Meta is fully pivoting to a closed-source model, with the new Avocado model set to be launched in the coming spring.
BlockBeats News, December 10th - After investing billions of dollars to build the most expensive team in tech history over the past few months, Meta CEO Mark Zuckerberg is now deeply involved in day-to-day development and is driving the company's strategic shift towards AI models that can be directly monetized. According to insiders, a new model codenamed "Avocado" is expected to be released in the spring of 2026 and may be launched in a closed-source form (meaning Meta will tightly control and sell access rights to the outside).
This move marks a significant departure for Meta from its long-standing emphasis on the open-source path. Zuckerberg is dedicating a significant amount of time to a core team called TBD Lab, which, in training Avocado, has even integrated third-party models including Google's Gemma, OpenAI's GPT-oss, and Alibaba's Qwen.
At the same time, Meta is significantly realigning its resource allocation, reducing investment in the metaverse and virtual reality, shifting funds to hardware like AI glasses, and planning to invest $600 billion in AI infrastructure in the United States over the next three years. (Jinshi)
You may also like

Ten Thousand Words Interpretation of STRC: Strategy for Making Money to Buy Coins New Magic

What competitive advantages are still defensible in the AI era?

For Whom the Bell Tolls, For Whom the Lobster Feeds? A Dark Forest Survival Guide for the 2026 Agent Player

Circle CEO's Latest Interview: Stablecoins Are Not Cryptocurrency

Deconstructing the Public Chain Pharos Capital Game: Is a $950 million valuation supported by assets like photovoltaics just a shell transaction under layers of betting?

a16z: AI is making everyone 10x more productive, but the true winner has yet to emerge

Why did the star Web3 project Across Protocol choose to abandon DAO?

In fact, ETH scaling is a major benefit for L2

Memories: 10 Key Contributions of the TON Core Team That Few People Knew in the Early Days

2025 South Korea CEX Listing Post-Mortem: Investing in New Coins = 70% Loss?

BIP-360 Analysis: Bitcoin's First Step Towards Quantum Immunity, But Why Only the "First Step"?

50 million USDT exchanged for 35,000 USD AAVE: How did the disaster happen? Who should we blame?

The Cryptographic Past of the Middle East

Resolving the Intergenerational Prisoner's Dilemma: The Inevitable Path of Nomadic Capital Bitcoin

Who Will Control AI? Why Decentralized AI May Be the Only Alternative to Government and Big Tech
AI has become critical infrastructure, and governments and corporations are competing to control it. Centralized development and regulation are entrenching existing power structures. The Web3 community is building a decentralized alternative — distributed compute, token incentives, and community governance — before that window closes.

Vitalik wrote a proposal teaching you how to secretly use AI large models

On the eve of the explosion of on-chain options

WEEX AI Hackathon: How Did This AI Trading Winner Succeed?
A self-taught AI trading enthusiast achieved top-10 results at the WEEX AI Hackathon. Learn about the mindset, AI tools, and lessons behind this impressive performance.