Author: Jerry Ng

I. The Payments Value Chain: Layers, Drivers, and Key Edges

Think of the payments ecosystem as a stack, from foundational network rails up to merchant-facing software. Each layer has its key drivers — capabilities that confer an edge in the market — and specific market share leaders or standouts.

1. Payment Networks & Issuers
   • Examples: Visa, Mastercard, AmEx, Discover, plus issuing banks (Chase, Citi).
   • Key Drivers/Edges:
     • Global Acceptance & Trust: The biggest networks have decades of reliability, brand recognition, established regulatory compliance.
     • Interchange & Fee Structures: They wield substantial pricing power to set interchange rates.
     • Fraud & Risk Tools: Investment in security protocols like EMV, tokenization.
   • Market Share: Visa and Mastercard together hold the lion’s share of global card-based payments (often cited at 70-80%+ in many markets).
2. Merchant Acquirers & Processors
   • Examples: FIS, Fiserv (First Data), Global Payments, Chase Merchant Services, Adyen, Stripe, PayPal/Braintree, dLocal.
   • Key Drivers/Edges:
     • Scale & Reliability: Handling billions in GMV, ensuring minimal downtime.
     • Global vs. Local Coverage: Some excel at cross-border (Adyen, dLocal); others focus on US scale (Stripe, PayPal).
     • Pricing & Value-Added Services: Fraud detection, BNPL, analytics.
   • Market Share:
     • Adyen: ~2–3% of total global acquiring volume, but ~20% share in large enterprise e-commerce (cross-border).
     • Stripe: rumored $1T+ in annual processing; strong with SMBs and developer-led integrations.
     • PayPal: $400B+ quarterly TPV includes branded & unbranded (Braintree).
     • dLocal: specialized in emerging markets, smaller absolute share, but rapidly growing (~40% yoy revenue).
3. Commerce Enablement Software
   • Examples: Shopify, NCR Voyix (retail/hospitality), Toast (restaurants), BigCommerce, Salesforce Commerce Cloud, WooCommerce.
   • Key Drivers/Edges:
     • Ease-of-Use & Integrations: For SMBs, user-friendliness + plug-ins. For enterprise, robust APIs and composability.
     • Vertical Depth: Toast dominates restaurants with features like menu engineering, table management; NCR excels with large chain retailers.
     • Ecosystem & Network Effects: Shopify has an extensive app ecosystem, driving lock-in for merchants.
   • Market Share (rough approximations):
     • Shopify: ~10% of U.S. e-commerce GMV, behind Amazon’s ~40%.
     • Toast: Possibly ~10–15% of U.S. restaurant POS in the SMB/mid-market range.
     • NCR: A legacy giant for big retailers and quick-service restaurants (QSR), with an extensive installed base of terminals and enterprise relationships.
4. Marketplaces & Aggregators
   • Examples: Amazon, eBay, Etsy.
   • Key Drivers/Edges:
     • Massive Consumer Demand: Aggregators centralize product search, user trust, and logistics.
     • End-to-End Control: Amazon’s logistics (Fulfillment by Amazon) + ad platform.
   • Market Share:
     • Amazon controls ~38–40% of U.S. e-commerce market share, making it the leading aggregator.
5. Stablecoins & Crypto Rails(new/emerging layer)
   • Examples: Tether, USDC, Stripe’s Bridge.
   • Key Drivers/Edges:
     • Instant, Low-Fee Cross-Border: Potential to bypass slow bank wires in developing or high-inflation markets.
     • Programmability: Could enable micropayments or automated escrow.
   • Market Share: Hard to quantify, but Tether (USDT) and USDC dominate stablecoin usage. Stripe’s Bridge aims to make stablecoin usage seamless for mainstream businesses.

II. Recent M&A: Motivations from Both Sellers and Buyers

Below are key deals illustrating why each side transacted and how it aligns with (or departs from) their core strategies. After each listing, we’ll connect it to the value chain layer.

1. Shopify & Deliverr (Sold to Flexport)

• Original Deal:
  • Buyer (2022): Shopify paid $2.1B for Deliverr, a fulfillment startup.
  • Seller: Deliverr specialized in “fast shipping” software + network for small merchants.
  • Buyer’s Rationale: Shopify wanted to build its own logistics network (vertical integration) to match Amazon-level fulfillment for Shopify merchants.
• Later Divestiture:
  • In 2023, Shopify sold the bulk of its logistics assets (including Deliverr) to Flexport.
  • Seller’s (Shopify) Rationale: Realized that logistics is capital-intensive with less synergy for a purely software-based model. Freed up resources to focus on commerce software.
  • Buyer’s (Flexport) Strategy: Flexport is a global freight forwarder and supply chain software firm. Acquiring Deliverr fits their ambition to manage end-to-end shipments for e-commerce sellers.
• Layer: Commerce Enablement + Logistics Side Quest.
• Why It Makes Sense:
  • Shopify = “asset-light” pivot.
  • Flexport = deepening e-commerce fulfillment capabilities beyond freight forwarding.

2. Stripe & Bridge

• Buyer: Stripe, a leading online payment processor.
• Seller: Bridge, a stablecoin infrastructure startup with multi-jurisdictional licenses (in 48 U.S. states + EU).
• Deal Value: $1.1B, rumored ~20–25x forward revenue.
• Buyer’s Rationale:
  • Accelerate cross-border stablecoin adoption (Bridge’s compliance + licensing = huge head start).
  • Expand Stripe’s “platform of platforms” strategy into near-instant settlement, micropayments, emerging market payouts.
• Seller’s Rationale:
  • Gains Stripe’s global scale and huge developer base, ensuring Bridge’s tech is quickly commercialized.
• Layer: Merchant Acquirer/Processor + Stablecoin Rails.
• Why It Makes Sense:
  • Stripe invests in next-gen payment rails to strengthen developer offerings, while Bridge needs a big distribution partner to achieve mainstream impact.

