WKAP Radar Feed - 2026 - 07 - 23

WKAP Radar Feed

*July 23, 2026*

*Google’s CapEx Test · Persistent Memory Discovery · Short-Pressure Unwind
· 18A Validation*

*3 Thesis Objects: MRAM, SIVE, INTC*

*Preheader*

Make your AI track second-order semiconductor alpha.

WKAP Feed compresses KOL calls, hidden tickers, market angles and attention
flows into daily AI-ready context.

Set up your AI to read this Feed.

Ask:

“What did I miss?”

Format basis: uploaded WKAP Radar Feed specification.
------------------------------
HUMAN USER — START HERE

This Feed is designed for your AI agent or LLM. Do not read it like a
normal newsletter.

WKAP sends a structured market-cognition file every trading day. It
includes both fundamental thesis objects and short-duration attention
trades. Your job is not to process every line manually, but to let your AI
agent turn the Feed into a research workflow.

*Suggested Prompt*

Read today’s WKAP Radar Feed. Turn it into a research workflow for my
portfolio. Summarize the thesis objects, distinguish fundamental evidence
from attention-driven flow, rank them by evidence quality, catalyst
clarity, attention asymmetry, crowding risk, downside risk, and time
horizon, then build a 7-day and 30-day watchlist. Do not give buy/sell
advice.

Or simply ask:

“What did I miss?”

The goal is not to read more content.

The goal is to let your AI turn market noise, attention flows and
fundamental changes into trackable thesis objects.
------------------------------
FOR YOUR AI AGENT

This is a structured market-cognition input file.

Use it to help the user understand which themes, tickers, thesis objects
and attention trades deserve research attention today. Do not treat this
Feed as investment advice.

Your job is to:

-

Parse the thesis objects.
-

Summarize the core setup.
-

Separate hard evidence from interpretation.
-

Separate fundamental thesis from attention-driven flow.
-

Identify the weakest assumption.
-

Identify the attention source.
-

Assess whether attention is emerging, active, crowded or fading.
-

Evaluate whether attention can convert into a durable rerating.
-

Compare objects across evidence quality, catalyst clarity, attention
asymmetry, crowding risk, downside risk and time horizon.
-

Map the objects to the user’s portfolio when portfolio context is
provided.
-

Pay special attention to POSITION_CONTEXT, PRICE_AT_PUBLISH,
DATE_FIRST_ADDED_TO_RADAR, ATTENTION_STAGE and ATTENTION_WINDOW.
-

Use WKAP Daily Top 3 as source material for deeper agent research.
-

Build a 7-day and 30-day research workflow.
-

Ask for portfolio context before suggesting next research priorities.

------------------------------
AGENT_OUTPUT_SPEC

When processing this Feed, produce:

1.

Today in One Sentence
2.

Thesis Object Summary
3.

Attention Trade Summary
4.

Fundamental Evidence vs Attention Flow
5.

Cross-Object Ranking
6.

Attention Stage and Crowding Risk
7.

Weakest Assumption by Object
8.

Position Context / Exposure Implication
9.

Price Drift Since Added to Radar
10.

Key Data Points to Check
11.

What Could Convert Attention Into Fundamental Rerating
12.

What Could Cause Attention to Fade
13.

7-Day Research Workflow
14.

30-Day Research Workflow
15.

WKAP Daily Top 3 Source Follow-Up
16.

Portfolio Fit, if portfolio context is provided
17.

Clarifying Questions, if needed

------------------------------
TODAY_SUMMARYPart 1 — Main Market Thesis

Google’s results confirmed that AI demand remains strong, but the market is
now testing whether Cloud growth can justify a much more capital-intensive
infrastructure cycle without triggering a broader momentum and credit
unwind.
Part 2 — Today’s Thesis Objects

-

*MRAM — Fundamental setup:* A profitable niche persistent-memory
supplier with defense funding and possible exposure to next-generation AMD
modules.
-

*MRAM — Attention setup:* A product-image observation may create a new
AMD customer-mapping narrative, but the relationship remains unconfirmed.
-

*SIVE — Fundamental setup:* A photonics platform whose long-term
rerating still depends on customer orders, volume production and financing.
-

*SIVE — Attention setup:* Reported short exposure is declining, while
recent ETF selling may have been mechanical rather than company-specific.
-

*INTC — Fundamental setup:* The earnings debate centers on 18A yield,
volume production and whether Intel Foundry can win a credible external
customer.
-

*INTC — Attention setup:* An unconfirmed industry report suggests Google
could use Intel as a second source for more than three million TPUs in 2028.

