Apple (AAPL) HOLD^* Verdict

The setup

On June 8, 2026, Apple closed its WWDC keynote at $313.97. Twenty-four hours later it closed at $290.55 — a −7.5% repricing in a single day. The Occam's Forensic Jury — a multi-LLM forensic equity research methodology built by John Gillespie at InsightfulAgents.AI LLC — had already converged. Where a week earlier two of four lenses dissented, by the June 9 re-run all four lenses had landed on SELL. The methodology read the disclosure stack before the market repriced. This page is the audit trail of what it read.

The four lenses

Each lens runs independently against the same source material (SEC filings, earnings transcripts, third-party financial data) and produces its own verdict. The lenses do not see each other's outputs during analysis. After all four return verdicts, a cross-method composition produces the final verdict; when the methods disagree, the conflict is flagged on the record and the jury dissent is preserved verbatim rather than averaged away.

• The Auditor reads the books — cash flow, balance sheet, capital efficiency, the Cash Rule of 40.
• The Architect reads the words — disclosure language, hedging patterns, what management does and doesn't admit in its prose.
• The Storyteller reads the absences — what's missing from the narrative; the omissions tell you something.
• The Sentinel reads the outside world — analyst targets, ownership filings, legal proceedings, anything outside the SEC filing.

00:55What signal is the headline missing?

Apple's headline numbers in the most recent quarter were strong: Services hit an all-time revenue record, the company authorized a fresh $100 billion buyback plus a dividend raise, Greater China returned to double-digit growth, and gross margin has run above the high end of guidance in recent quarters. None of that is in dispute. The question the methodology asks is different: what is the disclosure language preparing for that the headline numbers don't explain?

Five threads triangulate to the same answer — an emerging Apple-on-Google dependency stack that the filings were quietly documenting and that WWDC 2026 made explicit.

1. The Apple Intelligence MD&A demotion

Management's Discussion and Analysis is the section where management characterizes its own business and growth drivers. Across the three most recent comparable 10-Q filings — all covering the calendar quarter ending March 2026 — the count of AI-related mention-zones in MD&A was Microsoft 62, Alphabet 9, Apple 0. Apple Intelligence appears zero times in the FY25 10-K MD&A and zero times in the Q2 FY26 10-Q MD&A. Apple's AI language lives in the risk-factor section — the place a company flags concerns, not the place it describes how the business runs. Meanwhile, when the FY25 10-K names what drove Services growth, it names three things: “advertising, the App Store and cloud services.” None of them is AI.

2. The earnings-call confirmation

A second, independent instrument tells the same story. Apple Intelligence mentions on earnings calls ran 3–6 before launch, spiked to 34 at the FY24 Q3 announcement, then cooled in lockstep: 31, then 11, then 6, then 5. Management's spoken commentary on AI faded in step with the filing-language demotion. Two instruments, one direction.

3. The Berkshire trim, timestamped against the disclosure stack

Berkshire Hathaway reduced its Apple position by roughly 75% over nine quarters — from 915.6 million shares (Q2 2023) to 227.9 million (Q1 2026). The three largest trims were 13F-printed within about two weeks of the Apple filings carrying these disclosure signals. The methodology does not claim Berkshire's reasoning. It documents the timestamp pattern: whatever the filings contained was on the public record when the position-change prints landed.

4. Google search distribution disclosed as a material revenue risk

Apple's Q2 FY26 10-Q adds new emphasis to a risk it now spells out directly: a reversal on appeal of the DOJ antitrust remedy against Google could result in remedies “prohibiting Google from offering the Company commercial terms for search distribution… [which] could materially adversely affect the Company's ability to earn revenue from such licensing arrangements.” The largest single component of Services revenue is widely reported to be Google search-distribution payments — on the order of $20 billion or more per year. The filing now flags that this engine has a single-customer dependency subject to a third-party legal outcome.