3. NCR ATM Sale to Veritas Capital

• Seller: NCR, rebranded as NCR Voyix, historically known for hardware (ATMs, POS).
• Buyer: Veritas Capital, a private equity firm focusing on technology-enabled businesses.
• Deal Size: Estimated around $2.5B in 2023.
• Seller’s (NCR) Rationale:
  • Unloading the ATM business to become more SaaS and software-driven in retail/hospitality.
  • Aims for higher valuation multiples typical of software-led businesses.
• Buyer’s (Veritas) Strategy:
  • Acquire the stable, cash-generating ATM unit. Possibly streamline operations or combine with other portfolio assets in financial services.
• Layer: Commerce Enablement hardware vs. pure software pivot.
• Why It Makes Sense:
  • NCR can now focus on recurring-revenue POS solutions (Voyix) for hospitality/retail.
  • Veritas sees potential in an established ATM network that can still yield steady returns.

4. PayPal & Braintree

• Buyer: PayPal (2013).
• Seller: Braintree, a fast-growing unbranded processor powering Uber, Airbnb, etc.
• Deal Size: $800M (~19x forward revenue at the time).
• Buyer’s Rationale:
  • Gain “unbranded” gateway for enterprise merchants, funnel them into PayPal’s ecosystem for potential upsell (PayPal wallet, Venmo).
• Seller’s Rationale:
  • Access PayPal’s large merchant base, global scale, compliance resources. Braintree was smaller and could accelerate expansion with PayPal backing.
• Layer: Merchant Acquirer/Processor.
• Why It Makes Sense:
  • PayPal addressed a gap in high-volume, enterprise processing. Braintree overcame scaling challenges and gained brand synergy.

5. Toast’s Small Tuck-Ins

• Buyer: Toast, the restaurant-focused SaaS + payments provider.
• Sellers: Various small software startups in scheduling, inventory, back-of-house automation. Often deals <$100M.
• Buyer’s Rationale:
  • Bolster its vertical stack to be the “one-stop shop” for restaurants—beyond just payments/POS, also staff scheduling, menu analytics.
• Sellers’ Rationale:
  • Integrate into a leading restaurant POS ecosystem for broader distribution.
• Layer: Commerce Enablement (vertical).
• Why It Makes Sense:
  • Toast cements its advantage in a specialized domain, while acquired startups get scale and synergy.

6. Adyen: No Major M&A

• Strategy:
  • Prefers organic expansion. Adyen invests internally in R&D for a unified, single platform.
• Rationale:
  • Maintains control over technology stack, ensuring consistent codebase and margins.
• Layer: Merchant Acquirer/Processor.
• Why This Makes Sense:
  • Adyen’s brand is about reliability, simplicity, and synergy across geographies. M&A can introduce integration complexity.

III. Key Edges/Drivers in Each Layer — Plus Standouts

Payment Networks/Issuers:

• Drivers: trust, acceptance, network scale, interchange leverage.
• Standouts: Visa (~50% share of U.S. credit cards), Mastercard (~31%).
• Edge: Ubiquity. Hard to displace, though stablecoins might nibble at cross-border usage.

Merchant Acquirers/Processors:

• Drivers: broad coverage, advanced risk/fraud, local payment support, developer APIs.
• Standouts:
  • Stripe (SMB + developer-friendly, strong brand).
  • Adyen (global enterprise, high authorization rates).
  • PayPal/Braintree (massive user base, brand trust).
  • dLocal (emerging markets, local compliance).
• Edge: Speed, reliability, specialized features (e.g., stablecoins with Stripe, cross-border local methods with dLocal).

Commerce Enablement Software:

• Drivers: user experience, SaaS ecosystem, vertical integration (inventory, order mgmt.), channel expansion.
• Standouts:
  • Shopify: major e-commerce platform with ~10% US e-comm share, strong brand in SMB and moving upmarket.
  • Toast: 10–15% share in restaurant POS, highly tailored features.
  • NCR Voyix: large chain retailers/fast-food presence, pivoting to SaaS.
• Edge: deeper operational solutions = loyalty from merchants. R&D or acquisitions can further lock in vertical needs.

Marketplaces & Aggregators:

• Drivers: huge shopper base, frictionless shipping, trust.
• Standouts:
  • Amazon: ~40% of US e-commerce, overshadowing smaller marketplaces.
  • eBay, Etsy: niche or specialized aggregator.
• Edge: scale, brand, integrated ad & logistics. Potential to become a “payments-as-a-service” (e.g., Buy with Prime).

Stablecoins & Crypto Rails:

• Drivers: compliance, low-cost cross-border, speed, bridging on/off fiat ramps.
• Standouts: Tether (USDT), USDC, now Stripe Bridge bridging stablecoin utility for mainstream.
• Edge: early compliance and licensing can create a moat if mainstream adoption grows.

IV. Market Share Summaries

• Global E-Commerce: ~$5 trillion total in 2022, Amazon capturing ~40% in the US. Shopify is the #2 in the US with ~10%.
• Merchant Acquiring: Traditional incumbents (Fiserv, Global Payments, FIS) hold large slices but are less visible in pure online. Stripe and Adyen are strong in e-commerce, each with billions in GMV.
• Restaurant POS: Toast competes with Square/Block and legacy NCR. Toast is rapidly gaining share (~10–15%), Square is also popular in micro-merchants, while NCR still dominates big chains.

V. Why These Insights Make Analyzing Companies Easier

1. Layer Identification:
   • First, identify which layer(s) a company operates in (networks, acquiring, commerce software, marketplace, or stablecoins).
   • Check synergy: is it purely software (Shopify)? Is it bridging multiple layers (Amazon, Stripe)?
2. Key Drivers/Edges:
   • Evaluate if a firm’s “edge” is aligned with its growth strategy. (e.g., Toast’s vertical integration -> deeper restaurant loyalty; Stripe’s stablecoin push -> cross-border expansion.)
3. M&A and Divestitures:
   • Are they strengthening a core advantage (Stripe acquiring Bridge) or retreating from a low-synergy business (Shopify’s logistics exit, NCR’s ATM divestiture)?
   • Do the multiples paid or received seem justified by the synergy or capital relief?
4. Market Share & Standouts:
   • Understand if a firm is a dominant player (Amazon, Visa) or a specialist (dLocal, Toast) with high growth potential.
   • Market share can indicate moat (e.g., Amazon’s aggregator model) or a niche that’s under-penetrated (dLocal in emerging markets).

