Author: Jacob Zhao @ IOSG
Today, "the first stock of domestic storage," Changxin Storage, officially landed on the Growth Enterprise Market, exploding the venue with an astonishing rise of 500%. Although the storage sector is still disturbed by the recent wave of corrections, AI storage continues to be crazily revalued by capital in the current wave of technological narratives. Meanwhile, decentralized storage in the Web3 field is trapped in long-term silence and loss. Why is it that both are labeled as "storage," yet their market performances are worlds apart? The fundamental answer lies in the complete divergence of underlying value functions.
The revaluation of storage in the AI era is essentially a celebration of "hot data efficiency," serving the ultimate maximization of computing power utilization and commercial monetization; whereas decentralized storage upholds the value proposition of "cold data trustworthiness," advocating for data fairness, censorship resistance, and the long-term memory of human civilization. The former is an efficiency system for hot data, while the latter is a trustworthy system for cold data. The current capital market undoubtedly firmly stands on the "efficiency" side, but human civilization ultimately still requires an unalterable memory foundation. The long-term value of trustworthy cold storage has never vanished; it merely lies dormant on the dark side of cycles, waiting to be repriced by the times.
Why Storage Has Become the Focus of the AI Industry Chain Again
In the traditional IT era, storage was a "capacity business." CIOs focused on unit capacity costs, hard disk reliability, disaster recovery plans, archiving strategies, and device upgrade cycles lasting 3–5 years. Storage was seen as an accessory following server purchases.
This round of storage enthusiasm is not a traditional cycle recovery, but rather a repricing of data mobility driven by AI. In the era of large models, storage logic has transformed from "capacity first" to "efficiency above all," stubbornly pursuing extreme indicators such as GPU feeding rates, Checkpoint writing, and RAG ultra-low latency. This marks a leap in storage value from "the ultimate resting place of data" to "the high-speed passage of data into computation."
The evolution of resource bottlenecks in AI infrastructure is essentially a battle to supplement the "barrel effect." The true utilization rate of computing power is not a linear summation of individual assets but rather a stringent multiplicative effect: true computing utilization = GPU × HBM × DRAM × SSD × network × file system; any weak link will cause the overall computing utilization to collapse. In the AI era, storage has transformed for the first time from a "cost center" to an "efficiency engine." This is the fundamental logic behind the repricing of storage.
Panorama of AI Storage Architecture: From HBM Bandwidth Organs to Data Lake Foundations
AI storage is not merely an accumulation of single hardware but rather a tightly coupled, hierarchically scheduled complex system. In this system, industry value and capital focus are highly concentrated on HBM, enterprise-grade SSDs, SSD controllers, NVMe/CXL protocols, and high-performance storage systems. To clearly dissect its value flow, we will divide the AI storage architecture into four core levels from top to bottom:
Compute Proximity Memory Layer (Bandwidth Core): Dominated by HBM, supplemented by DRAM and CXL memory pooling technologies. This layer directly bonds with GPU/CPU packages or buses, aiming to break the "memory wall"; it is the first barrier determining whether computing power can be fully unleashed.
High-Speed Persistent Storage Layer (IO Hub): The core logic is Enterprise-grade SSD = NAND particles + SSD controller + NVMe/PCIe data pathway. This layer handles high-frequency Checkpoint writing, massive training set loading, and RAG hot data caching, representing the most distinct persistence storage increment in AI data centers.
Low-Cost Large Capacity Storage Layer (Capacity Foundation): Composed of HDDs, cold storage, and data lake archiving systems. Facing the exponentially expanding multi-modal raw data, historical logs, and compliance backups, this layer still provides an irreplaceable TCO (total cost of ownership) advantage.
AI Storage System and Data Software (Scheduling Brain): Includes high-performance parallel file systems, distributed object storage, vector databases, and RAG data governance layers. What AI truly consumes is not bare hardware but the data availability efficiently organized, indexed, and authorized by the software stack.
