Is EMC a value pit in this bull market under the DeAI craze?

What kind of AI projects can emerge in the crypto space? The answers can be various and diverse. But let's set aside more fantasies, fictions, and narratives, and seek answers from the world's top and most valuable AI companies.

NVIDIA, one of only three companies in the world with a market capitalization exceeding $3 trillion, is not only a core pillar of global AI applications but also the undisputed leader in GPU computing power. So, how does NVIDIA make money? The answer may not be complicated— the core of AI projects lies in computing power, especially the competition for GPU computing power. The logic of today's AI industry is merely a different form of the "selling shovels" business.

Recently, NVIDIA released its Q3 2024 financial report, and the data remains impressive: $35.08 billion in revenue, a year-on-year increase of 94%, exceeding analysts' expectations of $33.25 billion and also higher than the company's own performance guidance ($31.85 billion to $33.15 billion). Net profit reached $19.309 billion, a new high compared to the second quarter's "revenue of $30 billion and net profit of $16.59 billion." These figures not only refreshed market expectations but also reaffirmed NVIDIA's irreplaceability in the field of AI computing power.

Not only in traditional finance, but returning to the crypto space, the AI MEME craze is also sweeping in. From the initial AI MEME to the current AI Agents, more and more large-scale AI-based applications are emerging, attracting significant attention from capital and developers.

Recently released from prison, CZ is also very focused on AI. His recent tweets discussed the potential of combining AI with blockchain: "AI tagging (or AI data processing) is very suitable for being done on-chain. It can leverage the global low-cost labor force, eliminate regional biases, and enable instant payments through cryptocurrency." Meanwhile, Musk also announced that xAI will establish an AI game studio with the goal of "making games great again."

For a market like crypto, which is full of "beneficial myths," replicating a 1:1 NVIDIA might be overly idealistic, after all, this is a giant with significant technological accumulation and market position. However, if one were to replicate NVIDIA's profit-making "playbook"—that is, leveraging powerful computing power and the core competitiveness of GPUs to expand the market—perhaps it wouldn't be so difficult. Because behind all AI Agents projects are supported by massive AI models, and behind these models is the support of GPU computing power. Therefore, computing power remains the core driving force for the success of AI projects in the crypto space.

The AI Computing Power Lowland in Crypto

In this context, a project that possesses GPU innovation and basically covers all the AI applications and gameplay you can think of in the current market: Layer 1, computing power leasing, decentralized AI + Web3 creation platform, and AI agents—Edge Matrix Chain can be considered a very good ambush target, even a value lowland.

Lifting the "Heart" of AI

Traditional Layer 1 blockchains have demonstrated strong capabilities in many fields, but when you turn your attention to AI applications that require massive computing power, you'll find that their computational limits cannot support these enormous demands. AI, especially giants like large language models (LLMs), requires not ordinary computing power, but the unparalleled computing capabilities of GPUs. Without GPUs, the future of AI amounts to nothing.

EMC recognizes this and has developed a distributed GPU computing network aimed at providing sufficient computing power support for AI applications. Each EMC node not only provides ordinary computing power but also offers robust GPU support, allowing developers to run complex AI models seamlessly in a decentralized environment. Even more surprisingly, all this computation can be done at a cost lower than traditional cluster computing. Yes, handling massive computational tasks is no longer the exclusive domain of cloud computing giants, nor does it require exorbitant fees.

Moreover, EMC's network architecture further disrupts the limitations of traditional blockchains. It not only boasts low latency and high throughput advantages, ensuring rapid data storage and efficient retrieval, but more importantly, it provides an ideal stage for AI applications that require frequent data exchanges. Additionally, EMC adopts a modular design, allowing developers to flexibly build protocol layers, network layers, application layers, and asset layers according to their needs. This design enables developers to focus on innovation without worrying about the constraints of underlying architecture, laying a foundation for the widespread application of decentralized AI.

Speaking of innovation, EMC does not stop at the technical level; it has also introduced a revolutionary product—EMC Hub. This decentralized AI computing power scheduling platform offers unprecedented support for AI developers. Through EMC Hub, developers can create commercial AI services similar to Midjourney in just a few hours, at a cost only 30% of traditional development methods.

Not only that, but EMC has also closely collaborated with JarvisBot, further pushing decentralized AI capabilities into the market. JarvisBot, as an AI agent strongly supported by the EMC network, has successfully integrated into the TON ecosystem and provides users with a variety of AI services: from article writing, image generation, to content summarization and video creation. This AI agent was launched in June this year, leading the market trend by more than half a year.

He who has GPUs holds the world, possessing absolute hard currency

NVIDIA's GPUs are almost the lifeblood of the entire industry, especially the NVIDIA H100 GPU, the latest GPU model.

Because of its critical importance, GPUs have become an extremely scarce resource. Especially as AI gradually enters the "big model era," the demand for GPUs has surged, and the supply-demand relationship in the market has become increasingly tense. Obtaining high-performance GPU resources, especially top-tier chips like the NVIDIA H100, has become a form of "hard currency." For many AI companies, especially those in the early stages, acquiring sufficient GPU resources is almost as difficult as reaching the sky.

This also explains why large investment institutions, such as a16z, are willing to spend heavily to hoard a large number of H100 GPUs, renting GPU resources to invested companies in exchange for equity or other benefits. It can be said that he who has GPUs holds the world. However, just as market demand is so urgent, supply is extremely insufficient, the "scarcity" of NVIDIA H100 GPUs has also provided investment opportunities for many industry pioneers.

