Hackathons usually look clean in retrospect. You see the opening announcement, a few screenshots, finalists on a livestream, then a winners post with trophy emojis.
The actual build is much less tidy.
APIs fail. Product scope grows faster than the codebase. Teams discover that the feature they spent four days building does not solve the problem they started with. Someone eventually asks the uncomfortable question: would this product lose anything if we removed the blockchain entirely?
HackCanton Season 2 spent a lot of time in that territory.
More than 360 builders from 60+ countries entered the program, with 75+ teams appearing across the wider participation funnel. Thirty projects reached judging. Eighteen made the live Grand Final.
The strongest submissions had one thing in common: Canton was doing real architectural work.
A Daml contract buried somewhere in the backend was not enough. Teams had to show why privacy, selective visibility, authorization, or settlement changed the product.
That requirement produced private perpetual markets, institutional OTC desks, confidential RFQs, tokenized fund infrastructure, Bitcoin yield strategies, covered-call vaults, RWA products, analytics, and developer tooling.
Season 2 also distributed $10,000 in cash, $25,200 in NaaS credits, and 100,000 CC through the BitSafe and OnRails ecosystem challenges. For builders, though, the more useful result is a growing library of working Canton patterns that can now be studied, forked, challenged, and improved.
HackCanton Season 2 in Numbers
| Metric | Result |
|---|---|
| Builders | 360+ |
| Countries | 60+ |
| Teams across the wider funnel | 75+ |
| Projects in judging | 30 |
| Grand Finalists | 18 |
| Live Grand Final pitches | 18 |
| Core tracks | 5 |
| Ecosystem challenges | 2 |
| Cash awarded | $10,000 |
| NaaS credits | $25,200 |
| Partner rewards | 100,000 CC |
| Season 2 videos | 11 |
The funnel tells more than the registration count. Signing up takes minutes. Reaching judging means a team has made enough product and technical progress for somebody else to inspect the work.
Thirty projects crossed that line. Eighteen survived the next cut and had to defend their products live.
That is a much better measure of a hackathon than raw registrations.
The Main Constraint: Canton Had to Matter
A useful Canton project needs a more precise argument than “privacy is good.”
In financial software, the real question is usually which information should be private, from whom, and for how long.
Take an institutional RFQ. A buyer may need quotes from several dealers without letting those dealers see one another's prices. The wider market should not necessarily know that a large order exists before execution. Once the trade settles, auditors may need information that competing market participants still should not receive.
Those are product requirements, not abstract blockchain properties.
Season 2 produced several different versions of that problem.
| Project | Workflow problem | How Canton was used |
|---|---|---|
| Rocky Exchange | Perp positions can leak trading strategy | Private positions and on-chain settlement |
| Umbra | OTC dealers should not see competing quotes | Selective visibility and atomic DvP |
| Tirai | Large RFQs can reveal market-moving intent | Confidential pre-trade workflows |
| Pulla | Public pack openings make card counting possible | Controlled visibility of individual pulls |
| CrossDesk | Tokenized assets still need market infrastructure | Private fund and auction workflows |
The projects are very different on the surface. What connects them is information design.
Rocky cares about who can see a trading position. Umbra cares about who can see a dealer quote. Pulla cares about whether an observer can reconstruct the remaining contents of a blind-box pool.
A practical test came out of those examples:
If this product moved to a globally transparent chain tomorrow, what would stop working properly?
If the answer is “nothing,” the architecture probably needs another pass.
The Build Loop Started With the Product, Not the Contract
HackCanton teams moved through six stages:
Problem → ICP → Validation → GTM → MVP → Pitch
That structure pulled product decisions into the engineering process instead of postponing them until demo day.
| Stage | Question |
|---|---|
| Problem | What is broken today? |
| ICP | Who feels that problem most clearly? |
| Validation | What evidence shows the problem is real? |
| GTM | How could the first users reach the product? |
| MVP | What is the smallest workflow that proves the idea? |
| Pitch | Can the team explain the product under pressure? |
Each project also maintained an AI-guided journal inside AppsFactory. In practice, the journal was useful because it preserved the parts of a build that normally disappear from the final presentation: dead ends, technical blockers, abandoned features, scope changes, and assumptions that failed.
That history matters during a three-week build. A team can spend the first week designing a broad platform and discover halfway through that only one workflow is worth shipping.
The journal gave mentors a way to see that change rather than judging only the polished endpoint.
Scope was often the harder problem anyway. Three weeks is plenty of time to build one convincing workflow, but also plenty of time to accumulate a graveyard of half-connected features.