As an ecological extension, decentralized storage does not directly engage in the millisecond-level sprint of AI hot data; instead, it focuses on public data set preservation, AI training data provenance and long-term cold memory archiving, establishing its unique ecological niche as a "trustworthy cold layer."
HBM: The "Bandwidth Organ" Closest to Computing Power in the AI Storage Chain
High Bandwidth Memory is not traditional storage but high-bandwidth memory near the GPU. Its core mission is not to store data but to continuously "feed" data to computing power at extremely high bandwidth. HBM is the closest and most deterministic link in the AI storage chain, directly determining whether the GPU can be "fed well" and is currently the most critical supply chain bottleneck.
The core architecture of HBM is "3D DRAM stacking + 2.5D advanced packaging": through TSV vertical stacking and CoWoS heterogeneous integration, it greatly compresses the distance between storage and computing to achieve a leap in bandwidth. Its industrial barrier is not just DRAM design but also the system engineering of DRAM process, TSV, ultra-thin stacking, packaging, heat dissipation, testing, and customer certification. Any defect in yield in any link can lead to the scrapping of the entire HBM stack.
Currently, only SK Hynix, Samsung, and Micron among the three giants can stably mass-produce, establishing a triple moat of top-level DRAM processes, packaging capabilities, and NVIDIA/AMD customer certifications.
DRAM and CXL: System Memory Foundation and Memory Pooling Engine
HBM addresses extreme bandwidth at the GPU proximity, DRAM solidifies the server system memory foundation, while CXL attempts to break physical boundaries and reconstruct the organization of memory resources within data centers.
DRAM: Mainly carries CPU-side cache, data preprocessing, intermediate state temporary storage, and system operation, serving as the most fundamental system memory layer for servers. The global DRAM market is highly concentrated in the three giants: SK Hynix, Samsung, and Micron; Changxin Storage (CXMT) is a key variable in China's DRAM domestic replacement.
CXL (Compute Express Link): Is a new generation of cache coherence interconnect protocol aimed at data centers, designed to break the limitations of traditional DIMM slots, local memory capacity, and server memory resource islands, driving memory architecture towards expansion, pooling, and sharing. Currently, CXL is still in the early stages of transitioning from platform support to large-scale deployment, with high long-term architectural value; core companies include Astera Labs and Lanqi Technology.
Enterprise-grade SSD: The Data Hub Built from NAND, Controllers, and NVMe
Enterprise-grade SSD is the most core high-throughput, persistent increment in AI data centers, continuously "feeding" data to GPUs with extremely high throughput, extremely low latency, and stable QoS, spanning the full lifecycle from training data loading, Checkpoint writing, RAG retrieval, inference caching to log returning.
In the AI storage architecture, SSDs are not isolated hardware but part of a highly coupled system, distilled into the industry formula: Enterprise-grade SSD = NAND particles + SSD controllers + NVMe/PCIe data pathways. The three layers represent independent industrial chain links:
NAND Particles (Raw Material Layer): Determine storage density and unit cost while controllers manage performance release and lifespan. Representative companies: Samsung, SK Hynix (Solidigm), Micron, Kioxia, Western Digital, Yangtze Memory Technologies.
SSD Controller (Performance Empowerment Layer): Determines performance release, data error correction, QoS stability, and wear leveling. Representative companies: Phison (Qunlian), Silicon Motion (Hui Rong), Marvell, Maxio (Lianyun).
NVMe/PCIe (Data Pathway Layer): Determines the efficiency of data transmission from storage to computation. Combined with GPUDirect Storage technology, it reduces CPU memory bounce buffer and CPU involvement, significantly alleviating I/O bottlenecks. Representative companies: Broadcom, Marvell, Astera Labs.
HDD / Cold Storage / Archiving: The Low-Cost Foundation of AI Data Lakes
AI will not eliminate HDDs. As multi-modal large models demand video and image data and the exponential expansion of enterprise compliance logs and historical data sets occur, the demand for low-cost cold data storage is surging simultaneously. In the AI storage architecture, SSDs and HDDs layer and collaborate based on business value: SSDs handle hot data and high throughput while HDDs manage low cost and long-cycle preservation. Representative companies include Seagate, Western Digital, and Toshiba.
