Oracle: The $553B Revenue Backlog That’s Bigger Than OpenAI’s Entire Runway What RPO actually is: Remaining Performance Obligations = contracted future revenue that hasn’t been recognized yet. For SaaS/AI infrastructure, it’s cash already committed, usually via multi-year deals. Think of it as a “revenue pipeline in the bank.” Oracle’s numbers, Q3 FY2026 reported Mar 10, 2026: • RPO: $553 billion, up 325% YoY. For context, RPO was $138B at end-FY2025, then $455B in Q1 FY2026, $523B in Q2, $553B in Q3. That’s +$415B in 9 months. • Driven by AI infrastructure: “Large multi-year AI infrastructure contracts drove the surge”. Oracle signed a $30B/year deal with OpenAI in July 2025 and “four multi-billion-dollar contracts with three customers in Q1”. • Conversion already happening: Cloud infrastructure revenue grew 52%, then 55%, then 68%, then 84% YoY over the last 4 quarters. Total cloud revenue hit $8.9B in Q3, up 43.5% YoY. • Guidance: Q4 cloud growth 46-50%, FY2026 revenue $67B, FY2027 target raised to $90B. OCI alone guided from $18B FY2026 to $144B by FY2030. Why RPO matters vs. cash burn: Oracle’s customer-prepayment model means “most equipment needed is either funded upfront via customer prepayments so Oracle can purchase the GPUs, or the customer buys the GPUs and supplies them to Oracle”. That reduces balance sheet risk while locking in demand. OpenAI: Big User Numbers, But Revenue Targets Are Wobbling OpenAI’s disclosed scale: • Revenue: $2B monthly as of end-March 2026, implying ∼$24B annualized run rate. Sacra estimates $25B ARR Feb 2026. OpenAI generated $4.3B in H1 2025, +16% vs all of 2024. • Users: 900M+ weekly active ChatGPT users, 50M+ paying subscribers. Only ∼5.5% of 900M users pay. • Targets: Expected to 3x revenue to $12.7B in 2025 and double again to $29.4B in 2026. Cash-flow positive not until 2029 at $125B revenue. The misses: • WSJ reported OpenAI “missing its own revenue targets, falling short of user benchmarks” and “missed multiple monthly revenue targets in early 2026”.

Why Wall Street is Wrong About META (And Why It’s Undervalued) Every time Mark Zuckerberg announces a massive infrastructure spend, Wall Street panics, the stock dips, and intelligent investors get a massive buying opportunity. We saw it during the 2023 "Year of Efficiency" setup, and we are seeing it again right now. Despite a massive run over the past 2-3 years, the data shows that META is fundamentally undervalued based on its future earnings growth. Here is the breakdown of why the market is mispricing this phenomenal company. 1. The "Capital Expenditure Anxiety" Discount The market is currently punishing META because management raised 2026 capital expenditure (CapEx) guidance to a staggering $125B – $145B, driven heavily by investments into Meta Superintelligence Labs and their new "Muse" AI models. Short-sighted traders view this as a cash drain. But here is the reality: This isn't burning cash; it's an aggressive moat expansion. My Bullish Argument: Unlike its mega-cap peers, META doesn't have to borrow money to build AI infrastructure. So-far they are funding this entirely out of free cash flow generated by the Family of Apps (Facebook, Instagram, WhatsApp), which currently pulls in over 3.56 billion daily active users. 2. The Core Numbers (Growth vs. Multiple) When evaluating high-growth tech, you always want to look at what you are paying for every dollar of future growth. - The Trailing P/E: META currently sits at roughly a 22.6x trailing P/E ratio. For context, its 10-year historical average is closer to 26.8x. You are buying it at a 16% discount to its historical self. - The Forward P/E: Looking at consensus estimates for the end of 2026, META’s forward P/E drops to a microscopic 19.0x. - Let’s put that into perspective against its mega-cap peers: Microsoft (MSFT) Forward PE ~25.4x EPS Growth ~12-14% Amazon (AMZN) Forward PE ~30.8x EPS Growth ~11-13% Meta Platforms (META) Forward PE ~19.0x EPS Growth ~15.7% The Disconnect: META is projected to grow its top-line revenue faster than Microsoft, yet it trades at a massive valuation discount. You are essentially getting a hyper-growth AI leader at a legacy *value multiple*.