Verdict: KEEP | LR 1.2 | Mega-cap anchor, 6.79% combined weight

Alphabet is the #4 QQQ constituent at 6.79% combined weight (GOOGL 3.53% + GOOG 3.26%). Mega-cap anchor — held at benchmark weight by construction. No removal consideration. This is informational coverage with read-through value for the AI infrastructure complex.

Fundamentals are accelerating across all three segments simultaneously. The stock is deeply oversold (RSI 20.3, down 12.5% in a month) at 20.4x forward earnings — cheap for a company growing revenue +15% with expanding margins. Q1 earnings April 23 are the next catalyst. The primary tension is the capex/depreciation ramp vs revenue acceleration, and the market is currently pricing the former while underweighting the latter.

Financial Overview (FY 2025)

Q4 2025: Revenue $113.8B (+18% reported, +17% CC), EPS $2.82, OCF $52.4B (record quarter). $2.1B Waymo SBC charge inflated expenses ≈$0.17/share.

Segment Detail

Search ($224.5B, +13% FY / +17% Q4): Accelerating, not decelerating — the opposite of the AI disruption narrative. AI Overviews reaching 2B+ users, appearing in 18% of searches and 57% of long-tail queries. AI mode queries per user doubled since launch, sessions 3x longer. Gemini chatbot market share surged from 5.7% to 21.5% (Jan 2025 to Jan 2026) while ChatGPT dropped from 86.7% to 64.5%. Google's 81.6% global search share stable. OpenAI's ChatGPT Instant Checkout abandoned after near-zero completion rates.

YouTube ($40.4B ads, +12%): Q4 was +9% on election comp lapping. Annual revenue surpassed $60B across ads + subs. Shorts at 200B+ daily views.

Cloud ($58.7B, +36% FY / +48% Q4): The star segment. Margin expansion: 5.2% (FY23) to 14.1% (FY24) to 23.7% (FY25). Q4 hit 30.1%. Backlog $240B (+55% QoQ, doubled YoY). 14 product lines >$1B annual revenue. $1B+ deals in 2025 surpassed prior three years combined. Customers outpacing initial commitments by >30%. Apple partnership as preferred cloud provider + developing next-gen Apple Foundation Models on Gemini.

The Capex/Depreciation Tension

Central question for any GOOGL investor right now. The capex number is known. The depreciation schedule is not.

The capex: FY25 $91.4B. FY26 guided $175-185B. Nearly doubling. 60% servers (4-5 year life) / 40% data centers + networking (15-25 year life).

The depreciation: FY25 $21.1B (+38% YoY), Q4 run-rate ≈$5.3B/quarter. CFO Ashkenazi warned: "growth rate in 2026 depreciation to accelerate in Q1 and meaningfully increase for full year." She gave an adjective, not a number. That's a tell — you give numbers when the news is manageable.

Depreciation estimate for FY26:

At midpoint ($30B): +$9B incremental depreciation. After tax (≈21%): $7B net income drag, or ≈$0.57/share.

Forward P/E of 20.4x implies FY26 consensus EPS of ≈$13.42 — requiring +24% EPS growth. Revenue growing 15%. The remaining 9% must come from Cloud margin expansion (plausible — Q4 hit 30.1%) and operating leverage, but depreciation works against both. If depreciation comes in at the high end ($33B+), the +24% growth is very tight. If the low end ($26B), the "cheap at 20x" narrative is correct.

This resolves April 23. Q1 depreciation will reveal the FY26 trajectory.

Capital structure response: LT debt surged from $10.9B to $46.5B in one year. Feb 2026: additional ≈$27B issuance including a 100-year bond at 6.125% — extending duration aggressively. Buybacks slowed 27% ($45.4B vs $62.2B). Net cash ≈$74B ($120.8B cash less $46.5B debt), but 2026 capex ramp erodes this: if OCF ≈$190B and capex $180B, FCF drops to ≈$10B, far below buybacks + dividends (≈$50B). Net cash likely declines $30-40B in 2026.

The FCF compression nobody's discussing: FY25 FCF $73.3B (flat YoY — all OCF improvement eaten by capex). FY26 FCF could compress to ≈$10B. At $3.3T market cap, FCF yield goes from 2.2% to ≈0.3%. The sell-side values GOOG on P/E, not P/FCF — treating it like Amazon in an investment cycle. They may be right. But if the capex cycle extends to 2027-2028, the buyback machine that returned $107.6B over 2 years grinds to a halt.

Efficiency offset: 78% reduction in Gemini serving unit costs over 2025 (Pichai, Q4 call). Revenue from AI-built products +400% YoY in Q4 (Schindler, Q4 call). 10B+ tokens/min served (up from 7B Q3). Approximately 50% of code now written by AI coding agents (Ashkenazi, Q4 call). Unit economics improving dramatically even as capex doubles.

Tax law tailwind: July 4, 2025 law provides immediate R&D expensing + accelerated depreciation on eligible capex. Cash flow benefit, not P&L — partly explains OCF surge to $164.7B. In 2026, accelerated depreciation on $175-185B capex amplifies cash tax savings further.

