$Tesla(TSLA.US) performed excellently today, and so did $NVIDIA(NVDA.US). Looks like the bros have a good eye!

🚨$Meta Platforms(META.US) promises to build "Personal Superintelligence," but reality is forcing it to consider renting $Alphabet(GOOGL.US)'s brain

Not long ago, Mark Zuckerberg stood before a global audience and announced Meta's goal is to build Personal Superintelligence for every individual.

He emphasized that Meta possesses vast computing infrastructure, top-tier research teams, and tens of billions of dollars in funding, capable of bringing this future to reality ahead of schedule.

But this grand ambition soon faced a reality check.

Meta's flagship AI model, codenamed "Avocado," has been delayed.

According to multiple reports, this model lags behind major competitors in some of the most fundamental capabilities—such as logical reasoning, code generation, and complex writing.

Originally planned for release this month, Avocado is now postponed to May at the earliest.

Even more shockingly, Meta is internally discussing an option that was almost unimaginable in the past:

Licensing Google's Gemini model for Meta AI products.

In other words, a platform that once declared its intention to build superintelligence is now considering renting a competitor's AI engine.

This matter has attracted significant attention because Meta's investment in AI is staggering.

The company has committed to capital expenditures of $115 billion to $135 billion in 2026, nearly double the previous year.

Zuckerberg even told analysts that this year will be "the year AI fundamentally changes our workflows."

But Avocado's performance still seems to lag behind competitors.

Reportedly, it doesn't even clearly surpass Google's Gemini 3.0 released in November last year.

Although Avocado slightly outperforms Google's Gemini 2.5, launched in March this year, in some tests, it still lags behind the current leading models in key benchmarks overall.

And Meta's investment in AI can be described as unprecedented.

Zuckerberg personally poached over 20 researchers from OpenAI.
He also spent $14.3 billion to acquire Scale AI and appointed its 28-year-old founder, Alexandr Wang, as Chief AI Officer.

Simultaneously, Meta established a new department called "Meta Superintelligence Labs."

Reportedly, some top researchers received signing bonuses as high as $100 million.

After investing such massive resources, the currently public core projects are mainly two models:

Avocado — responsible for text capabilities
Mango — responsible for image generation

But the release dates for both models are constantly being pushed back.

Avocado's release date has shifted from late 2025 → early 2026 → mid-March → now May at the earliest.

Meanwhile, Google's position is becoming increasingly solid.

Google is gradually becoming the default supplier for global AI platforms.

Apple has signed a multi-year cooperation agreement with Gemini for Siri and Apple Intelligence.

Analysts estimate this deal could be worth up to $5 billion.

And Samsung's Galaxy AI is already running on Gemini.

If Meta ultimately also chooses to integrate with Gemini, then Google's AI model could potentially provide core capabilities for platforms with over 5 billion users.

This includes:

Apple's 1.5 billion devices
Samsung's global user base
And Meta's 3.6 billion monthly active users.

If this scenario truly materializes, the landscape of the AI industry could rapidly consolidate.

Questions also arise:

If only a few companies possess the strongest AI models in the future, will the tech industry enter a power structure more concentrated than the internet era?

When AI becomes infrastructure, whoever controls the models may control the entire ecosystem.