What Builders Actually Had to Ship
Once a team had a product worth testing, the engineering work became concrete.
Season 2 builders worked across Daml contracts, parties, Canton APIs, DAR deployment, DevNet, wallets, backend services, CBTC, cETH, and frontend integration.
This is one reason the workshop program leaned heavily toward implementation rather than broad ecosystem talks.
Opening Ceremony: Setting the Build Rules
Season 2 opened with more than logistics. The session introduced the AppsFactory workflow, project milestones, tracks, judging, team formation, and the partner challenges builders would work with during the season.
More importantly, it set the technical standard for the whole build:
Using Daml was not enough. Teams had to show why Canton improved the workflow.
That pushed builders to think about privacy, authorization, settlement, and visibility before they disappeared into implementation.
Workshop Track
| Session | What builders got from it |
|---|---|
| CBTC Bounty Challenge Kickoff | BitSafe introduced CBTC and Bitcoin-backed use cases around yield, trading, and settlement. |
| Canton Node + NaaS Workshop | NODERS covered DevNet, parties, APIs, DAR deployment, logs, debugging, and managed infrastructure. |
| How to Build a Complete dApp Using Ginie and dApp SDK | Ginie showed a workflow from plain-English requirements to Daml, testing, deployment, and frontend integration. |
| cETH Track by OnRails: The Basics | OnRails covered cETH use cases across trading, liquidity, lending, collateral, and treasury workflows. |
| Intro to Daml and Canton Architecture, DPM | Jatin Pandya from Canton Foundation covered Daml, parties, DPM, DAR packaging, and deployment. |
| Go-DAML & Go-Wallet Workshop | NODERS showed how Go backends can interact with Canton contracts, wallets, and asset flows. |
| What Comes Next for the Best Finalists | Finalists learned about infrastructure support and ways to continue after the competition. |
The workshop list also gives a decent picture of what a Canton application looks like once it leaves the smart contract editor.
There is contract logic, but there are also permissions, infrastructure, wallets, backend services, assets, and frontend state. A serious financial application has to connect all of those layers without losing sight of who can see or authorize each action.
Full Season 2 playlist:
https://www.youtube.com/playlist?list=PLcruYrU6F49c
Why Async Judging Came Before the Live Pitch
Thirty projects entered async judging before the Grand Final.
That order helps separate product quality from presentation skill. A polished speaker can make an unfinished prototype sound credible for a few minutes; asynchronous review gives judges time to inspect the implementation, project materials, Canton integration, documented progress, and actual workflow.
Only then did the live stage enter the equation.
The 18 finalists were:
| Auctions.Monster | 4lpha AI | CCX |
| Atrium | Trezia | Tirai |
| Pulla | BasisYield | Overwrite |
| Fourcast | Umbra | Rocky Exchange |
| Menzel | Continuum | Vanton |
| Veilio | CrossDesk | Daml Lens |
On August 5, those teams pitched live and answered questions from the judges.
For infrastructure projects, this is a useful second test. Architecture can absorb a lot of complexity on paper. A live pitch forces the team to compress it into a problem, a workflow, and a reason the system deserves to exist.
The async stage checked the work. The final checked whether the team understood the work well enough to explain it.
The Overall Winners
🥇 Rocky Exchange: Private Perpetuals Without Broadcasting the Trade
Rocky Exchange won HackCanton Season 2 overall and also took first place in the OnRails cETH Challenge.
The project is a privacy-preserving perpetual futures DEX built natively on Canton. Its design starts from a familiar problem in transparent derivatives markets: large positions can become useful information for other traders.
Rocky combined private positions with a Rust-based CLOB, Canton contract logic, and atomic settlement. Privacy was therefore part of the market mechanics rather than a setting in the UI.
| Reward | Result |
|---|---|
| Cash | $4,000 |
| NaaS | 3 months at 75% off |
| Ambassador Program | 3 months |
| cETH Challenge | 25,000 CC |
“Winning felt unreal. Building Rocky on Canton and seeing the whole community vibe with it was the best part. Grateful for the opportunity and everyone who supported us along the way.”
Watch the pitch:
https://youtu.be/U_bZDNHiXyg?t=4864
🥈 Umbra: Private OTC Settlement, Built Solo
Umbra finished second overall and also placed second in both the CBTC and cETH challenges.
It is a private OTC venue for institutional block trades. Buyers can request quotes from several dealers without revealing competing quotes, then settle the selected transaction atomically.
The project demonstrated private CBTC ↔ cETH settlement on DevNet, so the cross-asset workflow existed as more than a sequence diagram.
