GRASS Token Surges 100% in Two Weeks as Grass Shifts Narrative to AI Real-Time Data Layer

Author: Gemini, DeepChain TechFlow

On September 28, the Grass official released a lengthy manifesto by Andrej, CEO of Wynd Labs, titled Abundant Intelligence. The article did not announce any new client lists or concrete buyback plans, but it provided a coherent rationale for the recent price movements of the GRASS token.

Over the past two weeks, the price of GRASS has risen from around $0.35 to above $0.70, nearly doubling.

Funds had already priced in the movement on the market, and this lengthy article served more like a post-hoc explanation: Grass attempted to reframe its narrative from a DePIN project selling training data to an AI real-time information layer with a much larger valuation model.

Whether this new narrative can sustain the current market capitalization in the long term remains to be seen, as it must reconcile the gap between its millions of dollars in real cash flow and the long-term disconnect in token capture mechanisms.

Over the past year, market understanding of Grass's business has had a clear ceiling: scraping publicly accessible web pages using residential IPs and packaging them as snapshots for sale to AI labs.

Although this business generates cash flow, the industry consensus that high-quality training data is about to run out often leads the market to view it as a temporary, transitional business.

The core demand of yesterday's long post by the Grass CEO, in the author's understanding, is to break through this ceiling of expectations and create a new narrative. To understand Andrej’s proposed new direction, it is first necessary to clarify a critical weakness of current large models: they are outdated upon release.

No matter how many trillions of tokens were fed into it during training, once the weights are frozen and the model is released, its understanding of the world freezes at the moment training ended. When users ask about real-time news or compare the latest product prices in the chat interface, the model must connect to the internet during the "inference" phase to fetch the most current publicly available web pages.

However, large models themselves cannot access the internet.

Most major websites today have deployed strict anti-bot firewalls such as Cloudflare and Datadome. When AI labs attempt to scrape web pages in real time using data center servers, requests are aggressively blocked due to the distinctive characteristics of data center IPs. Cloudflare has grown into a nearly $100 billion company by successfully keeping automated bots at bay.

And this is precisely the business narrative that Grass is trying to reconstruct.

Grass originally built its foundation on real residential IP networks shared by millions of ordinary users around the world running devices at home. In the past, Grass simply used this network to silently scrape and clean web pages in bulk, packaging them into offline corpora for sale to AI labs; now, Andrej plans to turn this network directly into a real-time proxy browsing channel for AI models.

In this roadmap, the official plan outlines a complete toolchain for integrating AI into the real world:

• Search API (solving "where to find"): Over the course of long-term web-wide data crawling, Grass has accumulated a vast collection of public web copies; this API functions like a dedicated search engine built for AI, enabling models to first horizontally index the most relevant target webpages when faced with complex queries.
• Contents API (solving "how to view"): After locating the target webpage, the model no longer directly attempts to access it via data center requests. Instead, it initiates access through millions of real residential nodes behind Grass. To the target website’s firewall, this appears as a normal browsing session originating from an ordinary American household, thereby bypassing interception mechanisms like Cloudflare and directly retrieving and fully rendering the webpage’s content for the model.
• Multimodal retrieval (solving "how to understand complex content"): The real internet contains more than just text—it's filled with images, PDF reports, and embedded videos. This module enables models invoked in real time to directly understand non-textual information.

Following this logic, it’s easy to understand the current views of some overseas crypto analysts on GRASS.

If Cloudflare’s business model is “charge websites to block AI crawlers,” then Grass is trying to build the opposite model: “charge AI companies to use real people’s home networks as disguise channels to let AI in.”

At this point, Grass’s capital narrative has undergone a complete reanchoring: no longer is it positioned as an “off-chain data outsourcing provider” trading at a discount relative to traditional DePIN hardware multiples, but rather it has rebranded itself as an “AI real-time retrieval layer and agent gateway,” akin to Exa and Parallel, which command high valuation multiples in the primary market.

This is why, even without any press release announcing new substantive client acquisitions, this internally consistent roadmap alone was enough to spark market speculation in secondary markets over this shift in narrative.

This version is preferred by users for projects with real income and internally consistent logic.

