Antler 創辦人 Magnus:亞洲將主導 AI 與加密創新十年
Asia's AI & Crypto Decade
- Magnus Grimeland Antler 創辦人暨執行長
影片
影片與逐字稿來源:TOKEN2049
場次摘要
Antler 創辦人暨執行長 Magnus Grimeland 分享為何亞洲正成為 AI 與區塊鏈創新的全球重心。從 AI 運算成本、晶片製造到能源優勢,亞洲具備推動下一波 AI 與 agent-to-agent 經濟的關鍵條件。Antler 也積極投資全球早期新創,歡迎有志於 AI、區塊鏈與金融科技的創業者加入這波浪潮。
值得記下的 5 件事
- 早期創業支援可串接共同創辦人、商業模式、首筆資金與客戶,降低創業初期的資源缺口。
- 區塊鏈的獎勵機制可用於航班追蹤網路,讓提供設備與資料的人取得報酬。
- 較低成本的開放權重模型,可能讓亞洲團隊發展出更多大規模 AI 應用。
- 企業採用開放模型的考量除了成本,也包括避免把專有資料交給大型模型供應商。
- 晶片成本與模型效率改善之後,電力可能成為 AI 運算成本的重要限制。
逐字稿
Antler 全球早期創投布局
大家好,很高興見到大家,非常榮幸能在這裡。一直以來我都是 TOKEN2049 的忠實粉絲。今天我想跟大家分享一下 Antler 在做什麼。我們現在是全球最活躍的 AI 投資者,同時也會談談我們怎麼看亞洲,以及 AI 和區塊鏈領域的最新發展。Hey guys, great to see everyone. Wonderful to be here. Always been a big fan of Token 2049. I'm going to talk a little bit about what we do at Antler. We are now the most active AI investor in the world. And also how we look at what's happening in Asia and what's happening in the AI slash blockchain space.
Antler 現在是全球最活躍的早期投資機構之一。我們在全球大約 30 個城市設有辦公室,基金遍佈世界各地。目前我們的投資組合總價值約 260 億美元,累積投資了 2,000 家新創公司。每個據點都會從創業第一天就陪伴創辦人,並一路支持到 Series C 輪。So Antler, we are now the most active early stage investor in the world. We have offices in about 30 locations globally. We have funds all over the globe. We have about $26 billion portfolio value. We have 2,000 portfolio companies. In each of these locations, we work with founders and back them from day zero and continue to back them up until Series C.
我們能做的投資範圍很廣,從幾十萬美元的小額投資到一億美元的大額投資都沒問題。Antler 擁有 240 位全職投資專業人士。如果你正在創業,或是有興趣創業、正在尋找 pre-seed 輪資金,都可以找我們。我們幾乎可以在任何階段參與投資。So you can basically do everything from a few hundred K checks to $100 million checks all over the world. We have 240 full-time investment professionals. So if you're building something, or if you're excited about building something, or you are raising capital pre-seed round to us, we can invest at almost any stage.
全球 AI 投資與創業生態
我們目前是全球最活躍的 AI 投資者,這是來自 Dealroom 的數據。我們不只投資美國和歐美的優秀公司,也積極布局亞洲、非洲和南美洲。每年有超過 16 萬名創辦人申請加入 Antler,我們實際會和其中約 3,000 人合作。We are the most active AI investor in the world right now. This is data from Dealroom. We back not only great companies out of the U.S. and the West, but also all over Asia, Africa, and South America. More than 160,000 founders apply to Antler every year, and we work with about 3,000 of them.
如果你需要創業空間,我們在剛剛提到的所有城市都有辦公室。我們會協助你對接頂尖創投、尋找共同創辦人、調整商業模式、提供第一筆資金和客戶,任何你需要的資源都能協助。每年我們會投資 400 到 500 家新公司,並且一路支持到公司退出。So if you need a space to work on building your business, we have offices in all these locations I showed earlier. And we'll work with you and provide access to great VCs, co-founders, iterating the business model, first capital, and customers, anything you need help with. And then we back about 400 to 500 new companies every year. I continue to back companies all the way to their exit, basically.
我們的目標是長期投入,讓你的公司有機會獲得數億美元的資金。這裡有一些我們投資過的公司,像 Lovable 是史上成長最快的公司之一,10 個月內營收就破億美金。Airalo 也是大家可能用過的產品,是全球最大的 eSIM 市場,從新加坡辦公室孵化出來。我們從第一輪到第六輪都一路支持他們。The goal is to put hundreds of millions of dollars to work in your company over time. These are some of our portfolio companies. You might know a few of them. Lovable, fastest growing company in history. Actually, I think they grew faster to $100 million than any other business ever. Airalo, few of you are probably using that product. It's the biggest eSIM marketplace in the world. That was built out of Singapore in our offices. We were the first investor, second investor, third investor, fourth investor, fifth investor, sixth investor. We backed them on their entire journey.