The entire project was built by one person.
| Reward | Result |
|---|---|
| Cash | $2,500 |
| NaaS | 2 months at 75% off |
| Ambassador Program | 2 months |
| CBTC Challenge | 15,000 CC |
| cETH Challenge | 15,000 CC |
“Umbra was a solo build, one person trying to prove you could settle real CBTC against real cETH, privately and atomically on Canton.”
Three second-place finishes made Umbra one of the strongest technical performances of the season, but the more interesting result is the workflow itself: private negotiation followed by atomic settlement across real ecosystem assets.
Watch the pitch:
https://youtu.be/U_bZDNHiXyg?t=4440
🥉 Tirai: Keeping the RFQ Private Before the Trade Happens
Tirai finished third with a confidential multi-dealer RFQ and OTC desk.
The project focused on something that often gets lost in discussions about blockchain settlement: the information leak can happen before settlement begins.
If an institution broadcasts the size and direction of a large intended trade, the market can react before execution. Tirai keeps RFQs and competing dealer quotes private, then moves the final transaction through atomic DvP settlement.
Its design supports Canton ecosystem assets including cETH, CBTC, and Canton Coin.
| Reward | Result |
|---|---|
| Cash | $1,500 |
| NaaS | 1 month at 75% off |
| Ambassador Program | 1 month |
“Tirai came out of a lot of late nights and a bet on an idea I really believed in, so seeing it win felt unreal.”
Watch the pitch:
https://youtu.be/U_bZDNHiXyg?t=2239
Two Track Winners With Very Different Products
CrossDesk: What Happens After Tokenization?
CrossDesk won the DeFi / Financial Applications track with an on-chain fund factory and sealed auction desk.
Token issuance is easy to show in a demo. What happens afterward is harder. Funds need operational logic, ownership flows, market infrastructure, and ways for assets to change hands without the whole product collapsing into a token contract plus a dashboard.
CrossDesk tackled that second layer and demonstrated real transactions on DevNet.
Prize: $1,000
“I enjoyed learning a ton throughout the process. The part I'm excited about is what funds will be able to do with it.”
Its next interesting problem is no longer whether the MVP runs. It is whether the fund workflows survive contact with real users.
Watch the pitch:
https://youtu.be/U_bZDNHiXyg?t=7065
Pulla: Privacy in a Collectibles Market
Pulla won the RWA & Business Workflows track with a marketplace for blind-box packs containing real PSA-graded collectible cards.
The product looks playful, but the market-design problem underneath it is not.
If every pack opening is public, observers can track which rare cards have already been pulled and estimate what remains. Once that information becomes public state, the economics of the blind box start to change.
Pulla uses Canton’s visibility model to prevent outside observers from tracking individual pulls while keeping the final outcome verifiable.
Prize: $1,000
It is a useful counterexample to the idea that Canton privacy only makes sense for bank software. The same architecture can matter anywhere the value of a market depends on information not becoming public too early.
Watch the pitch:
https://youtu.be/U_bZDNHiXyg?t=2638
BitSafe CBTC Challenge: CBTC Across Three Financial Workflows
BitSafe backed a 50,000 CC challenge for projects building around CBTC.
| Place | Project | Reward |
|---|---|---|
| 🥇 | BasisYield | 30,000 CC |
| 🥈 | Umbra | 15,000 CC |
| 🥉 | Overwrite | 5,000 CC |
- BasisYield used CBTC inside a private, market-neutral yield strategy built around perpetual-futures funding.
- Umbra used CBTC as a settlement asset in confidential OTC trades, including atomic swaps against cETH.
- Overwrite used deposited CBTC to back weekly physically settled covered calls implemented with Daml contracts.
The challenge also came with technical guidance, mentoring, and a dedicated kickoff from BitSafe. The result was three distinct financial uses for CBTC: yield, private settlement, and options infrastructure.
OnRails cETH Challenge: cETH Inside Live Market Workflows
OnRails backed another 50,000 CC, this time around cETH.
| Place | Project | Reward |
|---|---|---|
| 🥇 | Rocky Exchange | 25,000 CC |
| 🥈 | Umbra | 15,000 CC |
| 🥉 | Ibex | 10,000 CC |
- Rocky integrated cETH into its private perpetual trading stack.
- Umbra used it in atomic OTC settlement against CBTC.
- Ibex brought cETH into confidential perpetual markets.
The challenge also came with a dedicated workshop, mentoring, and direct technical support during implementation. For a builder, that difference is substantial. A prize may be enough to open the documentation; getting a live integration working usually requires more than that.
