To gauge where retail sentiment stands on major AI stocks, we scoured Reddit investing communities where retail traders debate stock picks and swap theses in real time. Meta Platforms (NASDAQ:META) was a clear favorite of many long-term retail investors. The stock is down roughly 4% this year, but many Redditors believe now is the time to buy.
What Meta Does and Why It’s Down
Meta owns Facebook, Instagram, and WhatsApp, and makes almost all its money selling ads across them. The drop this year traces to one thing: Mark Zuckerberg is pouring billions into AI data centers, and investors aren’t sure it pays off.
Why Bulls Are Buying
The simple case: Meta’s revenue grew 33% year over year last quarter. One Redditor put it bluntly: “I’m at $300 cost basis and am never selling META. I’ll just buy the dip and keep buying.” Bulls argue the market is punishing capex fear, not the actual business.
There’s a more specific version of this bull case too. Meta is reportedly exploring leasing excess GPU capacity to third parties and has already struck a $21 billion expanded AI infrastructure deal with CoreWeave, plus talks for a compute deal worth up to $10 billion with Anthropic over two years. Bulls see this as Meta turning a cost center into a second revenue stream, becoming a “neocloud” instead of just a spender. That matters because the market has historically distrusted Zuckerberg’s capital allocation after the Metaverse bet, and some bulls argue this pivot is the first capex-related news the market has actually cheered rather than punished. Price targets from bulls run as high as $1,011 and around $1,000, with Meta Glasses flagged as a second catalyst as smart-glasses adoption ramps.
Why Bears Aren’t Buying It
The bear case centers on a competitive threat from China. Moonshot AI, a Beijing-based lab, just released Kimi K3, a new 2.8-trillion-parameter AI model with a 1-million-token context window and native image understanding, priced at $3 per million input tokens and $15 per million output tokens. Moonshot is making the model “open-weight,” meaning anyone will be able to download and run it themselves once the full weights are released on July 27. Benchmark testing shows Meta’s own AI model, Muse Spark, scoring below Kimi K3, as well as below GPT-5.6 and Claude Fable 5. Bears argue Kimi K3’s release accelerates the commoditization of mid-tier AI models and undercuts Meta’s pricing power in anything beyond its own ad platform.
Bears also point to history: Reality Labs, the metaverse division Zuckerberg renamed the company after, has burned through more than $80 billion in cumulative losses since 2020. Stock-based compensation ran $20.4 billion in 2025, close to 44% of free cash flow. One Redditor summed it up: Meta is “dead money” until capex stops. One bear rating puts fair value at just $570 to $635, well below where the stock trades today.
