我們很高興宣佈 AppWorks 投資了 Sanctum 的原生代幣 $CLOUD。這項投資象徵著我們對 Sanctum 創始團隊的長期信念，以及對於他們多年來建立的強大護城河的認可。

在這篇文章中，我們將分享為什麼我們認為 Sanctum 是 Solana 生態系中最重要的基礎建設之一；它將如何重塑加密貨幣領域中最被誤解的垂直賽道，以及為什麼現在 Sanctum 正處於爆發前夕。

Sanctum：打破常規的 Solana LST

要理解 Sanctum，我們首先需要了解 Solana 流動性質押代幣（LST）的歷史。

很長一段時間以來，大多數人假設 Solana 的流動性質押市場會遵循以太坊的劇本：出現一個主導者（如 Lido），而其他競爭者的市占率將逐漸被該主導者蠶食鯨吞。因此，在 Solana 流動性質押市場的早期，許多協議爭相複製 Lido 模式，試圖在早期達到規模經濟並統治該賽道。

然而，最具洞察力的創辦人往往能看見他人忽視的細節。

Sanctum 的創辦人 FP Lee 作為設計 Solana 質押帳戶合約的核心貢獻者，很早就知道 Solana 的架構設定與以太坊有著根本上的不同。每一個質押帳戶（Stake Account）本質上都是模組化、可轉移且可組合的。這意味著所有 Solana 上的 LST 在核心層面上本質是同質的（Fungible），它們只是同一個底層質押帳戶的不同包裝。

有了這個認知，下一個問題顯而易見：假設一個有著多元 LST 的世界注定將發生，那我們還缺少什麼？

仔細觀察，質押池（Stake Pool）服務於三個不同的目的：

1. 代幣化（Tokenize）： 將原生質押帳戶轉換為生息的 SPL 代幣。
2. 委託（Delegate）： 實現高質押收益並分散網絡去中心化。
3. 流動性（Liquidate）： 為質押的 SOL 提供即時流動性。

對 FP 而言，他的答案是「流動性」。

原因很簡單：以太坊已經進行過這個實驗。Lido 之所以主導以太坊，很大程度上是因為 stETH 整合進了所有主要的 DeFi 協議，並具有極高的可組合性。多年的整合使其成為網路上最具可組合性的資產，其在派發上的優勢逐漸如滾雪球般幫助 Lido 達到了統治地位。

在 Solana 上，同樣的結論更加清晰：質押帳戶是可以互換的；每一顆質押的 SOL 在交易流中都承載著相同的經濟權重。在這個世界裡，「多 LST 並存」的未來是預設狀態，而誰能協調流動性，誰就能贏得這場競爭。

因此，FP 沒有試圖爭奪成為單一最大的質押池，而是看到了更大的機會：成為連結 Solana 上所有 LST 的流動性層。

這就是 Sanctum 的起點。

他們破解這個難題的方式與上述洞察息息相關。首先，他們創建了 「Sanctum Router」，用於在成千上萬種 LST 之間進行交換。這種交換只需簡單地「unwrap」和「re-wrap」底層質押帳戶即可完成。

必要時，這些 Swap 也可以導流到 Infinity Pool，其這可以被視為一籃子由 Sanctum 管理的 LST 組合，且所有這些 LST 都在 Sanctum 的 LST 發行基礎設施下運作。當 Swap 被導流進入 Infinity 時，交易費用將分配給 Infinity Pool 的流動性提供者（即 $INF 持有者）。

這種設計增強了整個 Sanctum LST 生態系統的流動性，同時將 $INF 本身變成了 Solana 上最具吸引力的 LST 之一。

最後，Sanctum 管理著規模可觀的原生 $SOL 儲備，這由所有 Sanctum 品牌的 LST 共享。對於任何項目來說，他們無需自行建立數百萬美元的流動性池，只需透過 Sanctum 發行他們的 LST，而不必擔心流動性的冷啟動問題。

在不到兩年的時間裡，Sanctum 已經和 Bybit、Crypto.com、Backpack、Jupiter 等領域內的主要參與者合作，管理著超過 20 億美元的 $SOL。截至 2025 年底，Sanctum 以創紀錄的指標結束了該年度：總管理資產達 1,436 萬顆 SOL（年增 96%），年營收達到 585 萬美元（年增 226%）。

另一個令人印象深刻的指標是 Sanctum 在 2025 年對 Solana LST 類別的統治力；Sanctum 實現了近 800 萬顆 $SOL 的淨累積流入，而其兩大競爭對手 Jito 和 Marinade 的 LST 產品則以淨流出結束了這一年。這種差距凸顯了 Sanctum 在 LST 市場中流動性護城河的複利競爭優勢。

回過頭看，人們很容易認為這一切都順理成章，彷彿 Sanctum 的崛起是注定的。但事實是，這是Sanctum 團隊於過去五年來堅持不懈、不斷迭代，並在市場寒冬中保持信念才達成的結果。這將我們引向第一個，也是最重要的投資原因：創辦人 FP Lee。

投資論點 #1：我們對創辦人 FP Lee 的核心信念

 

尋找非共識的洞見

我們在 FP 身上看到一種獨特的能力：他能看見別人錯過的東西，並深入鑽研一個概念，直到浮現出反向的洞察。

2023 年，當許多人競相打造「Solana 版的 Lido」時，FP 注意到了上游的一些東西：Solana 的質押帳戶在設計上是模組化且同質的。這單一的觀察顛覆了整個劇本。他沒有去競爭成為最大的質押池，而是將 Sanctum 打造為連結每一個 LST 的流動性層，這徹底改變了遊戲規則。