VI. Final Thoughts: Putting It All Together

• Focus on Core: Many “purely software” players prefer asset-light strategies to preserve margins, seen with Shopify and NCR’s spin-offs.
• M&A That Makes Sense: Aligns with the buyer’s main quest (e.g., Stripe + stablecoin infrastructure, Toast + restaurant-specific tuck-ins) and offers the seller scale or a capital exit.
• Ecosystem Partnerships: Common in commerce enablement (Shopify & Stripe) but fierce competition within each layer (Stripe vs. Adyen vs. PayPal for the same merchants).
• Stablecoin Outlook: Stripe’s Bridge signals an emerging push for frictionless cross-border, though broad adoption depends on regulatory clarity.

By mapping who does what in the value chain and why they acquire or sell certain assets, it becomes far easier to analyze each company’s strategy. You can quickly see if they’re reinforcing their profitable core (Shopify’s divestiture, NCR’s pivot, Stripe’s stablecoin bet) or venturing into new territory that either expands synergy—or risks a distraction.

In Summary

• Value Chain Layers: From card networks to stablecoins, each layer has unique success drivers.
• Companies: Many either focus on one layer (Toast for restaurants, dLocal for emerging-markets acquiring) or straddle multiple (Amazon, PayPal).
• M&A: Tends to succeed when it fortifies a company’s main strategic edge—like bridging stablecoins for Stripe, or adding specialized software for Toast—but can backfire if it drags a software firm into asset-heavy realms, as Shopify discovered.
• Market Shares: Amazon at ~40% US e-comm, Shopify ~10%, Stripe topping $1T+ in processed payments, Adyen strong in enterprise cross-border, Toast & NCR battling for restaurant/hospitality POS, etc.

Armed with these distinctions, you should find it simpler to see where each firm adds value, how they compete or cooperate, and whether new M&A deals or divestitures truly align with their strategic sweet spot.

Sources of Information

• Stratechery Articles and Interviews
  • Shopify Earnings, Software Self-Awareness, Rebels and the Arms Dealer
  • Stripe Acquires Bridge, Stablecoins, Platform of Platforms
  • Adyen Earnings, Adyen’s European Context, Adyen vs. Stripe
  • An Interview with Lisa Ellis about Payments
  • An Interview with Michael Morton About E-Commerce Winners and Losers
• Company Earnings & Press Releases
  • Shopify Q3 2024 Earnings Call and Bloomberg coverage
  • PayPal Q3 2024 Financial Releases
  • Public statements from Stripe (Bridge acquisition), Adyen (H1 2024 results), Toast, and NCR Voyix (ATM sale to Veritas Capital)
• Industry Data & Market Estimates
  • eMarketer / U.S. Census Bureau statistics on e-commerce (2019–2024)
  • Third-party analysts (e.g., The Information, Bloomberg) for private company data (Stripe, Deliverr)
  • App tracking changes post-Apple’s ATT from Meta (public statements)

Fundamental Question #1: Objective – Generate Sustainable Free Cash Flow (FCF)

Core Idea:
dLocal’s ultimate goal is to produce stable and growing free cash flows by leveraging its niche position in emerging markets (EM) online commerce. The company has historically shown strong EBITDA-to-FCF conversion due to low capital intensity and minimal debt. Its challenge now is maintaining FCF growth amidst intensifying competition and shifting mix.

Recent and Projected FCF Figures:

• Historically, FCF has been minimal but positive. In 2023, adjusted net income reached $149 million (22.9% net margin), illustrating strong earnings power.
• By 2025, we estimate adjusted net income of ~$162 million (+25% Y/Y), enabling strong internal cash generation. However, free cash flow (FCFF) is slightly negative in the near term (e.g., -$7 million in 2025E) due to incremental working capital and strategic investments in new geographies and product lines.

FCF Expansion Constraints & Drivers:

• Constraints: Slight net working capital usage due to rapid volume growth, potential foreign exchange hedging costs, and minor capital expenditures for technology infrastructure and compliance.
• Drivers: As revenue surpasses ~$900 million by 2025 (vs. $650 million in 2023), scaling overhead and stabilizing tax rates (~20% by 2025, up from ~16.5% in 2023) still allow significant profit retention.

The aim is that as dLocal matures, economies of scale and process automation boost EBITDA margins back into the low-20% range and eventually translate into stronger positive FCF beyond 2025–2026.

Fundamental Question #2: How Does This Business Make Money?

Revenue Model & Pricing Dynamics:

• Pay-In Transactions: dLocal charges a merchant discount rate (MDR) on each approved transaction. Initially around 4–5%, these take rates are expected to compress towards ~3.5–4.0% by 2025 as larger merchants gain pricing leverage and local-to-local volumes, which carry lower margins, increase.
• Pay-Out Transactions: dLocal earns fees for distributing funds (e.g., marketplace seller payments). Although essential for stickiness, pay-outs often have lower take rates, pulling blended margins down.
• FX Spread Fees: Historically, EM cross-border flows allowed dLocal to capture healthy FX spreads. For instance, if FX spreads contributed around 60–80bps to revenue historically, competition and improved liquidity could narrow that by ~10–20bps by 2025. This narrowing could shave a few percentage points off revenue growth. For example, if FX revenue contributed ~10–15% of total revenue at peak, a 20-30% reduction in FX margin might trim total revenue growth by ~2–3 percentage points annually.