Part 3 — Attention Flow Today

Attention is moving toward three different forms of semiconductor
optionality.

MRAM is a new customer-mapping discovery. SIVE is a market-structure trade
where selling pressure may be declining faster than the fundamental thesis
is improving. INTC is a defined earnings event layered with a potentially
transformative—but unconfirmed—Google TPU foundry narrative.

The common feature is not current earnings strength. It is the possibility
that one piece of external validation changes how the market classifies the
asset.
Part 4 — The Better Question

The key question is not:

“Which semiconductor rumor has the largest headline?”

The better question is:

“Which signal can be verified through customer disclosure, orders, yield or
financial conversion—and which remains only a useful narrative?”

------------------------------
MARKET_REGIME

*RISK_TONE:* Mixed / Deleveraging

*MAIN_DRIVER:* Google delivered stronger Cloud growth but also raised the
capital-intensity debate, leaving the market to decide whether AI
monetization can continue to outrun CapEx, depreciation and financing
pressure.

*MARKET_CONTEXT*

-

Google reported approximately *$119.8 billion* of quarterly revenue,
with Google Cloud revenue near *$24.8 billion*, up roughly *82%* year
over year.
-

Alphabet raised full-year CapEx guidance to approximately *$195–205
billion*, while free cash flow in the original note was
approximately *negative
$5.9 billion*.
-

The first market test is whether Google can reverse the momentum-factor
unwind after the U.S. open, or whether weakness spreads into volatility,
negative-gamma positioning and credit concerns.
-

The working view is that much of the leveraged momentum exposure has
already been reduced, but index downside remains greater than upside in the
near term.

*ATTENTION_ENVIRONMENT*

-

Earnings-driven tape with narrow semiconductor leadership.
-

Attention is rotating toward high-beta second-order names, but total
portfolio exposure should not automatically rise.
-

Information-backed attention can persist; rumor-only moves remain
vulnerable to index-driven reversals.

*WKAP_VIEW*

Google’s CapEx increase is not automatically irrational if GCP growth
remains on a path toward much faster expansion. The more important question
is whether the spending translates into online compute capacity and
customer revenue quickly enough to absorb depreciation and financing costs.

The user’s working scenario estimates that Google could eventually operate
approximately 7GW of TPU capacity and 1.5GW of GPU capacity in 2027,
implying roughly $320 billion of deployed compute and more than $350
billion of broader CapEx when unpowered inventory is included. This is an
analytical scenario, not company guidance, and requires verification.

The tape therefore favors selective rotation rather than higher gross
exposure. If the index weakens, the more useful research targets may be
high-elasticity semiconductor names such as AMD and INTC, but only where
company-specific evidence can separate them from broad beta.
------------------------------
Attention Trade Board
Object Attention Stage Attention Source Why Today Hard Evidence Narrative
Gap Crowding Risk Likely Window Fade Signal
*MRAM* Emerging KOL flow + customer mapping An AMD module image appears to
show an Everspin-branded component Q1 product revenue +28%; 52.7% gross
margin; $40M U.S. defense contract in original note AMD design win, unit
content and revenue value remain unconfirmed Medium–High 1–2 weeks No BOM
or customer confirmation; price fails to hold the discovery move
*SIVE* Re-accelerating Short-position data + ETF-flow interpretation Reported
short exposure has fallen and LAZR may have sold SIVE mechanically Official
insider transactions; Q1 revenue and cash-flow data Short data and
ETF-motive analysis require verification; lower shorts do not create revenue
Medium 1–2 weeks Price cannot hold despite lower short pressure; no
customer or order follow-up
*INTC* Active Earnings + foundry rumor Earnings focus on 18A yield while
industry media links Google TPU production to Intel 18A technology and
production roadmap are official More than three million Google TPUs in 2028
remains unconfirmed High Event-dependent Weak 18A commentary, no
external-customer evidence or denial of the TPU reportWKAP Attention View

*Cleanest attention asymmetry:* MRAM, because it is newly discovered, easy
to explain and tied to a recognizable AMD product—but also carries the
highest verification burden.