如今，在 Solana 生態系中建構多年後，他又發現了另一個鮮少人注意到的低效之處：Solana 交易層的不可預測性。簡而言之，我們相信他近期收購 Ironforge 並推出 Gateway 的舉動反映了同樣的模式：鑽得夠深，找到結構性的瓶頸，然後構建出其他人往往視為理所當然的缺失基礎設施。

與社群的深度連結 

我們認為創辦人在 Crypto 世界裡建立社群的能力常常被低估。對我們而言，FP 是少數不只是「管理」社群，而是與社群「共同建設」的創辦人之一。

Sanctum 在代幣生成事件（TGE）前舉辦的 Wonderland 活動，不僅僅是另一個積分農場（Points farm）；Sanctum 將其變成了一個協作的、遊戲化的任務層，用戶質押 LST 以賺取經驗值並共同完成社群挑戰，這使得參與度具有黏性，而非僅是為了逐利。

結果令人驚艷，Sanctum 的 TVL 在 38 天內增長了 171%，社群論壇衍生出了常態性活動（如 Cloudfam Emoji Clan 競賽），讓成員在初次空投後很久仍持續創作與招募，展現了有機的動能而非付費買來的流量。其社群（後來被親切地稱為 CloudFam）成為了 Solana 上最強大的草根運動之一。

大多數人不知道的是，FP 和團隊為了展現這份在乎付出了多少努力。當 $CLOUD 空投的時候，他們不僅僅是跑個程式腳本或在試算表裡算數字。他們手動篩選了成千上萬個 Discord 頻道和 Telegram 聊天記錄，以識別真正的死忠用戶，並給予他們遠比大多數項目更慷慨的獎勵。

代幣終究是創辦人的第二個產品。把「代幣投資人關係（Token IR）」做好不僅僅是管理價格；更是建立信任、透明度和歸屬感。許多創辦人說他們在乎代幣持有者，但很少有人真正言行一致。

FP 是少數言行合一的創辦人之一，從發布月度更新到在 Discord 中持續與用戶互動，他將社群建設視為一門他仍持續優化、且熱愛的事情。隨著時間推移，這種真實性會產生複利效應，將用戶群轉化為更接近「運動」的力量。我們相信這將仍然是 Sanctum 最強大的長期差異化優勢之一。

找到天命

FP 最初並沒打算成為創辦人。在進入加密貨幣領域之前，他只是一個試圖尋找方向的年輕軟體工程師。他是出於好奇和對成長的渴望而跌跌撞撞進入 Solana 的，而他之所以留下來，是因為在旅途中，他找到了天命。

這種天命的發現來自於從事真正熱愛的事情。在加密領域建設多年後，FP 經歷了許多事情：FTX 的崩潰、許多創辦人對用戶撤資跑路、項目魯莽地賭光整個國庫等等。在一個常被短期主義和貪婪籠罩的行業裡，FP 想要證明：在 2026 的今天，你仍然可以打造一個偉大的產品、保持透明，並以正確的方式成功。

雖然聽起來有些虛無飄渺，甚至有些老套，但我們相信這是這個產業中最欠缺，卻也強大的驅動力之一，這也正是我們想要支持的使命。

投資論點 #2：許多對於 Sanctum LST 的利好正在醞釀

 

我們看見 Sanctum LST 的幾個催化劑，並堅信現在萬事俱備，只欠東風。

Solana LST 仍處於早期階段並在成長中

Solana LST 的滲透率在兩年內增長了 4.4 倍，從 3% 增長到 13.3%。這種加速是由像 Sanctum 這樣的生態系基礎建設的出現，以及節點運營者對搶佔 LSTfi 市場份額日益增長的需求所積極推動的。

然而，這 13.3% 的滲透率意味著大多數質押的 $SOL 仍然是非流動的。對比以太坊更成熟的市場（接近 35-40% 的 LST 滲透率），顯然 Solana 的流動性質押層——這個解鎖資本效率和可組合性的關鍵——仍處於早期局勢。

在此背景下，我們相信 Sanctum 正處於那個轉折點。Sanctum 的產品套件（Router、Reserve、Infinity）被明確設計為這種中立的流動性中心。這創造了一個強大的護城河，使 Sanctum 成為這個快速擴張且碎片化市場的預設聚合點。

LST 的網路效應

對於任何發行新 LST 的協議來說，他們不想自行建立數百萬美元的流動性池。相反，最簡單的方法是接入 Sanctum，並只專注於品牌、核心產品和分發；Sanctum 提供流動性層（Router + Reserve + Infinity），使這些 LST 在第一天就能與 Solana 的其餘部分互換。這就是我們所見證的。這個聯盟每天都在變得越來越強大。迄今為止，Sanctum 正在服務所有最大的客戶，如 Jupiter、Bybit、Backpack。

更多整合 → 更多支援的 LST → 更多可互換的路徑 → 跨 DeFi 的更深層效用。 這個飛輪效應是很難被追上的。

機構也正在採用 Solana LST

對 Solana 收益的競爭正迅速演變，超越了 DeFi 原生協議的範疇。新一波規模更大的成熟資本，從傳統金融的資產管理公司、潛在的 SOL 質押 ETF 到上市公司（如 DATs），正進入這個競技場。