Customer Concentration & Contract Duration:

• Top 10 clients ~60–65% of revenues. While no major churn reported to date, reliance on a few large digital merchants (e.g., Amazon, Netflix) poses a risk if they renegotiate terms or in-house some capabilities.
• Contracts tend to be multi-year with complex integrations, discouraging swift vendor switching. This complexity supports dLocal’s stable baseline of volumes.

Fundamental Question #3: Nature of the Cost Structure

Cost Structure & Operating Leverage:

• Fixed vs. Variable Costs:
  • Variable: Interchange, network fees, and local partner commissions scale with volumes. These are recognized in COGS.
  • Fixed: R&D, compliance, overhead, and SG&A scale more slowly, offering operating leverage as revenues grow.
• Current vs. Future State:
  In 2023, EBITDA margin at ~29.5%. By 2024/2025, due to pricing adjustments and mix shifts, margins dip (~21.7% in 2024E, rising to ~23.0% in 2025E).
  Despite slight margin compression, absolute EBITDA grows as TPV expands. The efficiency gains (one integration for multiple countries) eventually lower per-unit overhead costs.

FX and Processing Costs Impact on Unit Economics:

• As competition intensifies (Adyen, PayU, Ebanx, Stripe) and local acquiring improves, dLocal may pay slightly higher partner fees or accept narrower spreads.
• If FX spreads tighten from, say, ~80bps to ~60bps by 2025, that 20bps reduction can mean a loss of ~$10–15 million in potential revenue (on ~$9 billion TPV), requiring volume growth to offset.

Fundamental Question #4: Key Drivers of the Business

Academic Drivers – Growth, Margins, Capital Efficiency:

1. Organic Revenue Growth:
   Driven by TPV expansion: from ~$2bn in 2020 to $6–9bn range by 2025. With $922 million revenue forecast by 2025 (+25% Y/Y from $738m in 2024), volume growth remains the main revenue driver despite lower take rates.
2. Margin Trajectory:
   EBITDA margins compressed from ~30% in 2023 to ~22–23% in 2024–2025. Long term, better cost control and scaled operations could lift margins back to mid-20%.
3. Capital Intensity & ROIC:
   Minimal capex and asset-light model yield high ROIC (>25%). This capital efficiency remains intact, making dLocal a structurally profitable growth story.

Comparative Valuation & Market Perception:

• dLocal Multiples: Trades at ~21x EV/2024E EBITDA and ~16x EV/2025E EBITDA.
• Comps:
  • Adyen: Trades around ~22–25x 2025E EV/EBITDA, but is a more established player in developed markets.
  • PayPal: Closer to ~13–15x 2025E EV/EBITDA, with a broader ecosystem but slower growth.
  • dLocal’s multiples sit between PayPal’s lower multiple (due to mature growth) and Adyen’s premium (due to best-in-class global profile). dLocal’s EM focus and strong growth justify a premium to PayPal but not as high as Adyen, given margin pressures and execution risks.

Model Sensitivities – What Moves the Needle?

• Upside: Faster merchant expansion, stable FX spreads, or quicker margin recovery.
• Downside: Larger merchants securing fee concessions, accelerated FX margin compression, or losing a key client.

Fundamental Question #5: Business Momentum

Momentum Indicators & Current Trends:

• After years of triple-digit volume growth, the pace moderates to 50%+ TPV growth, still robust but normalizing.
• Revenue growth decelerates from 55% in 2023 to ~13.5% in 2024E and re-accelerates to 25% in 2025E as mix stabilizes.
• 2024 earnings dip slightly (-12% Y/Y in EPS) due to margin compression and transition costs, but by 2025 EPS rebounds (+26% Y/Y).

FX Impact & Normalization:

• Previously, FX fees were a strong tailwind. Now, more competition and liquidity reduce FX margins by an estimated 20-30bps, trimming revenue growth by a couple of percentage points. Merchants value stable, reliable processing over high FX spreads, so dLocal prioritizes relationship longevity over short-term FX gains.

Long-Term Narrative:

• The story evolves from hyper-growth, high-margin startup to a more mature, scaled EM payments enabler. While early valuation premiums may fade, steady expansion, ongoing product development (e.g., pay-outs, marketplace solutions), and incremental operating leverage support a constructive long-term outlook.

Where is the Debate?

• Investors debate the sustainability of dLocal’s pricing power, the intensity of competition, and the realistic long-term margin trajectory.
• The company’s ability to execute on growth in multiple EM markets simultaneously—maintaining quality as volumes surge—remains top-of-mind.

Conclusion

dLocal’s objective is sustainable FCF generation from its niche platform that simplifies EM commerce. Its revenue model—integrating pay-ins, pay-outs, and FX spreads—ensures diversified income streams, though take rates and FX fees are under pressure. The cost structure evolves as the business scales, and while margins dip in the near term, long-term economics remain compelling due to low capital intensity and high ROIC.

Key value drivers include robust volume growth, partial margin recovery, and capital efficiency. Valuation is currently moderate, with an EV/EBITDA multiple between established (PayPal) and hyper-growth (Adyen) peers. Momentum, while still positive, is moderating, and short-term earnings softness challenges near-term sentiment.

In short, dLocal transitions from an early hyper-growth story to a more balanced growth and profitability narrative. With a Dec-2025 PT of $14 and a Neutral rating, the outlook reflects optimism on long-term expansion balanced by caution over near-term margin and pricing headwinds.

Backgrounds: we went through Meta’s recent fundamentals in the last blog, and I wanted to talk about my recent thoughts on how Meta is able to capture the most growth in digital advertising within this generative AI Booming

Meta stands at the intersection of two major shifts in digital advertising: the recovery and reshaping of online ad measurement post-ATT, and the incorporation of advanced AI models that promise to make targeting, creative, and conversion optimization dramatically more efficient and effective. While every major ad platform is investing in AI, Meta’s advantages are structural, cultural, and cumulative—leading to a scenario where the benefits of AI are not just incremental, but multiplicative. In other words, Meta is not simply catching up with AI; it is poised to use AI to generate a kind of abundance that its rivals cannot easily match.

