everyday devices, while adding ZK diversity to Ethereum L1
And Starknet itself will soon adopt it. The S‑Two MVP is expected to reduce proving time from 24 minutes to under 3, while cutting costs by 2x.
This marks a major leap forward for Starknet performance. And it’s only the MVP.
c. Starknet is now a Stage 1 ZK Rollup
In May, Starknet officially reached Stage 1 on L2Beat, becoming the first pure ZK Rollup to get there.
What does that mean?
Stage 0: centralized infra, data posted on L1
Stage 1: operators replaced by contracts, permissionless exits, multi-party verification, constrained security council
Stage 2 (coming): full trustlessness, Ethereum-grade guarantees
You can dive deeper into the Stage 1 criteria here.
Now that the training wheels are off, the Starknet ecosystem is doubling down on becoming a Stage 2 rollup as soon as possible.
 |
Starknet will soon run faster, feel faster and prove faster on Mainnet. And it’s doing it all while pushing decentralization forward.
|
IV. Staking v2: toward economic decentralization
Speaking of decentralization, staking is the cornerstone of Starknet’s transition toward a fully decentralized network. And in Q2, real progress was made with the rollout of Staking v2, introducing a new layer of validator responsibilities and incentives.
Indeed, staking v2 went live on Mainnet in June, bringing two major improvements over Phase 1:
Validator Block Attestation: Starknet now operates in epochs. During each epoch, validators must attest to randomly selected blocks. If they don’t, they lose their rewards, and so do their delegators. This improves transparency and ensures validators are active before full consensus duties arrive.
Commission Changes: Validators can now increase their commission, but only under strict conditions. This adds flexibility without sacrificing fairness.
These upgrades make validator behavior more accountable and visible, both for the protocol and for delegators.
 |
As of current metrics (as of July 22):
430M STRK staked on Starknet
~12% of the circulating supply
119/151 actively attesting to blocks
 |
You can track all these metrics in real time here.
Note: both StarkWare and the Starknet Foundation have launched their own delegation programs, meaning they are delegating their STRK to external validators.
The good news? You can apply to both programs at the same time: StarkWare Delegation Program (round 2) and Starknet Foundation Delegation Program.
|
V. Builders, L3s, and full-stack creativity
Starknet isn’t just scaling throughput, it’s also scaling what’s possible for developers.
In Q2, the ecosystem shipped new primitives, tools, and grant programs that open up fresh design space across L2, appchains, and L3s.
a. The L3 building block is now available on Starknet
In April, Karnot launched a 14-hour ephemeral L3 called Gridy, built on top of Starknet. This short-lived “gamechain” proved that L3s on Starknet are now fully viable, and extremely efficient.
Sustained 220 UOPS, peaking at over 400 TPS
But the real unlock was the full abstraction of the L1 → L2 → L3 flow that Karnot delivered. No matter where you were (Ethereum L1, Starknet L2, or the Gridy L3) you didn’t need to download a new wallet, bridge funds, or switch networks. You just… played. Directly. Seamlessly. That’s what blockchain UX is supposed to feel like.
With this, Starknet now offers builders a composable stack with multiple flavors:
Starknet L2: maximum composability
Starknet L2 appchains: flexibility + autonomy
Starknet L3s: hyperscaling and use-case-specific adaptation
Pick your layer, launch your project, abstract the pain, and scale your business.
 |
b. Dev on Starknet… without writing a single line of Cairo
Cairo Coder is now live, a new tool from Kasar Labs that lets anyone generate Cairo contracts and programs directly from Cursor, their AI-assisted IDE.
It’s fully compatible with the Starknet dev stack and always up to date with the latest tools and features.
With Cairo Coder, you can now build whatever you want using Cairo, the most powerful and flexible provable language out there, without writing a single line of it.
So, whatever you have in mind: just build. Starknet handles the rest.
Here’s a real example of what devs are already building with it.
Want to try it? Start here.
c. Builder support is scaling too
The Starknet Foundation has launched two new initiatives to help projects start, grow, and thrive on Starknet:
Propulsion v2: covers gas fees for eligible dApp users (up to $1M total).
Payments Builder Grants: STRK grants for teams building payment-related projects, with support for user acquisition and traction.
The goal? Help good ideas go from 0 to live, and remove the friction along the way.
Devs (and non-devs) now have everything they need to build whatever they imagine, find PMF, and thrive on Starknet. Time to build.
|
VI. Extended: a next-gen Perp DEX goes native on Starknet
One of the biggest unlocks in June? Extended, a high-performance Perp DEX, announced its migration from StarkEx to Starknet L2, and it’s happening this summer.
Extended is arguably the most complete Perp product ever launched on Starknet. Think: Hyperliquid-level UX and features, but built natively on Starknet. The team behind Extended includes ex-Revolut builders, bringing TradFi-grade execution into crypto-native DeFi.
In short, Extended is bringing to Starknet:
Perp trading with up to x100 leverage
50+ assets, including crypto pairs and TradFi markets (Gold, S&P500, Nasdaq, Brent, EUR…)
A community vault yielding ~22–30% APR in USDC for LPs
An active XP campaign (Season 1 already live)
Full EVM wallet support: MetaMask, Rabby, Phantom, and more
Their migration plan is already live, and the protocol is set to launch on Starknet Mainnet by late July or early August. Everything has been designed for a seamless transition for existing users.
In my humble opinion? This is the biggest DeFi launch on Starknet since Ekubo and Vesu. And it’s exactly the kind of flagship product Starknet needs as its L2 liquidity layer matures.
Nhận xét
Đăng nhận xét