對於這些實體而言，基礎質押收益僅僅是起點。為了吸引龐大的資產管理規模（AUM），他們必須提供合規、品牌化且高度差異化的流動性產品。這需要在基礎 LST 之上疊加獨特的驗證者策略、透明的報告和深度的 DeFi 整合。

這正是 Sanctum 成為關鍵賦能基礎設施的地方。作為一個中立、模組化的平台，它提供了全端式的「質押即服務（Staking-as-a-Service）」，讓任何實體都能發行自己的 LST。它允許他們保留對財務策略和品牌的完全控制權，同時將深層的技術複雜性外包給 Sanctum 經過驗證的流動性層和可組合後端。

我們已經看到了強有力的驗證。2025 年 5 月，上市金融公司 DeFi Development Corp (NASDAQ: DFDV) 使用 Sanctum 的基礎設施推出了 dfdvSOL。這標誌著上市公司首次持有並使用定制的 Solana LST。他們的明確目標是將流動性貢獻回 Solana DeFi 生態系統，創造一種也能讓自家公司資產負債表受益的飛輪效應。

我們的核心信念是，隨著更多結構化的 SOL 產品出現以競爭這波新資本，Sanctum 將成為他們的預設發行合作夥伴和流動性骨幹。

我們的論點很明確：在中短期內，Sanctum 是這波機構浪潮中最被低估的基礎設施標的，他們正悄悄成為每個主要生息 Solana 產品所依賴的必要管線。

投資論點 #3：Sanctum App 將成為 Solana 的實質入口

在 2025 年 Solana Breakpoint 大會上，Sanctum 揭曉了其最新力作：Sanctum App。這是一個輕量級的手機端應用程式，旨在幫助用戶在一個有趣、遊戲化的環境中，透過每日獎勵輕鬆增長他們的 $SOL 持有量。

回到我們最初的論點，Sanctum LST 尚未達到市場滲透率峰值的主要原因是當前 DeFi 用戶體驗（UX）固有的摩擦。人們很容易忘記，許多 $SOL 持有者完全存在於「DeFi 泡沫」之外。他們對 Solana 的技術內部結構或 LST 收益背後的複雜數學不感興趣；接觸他們的唯一方法是將這些複雜性抽象化。

繼 Wonderland 的巨大成功之後，我們加倍確信 Sanctum 創始團隊是少數擁有這種「基因」的團隊之一：不僅能建立社群，更重要的是能贏得社群持久的信任。從團隊基因的角度來看，我們有信心他們能再次創造奇蹟。

從產品角度來看，這段旅程已經形成了一個完整的閉環。Sanctum 團隊是將複雜的質押帳戶概念轉化為易於使用的 LST 的先驅。現在，他們正將這一使命擴展到更廣泛的受眾。展望未來，我們預期 Sanctum App 將成為日常用戶的主要介面和入口，而 Sanctum LST 則繼續作為整個 Solana 基礎設施的基石。

總結

當我們看到 Sanctum 時，我們看到的不僅僅是一個協議，而是 Solana 下一個篇章的基礎：其核心是 FP，一位充滿長期思維，以信念以及熱忱建構產品的創辦人，他帶領 Sanctum 從一個簡單的想法發展成為 Solana 的流動性織物，現在更進一步擴展到交易基礎設施。除此之外，我們也預期 LST 業務將透過強大的網路效應、機構採用和跨生態系統的深度整合持續產生複利。

這就是我們投資的原因。FP 和 Sanctum 代表了我們所相信的一切。

About AppWorks

AppWorks 成立於 2010 年，是亞洲領先的創業加速器與創投機構。其生態系匯聚超過 2,000 位創業者，旗下基金管理規模達 3.86 億美元。AppWorks 加速器專注於協助早期新創實現從 0 到 1 的跨越，而 AppWorks 創投基金則支持團隊進行規模化成長。

We are delighted to share that AppWorks has invested in $CLOUD, the native token of Sanctum, as a demonstration of our long-term belief in Sanctum’s founding team and in the strong defensibility and moat they’ve built over the years.

Below, we will share why we think Sanctum is one of the most important companies being built on Solana; how it is reshaping one of the most misunderstood verticals in crypto, and why everything is finally clicking for them.

Sanctum – A Chronicle of Finding the Non Consensus and Right

To understand Sanctum, we need to first understand the history of Solana LST. 

For the longest time, most people assumed Solana’s liquid staking market would follow the Ethereum playbook: one dominant player, i.e. Lido, and everyone else fading into the background. Many protocols rushed to build copycats of the Lido model, believing the game was simply about scale and validator sets.

The most insightful founders see what others miss. FP Lee, Sanctum’s founder, being the core contributor who has designed Solana’s stake account contracts, knew early on that Solana’s setup is fundamentally different from Ethereum’s. Every stake account is modular, transferable, and composable. That means all LSTs on Solana are essentially fungible at their core. They are just wrappers of the same underlying stake accounts.

With that realization, the next question was obvious: in a world of many LSTs, what’s the missing primitive? 

Look closely, stake pools serve three distinct purposes: 

• Tokenize – convert native stake accounts into yield bearing SPL token
• Delegate – achieve high staking yield and decentralize the network 
• Liquidate – Provide instant liquidity for staked SOL

And FP bet on liquidity. 