AI 與區塊鏈創新案例
我們也投資了許多來自不同地區的 AI 新創,像是韓國的 Written、美國的 Micro1 和 Fluidstack,以及全球最強 AI 影片生成模型 Pixverse,總部就在這裡,在中國有大規模營運,客戶遍佈全球。還有東南亞的金融科技新創 Salmon,以及全球頂尖的資安 AI 防禦公司 Dream。這只是我們 2,000 家投資組合中的一小部分。Written, fast-growing AI company out of Korea. Micro1, one of the fast-growing AI companies out of the U.S. Same with the Fluidstack. Pixverse, by far the best AI video model in the world, based out of here. Big operations in China, customers all over the world. Salmon, fast-growing fintech here in Southeast Asia right now, actually. Very exciting business and Dream, probably the best cybersecurity AI defense in the world. These are just some of our portfolio companies. Out of 2000.
在區塊鏈和 AI 結合領域,我們最近也投資了三家非常有趣的公司。第一家叫 HiFi,從第一輪就參與,最近也參加了他們的大型融資。他們專注於全球法幣到加密貨幣的出入金基礎設施。第二家 Cream,則是用 AI 結合區塊鏈,解決金融產業的特定問題。第三家 Wingbits 更有趣,他們把全球航班追蹤雷達搬到區塊鏈上,任何人都可以買雷達加入網路,追蹤飛行物體並獲得報酬,這正是區塊鏈最適合的應用。In the space of kind of blockchain and AI, we have three very exciting companies we backed recently. One of them is called HiFi. We were there from the first round. We just took part in the latest massive round there. In the space of fiat to crypto onboarding and offboarding all over the world, basically building rails. Cream is, I think, a tremendous company who is using AI on the blockchain to solve very specific problems in the financial industry. Wingbits is super interesting, actually, because it's kind of a real-world blockchain case that works incredibly well.
大家應該都聽過 Flight Radar,就是追蹤全球航班的服務,也能追蹤無人機等各種飛行物,但這套系統其實很舊,是全球志工維護的。Wingbits 則把這整套搬到區塊鏈上,任何人都能買雷達、加入鏈上網路、開始追蹤飛行物並獲得報酬,現在他們還用 AI 偵測更小型的物體,成長速度非常驚人,幾乎一年內就追上了 Flight Radar 30 多年的規模。Probably most of you have heard about Flight Radar. Who's heard about Flight Radar? It's basically these radars to track all the flights across the globe. They also track drones. They track all types of flying objects, basically. But it's very outdated technology, run by this volunteer network all over the globe. And what Wingbits did is they put this whole thing on the blockchain. So basically, you can buy your own radar, join the chain, start tracking flying objects, and get paid for it, which is something blockchain is perfect for. And obviously now they're using AI to discover smaller and smaller objects all over the globe and growing like crazy. They basically caught up with Flightradar. I think it's been in business for like 30, 35 years. They caught up with them in about a year.
AI 時代的爆炸性成長
這些都是非常令人興奮的案例。我們都知道,AI 產業正經歷史上最爆炸性的收入成長,這不只在我們投資組合裡看到,公開數據也證實了。公司成長速度比以往任何時候都快。舉例來說,航空業從第一位乘客到 5,000 萬人搭飛機,花了 68 年;但 ChatGPT 只花 5 天就有 5,000 萬用戶,Pokémon Go 也只花 14 天。只要產品有市場適配,現在的成長速度是前所未有的。So, yeah, three very exciting use cases. I mean, we all know this, right? LMCCing the most massive growth in revenue that the world has ever seen. We see that in our portfolio. We see that in publicly reported numbers. And companies are scaling faster than ever, right? It took the airline industry. If you think about the first person who went on an airline, how long time do you think it took until 50 million people had been flying on an airplane? Anyone want to guess? 50 million people being airline customers. 50 years, very close. 68 years. So, the airline industry spent 68 years getting to 50 million customers. It took ChatGPT five days. It took Pokémon Go 14 days, right? If you have product-market fit in today's world, you can scale faster than ever.
