Aethir 共同創辦人 Mark Rydon 談去中心化計算的興起與中美地緣政治的挑戰

Aethir 的去中心化圖形處理單元 (GPU) 網路最初是作為雲端遊戲解決方案，但聯合創始人 Mark Rydon 表示，該公司現在專注於人工智慧

Aethir, a decentralized Graphics Processing Units (GPUs) network, initially focused on cloud gaming. However, co-founder Mark Rydon reveals that the company's attention has shifted towards artificial intelligence (AI). This pivot is driven by soaring demand for compute power in AI, which is increasing the value of GPUs and highlighting a clash between U.S.-China geopolitics.

Aethir 是一個去中心化圖形處理單元 (GPU) 網絡，最初專注於雲端遊戲。然而，聯合創始人馬克雷登透露，公司的注意力已經轉向人工智慧（AI）。這項轉變是由人工智慧運算能力需求飆升推動的，這增加了 GPU 的價值，並凸顯了中美地緣政治之間的衝突。

Founded in 2022, Aethir is a prominent protocol in the Decentralized Physical Infrastructure (DePIN) space, thanks to its real-world utility. DePIN protocols are networks where people collaborate to offer a service in exchange for a token.

Aethir 成立於 2022 年，因其在現實世界中的實用性而成為去中心化實體基礎設施 (DePIN) 領域的傑出協議。 DePIN 協議是人們協作提供服務以換取代幣的網路。

Over the past year, the DePIN space has experienced rapid growth, with CoinGecko reporting a $33 billion market cap for DePIN coins. A diverse range of DePINs has emerged, covering nearly every vertical imaginable, from flight tracking data to mapping.

在過去的一年裡，DePIN 領域經歷了快速成長，CoinGecko 報告 DePIN 代幣的市值為 330 億美元。各種各樣的 DePIN 已經出現，幾乎涵蓋了所有可以想像到的垂直領域，從航班追蹤數據到地圖繪製。

Aethir is a key project in decentralized compute, demonstrating the capability of distributed computing to power AI alongside large, centralized data centers.

Aethir 是去中心化運算領域的關鍵項目，展示了分散式運算與大型集中式資料中心一起為人工智慧提供動力的能力。

GPUs were initially used exclusively for gaming, with companies like Nvidia and AMD competing to produce more powerful GPU chips. The goal was to enhance the fidelity and resolution of virtual worlds for gamers.

GPU 最初專門用於遊戲，Nvidia 和 AMD 等公司競相生產更強大的 GPU 晶片。目標是提高遊戲玩家虛擬世界的保真度和解析度。

However, around 2006, researchers at Nvidia discovered an additional strength of GPUs: parallel computing. This capability, which enhances graphics, was also found to be effective in crunching massive amounts of data and could accelerate traditional computing workloads.

然而，2006 年左右，Nvidia 的研究人員發現了 GPU 的另一個優勢：平行運算。這種增強圖形功能的功能也被發現可以有效處理大量數據，並且可以加速傳統運算工作負載。

This series is brought to you by Consensus Hong Kong. Come and experience the most influential event in Web3 and Digital Assets, Feb.18-20. Register today and save 15% with the code CoinDesk15.

本系列由Consensus Hong Kong 為您帶來。 2 月 18 日至 20 日，快來體驗 Web3 和數位資產領域最具影響力的活動。立即註冊並使用代碼 CoinDesk15 可節省 15%。

But this was 2006, and verticals like AI were not yet on anyone's radar. As a result, Nvidia continued to trade for less than $1.

但那是 2006 年，像人工智慧這樣的垂直產業還沒有引起任何人的注意。結果，Nvidia 的交易價格繼續低於 1 美元。

Fast forward to today, and Nvidia is competing with Apple for the title of the world's most valuable company. In recent years, GPUs have become the driving force behind the AI revolution, powering everything from generative AI models to advanced machine learning algorithms.

快進到今天，英偉達正在與蘋果爭奪全球最有價值公司的頭銜。近年來，GPU 已成為人工智慧革命背後的驅動力，為從生成式人工智慧模型到高階機器學習演算法的一切提供動力。

Like the journey of Nvidia and GPUs, Aethir began by offering idle GPUs from data centers to bolster cloud gaming platforms. However, as co-founder Mark Rydon explained in an interview with CoinDesk, the company realized that the real money was in AI.