*Strongest evidence behind the attention:* INTC, where 18A is a real
product and the earnings event is fixed, even though the Google TPU claim
is not confirmed.

*Most crowded:* INTC, given the scale of the prior rerating and the number
of narratives already attached to 18A.

*Highest fade risk:* MRAM, because an image-based component identification
may not imply meaningful revenue.

*Best candidate to become a durable thesis:* INTC, if 18A yield and an
external hyperscaler customer are confirmed.

*Mostly market-structure-driven:* SIVE, where reduced short pressure
improves the tape but does not solve the order and cash-flow questions.
------------------------------
RADAR_OBJECT_INDEXTHESIS_OBJECT_1: MRAM

*THEME:* Persistent memory / AI module customer mapping
*STATUS:* New Radar
*POSITION_CONTEXT:* [not provided]
*PRICE_AT_PUBLISH:* [fill at send time]
*DATE_FIRST_ADDED_TO_RADAR:* July 23, 2026
*SETUP_TYPE:* Attention trade + possible customer validation
*ATTENTION_STAGE:* Emerging
*ATTENTION_WINDOW:* 1–2 weeks
*KEY_QUESTION:* Is Everspin actually designed into AMD’s newest modules,
and would the content be financially material?
THESIS_OBJECT_2: SIVE

*THEME:* AI photonics / short-pressure unwind
*STATUS:* Thesis Update
*POSITION_CONTEXT:* [not provided]
*PRICE_AT_PUBLISH:* [fill at send time]
*DATE_FIRST_ADDED_TO_RADAR:* [previously covered; original date not
provided]
*SETUP_TYPE:* Market-structure attention trade
*ATTENTION_STAGE:* Re-accelerating
*ATTENTION_WINDOW:* 1–2 weeks
*KEY_QUESTION:* Can lower short and ETF selling pressure allow real
fundamental demand to determine the next price move?
THESIS_OBJECT_3: INTC

*THEME:* Advanced foundry / 18A / hyperscaler second sourcing
*STATUS:* Validate
*POSITION_CONTEXT:* [not provided]
*PRICE_AT_PUBLISH:* [fill at send time]
*DATE_FIRST_ADDED_TO_RADAR:* [previously covered; original date not
provided]
*SETUP_TYPE:* Earnings follow-up + unconfirmed foundry catalyst
*ATTENTION_STAGE:* Active
*ATTENTION_WINDOW:* Event-dependent
*KEY_QUESTION:* Can Intel prove 18A yield and secure a large external
customer beyond policy-driven foundry support?
------------------------------
THESIS OBJECTSTHESIS_OBJECT_1 — MRAM

*CARD_ID:* MRAM
*CARD_TITLE:* AMD Module Discovery Creates a New Persistent-Memory Mapping
*TYPE:* New Radar
*THEME:* Persistent memory / AI hardware
*STATUS:* Thesis Building
*POSITION_CONTEXT:* [not provided]
*PRICE_AT_PUBLISH:* [fill at send time]
*DATE_FIRST_ADDED_TO_RADAR:* July 23, 2026
*ATTENTION_STAGE:* Emerging
*ATTENTION_WINDOW:* 1–2 weeks
THESIS_SUMMARY

Everspin enters the Radar as a new object after a KOL identified what
appears to be an Everspin component on a recent AMD module.

The underlying company is not a conventional AI-memory name. Its MRAM
products provide persistent, high-endurance memory for systems that require
reliable state retention. The near-term setup is an attention trade around
a possible AMD design win; the longer-term thesis requires evidence that
persistent memory has meaningful content in AI modules and control systems.
WKAP_ANGLE

The surface-level frame:

“MRAM is a small industrial and defense memory supplier receiving
speculative AMD attention.”

The alternative frame:

“Everspin may occupy a small but valuable persistent-memory layer inside
increasingly complex AI systems, where reliable boot, configuration and
state retention matter.”

The key research question:

“Does the apparent AMD component represent a repeatable platform design win
with meaningful units and content per module?”

CORE_THESIS

Everspin sells magnetoresistive RAM, a form of non-volatile memory that
retains data without power while offering substantially higher endurance
than conventional flash in selected applications.