我們投資的 Lovable 只花 10 個月就從零做到 1 億美金營收。現在正是創業最令人興奮的時代,AI 平台的能力每幾個月就翻倍。只要看看最新的模型,和幾個月前的版本相比,智慧程度已經大幅提升,而且每個 token 的運算成本越來越低,這代表你們現在腦中想到的各種應用,未來都會越來越可行。站在這波浪潮前端,機會真的非常多。It took our portfolio Lovable 10 months to get from zero to 100 million in revenue. And yeah, it's just the most exciting times to be alive. With the big AI platform shift. Also, the capability is just doubling every few months, right? If you look at the latest models that are coming out and you compare to just what we had available to us a few months from now, the intelligence level is way above what it was just a few months ago. And the cost per token gets lower and lower, which means that a lot of the things you guys might be thinking about or the use cases you're working on in your companies, they're going to be more and more possible going forward. And being on the front end of this is just creating a tremendous amount of opportunities.
亞洲成為 AI 創新重心
亞洲在這方面特別擅長的一個領域,就是每個代幣的成本。沒錯,你可以擁有幾乎相同水準的智慧,但成本卻大幅降低,這就創造了許多在其他地區看不到的大規模進步和擴展的機會。因此,我認為下一波的 AI 將會在亞洲誕生,對吧?我們正處於一個市場,許多開放權重模型都在這裡被打造,而且這些領先的模型成本遠低於美國的前沿實驗室。And one of the areas where Asia is particularly good when it comes to this is the cost per token. Right, you have almost the same level of intelligence, but the cost is drastically lower, which creates a lot of opportunities for mass progress and scale that you don't see in other parts of the world. The next wave of AI will therefore, I think, be built in Asia, right? We are in the marketplace where a lot of the open weight models are being built, and by far the leading ones, which have much lower costs than the frontier labs in the US.
我們也正處於所有晶片製造的地方,像是全世界最先進的晶片,以及最便宜的晶片和記憶體。而且,大多數用於支援 AI 系統的零組件,都是在亞洲這裡生產的。We're also in the place where all of the chips are being made, like all the most advanced chips in the world, and also the cheapest chips and the memory. And most of the things that go into the systems that are enabling AI are built here in Asia.
如果你有機會到亞洲各地的機器人或自動化工廠參觀,會發現這裡的技術真的領先全球。亞洲有機器人奧運、無人機競賽,各種自動化應用都走在世界前端。真正的創新重心正在往亞洲移動,不再只是複製過去歐美的模式。你現在就在正確的時間、正確的地點。And if you travel around the region and you went to some of the robotics factories and you compare with what you see elsewhere, this is way ahead, right? You have robot Olympics, you have drone competitions. Yeah, all of these things that robots are doing, which are just way ahead what you see in the rest of the world. So I think the center of gravity, when it comes to real innovation and not copying what happened in the past, is moving towards Asia. So you're in the right place at the right time.
開源模型與去中心化 AI
從 token 使用量來看,這是全球頂尖 AI 模型每週的 token 使用數據。綠色是美國模型,接著是中國的開源模型,最後是全球其他地區。這個趨勢一兩年前沒人預料到,但現在真的發生了。越來越多世界級公司選擇在開源模型上開發,因為他們不想把專有數據交給大型 frontier labs,也不相信這些資料真的會被隔離處理。What's happening with token usage, right? So these are the weekly tokens being used on the top models across the globe. These are the U.S. models in green. And then you see the Chinese models, which are all of them open source models. And then you see the rest of the world finally coming. This is a development that I think a lot of people didn't expect a year or two ago, but it's really happening right now. And you see big parts of the world, but also some of the world's best companies building on top of the open source models because they don't want their proprietary data to go to the big frontier labs. And they don't trust that even if they say it will be split out, that it actually is.
這不只是提供更便宜的 token,更是讓大家能去中心化地存取全球最強的智慧,這和 TOKEN2049 的精神非常契合。你可以用你想要的方式接觸這些模型,不需要被大公司決定怎麼用。So this is a huge opportunity not only to provide cheaper tokens, but also to provide decentralized access to the highest intelligence in the world, which I think fits very well in with the mantra and the modus of Token2049, that you can access these models. And build with them in the way that you want to do, and not dependent on some major corporation deciding how to use your model.
AI 能源競賽與亞洲優勢
推動這一切的關鍵就是電力。隨著晶片成本下降、模型效率提升,AI 智慧的成本會越來越接近電力成本。就像汽油或電力驅動的產業一樣,隨著零組件變便宜,最大成本會變成電費。現在中國和亞洲在電力競賽中領先,每年中國新增的電力,相當於美國過去十年新增的總量,這不只是過去的現象,未來還會持續下去。And obviously, what's driving all of this is power. And this is insane, right? If you think about as the cost of chips goes down, as models become more efficient, the cost of intelligence will get more and more equal to the cost of power, right? In the same way as over time, anything that uses gasoline or anything that uses electricity, as, you know, components and, you know, things gets cheaper. The biggest cost of running these things will be electricity, and it will be the same with AI. And clearly now, China and Asia is leading the power race, right? The amount of power being put online in China every year equates to what came online in the U.S. the last decade. Right, so this is not only something that has happened in the past, but it's continued to happening into the future.