Why? Because Ethereum already ran the experiment. Lido dominated Ethereum largely because stETH is integrated into all major DeFi protocols and has extremely high composability. Years of integrations made it the most composable asset on the network, and distribution compounded into dominance.

On Solana, the conclusion is even clearer. Stake accounts are interchangeable; every staked SOL carries the same economic weight in the transaction flow. In this world, a multi-LST future is the default, and whoever orchestrates liquidity wins the competition. 

Therefore, instead of fighting to become the single biggest stake pool, FP saw a bigger opportunity in becoming the liquidity layer that unites all LSTs on Solana.

And this was the start of Sanctum. 

The way they cracked this is deeply connected to the insight. First, they created “Sanctum Router” for swaps between thousands of LSTs. A swap takes place with simply unwrapping and re-wrapping the underlying stake account.

When necessary, swaps can also be routed into Infinity Pool, which can be regarded as a basket of LSTs that is managed by Sanctum, and all of these LSTs are under Sanctum’s flagship white labeled LST issuance infrastructure. When a swap is routed into Infinity, the trading fees are distributed to the liquidity providers of Infinity Pool, i.e. $INF holders. 

This design reinforces liquidity for the entire Sanctum LST ecosystem while turning $INF itself into one of the most compelling LSTs on Solana.

Lastly, a sizable native $SOL reserve managed by Sanctum is shared across all Sanctum branded LSTs. For any project, instead of spinning off another multi-million liquidity pool, they can just issue their LSTs with Sanctum without worrying about bootstrapping the liquidity. 

In less than 2 years, Sanctum is now managing over US$ 2B of $SOL with various key players in the space, including Bybit, Crypto.com, Backpack, Jupiter, etc. By the end of 2025, Sanctum closed the quarter with record-high metrics: 14.36M SOL (+96% YoY) under its umbrella and an annual revenue of US$5.85M (+226% YoY).  

Another impressive metric to look at is Sanctum’s dominance of the Solana LST category in 2025; Sanctum achieved a net cumulative inflow of nearly 8M $SOL, while its two largest competitors, Jito and Marinade, saw their LST offerings end the year with net negative inflows. This disparity underscores the compounding competitive advantage of Sanctum’s liquidity moat within the LST market.

Looking back, it is easy to think it was all smooth and inevitable, as if Sanctum’s rise was destined. But the truth is, this was the result of five years of persistence, iteration, and unshakable conviction through market winters. Every piece of today’s success was earned the hard way, block by block, decision by decision. And that leads us to our first and foremost thesis: the founder FP.

Thesis #1: Our core beliefs in FP Lee as a founder

Practiced Contrarianism

We see in FP that he has a unique ability to see what others miss and to dive deep enough into a concept until a contrarian insight emerges.

In 2023, while many were racing to build “the Lido of Solana,” FP noticed something upstream: Solana’s stake accounts were modular and fungible by design. That single observation flipped the playbook. Instead of competing to be the biggest pool, he built Sanctum as the liquidity layer that unites every LST, and that completely changed the game. 

And now, after years of living in Solana’s core, he’s spotted another inefficiency few others even noticed: the unreliable transaction layer. In short, we believe his latest move of acquiring Ironforge and launching Gateway reflects the same pattern: go deep enough, find the structural bottleneck, then build the missing piece of infrastructure that everyone else takes for granted.

Deep attachment to community

We believe founders’ ability to build a community in crypto is underrated. FP is one of the rare founders who doesn’t just “manage” a community but builds alongside them. 

The Wonderland campaign, which Sanctum team held before their TGE, wasn’t just another points farm; Sanctum turned it into a coordinated, gamified quest layer where users stake LSTs to earn EXP and complete community challenges together, which kept participation sticky instead of mercenary. 

The result was amazing, Sanctum’s TVL grew 171% in 38 days, and community forums spun up recurring events (like the Cloudfam Emoji Clan contests) that kept members creating and recruiting long after the initial drop, showing organic momentum rather than paid churn. Its community, later affectionately known as CloudFam, became one of the strongest grassroots movements on Solana.

What most people don’t know is how far FP and the team went to show that care. When it came time for the $CLOUD airdrop, they didn’t just run a script or crunch numbers in a spreadsheet. They manually sifted through thousands of Discord channels, Telegram chats to identify true evangelists and rewarded them far more generously than most projects would.

A token is ultimately a founder’s second product. Doing “token IR” right isn’t just about managing price; it’s about building trust, transparency, and belonging. Many founders say they care about token holders, but few actually walk the talk. 

FP is one of those few that from publishing monthly updates to constantly engaging users in Discord, he treats community building like an ongoing craft. Over time, that level of authenticity compounds, turning a user base into something closer to a movement. We believe this will remain one of Sanctum’s strongest long-term differentiators.

Found his mission along the way

FP didn’t set out to be a founder. Before crypto, he was just a young software engineer trying to find his way. He stumbled into Solana out of curiosity and a desire for growth, and he stayed because, somewhere along the journey, he found meaning.

That meaning came from experience. After years of building in crypto, FP has seen it all — the FTX collapse, teams rugging their users, projects recklessly gambling away entire treasuries. In an industry often clouded by short-termism and greed, FP wants to prove that it is still possible to build a great product, stay transparent, and win the right way.

As intangible, and perhaps as cliché as it sounds, we believe this is one of the strongest forms of drive in the industry — and it is precisely the kind of mission we want to back.

