Meta (META) on Solana

Meta has hired AI researcher Luke Metz to join its Superintelligence Labs, where he will report to Alexandr Wang, the former Scale AI CEO now leading Meta's AI efforts. The hire, confirmed to Axios in August 2026, adds another prominent researcher to Meta's fast-growing AI team as the company aggressively recruits top talent from competing labs.

Metz's recent trajectory reflects the intensity of competition for elite AI researchers: he left OpenAI in 2024 to join Mira Murati's Thinking Machines startup, returned to OpenAI earlier in 2026, and has now moved to Meta. The move continues a wave of high-profile additions at Meta since Alexandr Wang joined following a multibillion-dollar deal with Scale AI, with the company simultaneously launching multiple new models as part of its push toward superintelligence capabilities.

Opening arguments began August 20 in federal court before Judge Yvonne Gonzalez Rogers, as four state attorneys general — California, Colorado, Kentucky, and New Jersey — lead a coalition of roughly 30 states suing Meta for designing Instagram and Facebook to be addictive to young users. The states allege Meta violated COPPA by harvesting data from children under 13 without parental consent and misled the public about platform safety, demanding the court force removal of features including infinite scroll, autoplay video, disappearing content such as Instagram Stories, beauty filters, and algorithm-dominated recommendation feeds. Meta has denied all wrongdoing and called the $1.4 trillion damages figure "a money grab," arguing it already provides robust protections for minors.

The case functions as a test for litigation from 25 additional states and could set a precedent reshaping social media design industry-wide — analysts note that competitors like Snap would likely make preemptive changes to align with any court-mandated alterations. Meta has already lost two earlier youth-safety cases in Los Angeles and New Mexico, both currently under appeal, adding weight to the current trial's potential impact on the company's core consumer products.

Meta Platforms shares have fallen 27% year-to-date to around $544, placing the stock at just 73% of its 52-week high and pushing its trailing P/E to 21.2x — near the low end of its 10-year range of 12.6x to 44.9x. Jay Woods, Chief Global Strategist at Freedom Capital Markets and a regular CNBC contributor, warned this week that the apparent cheapness is deceptive: technical indicators are showing ominous signs, with implied volatility at 35 — in the 75th percentile of Meta's own trailing-year range — suggesting options markets are pricing in an unusually large move ahead.

The fundamental picture complicates the value thesis further. Meta's Q2 2026 capital expenditures reached $31.1 billion against $60.8 billion in revenue, leaving only $784 million in free cash flow. Management has guided $130–$145 billion in capex for the full year 2026 with no stated end date and declined to provide a 2027 budget, citing "highly dynamic" planning. GAAP operating income fell 8% year-to-date despite 28% revenue growth, as mounting infrastructure depreciation and legal headwinds — including a $942 million New Mexico judgment and ongoing 29-state youth addiction trials — continue to pressure the bottom line.

Meta is building data center capacity well beyond its current internal needs, and Zuckerberg has indicated that rival AI companies have already approached the company about purchasing compute access at premium pricing — a signal that Meta may be positioning itself to enter the AI cloud services market. The company spent roughly $30 billion on capital expenditure last quarter, nearly doubling its year-over-year pace, and raised its full-year capex guidance to $130–145 billion, the bulk of it earmarked for AI infrastructure.

Despite Q2 2026 revenue growing 28% year-over-year to $61 billion on the strength of advertising, Meta's operating margin compressed sharply — falling from 43% to 31% in the quarter — as depreciation from the infrastructure build-out weighs on earnings. The stock trades below $600 at a price-to-earnings ratio of around 22, below many large-cap tech peers, but the investment case hinges on whether Zuckerberg can convert surplus compute capacity into a meaningful external revenue stream to offset the margin pressure.

Meta's fair value estimate has been lowered from $828.80 to $754.10 following Q2 2026 results that revealed an earnings miss and a sharply higher capital spending commitment. The company posted $60.8 billion in quarterly revenue — up 33% year over year — but EPS of $6.18 fell short of the $7.14 analyst consensus. Management raised full-year capex guidance to $125–145 billion, almost entirely directed at AI data centers and infrastructure. The valuation revision reflects the knock-on effects: net profit margin contracted from 30.14% to 28.86%, the forward P/E multiple was trimmed from 25.11x to 22.83x, and the discount rate was nudged up to 8.91%. A court-ordered payment of roughly $942 million tied to mental health harm findings adds a further one-time drag.

Analyst views remain sharply divided on what the spending trajectory means for investors. BofA and Rothschild maintain a constructive stance, pointing to long-term AI monetization potential, small-business AI tool opportunities, and a nascent cloud computing business as reasons to look past near-term margin pressure. Scotiabank and others are more cautious, questioning whether free cash flow can stay durable as capital commitments scale ahead of earnings visibility. The core debate is one of timing: Meta is spending heavily now on infrastructure whose payoff remains uncertain in both scale and schedule.

