On January 16, the fees for launching tokens on the Bags platform exceeded $2 million, marking a thousandfold increase compared to December.

Source: DefiLlama.

According to data from The Block, the figure was over $100,000.

In December, the platform's average daily figure was around $640. However, since then, Bags has surpassed Pump.fun in the number of tokens issued.

Bags is a Solana-based platform for creating digital assets linked to internet trends and software. Recently, it has gained traction among open-source AI projects.

The surge in interest in the application correlates with positive trends in the blockchain itself. Over the past two weeks, the number of transactions and active addresses on the Solana network has increased by nearly 50%.

Speculation Scheme

The rise in interest in the application aligns with the positive dynamics of the blockchain. In the last two weeks, the number of transactions and active addresses on the Solana network has grown by almost 50%. Once a specific AI repository or utility attracts mass attention from the relevant community, the project comes under the radar of speculators.

Users launch tokens tied to trending AI repositories. In some cases, the project creator brands the digital asset, publicly acknowledging the connection. Following this, trading fees begin to flow into the developer's wallet. The market perceives this as a legitimization of the token, which drives up its price.

Traders attempt to get ahead by identifying promising AI projects that are likely to be officially recognized by their creators.

The sustainability of digital assets depends on whether developers will support speculative tokens. Otherwise, they face a downfall.

A notable example is the GAS asset. Initial support from the project creator propelled the token's market cap from $4 million to $60 million. However, it later plummeted below $1 million after criticism from the developer.

Price dynamics of GAS. Source: DEXTools.

Claude Code Becomes a Trend

Claude Code is rapidly gaining popularity: by July 6, 2025, it had been used by 115,000 programmers, generating 195 million lines of code weekly.

At Anthropic, they note that AI assistants significantly speed up development and automate routine tasks.

Key findings from the report include:

  • 79% of interactions with Claude Code involved automation, where AI performs tasks like writing code and fixing bugs;
  • AI is most commonly used for layout and frontend development (JavaScript, HTML, CSS);
  • The tool is more popular among startups than large companies.

Claude Code is also regarded as an excellent solution for vibe coding. A user named lukaslalinsky shared their experience:

“It has a learning curve. You need to learn how to set context correctly, break work into parts, etc. And of course, you need to know how to program.”

As a reminder, in January 2026, the AI startup Replit launched the Mobile Apps feature, allowing users to create applications through vibe coding and submit them for release directly from the editor.