Thesis #2: Major upcoming catalysts for Sanctum LST 

We see several catalysts for Sanctum LSTs, and we hold a strong conviction that everything is clicking for them now. 

Solana LST is still early-stage and growing

Solana’s liquid staking penetration has expanded 4.4x in two years, from 3% to 13.3%. This acceleration was actively driven by ecosystem catalysts like Sanctum and a growing appetite from validators to capture LSTfi market share.

However, this 13.3% penetration implies that most of the staked $SOL remains illiquid. Drawing a parallel with Ethereum’s more mature market (approaching 35-40% LST penetration), it’s clear Solana’s liquid staking layer, the key to unlocking capital efficiency and composability, is still in its early innings.

We believe in this context, Sanctum is positioned right at that inflection point. Sanctum’s product suite (Router, Reserve, Infinity) is explicitly designed to be this non-partisan liquidity backbone. This creates a powerful moat, positioning Sanctum as the default aggregation point for a rapidly expanding and fragmenting market.

Network effects of LSTs

For any protocols that are launching a new LST, they don’t want to spin up multi‑million‑dollar liquidity pools. Instead, the easiest way is to plug into Sanctum, and focus solely on branding, core product, and distribution; Sanctum supplies the liquidity layer (Router + Reserve + Infinity) so these LSTs are interchangeable with the rest of Solana on day one. This is what we have been seeing. The alliance has been getting stronger and stronger every day. To date, Sanctum is serving all the biggest clients like Jupiter, Bybit, Backpack. 

More integrations → more supported LSTs → more interchangeable routes → deeper utility across DeFi. 

That flywheel is hard to catch up to. 

The Institutional Staking Layer

The competition for Solana’s yield is rapidly evolving beyond DeFi-native protocols. A new, larger wave of sophisticated capital, ranging from TradFi asset managers and potential SOL staking ETFs to publicly traded companies (like DATs), is now entering the arena.

For these entities, base staking yield is merely the starting point. To attract serious AUM, they must offer compliant, branded, and highly differentiated liquid products. This requires layering unique validator strategies, transparent reporting, and deep DeFi integrations on top of a base LST.

This is precisely where Sanctum becomes the critical enabling infrastructure. As a neutral, modular platform, it provides the full-stack “Staking-as-a-Service” for any entity to launch its own LST. It allows them to retain full control over their financial strategy and branding while outsourcing the deep technical complexity to Sanctum’s proven liquidity layer and composable backend.

We are already seeing strong validation. In May 2025, DeFi Development Corp (NASDAQ: DFDV), a publicly traded financial firm, launched dfdvSOL using Sanctum’s infrastructure. This marked the first time a listed company held and utilized a bespoke Solana LST. Their explicit goal is to contribute liquidity back to the Solana DeFi ecosystem, creating a flywheel effect that also benefits their own corporate balance sheet.

It is our core belief that as more structured SOL products emerge to compete for this new wave of capital, Sanctum will be their default launch partner and liquidity backbone.

Our thesis is clear. In the short to medium term, Sanctum is the most underrated infrastructure play for this institutional wave, and they are quietly becoming the essential plumbing every major yield-bearing Solana product will depend on.

Thesis #3: Sanctum App will become the de facto entry point for Solana 

At the 2025 Solana Breakpoint, Sanctum unveiled its latest endeavor: the Sanctum App. This is a lightweight, delightful application designed to help users effortlessly grow their $SOL holdings through daily rewards in a fun, gamified environment.

This pivot stems from a first-principles approach. Referring back to our initial thesis, the primary reason Sanctum LSTs haven’t yet reached peak market penetration is the friction inherent in current DeFi UX. It is easy to forget that many $SOL holders exist entirely outside the “DeFi bubble.” They aren’t interested in the technical internals of Solana or the complex math behind LST yields; the only way to reach them is to abstract that complexity away. Following the massive success of Wonderland, we are doubling down on our belief that the Sanctum founding team is one of the few with the “DNA” to not only build a community but—more importantly—earn its lasting trust. From a team-gene perspective, we are confident they can capture lightning in a bottle once again.

From a product standpoint, the journey has come full circle. The Sanctum team were the pioneers who took the complicated concept of stake accounts and made LSTs accessible. Now, they are scaling that mission to a much wider audience. Moving forward, we expect the Sanctum App to serve as the primary interface and entry point for everyday users, while Sanctum LSTs continue to function as the foundational backbone of the entire Solana infrastructure.

Summary

When we look at Sanctum we see more than just a protocol, but the foundation of Solana’s next chapter: at its core is FP, a contrarian founder who builds with conviction and has led Sanctum from a simple idea into the liquidity fabric of Solana, and now it’s expanding even further into the transaction infrastructure. Apart from that, we also see the LST business continues to compound through strong network effects, institutional adoption, and deep integrations across the ecosystem. 

Ultimately, this is why we invested. FP & Sanctum represent everything we believe in. 

About AppWorks

Founded in 2009, AppWorks is a leading accelerator + VC  in Asia and Web3, managing US$ 368M across its funds. At AppWorks, we invest in the founders who are obsessed with true product-market fit and committed to playing the long game. If this resonates with you, let’s chat. 