5. WWDC 2026 made it explicit

At WWDC 2026, Apple licensed Google's 1.2-trillion-parameter Gemini model to power the load-bearing layer of the rebuilt Siri, paying Google an estimated $1 billion per year in a multi-year deal. Apple now has two commercial dependencies on Google pointing in opposite directions — receiving ~$20B+/year in search-distribution revenue in, paying ~$1B/year in Gemini licensing out — at the precise moment the Q2 FY26 disclosure stack flags the inbound stream as at DOJ-remedy risk. The “commercial relationships” language Apple added to its risk factors in FY25 reads now as forward preparation for exactly this.

Why the intrinsic value is $157.59

The single load-bearing input is the DCF growth rate. The methodology anchors it to Apple's most recently completed fiscal year — roughly 6.24%, a smoothed version of Apple's reported FY2024-to-FY2025 revenue growth of +6.43% ($391.0 billion to $416.2 billion). This is trailing-actual annual growth, not a forward forecast. Wall Street consensus projects forward at roughly 8–11% CAGR. The methodology chose observed performance over forward projection.

That conservative choice — paired with an 8.05% WACC and a 3.00% terminal growth rate — produces the $157.59 base intrinsic value and the −45.76% margin of safety. It is a deliberate, disclosed assumption, not a bearish thumb on the scale. If Apple sustains the higher forward growth rate the bull case assumes, the gap narrows materially. The DCF reports what the most recent actual results discount to.

The composition: how a unanimous SELL becomes a HOLD*

All four lenses dissented to SELL. The discounted-cash-flow method returned STRONG SELL. Yet the final verdict is HOLD with an asterisk. That is not a contradiction — it is the cross-method composition working as designed.

Alongside the absolute DCF, the methodology runs a parallel relative-valuation method that ranks Apple against its mega-cap peer cohort. On that method Apple sits at roughly the 65th percentile (FAIR) with 88th-percentile quality: the market has already priced Apple's operational superiority, so no relative discount is available even as the absolute DCF says the stock is expensive. The two methods disagree, and the methodology does not paper over that disagreement — it composes to HOLD*, and the asterisk carries the unanimous jury dissent onto the record. Conviction is LOW (35/100) precisely because the methods diverge this widely.

Usually the asterisk preserves a split jury — two lenses one way, two the other. This cycle it does something different: it preserves a jury that is unanimous in dissent against the composed verdict. That is the asterisk doing more work than usual, and it is the reason this verdict is worth reading rather than rounding to a label.

What would change the verdict

The Forensic Jury's bear-leaning read on Apple is conditional. It would weaken or reverse if any of the following materialize:

1. Apple sustains forward revenue growth meaningfully above the 6.24% trailing-actual baseline over multiple quarters, replacing the conservative DCF input.
2. Apple begins to characterize its AI roadmap as a growth driver in MD&A — not just in risk factors — with revenue attribution to match.
3. The DOJ antitrust appeal resolves in a way that removes the disclosed risk to Apple's Google search-distribution revenue.
4. The Gemini-powered Siri demonstrably converts into Services monetization that offsets the licensing cost and the AI-narrative gap.

Sources

AI mention-zone counts come from each company's most recent 10-Q, all covering the calendar quarter ending March 2026, deduplicated within the Management's Discussion and Analysis section. Source documents at SEC EDGAR:

• Apple Inc. Form 10-Q, fiscal quarter ended March 28, 2026 — filed May 1, 2026 (accession 0000320193-26-000013)
• Alphabet Inc. Form 10-Q, fiscal quarter ended March 31, 2026 — filed April 30, 2026 (accession 0001652044-26-000048)
• Microsoft Corporation Form 10-Q, fiscal quarter ended March 31, 2026 — filed April 29, 2026 (accession 0001193125-26-191507)
• Apple Inc. Form 10-K, fiscal year ended September 2025 (Services growth attribution; risk-factor language); Berkshire Hathaway Form 13F filings (position-change timeline).