Prize Structure
Season 2 combined cash, infrastructure credits, ecosystem assets, and continuation support.
| Category | Distributed |
|---|---|
| Overall podium cash | $8,000 |
| Track winner cash | $2,000 |
| Total cash | $10,000 |
| Podium NaaS credit value | $18,000 |
| Finalist NaaS credit value | $7,200 |
| Total NaaS credits | $25,200 |
| BitSafe CBTC Challenge | 50,000 CC |
| OnRails cETH Challenge | 50,000 CC |
| Total ecosystem rewards | 100,000 CC |
The NaaS component addressed a problem that begins as soon as a hackathon ends: somebody still has to run the infrastructure.
Rocky, Umbra, and Tirai received NaaS at $1,000 per month instead of the $4,000 market price for three, two, and one months respectively. That creates $18,000 in podium credit value.
All 18 finalists also received one month at the founder rate of $3,600 instead of $4,000, adding another $7,200 in credits.
The combined reward package was therefore:
$10,000 in cash + $25,200 in NaaS credits + 100,000 CC
Ambassador Program support for the top three came in addition to those figures.
Partners Inside the Engineering Loop
Season 2 included Canton, Canton Foundation, BitSafe, OnRails, PixelPlex, CCView, Console Wallet, and Ginie in the wider partner ecosystem.
The contributions developers felt most directly were the ones tied to code, infrastructure, assets, or support.
| Partner | Contribution |
|---|---|
| BitSafe | CBTC challenge, 50K CC, technical guidance, mentoring |
| OnRails | cETH challenge, 50K CC, workshop, mentoring |
| Ginie | Daml generation and dApp tooling |
| PixelPlex | Wider technical and ecosystem support |
NODERS addressed another part of the stack through Canton node infrastructure, NaaS, Go-DAML, and Go-Wallet. Those pieces become important as soon as a project grows beyond an isolated contract and starts needing persistent backend services or wallet flows.
The practical difference between a logo sponsor and an engineering partner becomes obvious near a submission deadline. Developers remember the teams that are still answering integration questions when the demo is due in two days.
What Developers Can Take From Season 2
1. Define the leak before choosing the privacy architecture
“Private finance” is too vague to build against.
“Dealer A must not see Dealer B's quote before execution” is useful. So is “a large perp position should not become a public trading signal.”
Once the information leak is concrete, choices around visibility, authorization, and settlement become easier to reason about.
2. Model the whole transaction lifecycle
Several Season 2 projects became stronger because they treated settlement as part of the product rather than background plumbing.
In institutional workflows, agreement is only one step. Authorization, payment, delivery, counterparty risk, and auditability all affect whether the product is usable.
If an application models the trade but ignores how value actually changes hands, it may be modeling only half of the workflow.
3. A complete narrow workflow beats a broad unfinished platform
Hackathons make feature creep unusually tempting because every new idea feels like something that could impress judges.
Usually the opposite happens.
A small workflow that can be demonstrated end to end gives judges something concrete to evaluate. A large platform filled with disconnected screens creates more surface area for questions.
4. Real assets expose architectural weaknesses quickly
CBTC and cETH forced teams to deal with questions that remain theoretical in a toy demo.
Who holds the asset? Who can transfer it? What information is visible during execution? How does settlement happen? What fails if one side of the transaction does not complete?
Once value actually moves through the system, the architecture has nowhere to hide.
Why 18 Is the Number Worth Remembering
Season 2 started with more than 360 builders and 75 teams. Those numbers show reach.
The more useful number is 18.
Those teams made it through the part of the hackathon that never looks good in a promo video. They narrowed scope, changed assumptions, dealt with unfamiliar infrastructure, finished enough of the product for async judging, and then stood in front of a live audience with something they could defend.
That is where a builder program proves itself.
HackCanton Season 3 now begins with two seasons of examples behind it. New teams can study which ideas survived, where privacy actually changed product design, how ecosystem assets were integrated, and which workflows judges found convincing.
A useful starting question remains:
What becomes better, safer, or newly possible because this workflow runs on Canton?
A precise answer usually leads to a better MVP.
HackCanton Season 3
HackCanton Season 3 is now open.
5 weeks. Online. Solo or team.
Registration:
https://appsfactory.cc/hackathons
Resources
- HackCanton platform: https://hackathon.appsfactory.cc/
- AppsFactory: https://appsfactory.cc/
- AppsFactory X: https://x.com/appsfactory_cc
- AppsFactory TG: https://t.me/appsfactory_cc
- AppsFactory Discord: https://discord.com/invite/v8ESnCnhJp
Top comments (0)