Meta (META) on Solana

Meta Platforms shares fell roughly 10% on July 30 after the company reported Q2 earnings that missed profit expectations by a wide margin. Diluted EPS came in at $6.18 against the $7.22 analysts had forecast, as operating expenses surged 55% year-over-year to $42.03 billion — swollen by $2.40 billion in legal fees and $1.18 billion in severance from May's layoffs. Operating margin compressed sharply from 43% to 31%, net income fell 14% to $15.85 billion, and free cash flow collapsed 91% to just $784 million despite $31.86 billion in operating cash flow, reflecting the scale of capital being consumed by AI infrastructure.

The selloff deepened on forward guidance: Meta raised its full-year capital expenditure range to $130–145 billion and increased full-year expense guidance to $165–169 billion, signaling no near-term relief on spending. Q3 revenue guidance of $61–64 billion (midpoint $62.5 billion) also came in below the $63.15 billion Wall Street had penciled in. The results highlighted a diverging investor verdict on Big Tech AI bets — Microsoft jumped 8% the same day after Azure crossed $100 billion in annual revenue and Copilot reached 30 million paid seats, demonstrating clearer returns. For Meta, the market's question is whether CEO Mark Zuckerberg's aggressive AI buildout will translate into monetizable products before the cost structure erodes further.

Meta Platforms and BlackRock have announced a joint venture to develop a data center campus in El Paso, Texas, valued at approximately $14 billion. Under the deal, BlackRock-managed funds will hold an 80% stake while Meta retains 20%; BlackRock is contributing roughly $4.9 billion in cash and raising $12.5 billion in bond financing, while Meta is contributing land and in-progress construction assets worth about $2.3 billion and will receive a $1 billion distribution to align ownership stakes. The facility is designed to deliver one gigawatt of compute capacity and is expected to begin operations in 2028.

The partnership is part of Meta's broader push to scale AI infrastructure, with the company targeting $600 billion in data center investment by 2028. Meta is building multiple gigawatt-scale facilities across the United States, including a Louisiana project that could eventually expand to five gigawatts, as it pursues what it describes as personal superintelligence applications. The El Paso joint venture structure allows Meta to unlock capital from existing construction assets while securing long-term access to the compute needed for its AI operations.

Meta Platforms shares closed Friday at $595.19, slipping below the $600 level ahead of its Q2 2026 earnings report due after market close on Wednesday, July 29. The stock is now down roughly 25% from its 52-week high of $796.25 and trades at approximately 22 times earnings — a multiple below the S&P 500's ~28.5x, a notable inversion given Meta's stronger growth profile.

The valuation discount appears driven less by advertising weakness than by investor anxiety over AI infrastructure spending. Meta guided Q2 revenue of $58–61 billion (implying 22–28% year-over-year growth, down from Q1's 33% pace) while raising its full-year capex range to $125–145 billion. Q1 fundamentals remained solid — ad impressions rose 19%, average price per ad climbed 12%, family daily active users reached 3.56 billion, and the company generated $12.39 billion in free cash flow while holding $81.18 billion in cash and securities. Wednesday's report will be closely watched for any clarity on the return on that AI capital investment and whether revenue growth can sustain a pace that justifies the spending trajectory.

Meta has launched a dedicated Seller app for Facebook Marketplace, marking a significant escalation in the company's e-commerce ambitions. The standalone app targets professional sellers with enhanced tools and the prospect of lower fees, positioning Facebook Marketplace as a serious alternative to established platforms and signaling that Meta is moving beyond its social-commerce roots toward a more structured, seller-first marketplace model.

The move puts direct pressure on eBay's high-margin professional seller segment and raises broader questions about Meta's long-term monetization strategy. With billions of monthly active users already browsing Facebook and Instagram, Meta has a built-in audience advantage; a purpose-built seller experience could accelerate its share of the $1 trillion-plus global e-commerce market. Analysts will be watching whether Meta's entry into dedicated seller tooling translates into meaningful take-rate revenue, or whether it primarily serves as a retention and engagement play for its advertising business.

Meta joined more than 20 technology companies — including Nvidia, Microsoft, Palantir, IBM, Hugging Face, and Mistral — in a letter to the Trump administration urging policymakers to reject "premature restrictions" on open-weight artificial intelligence models. The coalition warned that broad curbs on open-weight models, which can be downloaded, modified, and deployed on independent infrastructure, would harm American competitiveness and risk driving AI innovation overseas. The letter came amid growing concern within Washington over Chinese AI development, including White House accusations that Moonshot AI engaged in "large-scale, covert industrial distillation" to build its Kimi K3 model.

As the developer of the Llama model family — among the most widely deployed open-weight AI systems in the world — Meta has a direct stake in how the US regulates open-source AI distribution. The joint letter argued that open-weight models allow a broad community of researchers to examine model behavior, identify vulnerabilities, and build safeguards, and that "concentrating advanced AI capabilities behind a small number of closed models" compounds rather than reduces security risk. Signatories called for targeted legal frameworks addressing IP theft rather than sweeping technical restrictions on open-weight development. OpenAI and Anthropic, both reportedly preparing IPO filings, did not sign the letter.