從大學生寫 App，到打造雙引擎的創業者

JTCG 的創辦人周益鑫 (James) 是 AppWorks 加速器第 13 屆校友，也是少數同時具備產品、成長與企業銷售全方位能力的創業者。2016 年加入加速器之後，讓他在創業初期便透過豐富的社群活動結識許多同後屆的創辦人與導師，之後更進一步加入 AppWorks「六次方」社群，與 20–30 人規模的新創 CEO 定期討論招募、薪酬結構與營運難題，讓他不論在招募、募資或是找客戶，都能從社群中獲得支持。

在大學期間，他就打造了 Colorgy，一款幫助大學生整理課表與社交互動的 App。這產品最早從他自己手動匯入大學課表開始，一邊寫程式、一邊和同學測試，不斷修正介面與使用流程。他發現大學生其實對社交更感興趣，課表只是入口，因此他把焦點放在社群功能設計上，比如能看到同學選課、揪團開課等模組，讓用戶之間自然形成互動。他也從早期的數據中觀察到，不同科系的使用習慣差異極大，於是設計了模組化的彈性架構，以支撐多樣化需求。 Colorgy 最高曾達到 200 萬註冊用戶，深受大學生族群歡迎，整個過程讓 James 累積了完整的產品開發與使用者研究歷練。

後來他注意到法國的新創 Zenly 正在崛起，做的是即時地圖分享與朋友定位的社交 App。因為經營過大學生課表產品，James 對這個年齡層的生活模式與校園擴散節奏非常熟悉，他敏銳地意識到 Zenly 如果能搭配台灣校園的地推與產品優化，有機會快速打開市場。於是他主動寫信給 Zenly 的創辦人，分析台灣市場的成長潛力與校園推廣策略，爭取加入團隊。他不只提出想法，還親自拉了一群學生實測不同的 onboarding 流程與功能 A/B 測試，最後正式接手台灣與韓國的在地化與增長工作。短短一年內，他帶隊把月活用戶從 5 萬推升到 120 萬，並把這套成長模式複製到韓國，實際操演一次跨國市場的 go-to-market 策略。

從大學就開始創業，也讓他洞悉自己對於企業正規軍的策略理解不足。當他看到 LINE 要收購 Choco TV，覺得這是一個難得的機會，可以親身觀察國際公司是怎麼做併購與整合的。剛好 Choco TV 的創辦人 Davidd 也是Colorgy 的天使投資人，兩人相談合拍下，James 就加入整合團隊。參與了整個過程，從資料與技術對接、流程重新設計，到如何讓兩邊團隊建立共識、對齊預期。這段經歷讓他對大企業如何運作、如何協調文化落差有了第一手的理解，這些經驗也成為他日後共同創辦 MetaCRM、處理企業銷售與 B2B 合作的重要底氣。

James 能從零打造產品，也能在面對大型企業與通路夥伴時穩健談判，並且擁有實戰經驗：從一個大學生寫 App 的工程師、到主導國際產品落地、再到參與企業併購與跨部門整合，最後又成立 B2B SaaS 軟體新創。這些經驗讓 JTCG 能夠同時運作 PLG（product-led growth）與 SLG（sales-led growth）這兩種成長引擎，使其 AI 產品能快速打入市場並有效擴張。

解痛點、建資料飛輪，打造 AI 客服的持續競爭優勢

James 在經營 Zendesk 代理商期間，完整接觸企業客服流程，從大量客服工單與報表看見了兩個關鍵問題。第一，傳統規則式聊天機器人resolution rate (解決率，指顧客問題能被解決) 通常不到一成，而客服人力卻有七成時間進行重複「查詢系統資料、確認資料、回覆客戶」的工作；其二，多數國際大型客服 AI 工具雖功能強大，但價格昂貴且不夠接地氣，對本地語言與系統整合也缺乏彈性。這些觀察讓他意識到一個結構性機會：當地的 SI 可以造自己的 AI agent，取代國際 SaaS 工具，取代昂貴的國際 SaaS 預算，並提供更貼近需求的在地服務。

這些第一線經驗促成了 Raccoon AI 的誕生。機器人不僅能對話，更能直接喚用「查訂單、取消、退款」等後端 API，把客戶真正需要的動作一次完成。舉例來說，顧客詢問『我的訂單出貨了嗎？』時，Raccoon AI 不只回覆流程，而是會即時查詢電商後台的訂單管理系統，調出該訂單狀態，確認是否已出貨並提供物流單號，若尚未出貨還能進一步協助取消或改期。這是與傳統規則式聊天機器人最大不同，透過串接公司後台資料庫讓機器人有更多 context ，進而提高工單 resolution rate。

Raccoon AI 自推出以來在實際專案中展現出明顯差異化優勢。以台灣電商與旅遊業為例，系統可自動解決六成以上的客服工單，而同類 LLM chatbot 的平均 resolution rate 通常不及三成。這部分優勢來自於 Raccoon 長期累積的在地產業語料，例如電商常見的「出貨查詢」「修改地址」「訂單取消」等對話範式，甚至包含繁體中文語境下的客服語氣、商品分類邏輯與品牌風格調性，這些內容被整理成超過一百萬筆的向量知識庫，能讓 AI 在第一天上線就接近人類客服的回答品質。加上即將推出的 low code 客服流程編輯器（V2 版本，目前進入 Beta 測試階段），新客戶可在兩天內完成概念驗證，並預期在四週內完成正式上線。

