Social media is now a massive liability for Meta, Google and the rest of Big Tech
For more than a decade, social platforms were Silicon Valley’s perpetual motion machine. The promise was simple and wildly profitable: aggregate billions of users, target them with precision, and sell advertisers measurable results on an unprecedented scale. That engine still throws off enormous cash. Yet the balance has flipped. For Meta, Google, and the broader Big Tech cohort, social media has morphed from crown jewel to strategic liability—financially, politically, and operationally.
The bargain that broke
The original bargain powering social media looked unbeatable:
– Users traded data for free services and endless content.
– Platforms used that data to build superior ad products.
– Advertisers enjoyed performance and reach at bargain prices.
Three things unraveled the deal. First, regulators tightened the privacy vise, restricting data flows and raising compliance costs. Second, user growth matured in key markets just as engagement patterns shifted to short-form video and private messaging—formats that are harder to monetize. Third, the signal that made targeted ads so effective degraded, thanks to platform privacy changes, looming cookie deprecation, and a hostile policy climate around tracking. The result is a business that still scales, but with thinner unit economics, higher risk, and a political target on its back.
The ad signal collapse
Apple’s App Tracking Transparency turned a knob and billions in performance ad signal disappeared overnight. Meta alone told investors it expected a roughly $10 billion revenue impact in 2022. Subsequent workarounds—modeled conversions, aggregated measurement, and on-device processing—are partial, expensive, and fragile. Google’s evolving Privacy Sandbox and the long and winding road toward third‑party cookie deprecation keep the ecosystem unsettled, depressing confidence in cross-site measurement.
Meanwhile, ad dollars are migrating. Retail media networks (Amazon, Walmart, and a parade of grocers and pharmacies) offer purchase data and closed-loop attribution without the privacy baggage. Connected TV gives brand advertisers scale with relatively cleaner pipes. And TikTok siphoned attention with an interest graph that doesn’t depend on your social circle—further decoupling the ad engine from the social graph Big Tech painstakingly built.
Social’s monetization mix has also deteriorated. Short-form video fuels usage but carries structurally lower ad prices and revenue shares. On Meta, Reels cannibalized higher-ARPU time in Feed and Stories before the monetization gap began to narrow. On YouTube, Shorts drove necessary defensive engagement at the cost of premium long-form viewing. Creator payouts and rev-share expectations keep rising. It all adds up to less efficient ad dollars per minute of attention.
From nuisance to balance-sheet risk: regulation
Compliance is no longer a background cost. It is a recurring balance-sheet risk with teeth.
– Europe’s Digital Services Act forces risk assessments, algorithmic transparency, researcher access, and faster illegal-content takedowns, backed by fines up to 6% of global turnover. The obligations scale with user base, making social giants uniquely exposed.
– GDPR enforcement has matured from warning shots to multibillion-euro fines, including for cross-border data transfers—core to ad targeting and content delivery.
– The UK’s Online Safety Act, various state-level US youth safety bills, and global age-verification pushes create hard product trade-offs around encryption, design, friction, and speech.
– Content moderation is whiplash politics: the left demands more removal, the right demands less. Texas- and Florida-style laws that pressure platforms to carry content collide with European requirements to take more down. There is no single compliance position that satisfies all jurisdictions.
– Section 230’s protections remain, but the political consensus that social platforms are “special” has faded. The next crisis could trigger real erosion and fresh liability theories—especially around product design harms to minors.
Every one of these regimes generates expensive internal systems: compliance staffing, localized moderation in hundreds of languages, auditing and reporting pipelines, appeals processes, and data access portals for researchers and regulators. In the same period, AI-generated content has multiplied the surface area of potential harm—and therefore the scope of required safety tooling.
The impossible politics of moderation
Social networks are stuck in a no-win triangle:
– Users want free expression, safety, and instant service—simultaneously.
– Advertisers want high engagement but low controversy and near-perfect brand safety.
– Governments want platform-scale broadcast power when it favors them, and firm controls when it doesn’t.
Election cycles magnify every tension. Deepfake political videos, automated astroturfing, and cross-border influence operations are cheaper and more convincing each year. Platforms are forced into roles that look and feel like editorial judgment, even as they argue they are neutral conduits. That is expensive, thankless, litigious work that erodes trust with every decision.
AI turns friction into flood
Generative AI supercharges production. Posts, comments, videos, synthetic voices, and images can be spun up at industrial scale. This is not simply more content—it is content with intent. Spam, scams, and influence campaigns become abundant and bespoke. Identity verification, provenance, and authenticity move from nice-to-have to existential.
The bill lands on social platforms. They must detect, label, throttle, or remove at petabyte pace without crushing legitimate creativity. Watermarking and content authenticity frameworks are promising but unevenly adopted. Over-moderation angers users and governments; under-moderation scares advertisers and invites fines. The automation arms race ratchets costs while user trust erodes.
The economics are slipping
Even when top-line revenue grows, the underlying economics of social are getting tougher:
– Time is shifting to private messaging and DMs—great for retention, poor for ads.
– Short-form video monetizes worse per minute than feeds or long-form.
– Creator economies require richer tools and bigger payouts to stay; otherwise talent churns to platforms with better rev shares or commerce rails.
– Subscription experiments can reduce ad load but rarely replace lost ad dollars at scale; paywalls kneecap virality.
– Ad load is near natural limits on the biggest platforms, leaving price and targeting efficiency as the remaining levers—both under pressure.