A Simply Wall St discounted cash flow analysis estimates Meta Platforms (META) trades roughly 36.6% below its intrinsic value of approximately $937.84 per share, based on the company's last-twelve-month free cash flow of about $60.6 billion. The analysis also notes Meta's P/E ratio of 22.3x sits well below a modeled fair P/E of 38.0x — two methods independently converging on undervaluation, which the report describes as relatively uncommon.

The market discount appears to reflect uncertainty rather than deteriorating fundamentals. A New Mexico court ordered Meta to pay approximately $942 million and implement teen safety measures, adding to a broader wave of youth-safety litigation that has kept sentiment cautious. Ongoing AI infrastructure spending and continued losses at Reality Labs — estimated at $19.19 billion in 2025 — are cited as additional risks that the current price may or may not be adequately pricing in.

Meta CEO Mark Zuckerberg announced the "Future Is For Everyone Fund," committing $1 billion toward communities hosting the company's AI data centers, with allocations directed at public services, job creation, schools, utility cost stability, and environmental stewardship. The move comes as a May 2026 Gallup poll found 70% of Americans oppose AI-dedicated data centers in their neighborhoods, reflecting growing public skepticism over whether host communities receive meaningful benefits from the infrastructure build-out.

The fund represents a modest portion of Meta's record $145 billion capital expenditure plan for 2026, the majority of which is earmarked for AI infrastructure across the United States — including a $10 billion Louisiana data center that has drawn local controversy. The company also intends to sell excess computing capacity to other AI firms. Meta's decision to formalize community investment signals a broader shift in how major tech companies must negotiate public approval alongside regulatory clearance for large-scale infrastructure projects.

A seven-week federal trial opened in August 2026 at the U.S. District Court in Oakland, California, with 29 states alleging that Meta deliberately engineered Facebook and Instagram to be addictive to minors, misled consumers about platform safety, and illegally collected and used children's data in violation of federal law. Led by California, Colorado, Kentucky, and New Jersey, the coordinated multistate action is the most expansive government legal challenge ever mounted against Meta's social platforms. Judge Yvonne Gonzalez Rogers is presiding over a proceeding that employs an advisory jury — meaning the jury will issue findings on specific questions, which the judge will then use to inform her own ruling rather than treating the verdict as binding.

Meta estimates its maximum damages exposure at up to $1.4 trillion, a figure approaching the company's entire $1.5 trillion market cap, making this the highest-stakes litigation in the company's history. Beyond financial damages, the states are seeking structural remedies that could force changes to Meta's core engagement model, including mandatory age verification, elimination of infinite scroll, and algorithmic modifications that would deprioritize engagement-maximizing features for users under 18. Meta has said it "strongly disagrees" with the allegations and is "confident evidence would show its longstanding commitment to supporting young people." The case traces back to whistleblower Frances Haugen's 2021 Senate testimony, which prompted 29 state attorneys general to file a coordinated complaint in 2023 — a separate but related action from the $567 million abatement fund ordered in New Mexico last week.

Meta released Muse Glimmer, a compact AI model designed for agentic tasks that can run on consumer hardware — a Mac or PC with a single GPU — without requiring cloud infrastructure. The model extends Meta's open-weight strategy, positioning the company against closed-source rivals such as OpenAI, Anthropic, and Google while competing directly with Chinese open-weight leaders including Alibaba's Qwen3.8-Max, Moonshot's Kimi K3, and DeepSeek's V4-Flash.

Alongside the launch, CEO Mark Zuckerberg called for U.S. policy changes to reduce regulatory friction on AI training data, arguing that "foreign labs currently hold several advantages here since American labs have to comply with many additional restrictions." Zuckerberg stated that restricting access to foreign open-source models was not an effective solution and that U.S. policy must lower barriers if domestic open-source models are to maintain a lead. Meta also disclosed plans to pursue AI model distillation and said it would give independent directors authority over model safety release criteria as part of new governance structures.

Following Meta's earnings miss, ARK Invest moved roughly $15 million into Nvidia across five ETFs — with $8.1 million channeled through ARKK alone — and another $14.7 million into Taiwan Semiconductor across four funds on consecutive trading days. Rather than buying Meta directly, Cathie Wood's team treated the results as a confirmation that AI infrastructure demand stays elevated: Meta raised its 2026 capex guidance to $130–145 billion (up from a prior range of $125–145 billion) and reported $31.1 billion in Q2 capital outlays, with management saying the company is "supply-constrained" as it builds out model training, agent infrastructure, and data center capacity.

The positioning reflects a supply-chain read-through on Meta's spending: Nvidia designs the GPUs; TSMC manufactures them. Both stocks were trading near 25 times forward earnings at the time of purchase, which ARK characterized as well below prior AI-cycle peaks. The institutional response underscores that Meta's capex ambition — even where it pressured margins and drove the EPS miss — is being interpreted by at least some fund managers as durable demand signal for the broader AI hardware stack, rather than a reason to fade the name.