AI Storage Software Stack: The Scheduling Hub of Data Availability
What AI truly consumes is not bare disks but “data services” meticulously organized by software stacks. This architecture transforms underlying hardware into knowledge assets that AI can directly invoke, divided into four layers:
High-Performance Storage Systems (Supply System): With concurrent throughput and low latency at its core, it solves the "data starvation" problem of GPU clusters through parallel file systems, ensuring rapid flow of training and inference. Representative companies: VAST Data, WEKA, Pure Storage.
Object Storage (Raw Data Lake): Focused on the management of Object, Key, and Metadata, it carries massive amounts of unstructured data. It does not pursue extreme low latency but instead builds a capacity foundation with low costs and cloud-native characteristics. Representative company: AWS S3.
Vector Database (Semantic Index Layer): Responsible for storing, indexing, and retrieving the vectors generated by embedding models, enabling AI to accurately locate relevant content from vast knowledge. Representative companies: Pinecone, Milvus.
RAG Data Layer (Knowledge Invocation Layer): Beyond single retrieval, it encompasses data slicing, cleaning, permission control, and citation provenance, ensuring that enterprise data can be safely, accurately, and traceably invoked by large models. Representative company: Databricks.
From AI Hot Storage to Decentralized Cold Memory: Maximizing Efficiency vs. Maximizing Trustworthiness
AI storage is an extreme efficiency-driven system, with its value function focusing on maximizing computational output. HBM bandwidth determines whether the GPU can be fed well, SSD throughput determines the read and write efficiency of data sets and Checkpoints, while low latency impacts RAG and inference real-time experience. These indicators ultimately converge to GPU utilization and unit Token cost, directly determining the commercial profitability of AI applications. The ultimate goal of AI storage is not to preserve but to accelerate, with the served object being productivity.
Decentralized storage, however, has a completely different value function. It inquires whether data will still exist a decade later, whether it will be tampered with, and whether it can resist single-point censorship. Through cryptographic proofs and distributed networks, it builds an open access and permanently preserved public data foundation. Its ultimate goal is to defend the absolute truth of data and sovereign independence, serving fairness, censorship resistance, and the memory of civilization.
AI storage provides fuel for future productivity as "hot storage," while decentralized storage preserves unremovable historical records for human civilization as "cold memory." The former serves efficiency, pursuing extreme speed; the latter serves credibility, defending silent memory. The former determines how fast the model runs, while the latter decides whether the memory will be erased. Currently, market mechanisms reward productivity efficiency unreservedly, putting AI storage in the spotlight, while decentralized storage seems to be experiencing a collapse in valuation and a pullback in narrative.
The Vision and Reality of Decentralized Storage
There are many decentralized storage projects, but according to industry mentality and ecological sedimentation, the core representatives are always Filecoin and Arweave. Although both belong to "decentralized storage," their underlying architectural philosophies are nearly two completely different paths—the former approaches AWS's elasticity through market contracts, while the latter approaches the eternity of libraries through a one-time social contract.
Filecoin: Builds the most comprehensive verifiable economic system through PoRep and PoSt. It should not continue to struggle with AWS on consumer-level cloud storage but shift towards AI data provenance, public data set hosting, and compliance archiving, providing verifiable chains for model audits and copyright proofs. A necessary path is to package as S3-compatible APIs and support fiat payments, upgrading from the "cheap storage market" to "verifiable computing infrastructure."
Arweave: With a narrative of "one-time payment, permanent storage," it compels miners to preserve and rapidly access as much, especially scarce historical data as possible through Blockweave and SPoRA mechanisms. Its best positioning is as the foundation of human public memory—preserving human rights records, war crime evidence, cultural classics, archiving legal and financial history, and providing permanently accessible long-term memories for AI agents. The value of Arweave lies not in speed but in its capacity to carry the memory of civilization beyond cycles.