Jefferies analyst Brent Thill reiterated a Buy rating and $825 price target on Meta, arguing its AI-powered smart glasses represent a $14–18 billion hardware revenue opportunity over the next several years. The estimate assumes 35–45 million unit sales at an average selling price of $400, roughly mirroring Apple Watch adoption curves. Thill flagged that daily active users of Meta's eyewear have tripled year-over-year, and EssilorLuxottica is doubling production capacity to 20 million units in 2026, with the lineup spanning Ray-Ban Meta Gen2 ($379), Oakley Meta ($499), and the Ray-Ban Display with Neural Band ($799).

Beyond device sales, Jefferies identified AI subscriptions, advertising integration, and commerce monetization as additional revenue layers tied to the glasses platform, potentially making wearables one of the largest sources of long-term upside for Meta. The analysis underscores how hardware is becoming a meaningful revenue diversifier for a company whose revenue base has historically been dominated by social media advertising.

Meta's stock surged to an intraday high of $686 before closing the week at $646, restoring Mark Zuckerberg's net worth to approximately $222 billion and pushing him to fifth place on Bloomberg's global wealth ranking, trailing only Elon Musk, Larry Ellison, Sergey Brin, and Jeff Bezos. Zuckerberg holds a 13.5% ownership stake in Meta, making him a direct beneficiary of the rally.

The rebound has been driven in part by reports that Meta is exploring a pivot toward monetizing its AI infrastructure by leasing compute capacity to other companies, with potential contracts valued at up to $10 billion annually. The shift addresses longstanding analyst concerns about Meta's return on its roughly $140 billion annual capital expenditure, which had weighed on free cash flow relative to reported net income.

Jury selection began July 20 in Nashville for a landmark trial brought by Tennessee Attorney General Jonathan Skrmetti, who alleges Meta violated the state's consumer protection law by intentionally engineering Instagram to drive compulsive use among teenagers. The suit claims Meta knowingly deployed features — including autoplay, Reels, aggressive notifications, and disappearing content — to maximize engagement while suppressing internal research showing those same features caused harm. Skrmetti's office alleges CEO Mark Zuckerberg was directly warned by employees about negative teen impacts and declined to fund mitigation efforts, and that the company publicly misled consumers about platform safety.

The Tennessee case is one of several converging legal threats: a separate trial involving an individual plaintiff against Meta and Snap is scheduled to begin July 27, and a multi-state federal trial in California is set for August 18. Nearly every U.S. state has filed similar claims, and a New Mexico jury already returned a $375 million judgment against Meta earlier this year for misleading consumers on the same safety issues. If the Tennessee jury finds a violation, a second phase will assess monetary penalties — potentially $1,000 per violation under state consumer protection law — and could require platform modifications. Meta is contesting liability on Section 230 grounds, arguing the law shields it from liability over user-posted content.

Meta Platforms gained 21% in July through July 16, adding roughly $270 billion in market capitalization and outpacing every other Magnificent Seven stock. The move reversed an 11% decline in June and pushed Meta's price-to-earnings multiple from 20.5x to 24.4x, reflecting a meaningful re-rating in investor expectations around the company's AI strategy.

Two themes are driving the recovery. First, Meta's massive AI infrastructure buildout is being reframed as a potential revenue source: the company is developing a "Meta Compute initiative" to sell excess compute capacity and host AI models for external developers, opening a cloud services line alongside its advertising core. Second, Meta has guided 2026 capital expenditures to $125–145 billion, with Q1 2026 spend alone reaching $19.8 billion—about 61% of operating cash flow. Investors appear willing to reward the ambition so long as management can articulate credible return timelines; guidance clarity on the cloud initiative is seen as the next key catalyst for sustaining the rally.

Meta is in preliminary talks to lease computing power to Anthropic in an arrangement that could be worth approximately $10 billion, according to reporting from CNBC and the New York Times. The discussions are at an early stage and terms could change, but the deal would mark a significant step in Meta's push to commercialize its AI infrastructure spending. CEO Mark Zuckerberg indicated as recently as May that companies routinely approach Meta asking to purchase spare compute capacity, and the company has since hired Dave Brown, a former senior AWS executive, as it explores a formal cloud business. Meta is on track to spend up to $145 billion in capital expenditure on AI infrastructure in 2026, making excess capacity monetization increasingly attractive to investors.

For Anthropic, the talks underscore how severe the compute bottleneck remains even as the company approaches a $1 trillion valuation. Anthropic already reached a roughly $1.25 billion per month agreement to use SpaceX's Colossus 1 data center GPUs, and still enforces usage caps on its most advanced Claude models due to chip scarcity. The potential Meta arrangement would add another large-scale GPU source outside the traditional hyperscaler route. The cross-competitive dynamic — Meta's Llama models directly compete with Anthropic's Claude — is notable but has become increasingly common as compute scarcity normalizes supply deals between AI rivals.