From this perspective, examining GRASS’s current fundamentals, it is a growth-oriented data company with tens of millions of dollars in real annual revenue. According to data disclosed at the July token holder call, its ability to generate cash flow has already moved beyond the early proof-of-concept stage:

1. Total revenue for 2025 was $17 million. In the first half of 2026 (H1) alone, revenue reached $17 million, representing nearly a 7-fold increase compared to the same period last year.
2. The company expects revenue from "training data" alone to reach $65 million to $75 million for the full year of 2026. The business exhibits clear seasonality, with some large orders originally scheduled for the first half of the year (approximately $15 million) being deferred to the third quarter due to the training cycles of AI labs.
3. The company is currently profitable on an accounting basis. Through self-built computing power and storage facilities in 2025, monthly operating expenses (Opex) have been reduced to $2 million to $3 million (primarily for infrastructure), significantly lower than its revenue inflow rate.

In its architectural design, Grass avoids the trap common in traditional Web3 projects where equity companies reap the profits while tokens are left with scraps. Customer contracts and business revenues flow directly into Grass DataCo and the foundation, while the underlying development team, Wynd Labs, exists solely as an engineering and business services provider.

In theory, this ensures that all protocol revenue belongs to the network itself.

The business is running smoothly, and the architecture is clean, but how profits flow back to the token remains Grass's biggest challenge.

Under the current mechanism, Grass follows the typical logic of a growth-stage startup tech company, but this directly clashes with the liquidity-driven "crypto logic" desired by secondary markets.

1. Extremely restrained buyback力度

For traders in the secondary market, genuine buybacks and burns are the only hard-core value capture mechanism. However, the Grass official team has shown extreme reluctance in this regard. Aside from a small public market buyback of approximately $350,000 in November 2025, large-scale buybacks have largely stalled.

The official logic is: reinvesting the vast majority of profits into expanding computing power, storage, and network infrastructure yields a far higher long-term return than directly supporting the token price on secondary markets. While this heavy asset investment strengthens the business moat, it removes the most direct buying support for the token in the short term.

1. The Real-World Game of Node Incentives

In Stage 2 of the network, Grass made a key change to its tokenomics: stopping the issuance of new GRASS tokens to subsidize nodes, and instead distributing rewards using real USDC revenue generated by the protocol.

From the perspective of token inflation, this cuts off the endless selling pressure on the circulating supply, making it an absolute positive. However, from the standpoint of the network’s physical supply side, contradictions are beginning to emerge.

Some ordinary nodes providing broadband connectivity, which themselves earn only a few cents to tens of cents in stablecoin per month, are beginning to shut down. Grass’s new narrative of “bypassing Cloudflare restrictions” rests on the foundation of these millions of real residential nodes; if low returns cause the underlying network to shrink, its business moat will be physically eroded.

1. Stakers feel insufficient returns

Although the July governance vote approved allocating a portion of USDC revenue to GRASS stakers, the actual scale of these fees and the distribution ratio remain insufficiently transparent in public details.

When token holders cannot feel the millions of dollars in protocol revenue through an intuitive APY statement, the token price can only oscillate between earnings expectations and early lock-up release fears—such as the tail-end investor unlocks at the end of October.

So, Grass has completed the contextual shift from a “static data packet company” to an “AI dynamic information layer” with a new roadmap.

In the cryptocurrency market's capital game, this kind of narrative upgrade—where established projects align themselves with trending sectors—is often the most effective short-term catalyst.

But once the short-term emotional release is over, the long-term pricing of GRASS will still be determined by real business fundamentals. The market will ultimately test not how advanced the API features described in lengthy articles are, but whether the company can deliver on its multi-million-dollar financial guidance over the next six months. The more fundamental question is: when will this cash-rich entity finally convert its book profits into actual GRASS token buying pressure in the secondary market?

Before the narrative fully aligns with real money, the CEO’s statement merely provided investors with a reason to re-enter. If there is no clear token buyback or profit distribution forthcoming, I believe this doubling rally, fueled by expectations, will ultimately prove to be a short-term capital game rather than a true reassessment of asset value.