Pull to refresh
Logo
Daily Brief
Following Why Sign Up
The race to build frontier AI

The race to build frontier AI

New Capabilities
By Newzino Staff |

Meta delays its flagship model and weighs licensing Google's technology as rivals pull ahead

2 days ago: Avocado delayed to May; Google Gemini licensing discussed

Overview

Meta Platforms has committed more money to artificial intelligence than any other company in history—up to $135 billion in 2026 alone, and $600 billion in American data center infrastructure by 2028. But money hasn't bought capability. The company's next-generation AI model, code-named Avocado, was delayed from March to at least May 2026 after internal tests showed it trailing systems already shipping from Google, OpenAI, and Anthropic on reasoning, coding, and writing tasks.

In This Story

3 Videos
3 People
Mark Zuckerberg Alexandr Wang Yann LeCun
2 Orgs
Meta Platforms Meta Superintelligence Labs
12 Events
4 Key Stats
$115–135B
4 Scenarios
Meta ships improved Avocado, narrows the gap with rivals
3 Context
Yahoo licenses Google search (2000–2004)

Key Indicators

$115–135B
Meta's 2026 AI capital expenditure
Nearly double the prior year's spending, driven by data centers, chips, and cloud infrastructure for AI development.
$600B
Pledged U.S. AI infrastructure through 2028
Meta's total committed American investment in data centers and related infrastructure over four years.
$14.3B
Scale AI investment to recruit Alexandr Wang
Meta acquired a 49% stake in Scale AI and hired its founder as chief AI officer in June 2025.
4.5%
Meta stock drop on Avocado delay
The largest single-day decline since October 2025, triggered by reports the model trails Google, OpenAI, and Anthropic.

Interactive

Exploring all sides of a story is often best achieved with Play.

Ever wondered what historical figures would say about today's headlines?

Sign up to generate historical perspectives on this story.

Sign Up

Debate Arena

Two rounds, two personas, one winner. You set the crossfire.

People Involved

Mark Zuckerberg
Mark Zuckerberg
Leading company's all-in AI transformation
 Alexandr Wang
Alexandr Wang
Under scrutiny as Avocado model underperforms; sidelining rumors denied by Meta
 Yann LeCun
Yann LeCun
Departed Meta; raised $1.03 billion for new AI startup

Organizations Involved

Meta Platforms
Meta Platforms
Technology conglomerate
Flagship AI model delayed; weighing licensing rival's technology

Parent company of Facebook, Instagram, and WhatsApp, now spending more on AI development than any other corporation.
Meta Superintelligence Labs
Meta Superintelligence Labs
AI research division
Responsible for Avocado development; under pressure after model underperforms

Meta's consolidated AI division, formed by merging its generative AI, Llama, and product teams under a single mandate to build superintelligence.

Timeline

  1. Avocado delayed to May; Google Gemini licensing discussed

    Product Delay

    Internal testing showed Avocado trailing Google, OpenAI, and Anthropic models on reasoning, coding, and writing. Meta pushed the launch from March to at least May. Leadership discussed temporarily licensing Google's Gemini to power Meta AI products. Meta shares fell 4.5%, the steepest drop in over four months.

  2. Google ships Gemini 3.1 Pro, widening performance lead

    Competitive

    Google released Gemini 3.1 Pro with a 77.1% score on the ARC-AGI-2 benchmark—more than doubling its predecessor. Combined with strong releases from OpenAI and Anthropic in February, the competitive bar rose sharply just as Meta was finalizing Avocado.

  3. Meta forecasts record capital spending of up to $135 billion

    Financial

    Meta announced 2026 capital expenditure guidance of $115–135 billion, nearly double the prior year. Chief financial officer Susan Li said the company remained "capacity constrained" and that compute demand was growing faster than supply.

  4. Meta confirms shift from open-source Llama to proprietary Avocado

    Corporate Strategy

    Reports confirmed that Meta's next flagship AI model, code-named Avocado, would be closed-source and proprietary—abandoning the open-weight philosophy that had defined the Llama series and attracted a large developer community.