每解決一張客服工單，系統會將意圖、回答與品牌語調回寫至該客戶專屬的向量知識庫，進一步優化客服機器人的涵蓋能力與回應品質。這種「資料飛輪」機制讓每家商戶的 AI 隨使用逐步成長，搭配 Raccoon 極短的導入時間，能快速開始累積語料、加速學習與優化，有效鞏固其 first mover advantage。

另一方面，James 已與 Zendesk、91APP、Omnichat、與 Shopline 等夥伴建立合作關係，也為 JTCG 建立起穩定且具規模化潛力的銷售管道。這些通路目前已貢獻 20–30% 的潛在客戶來源，讓團隊得以專注在產品與語料的持續打磨，不必從零開始冷啟動。

透過資料飛輪、快速導入與通路槓桿的三方協作，JTCG 已建立一套具有持續複利效應的擴張模式，清楚展現其在效率與速度上的結構性領先。

把握時機，搶占生成式 AI 客服的成長浪潮

相較於歐美大型客服 SaaS，JTCG 的優勢在於更「接地氣」。國際客服平台在亞洲的本地化落地仍有落差，例如多數歐美主流客服工具並不會優先支援 LINE 或台灣主流電商平台，對本地企業而言導入門檻相對較高。因此在客服與售後支援領域，JTCG 面對的是一個成熟產品未充分覆蓋的市場，正是一個可以快速切入、卡位成為關鍵供應商的機會。

我們在 2025 年第一季看到，JTCG 交出了年成長超過 100%的成績。基於團隊過往累積的產品能力與渠道優勢，以及生成式 AI 正全面改寫客服產業的市場時機，我們選擇領投 JTCG 的 Pre‑A 輪，加速 Self-Onboarding 產品開發與銷售團隊的建置，並成為亞洲 AI 客服的領先者。歡迎任何有意導入 AI 客服、自動化售後流程的品牌與 JTCG 或 AppWorks 聯繫，攜手打造新一代的客服體驗。

From College App Developer to a Dual‑Engine Entrepreneur

James Chou, Founder of JTCG, an AppWorks Accelerator Batch #13 alumni, stands out as one of the rare entrepreneurs who can seamlessly switch between product design, growth hacking, and enterprise sales.

In college, James built Colorgy, an app to help students manage class schedules and connect with peers. It started with him manually inputting his own timetable, coding at night, and testing during the day with classmates. Feedback led him to realize students cared more about the social side than schedules. So he pivoted, adding features to help them find classmates and form study groups. As usage patterns varied across departments, he developed a flexible backend to support different needs. At its peak, Colorgy reached over 2 million registered users and became a campus favorite among college students. The entire journey gave James deep, hands-on experience in product development and user research.

When French start‑up Zenly began to rise with real‑time friend‑location sharing, James noticed the perfect fit for Taiwan’s campus culture. Drawing on his Colorgy experience, he cold‑emailed Zenly’s founders, outlining growth playbooks and offline campus seeding tactics. He even assembled a student squad to A/B‑test onboarding flows. In just one year he led Zenly’s Taiwan & Korea localization and growth, pushing monthly actives from 50 K to 1.2 M and later duplicating the playbook in Korea—his first taste of cross‑border go‑to‑market.

When LINE TV moved to acquire Choco TV, James saw a chance to observe a global‑scale integration from the inside. Choco’s founder Davidd Liu—an early Colorgy angel—invited him onto the post‑merger task force. James managed data & tech hand‑offs, redesigned workflows, and brokered cultural alignment between the two teams. The experience gave him first‑hand insight into how large enterprises negotiate, integrate, and manage product roadmaps—knowledge he later applied to enterprise sales and B2B partnerships.

James knows how to take an idea from zero to one, and he’s just as comfortable pitching enterprise clients as he is building products from scratch. His journey covers everything from building student-focused apps and scaling international consumer platforms to navigating corporate M&A and launching a B2B SaaS company. That blend of experience is exactly what enables JTCG to run both product-led growth (PLG) and sales-led growth (SLG) side by side—helping their AI tools hit the market fast and scale even faster.

Solving Real Pain & Building a Data Flywheel—The Raccoon AI Edge

During his time as a Zendesk reseller, James reviewed thousands of support tickets and spotted two systemic gaps:

1. Rule-based chatbots had a resolution rate (the percentage of customer issues that are ultimately resolved) of less than 10%, yet 70 % of agent time was spent on repetitive look‑up → confirm → reply loops.
2. Global CX‑AI suites were pricey, inflexible, and poorly localized for Mandarin and Taiwanese integrations.

He realized that local system integrators (SIs) could build their own AI agents, under‑cutting global SaaS while serving local needs better. Hence Raccoon AI was born.

Raccoon AI not only chats but also acts: calling backend APIs such as Look Up Order, Cancel, Refund in real time. Example: When a shopper asks “Has my order shipped?”, the bot fetches status, delivers the tracking code, or even cancels/reschedules if needed—compressing what used to take 3–5 human interactions into a single exchange.

Since its launch, Raccoon AI has demonstrated a clear competitive edge in real-world projects. In Taiwan’s e-commerce and travel sectors, the system can autonomously resolve over 70% of customer service tickets, whereas the average resolution rate for comparable products typically falls below 30%. This advantage stems from Raccoon’s long-term accumulation of localized industry data, including common dialogue patterns such as “shipment inquiries,” “address changes,” and “order cancellations”—as well as nuanced Taiwanese customer service phrasing, product categorization logic, and brand tone preferences. These elements are compiled into a vector knowledge base of over one million entries, enabling the AI to deliver near-human response quality from day one. Additionally, with the upcoming release of the low-code customer service flow editor (V2 version, currently in Beta), new clients can complete proof-of-concept within two days and are expected to fully go live within four weeks.

