Author: 0xjacobzhao, IOSG

Recently, the "first domestic storage stock" Yangtze Memory Technologies officially landed on the Growth Enterprise Market, exploding the scene with an astonishing surge of 500%. Although the storage sector as a whole is still disturbed by the waves of recent pullbacks, AI storage is still being crazily reassessed by capital in the current wave of technological narrative. Meanwhile, decentralized storage in the Web3 domain has fallen into long-term silence and loss. Why does the market performance differ so greatly, despite both being labeled as "storage"? The fundamental answer lies in the complete differentiation of underlying value functions.

The reassessment of storage in the AI era is essentially a celebration of "hot data efficiency," serving the ultimate maximization of computational resource utilization and commercial monetization; while decentralized storage adheres to the value proposition of "cold data credibility," defending 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 credibility system for cold data. The current capital market undoubtedly firmly stands on the side of "efficiency," but human civilization ultimately still needs an immutable memory foundation. The long-term value of credible cold storage has never disappeared; it simply lies dormant on the dark side of the cycle, 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 drive reliability, disaster recovery solutions, archiving strategies, and a device update cycle of 3 to 5 years. Storage was seen as an accessory following server procurement.

This round of storage boom is not a traditional cyclical recovery but a repricing of data flow capabilities by AI. In the era of large models, the logic of storage has shifted from "capacity first" to "efficiency first," striving to hit extreme metrics like GPU feeding rate, Checkpoint writing, and extremely low latency for RAG. This marks a leap in storage value from being "the final resting place of data" to "the high-speed passage for data entering computation."

The evolution of resource bottlenecks in AI infrastructure is essentially a "barrel effect" compensatory battle. The true utilization rate of computational power is not a linear summation of single assets but a strict multiplicative effect: actual utilization rate = GPU × HBM × DRAM × SSD × Network × File System; any shortcoming in one link will lead to the collapse of overall computing utilization. In the AI era, storage has transformed from a "cost center" into an "efficiency engine". This is the fundamental logic behind storage being repriced.

AI Storage Architecture Overview: From HBM Bandwidth Organs to Data Lake Foundations

AI storage is by no means just a pile-up of single hardware but rather a closely coupled and hierarchically scheduled complex system. Within this system, industrial value and capital focus are highly concentrated on HBM, enterprise-grade SSDs, SSD controllers, NVMe/CXL protocols, and high-performance storage systems. To clearly deconstruct its value flow, we will divide the AI storage architecture into four core levels from top to bottom:

• Compute Proximal Memory Layer (Bandwidth Core): Dominated by HBM, supplemented by DRAM and CXL memory pooling technology. This layer directly interfaces with GPU/CPU packaging or buses, aiming to break the "memory wall," representing the first checkpoint that determines whether computational power can be fully released.

• High-Speed Persistent Storage Layer (IO Hub): The core logic is enterprise-grade SSD = NAND chips + SSD controllers + NVMe/PCIe data paths. This layer undertakes high-frequency Checkpoint writes, massive training set loads, and RAG hot data caching, being the most definitive persistent storage increment in AI data centers.

• Low-Cost Large-Capacity Storage Layer (Capacity Foundation): Composed of HDDs, cold storage, and data lake archiving systems. In the face of exponentially expanding multimodal raw data, historical logs, and compliance backups, this layer continues to provide irreplaceable TCO (Total Cost of Ownership) advantages.

• 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 actually consumes is not bare hardware but the availability of data that is efficiently organized, indexed, and authorized by the software stack.

As an ecological extension, decentralized storage does not directly engage in the millisecond-level race of AI hot data but anchors public data set preservation, AI training data provenance, and long-term cold memory archiving, establishing its unique ecological niche as the "credible cold layer."

HBM: The "Bandwidth Organ" Closest to Computational Power in the AI Storage Chain

High Bandwidth Memory (HBM) is not conventional storage but rather high bandwidth memory close to the GPU. Its core mission is not to store data but to continuously "feed" data to computation at extremely high bandwidth. HBM is the link in the AI storage chain closest to computation and has the highest certainty, directly determining whether the GPU can be "satisfied," making it the current core bottleneck of the supply chain.

The core architecture of HBM is "3D DRAM stacking + 2.5D advanced packaging": through TSV vertical stacking and CoWoS heterogeneous integration, it achieves extreme compression of storage-computation distances, resulting in leaps in bandwidth. Its industrial barrier is not just DRAM design but a systemic engineering encompassing DRAM manufacturing processes, TSV, ultrathin stacking, packaging, heat dissipation, testing, and customer certification. Any flaw in yield at any link may lead to the entire HBM stack being discarded.

Currently, only SK Hynix, Samsung, and Micron are the three giants that can stably mass-produce, creating a triple moat of top-tier DRAM manufacturing processes, packaging capabilities, and NVIDIA/AMD customer certifications.

DRAM and CXL: System Memory Foundation and Memory Pooling Engine

HBM addresses extreme bandwidth for the GPU's proximity, DRAM solidifies the server system memory foundation, and CXL attempts to break physical boundaries to reconstruct the organization of memory resources within data centers.