  5. Chief AI scientist Yann LeCun departs Meta

    Leadership Change

    LeCun, a Turing Award winner who founded Meta's AI research lab 12 years earlier, confirmed he would leave to build Advanced Machine Intelligence Labs, a startup pursuing a fundamentally different approach to AI.

  6. Zuckerberg signals retreat from fully open-source AI

    Corporate Strategy

    After years of championing open-source AI models, Zuckerberg said Meta would likely not open-source all of its superintelligence models, citing competitive and security concerns—partly driven by Chinese firm DeepSeek's use of Llama's architecture.

  7. Zuckerberg launches Meta Superintelligence Labs

    Organizational

    Zuckerberg announced the creation of Meta Superintelligence Labs in an internal memo, consolidating the company's AI research, model development, and product teams under a single division led by Wang.

  8. Meta pays $14.3 billion for Scale AI stake, hires Wang

    Corporate Strategy

    Meta acquired a 49% nonvoting stake in Scale AI and recruited its 28-year-old founder Alexandr Wang as Meta's first chief AI officer, making him one of the highest-paid employees in the technology industry.

  9. Meta splits AI division into two groups

    Organizational

    Chief product officer Chris Cox restructured Meta's AI teams into an AI Products group (led by Connor Hayes) and an AGI Foundations unit (co-led by Ahmad Al-Dahle and Amir Frenkel) to speed product development.

  10. Llama 4 Behemoth delayed indefinitely

    Product Delay

    Meta's largest planned model, Llama 4 Behemoth, was pushed from early summer to fall 2025 or later. Engineers grappled with whether its improvements over earlier versions justified a public release. The model was eventually shelved.

  11. Llama 4 launches amid benchmark manipulation allegations

    Product Launch

    Meta released Llama 4 Scout and Maverick, but the launch was overshadowed by accusations that a specially crafted, unreleased variant had been submitted to the LM Arena benchmark to inflate scores. Meta denied the claims.

  12. Meta's head of AI research announces departure

    Leadership Change

    Joelle Pineau, vice president of AI research who led Meta's Fundamental AI Research lab since 2023, announced she would leave the company. She later joined Cohere as chief AI officer.

Scenarios

1

Meta ships improved Avocado, narrows the gap with rivals

Discussed by: TipRanks analysts (39 of 44 maintain Buy ratings on Meta); StockTwits retail investors who view the delay as a short-term setback

Meta uses the additional months to close Avocado's performance gaps, launching a model by mid-2026 that is competitive with Google's Gemini 3.x and OpenAI's latest offerings. The Gemini licensing discussion is quietly dropped. This outcome depends on whether Avocado's shortcomings are tuning problems—fixable with more compute time and data refinement—or deeper architectural issues inherited from the failed Behemoth training runs. Meta's massive infrastructure gives it the raw compute to iterate quickly if the problems are tractable.

2

Meta licenses Google Gemini to power its AI products

Discussed by: Fortune, New York Times, and Trending Topics EU, citing internal Meta discussions; multiple analysts note the strategic implications

Meta signs a licensing agreement with Google to use Gemini across WhatsApp, Instagram, and Facebook while Avocado is refined. Users would interact with Google-built AI branded as Meta AI, with no visible indication of the switch. This buys time but hands Google a public validation of its model superiority and creates a dependency that could prove difficult to unwind—particularly if users notice no quality change when Meta eventually switches back to its own model.

3

Investors push Meta to slow AI spending after model struggles

Discussed by: Motley Fool, PYMNTS, and 24/7 Wall Street, all questioning whether $135 billion in annual spending is justified without a competitive model

If Avocado's May launch also disappoints, or if the Gemini licensing deal materializes, investor patience could erode. Meta's advertising business generates strong cash flow, but a sustained gap between spending and model quality would invite pressure to reduce capital expenditures or redirect funds to proven revenue drivers. The five Hold-rated analysts on TipRanks represent a minority view, but that minority could grow if Meta's 2026 AI roadmap continues to slip.