The dilemma of decentralized storage projects like Filecoin and Arweave does not stem from incorrect value propositions but from long-term mismatching of productization, retrieval experiences, real needs, and Token incentives. This reveals a huge gap from geek ideals to mainstream commercial applications:
Supply and Demand Incentive Mismatch: Early networks represented by Filecoin rapidly expanded through Tokens but did not build sufficient strong demand sides, leading to immense capacity but insufficient utilization and payment conversion. It rewards "I can store" rather than "I need to store."
Lack of Enterprise-Level Service Capability: AWS's barriers are not disks, but a "data operating system" comprised of APIs, SLAs, permission management, compliance audits, and technical support. Enterprises are purchasing "peace of mind," rather than needing to handle keys and node selection on their own as experimental infrastructure.
Retrieval Experience Shortcomings: "Putting in" does not equate to "retrieving out stably and with low latency." Dispersed nodes, complex topologies, and lack of unified SLAs make it hard to support AI hot data workflows, being better suited for trustworthy cold archiving and data provenance.
Inadequate Privacy Compliance: Enterprises cannot simply write private data into public permanent networks; the rights to deletion and permanent immutability inherently conflict. Decentralized storage is better suited for public data and long-term archives rather than indiscriminately accommodating core private data.
Token Economics Amplify Cycles: Bull market financialization masks insufficient demand, while bear market miner ROI decline reveals commercialization shortcomings. Tokens can cold-start supply but cannot automatically create demand and sustainable revenue.
Other decentralized storage projects often focus on specific ecologies or niche paths: Storj/Sia's cross-cycle industry mentality and Web3 narrative influence are weaker than Filecoin/Arweave; BNB Greenfield/Walrus are tied to specific public chain ecologies of BNB or SUI; Celestia/EigenDA belong to the data availability (DA) layer, serving Rollup transaction confirmations rather than long-term archiving; projects like 0G, which mix AI/DA narratives, attempt to integrate storage, data availability, computation, and AI agent settlement into a set of AI-native modular infrastructure, but their real needs, developer adoption, and commercialization loops remain to be verified.
Future Opportunities for Decentralized Storage: The Long-Term Pendulum of Efficiency and Trustworthiness
During a technological dividend explosion period, capital madly chases efficiency, with assets like GPU and HBM awarded extremely high premiums, naturally marginalizing decentralized storage that advocates "trustworthy and fair." However, the pendulum of history will not remain forever on the efficiency side. Unreasonable bans and content deletions by super platforms, the outbreak of AI copyright lawsuits demanding proof of data sources, geopolitical conflicts igniting data sovereignty disputes, monopolies causing public archives to disappear, and regulatory pressure on the compliance of model training data may all brew a repricing of "trustworthy storage," and the future opportunity of decentralized storage can still reflect unique value in the following directions:
AI Data Provenance: Build "data lineage proofs" combining cryptographic evidence to address regulatory and audit pressures.
Public Datasets and Civilizational Archives: Anchor scrutinized files and cultural heritage, constructing irreplaceable and unremovable memories.
Trustworthy Archiving and Compliance Evidence: Achieve trustworthy self-evidence through hashing, providing high-grade digital notarizations.
Integration of ZK/TEE/DID Technologies: Resolve privacy tensions, upgrading from a single "storage protocol" to a "trustworthy data infrastructure."
Invisible Product Routes: Provide S3-compatible APIs and fiat billing, allowing users to directly purchase "trustworthy archiving" services.
AI storage and decentralized storage represent one side striving for extreme efficiency to provide fuel for our future; the other defends silent memories, safeguarding our right to look back at the past. The current market unreservedly rewards efficiency, making decentralized storage appear silent or even collapsing; but as the AI era further amplifies data monopolies, copyright disputes, and the fragility of historical memory, decentralized storage may embrace revaluation with the stance of a "trustworthy cold layer." Memories that cannot be easily erased by platforms, companies, or any singular power may transition from romantic idealism and marginal faith into necessary infrastructure.
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