EMC (Edge Matrix Chain) is leveraging its innovative Layer 1 network to connect GPU computing power with DeFi platforms, launching an unprecedented GPU tokenization solution. This solution not only addresses the liquidity issues of GPU computing power but also allows these "high-performance hardware" to become tradable assets, thus helping more users participate in this AI computing power ecosystem.

At the Token2049 Conference on September 19, 2024, EMC showcased an innovative NVIDIA H100 SuperPOD Live Demo.

This modified machine, the NVIDIA H100 SuperPOD, consists of 1024 H100 GPUs, possessing powerful computing capabilities that can provide enormous computing power support for decentralized platforms. The computing power can not only support complex AI model training but also earn annual revenue of up to $15 million through the GPU resource leasing model.

Through this demonstration, EMC not only showcased the decentralization of AI computing resources but also revealed how tokenizing GPU computing power can create new economic opportunities and income sources for global users. What does this mean? This is not only an ultimate presentation of AI technology but also a deep exploration of the commercialization of AI computing power.

But this is not the most exciting part. By integrating these GPU computing resources into a decentralized platform, EMC proposed the idea of tokenizing these high-performance hardware. In other words, global investors and developers can use blockchain technology to decompose these computing resources into smaller units, turning GPU computing power into a tradable asset. This way, anyone can purchase and stake these GPU computing powers on the EMC platform, obtaining an annual yield rate (AYR) of about 12%, and directly participate in the wave of AI computing.

Currently, the RWA assets provided by the EMC platform mainly come from top-tier high-performance GPUs globally, such as NVIDIA A100, H100, and other models (e.g., 4090, 3090). These GPU computing power assets are provided by multiple ComputeDAO members, including important participants like NWG and exaBITS.

This innovation not only makes AI infrastructure more decentralized but also, through tokenization, breaks the traditional "high threshold" that originally belonged to a few large enterprises, allowing every ordinary user the chance to participate. Through EMC's technology, AI computing power has become a liquid asset, no longer just expensive hardware sitting in the hands of certain institutions.

Why is EMC undervalued?

After introducing what EMC is doing and its unique competitive points, let's return to the most important question: why do I think EMC is undervalued?

Let's start with the data. EMC currently covers multiple innovative fields: Layer 1 blockchain, computing power leasing, decentralized AI + Web3 creation platform, and AI agent tools. Just looking at these concepts and business modules, its annual revenue has already reached $15 million. What do you think a reasonable market value for EMC should be when you add these highlights together?

If you still find it hard to give a precise answer, let's look at the financing and valuation provided by VCs: On August 30 this year, the multi-chain infrastructure Edge Matrix Chain completed a $20 million financing round, led by P2 Ventures and Amber Group.

This means that from the perspective of these venture capital institutions, EMC's valuation is around $200 million. (Generally speaking, when venture capitalists provide valuations, they add about a 10x premium, and those who understand investments know how deep the insights are, haha.)

I believe this valuation is reasonable, especially since the market cap ceiling for GPU leasing projects—ionet, which focuses solely on GPU leasing, currently has a market cap of $300 million. And what about EMC? Not only does it have GPU leasing, but it also has a complete Layer 1 blockchain network, plus a decentralized AI creation platform and AI agent tools, which is a combination of four killer narratives.

Let's also take a look at the strength behind EMC. Recently, EMC collaborated with DoraHacks to host a high-profile hackathon, directly offering $10 million in grants and a $1 million prize pool to support innovative projects.

For such a project, not to mention $300 million, even $100 million seems entirely reasonable, especially since there are countless meme projects launching at $100 million from Pumpfun. In contrast, EMC's comprehensive advantages in technology, team, and market prospects can fully support a higher valuation. Currently, Edge Matrix Chain's market cap is $5 million, which is definitely hidden alpha.

Interestingly, while I was researching the EMC project, I unexpectedly found that EMC appeared in a Wall Street Journal article about GPUs. This article attracted widespread attention, and I have read it myself, but at the time, I did not pay special attention to the mention of EMC. The title was "China's AI Engineers Are Secretly Accessing Banned Nvidia Chips."

To cope with the shortage of computing power, Chinese AI engineers are collaborating with international computing power suppliers to access the required resources through a decentralized GPU network, and lawyers state that participating in accessing banned NVIDIA AI chips as buyers, sellers, and brokers in this manner does not violate any laws.

And EMC is one of the key platforms in this context. EMC allows developers to acquire the world's top H100 GPU resources at lower costs and with greater flexibility, resources that are widely used for training large language models (LLMs). This is undoubtedly a huge advantage for developers looking to participate in advanced AI projects.

At the same time, the Wall Street Journal also mentioned some technical details, such as how EMC protects the anonymity of transactions through smart contracts and ensures the fair distribution of computing power resources. These details not only showcase the depth of EMC's technology but also reflect its forward-thinking in data privacy and security. Especially in the context of the growing global demand for data security and privacy protection, EMC's decentralized model appears to align more closely with future trends.

Moreover, the Wall Street Journal also mentioned some technical details, such as how EMC protects the anonymity of transactions through smart contracts and ensures fair distribution of computing power resources. These details not only showcase the depth of EMC's technology but also reflect its forward-thinking in data privacy and security. Especially against the backdrop of the increasing global demand for data security and privacy protection, EMC's decentralized model seems particularly aligned with future development trends.

The now well-known AI projects have become too large in market capitalization. Once everyone realizes the potential of these projects, can they still be considered alpha?

What is truly worth exploring is often those projects with lower market capitalizations. Currently, EMC is undoubtedly such a hidden alpha and value lowland. If the project team collaborates on some marketing activities and waits for the wind to come, EMC should perform well.