Every solved ticket writes back intent, answer, and brand tone to that merchant’s private vector store, widening coverage and improving quality. Fast onboarding means data starts compounding early, reinforcing JTCG’s first‑mover advantage. Long-nurtured partnerships with Zendesk, Omnichat, Shopline, 91APP, and others now contribute 20–30% of inbound leads. Their ecosystem connections help the team avoid cold starts and stay focused on improving product and data quality. 

The trio of data flywheel, rapid deployment and channel leverage forms a structural lead in both efficiency and speed.

Seizing the Moment: Riding the Generative AI Customer Support Wave

Taiwan’s total addressable market includes approximately 71,000 support seats in the e-commerce, travel, and logistics sectors. Even automating just one-third of those would redirect an estimated NT$4.5–6 billion (US$140–190 million) annually toward AI customer service SaaS. Global CX giants still lack proper integration with platforms like LINE or local commerce infrastructure, creating friction for Taiwanese firms. Meanwhile, local MarTech players such as Super 8 and Appier primarily focus on marketing funnels rather than post-sale support, leaving this sector under-served and primed for rapid adoption.

In the first quarter of 2025, JTCG recorded over 100% year-over-year growth and a gross margin exceeding 60%. With a proven track record in product development, data moats, and robust distribution channels, we led the Pre-A round to accelerate self-onboarding product features and secure JTCG’s leadership position in Asia’s first wave of AI-driven customer support.

We welcome any brands interested in implementing AI customer service and automating after-sales processes to contact JTCG or AppWorks. Let’s work together to create a new generation of customer service experiences.

Some founders are relentless. Mike Hahn is one of them.

Over the past two decades, Mike has built five startups, completed multiple exits, and held intrapreneurial roles—all while demonstrating a rare tenacity to keep pushing, not for headlines or capital, but because he genuinely loves solving hard problems. With Vaudit, his sixth venture, Mike is applying everything he’s learned to build a more transparent and accountable digital advertising ecosystem.

The Hidden Leak in Ad Spend

Today, advertisers are spending more per click than ever. Yet a surprising portion of that spend never reaches real, verified users. During early customer discovery, Mike and his team uncovered that a large share of digital ad budgets was leaking into non-compliant, invalid traffic before it ever hit the final invoice. The lack of real-time auditing tools made this invisible to most.

Vaudit was born to address this head-on. By introducing a pre-bill audit layer, it helps advertisers reclaim wasted spend and restore integrity to the funnel. In a market where trust is assumed but rarely verified, this solution hits a nerve. In one case, an early customer discovered double-digit discrepancies, something they had never seen before.

Humble Beginnings, Global Impact 

Mike’s story begins in a trailer park in Indiana, the son of very young parents in a low-income household. Alongside his two brothers, he relied on athletic scholarships just to afford college. But when a serious injury ended his American football career, Mike was forced to pivot, redirecting all his energy toward entrepreneurship.

That shift led to a surprise introduction to tech, thanks to a referral from his brother. He landed an entry-level role at Oracle, where he quickly made his mark, designing a new go-to-market model that would scale into a global revenue engine. In just three years, he rose to lead a team of over a thousand people.

He left to solve what he saw as the next big inefficiency, client onboarding, by building a streamlined implementation platform. Oracle later brought that solution back in-house. At a legacy marketing firm, he led the digital transformation, developing new products and scaling a national franchise business. Later, he founded a data platform acquired by a global marketplace. 

After years of building in the U.S., Mike began exploring opportunities abroad. Through a mentorship role with PwC advising startup founders in Shanghai, he was exposed to the fast-growing innovation landscape across Asia. It was during this time that he recognized the immense, untapped potential of Southeast Asia and identified Vietnam and Thailand as rising hubs of technological excellence, an ideal launchpad for his next venture.

He went on to build a GenAI-native intent engine, successfully leading the company through product-market fit and into its growth phase. But Mike wasn’t done. He had identified yet another blind spot—a hidden inefficiency quietly draining ad budgets at scale. That insight pulled him back into the arena, ready to take on his next big challenge.

Through each chapter, a pattern emerges: Mike spots overlooked friction and builds systems to remove it. Not once, but repeatedly.

Rebuilding Trust in Digital Advertising

Digital advertising is a $600B+ industry running on assumed trust and partial verification. Small and mid-sized businesses, especially, lack tools to confirm what they’re paying for. As regulatory pressure grows and platform opacity deepens, the need for audit-grade visibility has never been clearer.

Mike has proven he can find product–market fit, scale meaningful solutions, and navigate uncertainty with clarity. With Vaudit, he’s channeling that experience into one of the most overlooked pain points in modern marketing. His vision is matched by a magnetic leadership style—so compelling that it drew AppWorks’ most charismatic principal, Jun, to join him full-time in building the company. It’s not often you see a VC trade the investor’s seat for a founder’s desk, and even less common to find a founder who can inspire that move. Since AppWorks’ pre-product investment in 2023, Jun has been side by side with Mike, helping bring Vaudit to life. We feel fortunate to back a founder whose blend of execution, insight, and leadership is exactly what it takes to build this missing layer of accountability.

We’re proud to support Mike in this next chapter.