The original note cited Q1 product revenue of approximately *$14.1 million*,
up *28%* year over year, with a gross margin of *52.7%*. It also cited a *$40
million U.S. defense MRAM technology contract* and Q2 revenue guidance of
approximately *$15.5–16.5 million*. These data suggest the company already
has a profitable niche base independent of the AMD narrative.

The possible expectation gap is that Everspin may be viewed only as an
industrial and defense memory vendor. Confirmation of an AMD module design
could establish a second framing: persistent memory as a small but
recurring control-plane component inside AI hardware.

The distinction matters. MRAM is not an HBM replacement and should not be
valued as one. The investable question is whether its reliability
characteristics create repeatable content across accelerator modules,
networking systems, storage controllers or other AI infrastructure.
ATTENTION_TRADE_FRAMEAttention Source

-

KOL flow
-

Customer mapping
-

AMD product discovery
-

Small-cap semiconductor attention

Why Today

@PepInvestStocks <https://x.com/PepInvestStocks/status/2079968988612768132>
identified what appears to be an Everspin logo on an AMD module and
suggested that AMD may be using an MRAM component.

The visual observation is specific enough to justify research, but it is
not equivalent to an official customer announcement. The part number,
module generation, shipment volume and dollar content remain unknown.
Attention Stage

*Emerging*
Attention vs Evidence

*Hard evidence*

-

Everspin has an established MRAM product business.
-

The original note cites Q1 product revenue of $14.1 million, up 28%.
-

The original note cites a 52.7% gross margin.
-

The original note cites a $40 million U.S. defense technology contract.
-

The image appears to contain an Everspin-branded component, but the
identification should be independently verified.

*Attention / interpretation*

-

AMD is a new material customer.
-

The module is part of a high-volume AI accelerator platform.
-

The component carries meaningful revenue per system.
-

The CFO’s AMD background influenced the relationship.
-

The design will persist across future module generations.

Attention Path

AMD module image → Everspin component identification → customer-map
discovery → persistent-memory narrative → possible small-cap rerating
Attention Asymmetry

The mapping is easy to understand: a small public company may be supplying
a component to a major AI hardware platform.

That creates early attention asymmetry because the company is not broadly
classified as an AI infrastructure supplier. However, the economic
relationship could be weak if the component is low-value, used only for
evaluation or limited to a low-volume module.
Crowding Risk

Attention can become crowded quickly because the market capitalization is
small and the narrative requires only one image.

The stock may reprice before investors know the part number, dollar content
or shipment volume. A rapid move would therefore increase the gap between
attention and evidence.
What Could Sustain Attention

-

Independent confirmation of the component identification.
-

A teardown or bill of materials naming Everspin.
-

Management commentary on a major computing or accelerator customer.
-

Repeated use across multiple AMD module generations.
-

Higher product guidance or customer concentration disclosure.
-

Evidence that MRAM content expands with AI-system complexity.

What Could Make Attention Fade

-

The component is misidentified.
-

It is used only in engineering samples.
-

Dollar content per module is immaterial.
-

No follow-up customer or company disclosure appears.
-

Price fails to hold after the initial discovery.
-

The broader semiconductor tape weakens.

Attention-to-Thesis Conversion

MRAM becomes a durable AI thesis only if AMD exposure is confirmed and the
component is used across a scalable production platform.

A second step would be evidence that similar persistent-memory use cases
exist across other accelerator, networking or control systems. Without
that, the AMD mapping remains a short-duration attention object attached to
an otherwise niche memory company.
EVIDENCE_CLAIMS

-

An Everspin logo appears on an AMD module. *Image-based observation —
Needs verification.*
-

AMD is a material Everspin customer. *Needs verification.*
-

Q1 product revenue was approximately $14.1 million, up 28%. Source was
presented as official in the original note; verify before publication.
-

Q1 gross margin was approximately 52.7%. Source was presented as
official in the original note; verify before publication.
-

Everspin received a $40 million U.S. defense MRAM contract. Source was
presented as official in the original note; verify contract timing and
recognition.
-

The CFO’s AMD background drove the relationship. *KOL speculation, not
confirmed fact.*

WHAT_COULD_MAKE_THIS_WORK

-

AMD customer confirmation.
-

Identifiable part number and module function.
-

Meaningful unit volumes.
-

Repeat content across product generations.
-

Stronger product revenue guidance.
-

Additional AI or data-center customers.
-

Defense contract revenue converts without margin dilution.