4

Meta's AI talent drain accelerates, compounding model delays

Discussed by: Fortune ("FAIR is dying a slow death" report); CNBC coverage of LeCun's departure and internal friction over Wang's leadership style

The departures of LeCun, Pineau, and 600 laid-off employees have already thinned Meta's AI research bench. If Avocado's struggles are perceived internally as a leadership or culture problem—Wang's 70-hour workweek demands, pay disparities between new and existing staff, computing resource bottlenecks—more senior researchers could leave for competitors or well-funded startups like LeCun's AMI Labs, which just raised over a billion dollars. A talent spiral would make each subsequent model harder to build.

Historical Context

Yahoo licenses Google search (2000–2004)

June 2000 – February 2004

What Happened

In June 2000, Yahoo replaced its in-house search technology with Google's engine, making Google the system powering Yahoo's search results. The deal gave Google a "Powered by Google" credit on one of the internet's most-visited pages, exposing millions of users to the Google brand for the first time.

Outcome

Short Term

Yahoo got a better search product and could focus on media and content. Google gained massive visibility and traffic.

Long Term

Google used the exposure to build its own destination site, launched AdWords, and became the dominant search engine. When Yahoo ended the deal in 2004 and built its own search, it was too late to reclaim the market.

Why It's Relevant Today

If Meta licenses Gemini, it risks a similar dynamic: Google gains public validation that its AI model is superior enough for a competitor to pay for, while Meta's own brand becomes associated with someone else's technology. The longer the licensing lasts, the harder it becomes to justify switching back.

Apple switches from PowerPC to Intel processors (2005)

June 2005 – August 2009

What Happened

Apple had built its Mac identity around PowerPC chips developed with partners IBM and Motorola. When those partners failed to deliver processors competitive with Intel's x86 chips—particularly on laptop power efficiency—Steve Jobs announced at the 2005 Worldwide Developers Conference that all Macs would transition to Intel hardware, a move many considered unthinkable.

Outcome

Short Term

Mac performance improved significantly, and the transition was completed faster than expected. Developers adapted within two years.

Long Term

Apple treated Intel as a bridge, not a destination. It quietly developed its own ARM-based chips and in 2020 launched Apple Silicon, leaving Intel behind entirely. The interim dependency funded the time needed to build a superior alternative.

Why It's Relevant Today

Meta's potential Gemini licensing could serve a similar bridge function—buying time to improve Avocado without shipping an inferior product. The question is whether Meta, like Apple, treats the dependency as temporary and invests in catching up, or whether it becomes a crutch.

IBM outsources its PC operating system to Microsoft (1981)

1980–1981

What Happened

IBM, the dominant computer company, needed an operating system for its new Personal Computer and licensed DOS from a small company called Microsoft rather than building its own. IBM believed hardware was the real competitive advantage and that the operating system was a commodity component. The deal gave Microsoft the right to license DOS to other manufacturers.

Outcome

Short Term

IBM shipped the PC on schedule and dominated the early personal computer market.

Long Term

Clone manufacturers used the same Intel chips and Microsoft software to build cheaper alternatives. IBM's hardware advantage eroded, and Microsoft's Windows became the industry's true platform. Microsoft eventually became the world's most valuable company.

Why It's Relevant Today

Meta's $600 billion infrastructure bet assumes data centers and compute are the durable competitive advantage. But if the AI model—the software layer—is where differentiation actually lives, Meta risks building enormously expensive infrastructure that any capable model provider could theoretically run on. The IBM precedent suggests that the company controlling the software platform, not the hardware, often captures the most value.

Sources

(14)
+8
Fortune Benzinga Daily Tech News Show PYMNTS TipRanks CNBC Axios WinBuzzer TechCrunch Fortune CNBC Fortune CNBC CNBC