A History of Scaling From the beginning, Meta’s approach to digital advertising has been predicated on massive scale, broad advertiser access, and relentless iteration. Over the last decade, the company combined user-generated content with sophisticated machine learning to match 10 million advertisers to billions of users worldwide. Even as Apple’s ATT disrupted deterministic tracking, Meta responded by applying AI to reconstruct missing signals, leaning into probabilistic models that would have seemed outlandish just a few years ago. The company’s swift adaptation shows that scale is more than just a number at Meta—it’s the foundation that lets them absorb shocks and re-tool quickly.

This scale is integral to how Meta’s AI models work. The company’s advantage in training complex targeting and measurement models is not just about having better algorithms; it’s about having more data to refine them on. Meta’s ability to observe billions of interactions across Facebook, Instagram, Messenger, WhatsApp, and now Threads provides a universe of signals to feed into its models. The result: a compounding effect, where each incremental user and advertiser enhances the accuracy of the system, leading to better performance for everyone.

AI as a Force Multiplier AI’s potential to improve targeting is well understood: better predictions about who will click or convert can justify higher bids and raise CPMs. What’s less appreciated is how AI can reshape ad creative generation, frequency management, ad relevance scoring, and even format innovation.

• Creative Optimization: In a world where generative AI can produce multiple variants of an ad, Meta’s vast data and feedback loops ensure it can quickly identify which creatives resonate best. Over time, this process becomes automated, so small advertisers can get results previously reserved for big brands with large creative teams.
• Dynamic Formats: Consider Reels. Initially challenging to monetize, it’s now at a $10B run-rate. Why? Because AI optimizes which Reels users see, ensuring engagement remains high. By applying AI to ad selection and pacing, Meta can close the initial CPM gap between Reels and feed ads. The same logic will eventually apply to Threads, messaging ads, or any new surface Meta explores.
• Integrated Ecosystem: While competitors excel in certain niches—TikTok in short-form video, Google in search, Snap in AR lenses—Meta sits across multiple paradigms. AI binds these disparate elements together, allowing learnings from one format to inform another. User interest gleaned from Instagram feed engagement might help improve ad relevance in WhatsApp’s click-to-message ads. This integration is uniquely Meta’s: a vast network of products unified by a single AI backbone, making every user interaction an input to improve the entire system.

The Uniqueness of Meta’s Opportunity It’s tempting to say AI benefits every ad platform. And it does—incrementally. However, Meta’s advantage is rooted in the combination of scale, data, and product diversity. That combination changes the slope of improvement. Other platforms have portions of Meta’s toolkit:

• Google: Owns massive data and powerful AI, but is concentrated on search and YouTube. While Google’s reach is enormous, its social engagement is limited. Its advantage in text-based ads is huge, but it must integrate generative AI into a different kind of user intent paradigm. The complexity of Google’s ecosystem—search, shopping, cloud—might slow focused iteration on social-like ads.
• TikTok: Remarkable engagement but comparatively immature monetization stack. AI helps, but without Meta’s decade-plus of ad auction refinement and cross-app data, gains may plateau sooner. TikTok’s user data is rich for entertainment but may lack the rich profile depth—interests, social graphs, and long-term behaviors—that Meta’s platforms have accumulated.
• Snap & Pinterest: Both have niche value propositions and unique ad products. AI will improve their targeting and creative too, but with smaller datasets and fewer global advertisers, the pace of improvement lags. Their revenue scale and margin profile limit how aggressively they can invest in cutting-edge AI infrastructure. Meta’s billions in CAPEX on AI data centers are amortized over a much larger revenue base, yielding better per-unit cost efficiencies.
• Amazon Ads: Great for commerce-based intent, but less effective for building brand awareness or capturing top-of-funnel interest. Amazon’s AI can refine sponsored product placements, but it doesn’t match Meta’s global audience graph or ability to integrate signals across various user contexts—personal feed, short-form video, messaging interactions.

Building a Moat of Complexity One underrated aspect of AI improvements is that they tend to accumulate complexity and sophistication over time. The more Meta invests, the better it integrates AI into every part of the ad stack—creative suggestions, audience expansion, bid optimization, frequency capping—the harder it becomes for newcomers to replicate. This complexity, when managed well, forms a moat. Other platforms can copy one feature or another, but Meta’s advantage is combinatorial. Every incremental improvement in targeting informs reel monetization, which informs click-to-message ads, which will eventually inform whatever Threads’ monetization becomes.

This complexity leads to abundance: abundant impressions, abundant ad performance improvements, abundant incremental revenue streams. Over time, Meta’s AI systems drive a virtuous cycle: better AI leads to better ad experiences, which leads to more advertiser demand and user engagement, which in turn enhances Meta’s training data, feeding the AI improvement loop.

Long-Term Cash Flow and Strategic Patience While the ultimate objective—sustained and growing FCF—has been previously outlined in detail, it’s worth re-emphasizing how AI’s second-order effects reinforce that goal. As the incremental cost of improving AI models and inference declines over time, operating leverage increases. Margins gradually expand, and a greater share of each incremental dollar of revenue turns into free cash flow.

This matters because it gives Meta optionality. With robust FCF:

• Meta can fund reality labs R&D for AR/VR without jeopardizing financial stability.
• It can continue to buy back shares or even consider dividends long-term.
• It can absorb macro shocks, regulatory changes, or potential new platform shifts like Apple’s rumored push into AR glasses.

In a scenario where all large tech companies employ AI, the incremental differences in scale, data, and product integration matter enormously. Meta’s position, despite all the scrutiny and challenges of the past two years, is growing stronger in relative terms. It’s not just that Meta recovers from ATT; it leaps ahead by fully embedding AI-driven improvements into its ecosystem, ensuring that future headwinds (like privacy changes, and competitor entries) are more manageable with the cushion of massive, AI-enhanced FCF generation.