這讓亞洲有機會成為未來十年 AI 領導企業的誕生地。相對來說,美國和歐洲要追趕很困難,因為沒有人想在自家後院蓋發電廠,即使是再生能源,像風力、太陽能也會遇到空間和美觀問題,法規障礙重重。但亞洲正在瘋狂建設電力基礎,這將成為未來企業的關鍵優勢。That creates huge opportunities for people to build here in the region, the leading AI companies of the decade. And it also means, obviously, that the U.S. and Europe has some real work to do. And it's difficult because building power station, nobody wants a power station in the backyard. Even if it's renewable energy, windmills and solar panels, they take up space and they might not be that amazing to look at. So it's hard to work through all that regulatory red tape, while in Asia they're building power like crazy, which will enable. The companies of the future.
亞洲金融科技與區塊鏈跳躍式發展
我對於金融科技區塊鏈領域正在發生的事情感到非常興奮,對吧?如果你思考一下這在亞洲到目前為止是如何發展的,其實我們跳過了一整個世代的金融軟體和金融科技,對吧?我們是直接從現金,跳到人們用手機支付,對吧,我們幾乎完全跳過了信用卡這一環。我之前在亞洲創辦過一家名為Solora的電子商務公司,我們真的有讓使用者在我們網站上用現金付款,或者用錢包支付。但當時在印尼,信用卡的滲透率只有3%。我認為現在同樣的情況正在發生。And I'm very excited about what's happening in the fintech blockchain space, right? If you think about how this has developed in Asia up until now, we actually skipped a whole generation of financial software and financial technology, right? We went straight from cash. To people paying with the phone, right, we skip cards more or less altogether. I built an e-commerce company before in Asia called Solora, and we literally had people pay with cash on our website, or they paid by wallets. But in Indonesia at the time, there was 3% credit card penetration. And I think the same is happening.
我認為我們現在正進入一個經濟時代,亞洲越來越多的經濟活動將在區塊鏈上運行,特別是在代理人對代理人的經濟模式中。我認為這很可能是當今金融產業中最大的機會。I think we'll get to an economy now where more and more of the economy in Asia will run on the blockchain, and particularly within the agent-to-agent economy. And I think this is probably the biggest opportunity in the financial industry today.
我認為這是全球最大的機會之一,未來會看到越來越多 AI agent 彼此直接交易,不再經過人類或銀行。我們投資組合裡已經有不少公司在解決這個問題,也知道現場有很多人正在努力。這是亞洲有機會贏下來、打造下一代領導企業的領域,規模可能是數兆美元。如果你對這個主題有興趣,歡迎來找我們聊聊。I think this is probably the biggest opportunity in the world today, right? That you'll see agents transacting with each other and not going through human beings or banks at all. And we already see that with a number of our portfolio companies. That are struggling with this particular problem. And I know there's a lot of companies out there who's working on it. Probably there are people here in the audience that are working on it. But this is, I think, a place where Asia can really win and build that next generation leading companies. And this is probably a multi-trillion dollar opportunity that we are working on. And if you guys in the room are working on it or interested in being part of that. Reach out to us. If we can be helpful with anything, let me know.
Antler 助力全球創業者
我們從創業最早期就支持優秀創辦人,一路到 Series C。甚至有些公司還沒想好要做什麼,只要創辦人夠強,我們也會投資。如果你現在正想創業,非常歡迎早期就跟我們合作,我們在全球各地都有辦公空間。如果想聯絡我,這裡有我的聯絡方式。We back great founders from the really early days up until Series C. We also back companies that don't have an idea yet, just based off the quality of the founder, right? So if you're thinking about building something now. We'd just love to work with you in the early days, and we have space in our offices all over the world. So if you want to get in contact, just reach out to me. These are my details.
如果你對 agent-to-agent 經濟特別有興趣,可以掃這個 QR code 報名,我們會主動聯絡你。我真的認為這是全球最大的機會之一,也是大家可以一起努力的領域。謝謝大家今天來聽,也期待未來能和在座許多人一起合作。現在正是創業、打造偉大公司的最佳時機,謝謝!And if you want to sign up and you're interested in particularly what I talked about earlier, which is the agent-to-agent economy, just scan this QR code and sign up, and we'll get in contact with you. I really think this is one of the biggest opportunities in the world. And it's a place where we all can work together. So yeah, thank you all for coming, and thanks for listening to me, and look forward to hopefully working with a lot of you going forward. Most exciting time ever to build great companies. Thank you.