Market Consensus

Sell-Side

62 Buy / 6 Hold / 0 Sell (91% bullish, 0% bearish). Median target $375 (+37%). Mean $360 (+31%). Range $185-$405. Not a single analyst has a Sell. Q1 EPS estimate $2.62.

Beat cadence: +39.8% (Q1) -> +5.1% (Q2) -> +26.9% (Q3) -> +6.8% (Q4). Narrowing, but still consistently beating. 4 straight quarters.

Options Market

Positioning and flow are diverging — the signature of a nervous long.

IV term structure across the earnings event (Apr 10: 35.2% -> Apr 17: 39.3% -> Apr 24: 38.1% -> May 1: 42.0%) implies a roughly +-6-8% earnings move. Wide for GOOG.

Insider Activity

Pichai selling ≈$10M/week (systematic 10b5-1 plan, not discretionary). Brin gifted 437K shares ($120M) — estate planning. No unusual pattern, no signal.

Price Action

RSI 20.3 (extreme oversold for a $3.3T company). Down 12.5% in one month. Below 50-DMA ($318), above 200-DMA ($260). Of the selloff, approximately 9% is beta drag (1.11 beta x Nasdaq correction) and 3-4% is idiosyncratic (LA social media verdict + capex anxiety).

What the Consensus Actually Is

The Street says buy aggressively at $375. The options market agrees directionally but is hedging hard — put flow, fat skew, elevated IV. The tape says something is wrong right now. The real consensus isn't "$375 with confidence." It's "probably right but nervous about the path" — exactly what the options positioning reveals.

Mispricing Assessment

No actionable mispricing for the basket.

The one tension worth monitoring: the Q1 depreciation number resolves the FY26 EPS trajectory. If depreciation steps up moderately ($6-6.5B in Q1 vs $5.3B Q4 run-rate), margins hold and the stock is genuinely cheap at 20x. If it steps up sharply ($7-8B+ in Q1), FY26 EPS estimates get cut and the "cheap" valuation is an illusion built on the wrong denominator.

The FCF compression from $73B to ≈$10B is a medium-term risk the sell-side is glossing over by valuing on P/E. If the capex cycle extends beyond 2027, net cash declines and the buyback machine stops. But the market is treating this as an investment cycle (correctly, based on AI revenue +400% and serving costs -78%), not a structural deterioration.

At $3.3T market cap with 50+ analysts and 63% idio variance, there is zero informational edge. The market is approximately right — with a wide error bar around depreciation that resolves April 23.

Antitrust / Legal

Three active proceedings, none resolving in basket window:

1. DOJ Search: Final judgment Dec 2025 (distribution restrictions, data sharing, GenAI covered). Google appealed Jan 2026. DOJ cross-appealed Feb 2026. D.C. Circuit hearing late 2026. Chrome/Android divestiture rejected. Apple TAC (≈$20B/yr) at risk.
2. DOJ Ad Tech: Mixed decision Apr 2025. DOJ seeking publisher-side structural remedies. Awaiting final judgment.
3. EC Ad Tech: EUR 3.0B fine Sep 2025. Appealed.
4. Social media litigation: LA jury found YouTube negligent in addiction case (Mar 25, 2026). $1.8M of $6M verdict. First jury validation of this legal theory. 2,000+ similar lawsuits pending. Aggregate exposure ≈$3.6B if all settle at similar levels — roughly 1 quarter's FCF. Financially immaterial.

For 15-week window: zero probability of resolution. Background noise, not catalysts.

Factor Profile

Idio variance ≈63% is below 75% target. This is a market-driven name — moves with QQQ. That's exactly what you want from a mega-cap anchor: high correlation to the benchmark means holding at benchmark weight adds zero tracking error.

Informational edge: zero. $3.3T cap, 62 analysts. Value to basket is as an anchor, not an alpha source.

15-Week Window Assessment

Q1 earnings (Apr 23) — the only event that matters:

• 4 consecutive beats (avg +19.7%), cadence narrowing (+39.8% -> +6.8%)
• Cloud momentum should continue (+48% Q4, supply-constrained, $240B backlog)
• Depreciation will be the headline — "accelerating" per CFO
• 70% probability of EPS beat
• Market pricing +-6-8% implied earnings move (elevated for GOOG)

Macro/tariff risk: 70%+ of revenue from advertising, which per 10-K "tends to correlate with overall economic conditions." April 2 tariff deadline could slow ad spending. Revenue is 48% US, 29% EMEA, 17% APAC. Advertising cyclicality means GOOG underperforms QQQ in sharp macro drawdowns. Beta 1.11 confirms.

Scenarios for basket window:

Expected excess return vs QQQ: roughly flat. Beta 1.11 means slight underperformance in continued down-tape, slight outperformance in recovery. Not a source of filtration alpha either way.

Verdict: KEEP

Mega-cap anchor at benchmark weight. Fundamentals accelerating. Cloud becoming second growth engine with tripling margins. Search defying AI disruption narrative (+17%). At 20x forward, a 15% grower at a discount to QQQ's multiple.

No edge. No removal consideration. Q1 Apr 23 resolves the depreciation uncertainty and is likely a positive catalyst (70% beat probability). Antitrust is multi-year noise.

Only risk is broad market selloff hitting beta-1.1 names. That's market risk, not idio — can't hedge by removing an anchor.