• DRAM: Primarily carries CPU-side caches, data preprocessing, intermediate state storage, and system operations, serving as the most fundamental system memory layer for servers. The global DRAM market is highly concentrated among the three giants SK Hynix, Samsung, and Micron; Yangtze Memory Technologies (CXMT) is the core variable in China's domestic DRAM substitution.

• CXL (Compute Express Link): A next-generation cache coherence interconnect protocol for data centers, aimed at breaking through the limitations of traditional DIMM slots, local memory capacity, and server memory resource silos, promoting memory architecture evolution toward extension, pooling, and sharing. Currently, CXL is still in the early stages of transitioning from platform support to large-scale deployment, with high value in the medium to long term; core companies include Astera Labs and Lanqi Technology.

Enterprise-Grade SSD: The Data Hub Constructed by NAND, Controllers, and NVMe

Enterprise-grade SSD is the most core high-throughput persistent increment in AI data centers, using extremely high throughput, very low latency, and stable QoS to continuously "feed" data to the GPU, spanning the full lifecycle of training data loading, Checkpoint writing, RAG retrieval, inference caching, and log backflow.

In the AI storage architecture, SSDs are not isolated hardware, but a highly coupled system, which can be distilled into the industrial formula: enterprise-grade SSD = NAND chips + SSD controllers + NVMe/PCIe data paths. The three layers represent independent segments of the industrial chain:

• NAND Chips (Raw Materials Layer): Determines storage density and unit cost; the controller manages performance release and lifespan. Representative companies include Samsung, SK Hynix (Solidigm), Micron, Kioxia, Western Digital, and Yangtze Memory Technologies.

• SSD Controller (Performance Empowerment Layer): Determines performance release, data error correction, QoS stability, and wear leveling. Representative companies include Phison (群联), Silicon Motion (慧荣), Marvell, Maxio (联芸).

• NVMe/PCIe (Data Path 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 include Broadcom, Marvell, Astera Labs.

HDD / Cold Storage / Archiving: The Low-Cost Foundation of AI Data Lakes

AI will not eliminate HDDs. With the demand for multimodal large models for video and image data, as well as the exponential growth of corporate compliance logs and historical data sets, the demand for low-cost cold data storage is skyrocketing simultaneously. In the AI storage architecture, the collaboration between SSD and HDD is layered based on business value: SSD handles hot data and high throughput, while HDD addresses low cost and long-term preservation. Representative companies include Seagate, Western Digital, and Toshiba.

AI Storage Software Stack: The Scheduling Core of Data Availability

What AI truly consumes is never bare disks, but "data services" that have been precisely organized by the software stack. This architecture transforms underlying hardware into knowledge assets that AI can directly call upon at the upper level, specifically divided into four layers:

• High-Performance Storage Systems (Supply System): Focused on concurrent throughput and low latency, it solves the "data hunger" problem of GPU clusters through parallel file systems, ensuring rapid circulation of training and inference. Representative companies include VAST Data, WEKA, Pure Storage.

• Object Storage (Raw Data Lake): Focused on managing Objects, Keys, and Metadata, it carries massive amounts of unstructured data. It does not pursue extreme low latency but builds a capacity foundation with low cost and cloud-native characteristics. Representative company: AWS S3.

• Vector Database (Semantic Indexing Layer): The vector database is responsible for storing, indexing, and retrieving vectors generated by embedding models, allowing AI to accurately pinpoint relevant content from a vast pool of knowledge. Representative companies: Pinecone, Milvus.

• RAG Data Layer (Knowledge Invocation Layer): Beyond simple retrieval, it encompasses data slicing, cleansing, permission control, and citation tracing, 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: Maximization of Efficiency vs. Maximization of Credibility

AI Storage is an extremely efficiency-driven system, with its value function focused on maximizing computational output. HBM bandwidth determines whether the GPU can be satisfied, SSD throughput determines the read and write efficiency of datasets and Checkpoints, and low latency is crucial for real-time experiences in RAG and inference. These metrics ultimately converge into GPU utilization and unit token costs, directly determining the commercial profitability of AI applications. The ultimate goal of AI storage is not preservation but acceleration, serving productivity.

In contrast, the value function of decentralized storage is entirely different. It questions whether the data will still exist in ten years, whether it has been tampered with, and whether it can withstand single-point censorship. Through cryptographic proof and distributed networks, it constructs an open-access and permanently preserved public data foundation. Its ultimate goal is to defend against data manipulation and ensure absolute authenticity and sovereign independence, serving fairness, demands for censorship resistance, and the memory of civilization.

AI storage is the "hot storage" that provides fuel for future productivity, while decentralized storage is the "cold memory" that preserves irremovable historical records for human civilization. The former serves efficiency, pursuing extreme speed; the latter serves credibility, defending silent memories. The former determines how fast models run, while the latter determines whether memories will be deleted. Currently, the market rewards efficiency without reservation, putting AI storage at the pinnacle, while decentralized storage seems to be experiencing valuation collapse and a silent narrative drain.

