Key Take‑Aways
- Meta Platforms, Inc. (ticker: META) is facing a major leadership shake‑up: its chief AI scientist, Yann LeCun, is reportedly planning to exit and launch his own startup.
- Meta has reorganised its entire AI effort under the new Meta Superintelligence Labs (MSL) division, signalling a strategic pivot towards large infrastructure, product‑scale AI and “superintelligence”.
- Despite growth in its core business, Meta’s AI strategy still lags key competitors in model‑leadership and monetisation, and its escalating infrastructure investment raises concerns about margins and execution.
- From an investor perspective, META remains a high‑potential yet high‑risk play: its legacy business is strong, but the AI transition and leadership turbulence increase uncertainty.
Why LeCun Is Leaving
Yann LeCun has long served as the face of Meta’s fundamental AI research—recruited in the FAIR days and a widely respected deep‑learning pioneer. The timing and nature of his departure raise important questions.
Internal Structural Shift
In June 2025 Meta announced the formation of Meta Superintelligence Labs, a unit intended to bring its foundational AI research, product‑AI efforts and infrastructure build‑out under one roof. Meta CEO Mark Zuckerberg named ex‑Scale AI chief Alexandr Wang to lead MSL, signalling a pivot toward high‑speed, product‑driven AI deployment. LeCun, who previously reported to the CPO, is now reported to have been moved under Wang’s domain — a change that suggests Meta is prioritising engineering scale and deployment power over pure research.
Philosophical and Strategic Mismatch
LeCun has publicly signalled caution toward the “bigger‑model, faster‑release” LLM race, favouring more foundational AI research. Meta’s renewed ambition—targeting superintelligence, massive infrastructure and rapid monetisation—may represent a divergence from his preferred path. The departure therefore may reflect deeper tension between research culture and product/scale urgency.
Autonomy and Timing
By launching his own venture, LeCun can pursue research with fewer institutional constraints. For Meta, the desire to build infrastructure, productise AI and deliver large‑scale monetisation may have conflicted with his research orientation. The optics of his exit reinforce concerns about talent turnover and alignment in Meta’s AI strategy.
Meta’s AI Trajectory and Where It Trails
Meta is clearly doubling down on AI — but bold ambitions come with risk and indicate areas of relative weakness.
Meta's Recent Moves
- In June 2025 the company formed the Meta Superintelligence Labs division to unify its AI efforts and accelerate work toward artificial general intelligence (AGI).
- Meta has committed to around $600 billion of U.S. infrastructure investment over the next three years, including data‑centres to fuel its AI ambitions.
- Its core business remains lucrative: the advertising platform is embedded with AI‑driven tools, and its scale (billions of users) remains a competitive advantage.
Where Meta's Nehind in the Field of AI
- Meta has lagged behind OpenAI and Alphabet Inc./DeepMind in the large‑language‑model race; the reception of Meta’s open‑source Llama‑4 was described as “poor”.
- The margin impact of Meta’s AI infrastructure spending has drawn scrutiny — growing capital‑expenditure and infrastructure investment are pressuring near‑term profitability.
- LeCun’s departure may exacerbate talent and leadership risks: when research luminaries leave, the signalling effect can be more than symbolic.
In short: Meta is betting big, but must convert scale and ambition into differentiated advantage. Simply spending more does not guarantee leadership.
What Meta Will Do Next — Strategic Pivot
Meta’s next phase appears to be shaped by a handful of tactical shifts and bets.
From Open‑Source to Closed Infrastructure
While Meta has historically embraced open‑sourcing models (e.g., Llama), its new approach appears to emphasise building proprietary compute, internal models and vertically integrated infrastructure through MSL. This reflects a strategic move from broad community contribution toward owning the full stack.
Monetising Existing Assets while Building for the Long Term
Meta must balance its legacy advertising business (which currently supports much of the AI investment) with building new revenue streams from AI features, products and infrastructure. Embedding AI into ads, Reels, WhatsApp, VR/AR and beyond is a near‑term priority, while next‑gen models and infrastructure build‑out represent the long‑term horizon.
Front‑loading Infrastructure Build‑Out
Meta is committing heavily now — building data‑centres, expanding compute capacity, hiring elite AI talent — with the assumption that compute will become the bottleneck in the coming AI era. While this may build future moat, it comes with near‑term cost risk and execution requirement.
Meta’s ability to navigate this transition—balancing legacy revenue, product ramp‑up, infrastructure cost and talent management—will define its success.
Meta’s AI outlook — Cautious Optimism
Looking ahead, Meta’s AI story has both compelling upside and clear risks.