與 Nvidia 和 GPU 的旅程一樣，Aethir 也從提供資料中心的閒置 GPU 開始，以支援雲端遊戲平台。然而，正如聯合創始人 Mark Rydon 在接受 CoinDesk 採訪時解釋的那樣，該公司意識到真正的錢在於人工智慧。

"We started Aethir within a gaming context," said Rydon. "But we quickly realized that the enterprise-capable GPU cloud we were building had massive relevance to AI sectors. AI runtime customers will pay a lot more than gamers, and the demand for compute in AI is just incredible."

「我們在遊戲環境中創建了 Aethir，」Rydon 說。 “但我們很快意識到，我們正在構建的企業級 GPU 雲與 AI 領域有著巨大的相關性。AI 運行時客戶將比遊戲玩家支付更多的費用，而且 AI 計算的需求令人難以置信。”

TrendForce, a market research firm, estimates the value of the AI server industry to be $205 billion, which is particularly impressive considering that they estimate the value of servers, as a whole, to be just over $300 billion.

市場研究公司 TrendForce 估計 AI 伺服器產業的價值為 2,050 億美元，考慮到他們估計伺服器的整體價值剛剛超過 3,000 億美元，這一數字尤其令人印象深刻。

Aethir operates as a marketplace for GPU compute, offering idle GPUs to businesses that require on-demand capacity without the overhead of owning or maintaining their own hardware.

Aethir 作為 GPU 運算市場運營，為需要按需容量的企業提供閒置 GPU，而無需承擔擁有或維護自己硬體的開銷。

This is especially significant for researchers outside of prestigious institutions, as it provides them with access to computing power without the massive capital expense of purchasing their own infrastructure.

這對於著名機構以外的研究人員來說尤其重要，因為它為他們提供了運算能力，而無需花費大量資金購買自己的基礎設施。

“This is the most expensive compute that there’s ever been, and if you’re not efficient in managing that compute, it’s a huge capital drain,” explained Rydon during an interview on the sidelines of Devcon in Bangkok last November. "Our decentralized approach allows us to scale into regions…that need compute the most but can least afford it."

「這是有史以​​來最昂貴的計算，如果你不能有效地管理該計算，就會造成巨大的資本流失，」Rydon 在去年 11 月曼谷 Devcon 間隙接受采訪時解釋道。 “我們的去中心化方法使我們能夠擴展到需要計算最多但負擔能力最差的區​​域。”

Given Nvidia's leadership in the AI space, Chinese researchers are eager to obtain access to these ultra-powerful chips, which are generations ahead of what is available domestically.

鑑於英偉達在人工智慧領域的領先地位，中國研究人員渴望獲得這些超強大的晶片，這些晶片比國內現有的晶片領先幾代。

However, geopolitics have intervened. The U.S. Department of Commerce has prohibited the sale of Nvidia's most advanced chips, such as the H100, to China in an effort to keep Beijing one step behind.

然而，地緣政治已經介入。美國商務部禁止向中國出售英偉達最先進的晶片，例如H100，以讓北京落後一步。

According to a report by the Wall Street Journal in August 2024, Chinese firms were bypassing U.S. export controls by renting decentralized GPU computing power abroad through companies like Aethir.

根據《華爾街日報》2024 年 8 月的一篇報道，中國企業透過 Aethir 等公司在國外租用去中心化 GPU 運算能力，從而繞過美國的出口管制。

“We don’t accept high performance computing [requests] from places like China. You can’t if you have H100s," said Rydon. "They’re not meant to be in China. We geofence and don’t allow that compute to onboard.”

「我們不接受來自中國等地的高效能運算[請求]。如果你有 H100，那就不行了，」Rydon 說。「它們不適合在中國。我們設置了地理圍欄，不允許該計算進入。

Despite the push towards decentralization in Web3, Rydon emphasizes that Aethir is not fully permissionless.

儘管 Web3 正在推動去中心化，但 Rydon 強調 Aethir 並不是完全無需許可的。

"This is necessary for $100 million deals to happen,” said Rydon.

「這對於達成 1 億美元的交易是必要的，」Rydon 說。

A Web2 layer is required to track service agreements that outline the performance and reliability standards for enterprise clients and to conduct Know Your Customer (KYC) processes, which verify the identities of participants to ensure regulatory compliance.

需要 Web2 層來追蹤概述企業客戶的效能和可靠性標準的服務協議，並執行了解您的客戶 (KYC) 流程，該流程驗證參與者的身份以確保法規遵循。

"I don’

“我不’