Conclusion The world of digital advertising is being reshaped by AI, and while multiple platforms stand to gain, Meta’s structural advantages uniquely position it to reap outsized benefits. The company’s scale, multi-format ecosystem, and deep integration of AI into every facet of the advertising lifecycle create a moat that competitors will struggle to overcome.

As we move forward, it’s likely that much of this progression will feel incremental: a few percentage points improvements in ROAS here, a slight increase in Reels CPMs there. But just as the Internet and mobile smartphones eventually accumulated into massive secular changes, these “small” AI-driven enhancements compound into something transformative. Over the next few years, Meta won’t just restore its pre-ATT profitability; it may well surpass it, achieving a level of AI-fueled abundance that cements its place at the center of global digital advertising.

In the final analysis, Meta’s readiness to harness AI isn’t merely a defensive maneuver to recover lost ground. It’s an offensive strategy to shape the next decade of digital advertising, ensuring Meta, more than anyone else, emerges as the platform of choice for advertisers seeking scale, performance, and innovation—and one that, by extension, sets the standard for what AI-driven abundance really means in this industry.

1. Their FCF generation path
2. How they monetize the business
3. Cost & spending structure
4. Business drivers
5. Business momentum

Fundamental Question #1: What is the Ultimate Objective? (Focus on FCF Generation)

Objective & Explanation: Meta’s primary long-term objective is to create sustainable, growing Free Cash Flow (FCF) from its global digital ecosystem. Historically, Meta’s core ad model generated robust operating margins (30–45%) and high FCF conversion, allowing massive reinvestment (in AI and the metaverse) and consistent share buybacks. Now, as Meta invests heavily in new platforms (e.g., Threads, AR/VR), its FCF objective involves balancing near-term spending with profitable growth lines, ensuring that after heavy CAPEX and R&D cycles, incremental revenue streams produce even larger FCF down the line.

Why FCF Matters for Meta:

• Flexible Capital Allocation: With strong FCF, Meta can fund R&D in AI, AR/VR, Reality Labs, and new ad formats without external financing. In 2023, CFO ~ $70B+ and net cash ~$42B provided ample liquidity.
• Shareholder Returns: Historically, Meta returned capital via buybacks (~$21B in last TTM before Q4:23). Healthy FCF ensures ongoing repurchases and potential dividends in the future.
• Resilience Against External Shocks: Regulatory changes (ATT/IDFA), macro slowdowns, or competitive threats (TikTok, link-in-bio fragmentation) require steady FCF to adapt product strategies and pivot quickly.

Short-Term vs. Long-Term FCF Outlook:

• Short Term (2024–2025): Elevated CAPEX ($30–35B in 2024) reduces near-term FCF growth, but still leaves tens of billions in annual FCF. Operating expense guidance ($94–99B) ensures disciplined cost management. AI-driven ad improvements offset some cost pressures.
• Long Term (2026–2027+): Once AI infrastructure scales, incremental cost per user declines, boosting FCF margins. Emerging revenue lines (Threads at $7–10B revenue by 2027, Reels at $10B run-rate) mature, requiring less incremental cost. Reality Labs, if successful by late-decade, adds non-ad revenue streams, further diversifying FCF sources.

In essence: Meta’s objective—steady FCF growth—is achieved by strategically investing now in AI and product diversification to ensure future revenue outstrips incremental costs, leading to a stable, expanded FCF base.

Fundamental Question #2: How Does the Business Make Money?

Core Revenue Model: Meta primarily sells targeted advertising to ~10 million advertisers aiming to reach 3B+ monthly users on Facebook, 2B+ on Instagram, and millions more on Messenger, WhatsApp, and Threads. The vast majority (~97%) of Meta’s ~$135B 2023E revenue comes from ad placements. Non-ad revenue (~3%) stems from hardware (Quest headsets) and digital sales in Reality Labs.

Deeper Dive into Ad Monetization:

• Ad Auction System: Advertisers bid to display sponsored posts, stories, or reels. Meta’s platform ranks ads using AI models that predict which ads users will likely engage with. This system sets a “cost per impression” or “cost per click” dynamically.
• ARPU & Pricing Power: User growth (+2–3% YOY) plus stable or rising engagement (DAU/MAU ratio ~67% on Facebook) allow more impressions. AI-driven relevance increases ad ROI, letting Meta charge higher effective CPMs. Over time, as Meta refines targeting post-ATT/IDFA, higher conversions justify premium pricing.
• Ad Formats & Diversification:
  • Feed & Stories (Mature formats): Established, highest ARPU channels.
  • Reels (Short-Form Video): Recently ramped to $10B run-rate. Initially less monetized than feeds (lower CPMs), but improved targeting now closes that gap. Brands pay for interactive vertical video ads as short video grows popular.
  • Click-to-Messaging (WhatsApp/Messenger): Advertisers pay to initiate a chat. This is a high-intent format; if a user clicks to message, conversion rates often surpass traditional ads. Over $10B run-rate combined in 2023. Messenger ~75% and WhatsApp ~25% of that total initially, but WhatsApp growing faster.
  • Threads Monetization (Upcoming): If Threads achieves 600–700M MAUs by 2027 and can reach ARPU of $21–30, it may generate $7–13B revenue. Threads likely offers text-based sponsored posts, brand-safe environments, and niche interest communities for native ads.

Reality Labs & Future Monetization:

• Hardware Sales (Quest VR): Sells at $300–500 per device. Revenue small (~$1–2B), but negative margin currently.
• Future AR/VR & Metaverse Commerce: In 5–10 years, Meta could levy platform fees, sell premium VR experiences or digital goods, and incorporate ads into immersive worlds.

Why Growth Happens:

• Global Ad Budget Shifts Online: With digital ads ~65% of global ad spend and still rising, advertisers follow users. More users + better tools = more spending.
• AI Tools Improve Targeting & Measurement: Post-ATT, better AI signals restore advertiser confidence, driving higher spending. Performance advertisers return as ROAS improves.
• New Formats (Reels, Messaging, Threads) Add Inventory & Engagement: These formats tap into user trends (short-form video, private messaging, interest-based communities) and convert them into ad opportunities.

