Skip to main content
← BACK TO HOMESector Research

Technology.

From platform economics to AI compute scaling laws — we analyze the largest technology companies through the lens of network effects, marginal cost structures, and capital allocation discipline that separates durable compounders from multiple-expansion stories.

Our Thesis

Large-cap technology is the most heavily covered yet frequently misunderstood sector in global equity markets. Consensus models focus on quarterly revenue beats and earnings multiples while missing the structural dynamics that actually drive long-term value creation: the depth of platform moats, the economics of cloud margin expansion, the return profile of AI capital expenditure cycles, and the regulatory vectors that could reshape market structure over a decade.

We approach technology differently. Every platform company is a system of interlocking network effects, switching costs, and data advantages that compound at different rates. Apple's ecosystem lock-in through hardware-software-services integration creates different durability characteristics than Google's information retrieval monopoly or Meta's social graph density. Understanding these moat architectures at a granular level — which are strengthening, which are eroding, and where new competitive vectors are emerging — is what allows us to distinguish between genuine compounders trading at fair multiples and consensus favorites priced for perfection.

The current technology landscape is being reshaped by the most significant platform shift since mobile: artificial intelligence. The magnitude of capital being deployed — over $200 billion in combined hyperscaler capex in 2025 alone — creates both enormous opportunity and significant risk. Our research focuses on identifying which companies will generate returns on this investment that exceed their cost of capital, and which are engaged in a capex arms race that will ultimately destroy shareholder value. History shows that not every dollar spent on transformative technology earns a return — the fiber optic buildout of 1999-2001 proved that infrastructure investment and equity returns are not the same thing.

Research Framework

Platform Moat Analysis

  • Network effect taxonomy — direct (social), indirect (marketplace), data (search/AI), ecosystem (Apple)
  • Switching cost measurement by user segment and product cohort
  • Multi-homing analysis — where users maintain parallel platform relationships
  • Developer ecosystem health metrics (API calls, third-party revenue, platform dependency ratios)
  • Competitive entry barriers: minimum viable scale requirements by vertical

Cloud & Infrastructure Economics

  • Cloud revenue decomposition: IaaS, PaaS, SaaS, AI-as-a-Service contribution margins
  • Gross margin trajectory modeling — hardware amortization cycles vs. software attach rate
  • Workload migration analysis: on-premise remaining addressable market by enterprise tier
  • Multi-cloud vs. single-cloud adoption trends and lock-in economics
  • Capital intensity ratios: capex/revenue trends, useful life assumptions, depreciation policy shifts

AI Capex Cycle Analysis

  • GPU cluster economics: cost per FLOP trends across training and inference workloads
  • Hyperscaler capex decomposition — AI vs. core cloud vs. maintenance spend
  • Return on incremental capital employed (ROICE) modeling for AI infrastructure
  • AI revenue attribution: direct (API/model access) vs. indirect (product enhancement, retention lift)
  • Inference cost curves: model efficiency gains vs. demand elasticity at lower price points

Valuation Methodologies

  • DCF with explicit modeling of reinvestment rate decline as platforms mature
  • SaaS metrics: ARR growth, net revenue retention, CAC payback, Rule of 40 scoring
  • Sum-of-the-parts for conglomerates (Alphabet segments, Amazon AWS vs. retail vs. ads)
  • Normalized free cash flow yield adjusted for stock-based compensation dilution
  • Regulatory risk discounting: antitrust scenario analysis with probability-weighted outcomes
Research Example

How We Analyze a Platform Company

Step 1: Moat Architecture Mapping

Before building a financial model, we map the complete moat architecture. A company like Microsoft operates multiple reinforcing moat systems simultaneously: enterprise switching costs (Active Directory, Office 365 integration depth), developer ecosystem lock-in (Azure DevOps, GitHub, VS Code), and data network effects (LinkedIn graph, Bing training data for Copilot). Each moat has a different durability profile and competitive exposure. Enterprise switching costs around identity management are nearly impregnable; cloud infrastructure margins face persistent competitive pressure from AWS and GCP. Understanding which moats are load-bearing and which are decorative is the foundation of our analysis.

Step 2: Unit Economics Decomposition

Reported segment financials often obscure the true economics. When Alphabet reports Cloud revenue, the mix of IaaS (low margin, capital intensive), PaaS (higher margin, stickier), and AI services (high margin but unproven retention) matters enormously for forward margin trajectories. We decompose each revenue stream to its unit economic level: revenue per user or per workload, gross margin by product, customer acquisition cost by channel, and cohort retention curves. This bottoms-up approach frequently reveals margin expansion or contraction dynamics that top-down models miss by two to three years.