The Vision and Reality of Decentralized Storage

There are many decentralized storage projects, but based on industry mindset and ecological sedimentation, the core representatives remain Filecoin and Arweave. Although both belong to "decentralized storage," their underlying architectural philosophies are almost two completely different paths — the former approaches AWS's elasticity through market contracts, while the latter approaches the eternity of libraries through one-time social contracts.

• Filecoin: Created the most complete verifiable economic systems through PoRep and PoSt. It should not continue to clash with AWS over consumer-grade cloud storage but instead pivot toward AI data provenance, hosting public datasets, and compliant archiving, providing verifiable chains for model auditing and copyright proof. A necessary path is to encapsulate as S3 compatible APIs and support fiat payments, upgrading from a "cheap storage market" to "verifiable computing infrastructure."

• Arweave: With the narrative of "one-time payment, permanent storage," it forces incentives through Blockweave and SPoRA mechanisms for miners to save and quickly access as much, especially scarce, historical data as possible. Its optimal position is as a foundation for public memory — preserving human rights records, war crime evidence, cultural scriptures, archiving legal and financial history, providing permanently accessible long-term memory for AI Agents. Arweave's value lies not in speed, but in its capacity to carry civilization's memory through cycles.

The dilemmas faced by decentralized storage projects such as Filecoin and Arweave do not stem from incorrect value propositions, but rather from a long-term mismatch in productization, retrieval experience, real demand, and token incentives. This reveals a vast chasm from geek ideology to mainstream commercial applications:

• Supply-Demand Incentive Mismatch: Early networks represented by Filecoin rapidly expanded through tokens but did not build a sufficiently strong paying demand side, leading to enormous capacity with inadequate utilization and payment conversion. They rewarded "I can store" rather than "I need to store."

• Absence of Enterprise-Level Service Capability: AWS's barrier is not hard drives, but the "data operating system" constituted by APIs, SLAs, permission management, compliance auditing, and technical support. Enterprises buy "peace of mind," not an experimental infrastructure needing them to handle keys and node selection themselves.

• Retrieval Experience Shortcomings: "Getting stored" does not equal "retrieving stably and with low latency." Decentralized nodes, complex topologies, and lack of unified SLAs make it difficult to handle AI hot data workflows, being more suitable for credible cold archiving and data provenance.

• Insufficient Privacy Compliance: Corporate private data cannot simply be written into public permanent networks; the right to delete and permanence issues create inherent conflicts. Decentralized storage is more suited for public data and long-term archives rather than indiscriminately handling core private data.

• Token Economics Amplifying Cycles: Bull market financialization obscures demand shortages, while bear market miner ROI declines expose commercialization shortcomings. Tokens can cold-start supply but cannot automatically create demand and sustainable revenue.

Other decentralized storage projects often focus on specific ecosystems or niche tracks: Storj/Sia have less industry mindset and Web3 narrative influence than Filecoin/Arweave; BNB Greenfield/Walrus binds to specific public chain ecosystems like BNB or SUI; Celestia/EigenDA belong to the data availability (DA) layer, serving Rollup transaction confirmations instead of long-term archiving; projects like 0G that weave AI/DA narratives attempt to integrate storage, data availability, computation, and AI agent settlements into a modular AI-native infrastructure, but their real demands, developer adoption, and commercialization closed loops still need to be validated.

The Future Opportunities of Decentralized Storage: Long-Term Pendulum of Efficiency and Credibility

During the surge of technical dividends, capital frantically chases efficiency, and assets like GPUs and HBM are given extremely high premiums while decentralized storage advocating "credibility and fairness" is naturally marginalized. However, the pendulum of history will not remain stuck on the efficiency side forever. The unreasonable bans and content deletions of super platforms, explosive AI copyright lawsuits pushing for proof of data sources, data sovereignty disputes triggered by geopolitical conflicts, public archival disappearances due to data monopolies, and regulatory pressures for compliance in model training data, among others, may brew a repricing of "credible storage." Future opportunities for decentralized storage still have the potential to showcase unique value in the following directions:

• AI Data Provenance: Building "data lineage proofs" through cryptographic verification to respond to regulatory and audit pressures.

• Public Datasets and Civilizational Archives: Anchoring censored archives and cultural heritage, constructing irreplaceable irreversible memories.

• Credible Archiving and Compliance Evidence: Achieving credible self-evidence through hash proofs, providing high-level digital notarization.

• ZK/TEE/DID Technology Integration: Resolving privacy tensions, upgrading from a single "storage protocol" to a "credible data infrastructure."

• Invisible Product Route: Providing S3 compatible APIs and fiat billing, allowing users to directly purchase "credible archiving" services.

AI storage and decentralized storage serve different yet significant roles; one strives for ultimate efficiency, fueling our journey into the future, while the other defends silent memories, preserving our right to reflect on the past. The current market rewards efficiency without reservation, thus making decentralized storage seem silent or even collapsing; however, as the AI era further amplifies the vulnerabilities of data monopolies, copyright disputes, and historical memories, decentralized storage may welcome a repricing of value in the guise of a "credible cold layer." Memories that cannot be easily erased by platforms, companies, or any single authority may transform from a romantic idealism, a 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。