Strengths
- The scale of Meta’s global user base (Facebook, Instagram, WhatsApp) provides a meaningful reach for deployment of AI‑driven features and monetisation at large scale.
- It remains profitable in its core business, providing the cash engine to support the AI pivot.
- The infrastructural commitment and strategic restructuring suggest it is positioning itself for the long‑term AI era.
Risks
- The timeline for achieving meaningful returns from these investments is uncertain — superintelligence, or even advanced large‑model leadership, is not a short‑term outcome.
- Leadership turbulence (e.g., LeCun’s exit) and internal realignment may dampen morale or slow progress.
- Intense competition from OpenAI, Google/DeepMind, Microsoft and others means Meta must not only invest but win distinct advantage.
- Near‑term margin and free‑cash‑flow pressure is real given high capex and hiring spend.
In summary: Meta has the pieces to be a major AI player—but the “crowded field + high stakes” combination makes this a high‑probability of big rewards or big disappointment scenario.
Is META a Buy now?
Fundamentals
Meta's Q3 2025 results: revenue US$51.24 billion (+26 % Y/Y) and free‑cash‑flow of US$10.62 billion. Meta's annual capital expenditure was raised to US$70‑72 billion for 2025. The company therefore continues to grow its core business while investing heavily for the future—but the high capex raises questions about near‑term returns.
Technical Outlook
While Meta’s momentum remains positive from a growth perspective, the elevated valuation (given the size of the company) and the uncertainty around execution make the risk‑reward less asymmetric than it may appear. The stock has been buoyed by AI narrative, but actual monetised AI revenue remains limited.
Peer Comparison: META vs. GOOGL (Alphabet) & MSFT (Microsoft)
Meta vs. Alphabet (GOOGL)
- According to one model, Meta’s expected revenue growth in 2025 is ~11.9% Y/Y, compared to Alphabet’s ~6.6%.
- Valuation‑wise, Meta trades at ~8.9× forward Price/Sales while Alphabet is lower, ~6.1× in the same metric.
- Performance wise, Meta’s year‑to‑date and 12‑month return trail Alphabet: Meta ~30.6% vs Alphabet ~54.9% in the latest 12‑month window.
Meta vs. Microsoft (MSFT)
- Over the past 12 months, Meta’s return (~8% according to one source) falls short of Microsoft’s ~20%.
- On valuation: Meta’s forward Price/Sales is lower than Microsoft’s (~8.6× for Meta vs ~10.9× for Microsoft) per one analysis.
- Risk profile: Meta shows higher volatility (≈13.6%) compared with Microsoft (~5.5%) in a volatility comparison.
What the Peer Data Implies
- Meta arguably offers higher growth expectations relative to Alphabet (but also at higher valuation) and somewhat competitive valuation relative to Microsoft.
- However, peers like Alphabet and Microsoft present less execution risk and broader business diversification (e.g., cloud, enterprise AI) compared with Meta’s heavier ad‑business + infrastructure pivot.
- Meta’s higher volatility and leadership / execution risk (as discussed earlier) make its premium make more sense only if the AI/infrastructure investments pay off ahead of peers.
Implied Growth Assumptions for Meta
Given Meta’s higher valuation relative to some peers, the market appears to assume:
- Continued strong growth in its core ad business (at >10% annual revenue growth)
- Successful monetisation of AI infrastructure and new business lines beyond ads
- Maintenance of high margins and profitability despite heavy infrastructure spend
- Leadership in AI infrastructure scale to justify the investment premium
If any of these assumptions falter — e.g., growth slows, margins hit by capex, execution delays — Meta is likely to under‑perform expectations.
Catalysts
Positive Triggers:
- Evidence of monetised AI features beyond the ad business (e.g., new subscription/enterprise AI revenue).
- Clear demonstration of infrastructure investment translating into improved margins or new revenue streams.
- Stable leadership and research talent retention post‑LeCun exit.
Negative Triggers:
- Delays or failures in AI monetisation, margin erosion from capex, talent attrition or strategy drift.
- Competitor breakthroughs that further widen Meta’s model‑leadership gap.
Investment Conclusion
BestStock AI leans Hold for most investors rather than immediate Buy. If you already hold META and believe in its long‑term AI transformation, staying invested may be justified given the scale, ambition and cash‑flow base. For new entries, however, patience is advisable: waiting for clearer monetisation signals, leadership stability and margin improvement would reduce risk.
Meta remains a compelling long‑term AI bet, but the near‑term path is bumpy and the premium you pay today demands execution. If you’re comfortable riding volatility and believe in the five‑to‑ten‑year horizon, taking a smaller position may make sense; otherwise, waiting for more visible delivery would be prudent.