Fundamental Question #3: What is the Nature of the Cost Structure?

Cost Composition:

• Cost of Revenue (~$31B in 2024E, ~20% of revenue):
  • Data Centers & AI Infrastructure: Servers, storage, network hardware, and electricity form a large part. AI model training (LLaMA/Llama 2) demands specialized, expensive GPUs. Depreciation of servers (7–9% of revenue) alone is substantial.
  • Content Review & Moderation: Thousands of contractors ensure brand safety, costing billions annually.
  • Partner Fees & Payment Processing: For ad transactions and commerce features.
• Operating Expenses (~$67B in 2024E, ~44% of revenue):
  • R&D (~$42B in 2024E, ~28% of revenue): Funds AI model development, VR/AR research, and product innovations. Payback might be 2–5 years. AI infrastructure investments improve ad yields within 1–2 years, while AR/VR may take 5+ years to pay off meaningfully.
  • Sales & Marketing (~$15B in 2024E, ~10% of revenue): Primarily headcount costs for sales reps, marketing staff, and user support. As Meta reduces headcount and relies more on self-serve ad platforms, this grows slower than revenue. Payback is quick; each sales rep can manage thousands of advertisers.
  • G&A (~$10.5B in 2024E, ~7% of revenue): Corporate overhead, legal, regulatory compliance. Stable overhead; scaling revenue reduces G&A% over time, boosting margins.

Capex & Payback Periods:

• Capex ($30–35B in 2024): Predominantly for AI servers and data centers. Payback ~2–3 years as better AI relevance translates into higher RPMs, while VR/AR spending might have a longer horizon (~5 years) before meaningful profitability.
• R&D vs. Revenue Growth: If AI increases conversion rates by a few percentage points, incremental revenue from higher ROAS can justify billion-dollar AI investments within 1–2 years.

Long-Term Implication of Costs: As infrastructure scales and AI models mature, the incremental cost per additional user or impression falls. Over a 3–5-year horizon, stable or declining cost ratios boost operating leverage, expanding EBIT margins and, thus, FCF.

Fundamental Question #4: What Are the Key Drivers of the Business?

Key Growth Drivers & Logic Chain:

1. AI-Enhanced Ad Targeting → Higher Advertiser ROI → More Ad Spend → Revenue Growth → Margin Expansion → Stronger FCF.
   Post-ATT, AI signal reconstruction lets Meta show highly relevant ads. Advertisers see better conversions, bid higher, raising CPMs and revenue, which scales margins and FCF.
2. Diversification of Ad Formats (Reels, Messaging, Threads) → More Inventory & Engagement → Higher ARPU & Lower Concentration Risk → Stable Growth.
   Short-form video (Reels) caters to younger audiences, messaging ads target high-intent users, and Threads could capture Twitter-like text engagement. Each format supports a new behavior and monetizes differently, cumulatively pushing revenue growth and reducing dependence on one format.
3. Global Digital Ad Market Expansion → Meta’s Scale Advantage → Gains in Market Share.
   The digital ad market (~8–10% annual growth) and Meta’s unmatched reach (3B+ MAUs) ensure it’s a top-of-mind platform for global brands and SMBs. As marketing budgets shift online, Meta gets a large share, reinforcing top-line momentum.
4. Reality Labs (Long-Term Bet) → New Ecosystem & Commerce → Diversified Revenue Streams in 5–10 Years.
   If metaverse platforms (Horizon Worlds, Quest headsets) succeed, Meta could introduce VR-commerce fees, AR ads, and digital goods sales. While uncertain, successful execution will unlock new billion-dollar revenue channels by 2030.

Impact on Financials:

• Sustained top-line growth at high single to low double digits, stable or improved EBITDA margins (55–60%), and balanced capex eventually translate into accelerating EPS and FCF growth. Over time, these drivers reduce reliance on one platform or format, improving revenue stability and investor confidence.

Fundamental Question #5: What Is the Business Momentum?

Momentum Indicators & Data:

• User Engagement & MAUs: Facebook stable at 3B MAUs, Instagram at 2B. Threads at ~275M MAUs by late 2024 up from ~100M in Q3:23, indicating rapid adoption. This top-of-funnel growth sets the stage for more impressions and future monetization.
• Time Spent: Reels accounts for >20% of Instagram time, growing at double-digit Y/Y. Messenger usage stable with billions of daily messages. WhatsApp strong in emerging markets.
• Short-Term Catalysts: Q4:23 Temu/Shein ad surge adds a 3–4% revenue tailwind. FX relief adds ~1–2% in Q4. Reels are now revenue-neutral, no longer diluting earnings. These near-term positives help beat consensus near-term.
• Challenges into 2024: 2H24 comps get tougher as Temu/Shein spend normalizes and FX stabilizes. Link-in-bio fragmentation may subtly reduce direct engagement if more user time shifts off-app. Competition from TikTok for younger demographics remains.

Logic Chain of Momentum:

• Rising Engagement in Reels & Messaging → Advertisers Confident → Higher Bids → Near-term Revenue Beat.
• Temu/Shein Spend (Q4:23) → One-time Boost → Strong Q4 Print → Higher Stock Price Short-term → But less of a lift in Q4:24.
• FX Tailwind Q4:23 → Extra 1–2% EPS → Investor Confidence → Slight Multiple Expansion.
• Threads Rapid MAU Growth → Future Monetization Potential → Investors Price in Long-term Upside.

Long-Term Momentum Implications:

• Q4:23/Q1:24 upside sets a positive near-term narrative. As comps toughen in 2H24, investors shift focus to 2025–2027 growth drivers (Threads revenue, AI-driven ad yields). Sustained user base expansion, stable or improving ARPU, and diversified formats keep momentum positive, albeit moderated after early 2024 peaks.