Operating leverage is also less reliable. Every incremental country and demographic cohort adds compliance overhead, trust-and-safety coverage, and political exposure. The marginal cost of a new user is no longer just hosting and bandwidth; it’s governance.
Geopolitics and fragmentation
The internet that made social platforms global is fragmenting. Data localization, government takedown orders, and demands for “local representation” complicate operations in India, Turkey, and Brazil. TikTok faces recurring threats of bans or forced divestiture in Western markets. Even where services continue, the cost of compliance rises and the scope of permissible content narrows. Social platforms are being carved into regional variants with different rules, products, and risks—destroying the efficiencies that once defined them.
Company by company
Meta: Still the archetypal social company, Meta has bought time with operational efficiency and world-class AI ranking that made Reels competitive with TikTok. But it sits under a thicket of lawsuits alleging harms to children, shoulder-tapped by regulators in every major jurisdiction, and tied to a legacy social graph model that matters less in a world of interest-first feeds. WhatsApp is Meta’s cleanest asset—massive, sticky, and relatively controversy-free—but remains under-monetized outside a few markets and commerce use cases. Reality Labs soaks up cash and headlines; the real drag on the multiple is the policy and legal overhang on the core social business.
Google: YouTube is both social and television. Shorts preserved share but compressed monetization and pushed Google deeper into creator politics. Kids content continues to raise COPPA and regional child-safety liabilities. Brand safety flare-ups periodically drive advertiser pullbacks. Add in DSA obligations in the EU and you have a property that’s strategically essential yet much messier than Search or Cloud. For an investor base increasingly captivated by AI and enterprise software margins, social video looks like a riskier, lower-multiple component.
Snap: A product-focused underdog that’s often first to feel industry shocks. Apple’s privacy changes hit Snap early and hard. AR bets remain interesting but capital intensive. The company’s audience is valuable but sensitive to any missteps in youth safety—creating a similar liability profile with fewer resources to absorb shocks.
X (formerly Twitter): Perhaps the starkest case study of social as liability. The dynamics of real-time public discourse, light policy infrastructure, and heavy political attention make brand safety elusive. Advertiser churn shows how quickly the value of reach erodes when trust breaks.
Reddit and Discord: Community models are sticky and valuable, but moderation at scale is a never-ending grind. Both have turned to data licensing for AI model training—an implicit acknowledgment that the data exhaust may be worth more, and less contentious, than traditional social ads.
LinkedIn: The partial exception proves the rule. A professional identity network with clear utility, subscription revenue, and lower-stakes discourse carries far less policy heat and steadier economics.
Why they can’t just walk away
If social is such a headache, why not exit? Because social is distribution. It is where cultural moments happen, where creators build audiences, where small businesses find customers, and where new products are discovered. For Meta and Google, social funnels demand into commerce and search; without it, their moats shrink. For Apple and Amazon, owning the rails is more attractive than owning the square, but even they rely on social to energize ecosystems. Big Tech can de-emphasize social, but it cannot ignore it.
Strategies to reduce the drag
– Shift the center of gravity to utility. Messaging, payments, and commerce carry fewer political fireworks and clearer user value. WhatsApp Business, click-to-message ads, and integrated storefronts are logical paths.
– Invest in authenticity infrastructure. Verified identities, provenance metadata, and watermarking for AI media won’t solve everything, but they harden the ecosystem against abuse and restore some advertiser confidence.
– Localize governance. Not just more moderators, but product features and enforcement calibrated to local law and culture, with transparent appeal paths and auditability.
– Rethink incentives with creators. Sustainable rev shares, commerce integrations, and financing tools keep premium content on-platform without racing to the bottom on payouts.
– Build privacy-first ad stacks. On-device modeling, cohort approaches, and first-party measurement can claw back performance without violating the spirit of privacy laws—albeit at lower ceiling efficiency.
– Consider structural separation or clearer reporting. If social is dragging the multiple, ring-fencing it—operationally or financially—can surface the value of less controversial businesses like cloud, enterprise software, or messaging.
The investor lens
The question is not whether social media will keep making money. It will. The question is what multiple that cash flow deserves given the policy, legal, and reputational tail risks; the rising cost of trust and safety; and the structural headwinds to ad efficiency. When paired with attractive, lower-risk growth stories—cloud, AI infrastructure, enterprise subscriptions—social now looks like the boat anchor rather than the sail.
What to watch:
– Enforcement, not just legislation: DSA actions, GDPR fines, and youth-safety cases will set precedents and price the risk.
– Measurement stability: Any credible, privacy-respecting replacement for cross-site tracking that wins regulator approval could lift social ad ROAS and sentiment.
– AI provenance standards: Broad adoption of content authenticity frameworks would contain some of the risk explosion.
– Creator economics: If platforms can make creators net better off without catastrophic margin erosion, the monetization mix improves.
– Geopolitical shocks: Another wave of bans, blocks, or forced divestitures would crystallize the fragility of global social businesses.
The paradox of scale
Social platforms still command human attention like few institutions on earth. But scale is now a double-edged sword: every additional user, language, and election amplifies both value and liability. For Meta, Google, and their peers, the social era is not ending; it’s aging. The profits remain, but the premium is gone. The wise strategy is to keep harvesting the cash while building businesses whose economics and politics are less combustible.
That is the new reality: social media is no longer the engine. It is the drag coefficient that every other Big Tech ambition must overcome.