Step 3: Capital Allocation Scorecard

Technology companies generate enormous free cash flow, and how they deploy it determines long-term returns. We score capital allocation across five dimensions: organic reinvestment ROI (R&D dollar productivity measured by revenue and margin contribution), M&A track record (acquisition IRR vs. stated hurdle rates), buyback discipline (are they buying back stock at reasonable valuations or just offsetting SBC dilution?), dividend policy (appropriate maturity signaling vs. premature payout), and balance sheet optimization (net cash position relative to strategic optionality needs). Apple's capital return program is a masterclass; Meta's Reality Labs spending is a cautionary example of conviction without accountability.

Step 4: Regulatory Scenario Modeling

Antitrust risk is the single largest unmodeled variable in large-cap tech. The DOJ's case against Google, the EU's Digital Markets Act enforcement, and potential legislative action on app store economics each represent scenarios with materially different equity outcomes. We probability-weight these scenarios: structural separation (low probability, high impact), behavioral remedies (moderate probability, moderate impact), and consent decrees that maintain the status quo (highest probability). A 10% probability of forced divestiture of Google's ad tech stack changes the risk-adjusted valuation meaningfully even if the base case assumes no action.

Step 5: Valuation & Entry Discipline

Technology stocks reward patience and punish consensus chasing. We use a multi-factor valuation framework: DCF with explicit reinvestment rate assumptions (declining as platforms mature), normalized FCF yield adjusted for SBC (which can represent 5-15% of revenue at growth-stage companies), and relative valuation against the company's own history and peer set. The entry discipline matters — buying Microsoft at 25x FCF during a cloud deceleration scare is fundamentally different from buying it at 35x during a Copilot hype cycle, even though the long-term thesis is identical. We wait for the market to offer quality at reasonable prices, not reasonable quality at any price.

Coverage Universe

MEGA-CAP PLATFORMS
  • Apple (AAPL)
  • Microsoft (MSFT)
  • Alphabet (GOOGL)
  • Amazon (AMZN)
  • Meta Platforms (META)
  • NVIDIA (NVDA)
CLOUD & ENTERPRISE
  • Salesforce (CRM)
  • ServiceNow (NOW)
  • Snowflake (SNOW)
  • Datadog (DDOG)
  • MongoDB (MDB)
  • Palantir (PLTR)
ADVERTISING & COMMERCE
  • The Trade Desk (TTD)
  • Pinterest (PINS)
  • Shopify (SHOP)
  • MercadoLibre (MELI)
  • Uber (UBER)
  • DoorDash (DASH)
CYBERSECURITY
  • CrowdStrike (CRWD)
  • Palo Alto Networks (PANW)
  • Zscaler (ZS)
  • Fortinet (FTNT)
  • SentinelOne (S)
FINTECH & PAYMENTS
  • Visa (V)
  • Mastercard (MA)
  • Block (SQ)
  • PayPal (PYPL)
  • Adyen (ADYEN)
ETFs & INDICES
  • QQQ (Nasdaq 100)
  • XLK (Tech Select)
  • IGV (Software)
  • SKYY (Cloud Computing)
  • HACK (Cybersecurity)

Current Themes

AI Monetization Inflection

The critical question in technology is no longer whether AI works but whether it generates returns on the $200B+ in annual hyperscaler capex. Microsoft's Copilot attach rates, Google's AI Overview ad monetization, and Meta's recommendation engine improvements are the earliest indicators. Companies that demonstrate measurable revenue uplift per dollar of AI infrastructure spend will separate from those engaged in capex without accountability. We track unit-level AI revenue attribution across every major platform.

Cloud Reacceleration

After a 2023-2024 optimization cycle where enterprises rationalized cloud spend, growth is reaccelerating as AI workloads drive new consumption. Azure, AWS, and GCP are all reporting improving growth trajectories, but the composition of that growth matters — AI inference revenue at 70%+ gross margins is structurally different from commodity IaaS at 30% margins. We model cloud revenue by workload type to identify which providers are capturing the highest-quality growth.

Antitrust Structural Risk

The DOJ's victory in the Google search case and ongoing enforcement against Apple's App Store and Meta's acquisitions represent the most significant regulatory risk to large-cap technology in two decades. Remedies could range from behavioral changes (opening default search agreements to competitive bidding) to structural separation (divesting Chrome or the ad tech stack). We probability-weight these outcomes and incorporate them into our valuation framework as discrete scenario adjustments.

Capital Return Maturation

As growth rates normalize, capital return programs become increasingly important to total shareholder return. Apple has returned over $700 billion through buybacks and dividends since 2012. Alphabet and Meta have initiated massive buyback programs. The quality of these programs — buying below intrinsic value vs. mechanically offsetting dilution — varies enormously and is an underappreciated driver of long-term returns. We score every major tech company's capital allocation discipline on a quarterly basis.

Interested in our technology research?

All inquiries are treated with strict confidentiality.

Investor Inquiries