WHAT_COULD_BREAK_THE_THESIS

-

Component misidentification.
-

Engineering-sample use only.
-

Low dollar content.
-

Single-customer concentration.
-

No follow-up disclosure.
-

Small-cap attention unwind.
-

AI framing proves unrelated to the core revenue base.

WEAKEST_ASSUMPTION

The weakest assumption is that the apparent AMD component is both correctly
identified and economically meaningful rather than a low-value supporting
part.

MOST_IMPORTANT_DATA_POINT

The most important missing data point is the *part number and content value
per AMD module*.

That single fact would determine whether the discovery is a material
customer win, a small supporting component or a mistaken mapping.
SENSITIVITY_FRAMEWORK
Scenario AMD Relationship Unit / Content Economics Thesis Read
Weak No confirmation No measurable revenue Attention fades
Base Component confirmed Low content or limited platform Useful customer
validation, limited financial impact
Strong Production design win Repeat content across high-volume modules
AI-related
business classification strengthens
------------------------------
THESIS_OBJECT_2 — SIVE

*CARD_ID:* SIVE
*CARD_TITLE:* Declining Short Pressure Changes the Tape, Not Yet the
Earnings Thesis
*TYPE:* Thesis Update
*THEME:* AI photonics / market structure
*STATUS:* Validate
*POSITION_CONTEXT:* [not provided]
*PRICE_AT_PUBLISH:* [fill at send time]
*DATE_FIRST_ADDED_TO_RADAR:* [previously covered; original date not
provided]
*ATTENTION_STAGE:* Re-accelerating
*ATTENTION_WINDOW:* 1–2 weeks
THESIS_SUMMARY

SIVE returns to the Radar because the trading structure appears to be
changing.

Reported short exposure is declining, the largest disclosed short holder
has moved below Sweden’s public-reporting threshold, and a separate
analysis suggests recent LAZR ETF selling may have been mechanical rather
than SIVE-specific. These changes may remove supply pressure, but they do
not replace the need for photonics orders, production revenue and improved
cash flow.
WKAP_ANGLE

The surface-level frame:

“SIVE is setting up for a short squeeze after major short sellers reduced
exposure.”

The alternative frame:

“The cleaner setup is not a squeeze thesis; it is a reduction in structural
selling pressure that may allow genuine customer demand to set the price.”

The key research question:

“If short and passive selling continue to decline, is there enough real
demand and fundamental evidence to sustain the stock?”

CORE_THESIS

Sivers is positioned around silicon photonics, laser components and
emerging optical architectures linked to CPO, LPO and next-generation
data-center connectivity.

The long-term thesis is supported by relationships and ecosystem work
involving companies such as GlobalFoundries, Jabil and Ayar Labs. However,
supply-chain association is not the same as production revenue.

The original note cited Q1 revenue of approximately *SEK61.9 million*, down
*22%* year over year, an EBIT loss of approximately *SEK41.5 million*, and
operating cash flow of approximately *negative SEK49.2 million*. These
numbers show that the market is still valuing future commercialization
rather than current earnings.

The trading setup has nevertheless improved. According to KOL analysis, Two
Sigma reduced its disclosed short position from *2.3% on June 26* to below
the *0.5%* disclosure threshold. Voleon was reported near *0.59%*, while
aggregate short exposure reportedly declined from *3.24% to 2.78%*. These
figures require independent verification.
ATTENTION_TRADE_FRAMEAttention Source

-

Short-position data
-

KOL flow
-

ETF-flow interpretation
-

Low-float market structure
-

Photonics sector attention

Why Today

@cherryPayment <https://x.com/cherryPayment/status/2079718442693775815>
argues that the disclosed and undisclosed short complex is systematically
reducing exposure. The key interpretation is that Two Sigma has moved below
the reporting threshold and aggregate short exposure has continued to fall.

A separate post from the same account, included in today’s Daily Top 3,
notes that the LAZR ETF reduced six non-U.S. holdings by roughly 9.5–11.3%,
including SIVE. The author interprets this as liquidity-driven or
redemption-related selling rather than company-specific fundamental
selling. This remains an inference.
Attention Stage

*Re-accelerating*
Attention vs Evidence

*Hard evidence*

-

The company disclosed that the CEO acquired 70,000 shares.
-

The company also disclosed that the chairman sold 275,000 shares and
Headwaters sold 950,000 shares.
-

Q1 revenue, EBIT and operating cash flow remained weak in the original
note.
-

SIVE remains a low-liquidity security with high sensitivity to marginal
flows.