Data Points Supporting Momentum:

• DAUs on Facebook are stable at ~67% of MAUs, and Instagram is similarly strong.
• Messaging ads at a $10B run rate (growing ~30–40% YOY) show marketers love conversational commerce.
• Reel ad load increases without destroying user experience, indicating skill in balancing engagement vs. monetization.

Conclusion

Meta Platforms stands at a positive inflection, balancing near-term tailwinds (AI, ad mix diversification) with long-term bets (AR/VR). Its top-line growth (~10%+), stable margins (~35% operating, ~55% EBITDA), and strong FCF underpin resilience. Although 2H24 faces tougher comps and subtle user engagement fragmentation via link-in-bio, the overall momentum remains constructive. The company’s broad ecosystem, successful Reels monetization, and promising Threads adoption ensure a robust pipeline of revenue drivers feeding into its ultimate objective: sustaining and increasing free cash flow over time.

Key Takeaways

• I want my life to experience infinitely more.
• Persist in learning by any means necessary.
• Even during seemingly meaningless moments in history, remain curious. CURIOSITY adds depth and dimension to life.

Key Opinion Quotes
“All of this is because I’ve never stopped learning. Back when I was in the Gobi Desert, I thought: if this is truly my fate—if I will never leave this desert—then I have nothing to complain about because fate is beyond my control. Throughout history, many people have been unlucky, and many have missed opportunities. So why should I complain?

“But if an opportunity appears before me one day and I fail to seize it, that is on me.”

Selected Quotes

On Knowledge and Learning

• “I realized how ignorant I was, and the only way to overcome ignorance is through continuous learning.”
• “Finance is highly technical and requires no specific cultural background; it’s like studying mathematics.”
• “The flexibility of the U.S. system suited me perfectly. Back home, entering university felt like scaling a mountain. There was almost no freedom of choice.”

On Opportunity

• “Opportunity is open to everyone. I just happened to seize it while others did not.”
• “My life philosophy is to keep learning and stay prepared. That way, when opportunities appear, I can grasp them.”
• “I can’t give up, no matter how great the setbacks. To give up is to commit a crime against myself. I must keep going, keep working hard, and wait for the next chance.”
• “Wasting time is like committing a sin against oneself.”
  • Notes to Self – My life needs both joy and moments of idle time tho. But to abandon curiosity is to commit a sin against myself.
• “Trying something out never hurts—it’s just like another practice exam.”
• “Figure out what the examiner wants—understand their perspective and meet their expectations.”

Building Goodwill and Connections

• “Fame can draw unwanted attention—just as a well-fed pig is always at risk of the butcher’s knife.”
• Practice humility, stay low-key, ‘hide your brightness,’ and quietly make your fortune.
• “People don’t like those who are too different from themselves.”
• “Make many friends—when you have a wide network, it’s harder for those in authority to cause trouble for you.”
• Students pursue top schools not only for knowledge but also for two other vital reasons:
  1. To gain the elite institution’s ‘brand label.’
  2. To build a valuable network of relationships that will serve them throughout their careers.

Dealing with Authority

• “Acting on blind faith or impulse leads to costly mistakes.”
• “Who could refuse a word of congratulations?”

Background:
These notes are drawn from my reading of Weijian Shan’s Out of the Gobi: My Story of China and America. Shan’s memoir recounts his extraordinary journey from the remote Gobi Desert—a place of exile during the Cultural Revolution—to becoming a successful economist and investor in the United States. His story highlights perseverance, the relentless pursuit of education, seizing opportunities, and the importance of flexibility and humility. Through these notes, I hope to capture some of his key insights and reflect on their relevance in our own lives.

After some of the insightful readings that I got done earlier this week, I decided to pick up this blog. Wanting to record this, I wanted to use this very first blog in two years to record my incentives to restart my blogs.

To get started with, there are a few readings that I got done that made me think getting writing stuff started would be something that is significantly helpful.

1. Howard Marks,
   • I’ve long been following Howard’s Memos and am always curious about how smart people like him are able to come up with these innovative ideas and able to make them so clear and understandable to the audience, wanting to get as smart one day.
   • This week, I went through one of his podcasts where he talked about
     • how he was able to maintain the hobby of writing and training himself being able to present complex ideas in clear and accessible language that can help engage and delight your audience
     • how this quality work can make an impact on his ideas.
2. George Soros
   • Started reading his book The Crash of 2008 & What It Means: The New Paradigm for Financial Markets per Point 72 Academy’s book list recommended
   • He mentioned that serious studies could help him in his investment decisions

My takes

1. I basically see this hobby of blogging being able to help me get started to thinking logistically
   • According to my traditional Chinese educational background, I always see myself as someone who is taking ideas instead of someone coming up with innovative opinions no whether on investment stuff or social justice affairs
   • Going deeper into the reasons, I’m finding myself lacking the ability to come up with complex logical ideas and convey them, while this is something super important in the world of business and investment.
   • I see writing something about myself as a perfect way to train myself on this.
2. I see this as a good way to help me to take notes
   • I recently self-identified myself as a financial nerd, loving reading stuff and practicing my financial modeling.
   • Finding a place to record all of my takes from different books seems to be more important.
   • I see blogging something helping myself releaseing the computing power of my brain through putting the notes and organizing them on a blog.
   • What is even better, I might be able to create a chronological archive seeing how my readings and takes evolved if I’m able to keep this blogging thing going.
3. Finally, I guess this could be a good way to help me think logically
   • Getting back to the “computing power” of my brain thing
     • I’m always someone doubting if I’m someone with ADHD
     • It is exhausting keeping everything I see, think, and want to memorize in my mind logically
   • As writing is something that I would be able to record everything on the page and being able to organize them logically, I believe this is something that can help me develop this ability to think logically
   • As my readings and my ability to think evolve, I might be able to check previous archives in this blog easily, instead of exhaustingly trying to retrieve something from my mind using my brain’s limited computing power.

So here I wanted to get this as a good point to restart my investment and social justice blog, recording my notes and thoughts on the readings that I’ve done recently.

Jerry Oct 18th, 2024