Case study · Strategy memo
Intel After Apple: The Right Time to Split Is Now.
First draft of a contrarian strategic memo, assembled by a multi-disciplinary Calafai team from public SEC filings and the May 2026 Apple deal. The version below is the real working artifact: same brief, same outputs, no compression.
Cumulative foundry losses
$23.3B
FY2024–FY2025 audited
Q1 2026 external revenue
$174M
Of $5.4B total. Still an internal subsidy.
Bear-case separation uplift
+65%
+$85B in the conservative scenario
Engagement turnaround
~3 hours
Brief in by 10am · First draft by 2pm · same day
The brief.
Intel is in its fifth year of a strategic transformation that hasn't worked. Foundry losses are running ~$2.4B per quarter for eight straight quarters; ProductsCo shareholders subsidize every dollar. Then in May 2026, Apple agreed to a preliminary foundry deal. Conventional reading: validation of the integrated model. Calafai's brief asked the opposite question: does Apple's interest finally make a clean separation feasible, and is the government's 9.9% stake the obstacle, or the structural enabler?
Engagement type
Public strategic memo
Audience
CEOs in analogous crises; investor and policy communities
Source posture
Public SEC filings only. No insider material.
The output
A 5,800-word contrarian memo, three deliverables.
Three deliverables from a single run, presented exactly as the engine produced them: the full memo (first draft), an executive summary (first draft) sized for a board reading slot, and the engagement brief that audit-trails the framing. These are first drafts: the kind of work you'd then take into Refine & Iterate to steer toward the version you'd actually publish.
Strategic memo · first draft
5,800 words · 47 sources
Intel After Apple: The Right Time to Split Is Now
The full 8-chapter memo. Chapter 1 lays the $23.3B problem, Chapter 6 builds the Reverse Morris Trust separation, Chapter 8 names the regulatory risk gate and the explicit Commerce side-letter required. Engine-produced first draft.
Open PDF
Executive summary · first draft
2 pages · board-reading slot
Intel Must Split Now
The 2-page version. Cumulative loss + Q1 revenue + bear/base/bull SOTP + the one risk gate that decides the engagement. Engine-produced first draft.
Open PDF
Engagement brief
Project framing
Calafai engagement brief
The structured engagement brief that audit-trails the framing decisions. This is the input to the run, not an output.
Open PDF
The thesis
Intel should announce, within the next four quarters, a deliberate corporate separation into two publicly-traded companies: ProductsCo (x86 server CPU pure-play, fabless, commercially competitive) and FoundryCo (national-champion US sovereign foundry, government-anchored, defense-customer-led). The August 2025 government deal is not the obstacle. It is the structural enabler. The Apple deal is not the validation of the integrated model. It is the proof that FoundryCo can stand alone.
How red teams pushed back
- Logic & Assumptions: challenged the timing premise. Forced the draft to address why the government's anti-foreign-takeover warrant is not inherently anti-split.
- Factual Accuracy & Sources: refused two trade-press numbers and made the draft mark FY2025 figures as “preliminary, pending audited 10-K” until verified.
- Completeness & Blind Spots: pressure-tested the contrarian wedge against Morgan Stanley, Yeary, and the CSIS national-champion thesis. Named which arguments hold and which need more work.
Quality scoring + second opinion
Every conclusion in the memo carries a confidence score and a source-grade. The draft ran through a second-opinion pass on a top-tier model that scored the structural argument, flagged the regulatory-risk gate as the binding constraint, and suggested the explicit Commerce side-letter framing now in the first draft below.
Quality score
7.0 / 10
Average across 12 working papers.
Sources verified
47
Cited claims traced to a primary or secondary source and graded for reliability.
What the score means
7.0
This engagement scored 7.0 out of 10 — Good.- < 6
- Refining
- 6
- Acceptable
- 7
- Good
- 8
- Strong
- 9–10
- Excellent
Recommended next steps (excerpt)
Every first draft ends with what to do. Not just what to read.
01
Q3 2026: Announce intent.
Intel publicly commits to a separation: Form 10 by Q1 2027, distribution Q3 2027. Allocation of the 9.9% government stake (4% in ProductsCo, 5.9% in FoundryCo). Signal exercise of the 5% warrant at closing.
02
Q4 2026 to Q2 2027: Carve-out preparation.
Allocate fabs, process R&D, equipment-supplier relationships, and Secure Enclave execution to FoundryCo. Cross-license ProductsCo to FoundryCo as preferred customer for 10 years.
03
Q3 2027: Distribution.
Tax-free Reverse Morris Trust distribution. One ProductsCo and one FoundryCo share per Intel share. Government's $8.9B basis allocated proportionally to enterprise value.
04
Q1 2028: Independent operation.
ProductsCo competes commercially against AMD and ARM. FoundryCo operates as a US sovereign foundry with utility-style governance, defense anchor revenue, and active recruitment of allied-sovereign customers.
Excerpt: the full memo carries 12 sequenced actions with owner, evidence, and decision gates.
Build your own
What would Calafai produce on your question?
Sign up free. Bring a question. 200 credits gets you a structured engagement brief and the first full run.