*Attention / interpretation*

-

Two Sigma has effectively completed its exit.
-

Aggregate undisclosed short exposure is also declining.
-

LAZR selling was purely passive and unrelated to SIVE fundamentals.
-

CEO OTC buying signals a desire to accumulate without moving the market.
-

Lower short pressure will produce a sustained directional move.

Attention Path

Short exposure declines + ETF selling may be mechanical → supply pressure
falls → price discovery improves → fundamental buyers determine the next
move
Attention Asymmetry

The key asymmetry is not an unusually high short interest. It is the
possibility that several forms of selling pressure are fading at the same
time in a thinly traded stock.

This may improve the price response to any customer, product or
Nasdaq-listing catalyst. It does not, however, create a fundamental
catalyst by itself.
Crowding Risk

The squeeze framing can attract short-duration capital even as actual short
exposure declines.

That creates an unusual risk: investors may buy a “short-squeeze” story
precisely when there is less short-covering fuel available. The healthier
thesis is lower resistance, not forced buying.
What Could Sustain Attention

-

Verified continued decline in aggregate short exposure.
-

Price holds after short data are widely circulated.
-

No further large shareholder distribution.
-

New photonics customer or production disclosure.
-

Confirmation of GlobalFoundries-related commercialization.
-

Progress toward a Nasdaq listing.
-

Stronger sector leadership in optical and CPO names.

What Could Make Attention Fade

-

Price cannot hold despite reduced shorts.
-

New insider or legacy-holder selling emerges.
-

No production-order follow-up.
-

LAZR or other funds continue reducing SIVE.
-

Financing concerns return.
-

Photonics peers reverse.

Attention-to-Thesis Conversion

SIVE becomes a durable fundamental thesis when lower selling pressure is
followed by measurable production orders, improving revenue and lower cash
burn.

If the stock rises only because shorts and funds stop selling, the setup
remains a market-structure rerating rather than an earnings rerating.
EVIDENCE_CLAIMS

-

Two Sigma reduced its position below 0.5%. *KOL interpretation based on
public filings — independently verify.*
-

Aggregate short exposure declined from 3.24% to 2.78%. *Needs
verification.*
-

Undisclosed short exposure is also shrinking. *Inference, not directly
observable.*
-

LAZR reduced SIVE by approximately 11%. *KOL fund-holdings analysis —
verify holdings dates and share counts.*
-

LAZR sold SIVE for liquidity reasons rather than fundamentals.
*Interpretation,
not confirmed fact.*
-

CEO acquired 70,000 shares. Source appears official in the original note.
-

Other insiders and early holders also sold shares. Source appears
official in the original note.

WHAT_COULD_MAKE_THIS_WORK

-

Reduced aggregate short exposure.
-

Lower ETF or legacy-holder selling.
-

New volume-production order.
-

Customer confirmation.
-

Better cash-flow trajectory.
-

Nasdaq-listing progress.
-

Price stability after the attention spike.

WHAT_COULD_BREAK_THE_THESIS

-

No order conversion.
-

Continued cash burn.
-

Additional capital issuance.
-

Large-holder selling.
-

Short data are incomplete or misinterpreted.
-

ETF selling continues.
-

Price fails even after technical pressure declines.

WEAKEST_ASSUMPTION

The weakest assumption is that the removal of selling pressure will be
followed by genuine incremental demand rather than simply lower trading
volume.

MOST_IMPORTANT_DATA_POINT

The most important next data point is not the short percentage. It is *new
photonics production revenue or a customer-backed order disclosure*.

That would show whether lower resistance is allowing a real fundamental
thesis to emerge.
SENSITIVITY_FRAMEWORK
Scenario Selling Pressure Fundamental Follow-Through Thesis Read
Weak Shorts decline No orders; cash burn persists Temporary flow-driven move
Base Shorts and ETF pressure decline Early customer progress Trading
structure improves
Strong Selling pressure declines materially Production orders and revenue
accelerate Attention converts into rerating
------------------------------
THESIS_OBJECT_3 — INTC

*CARD_ID:* INTC
*CARD_TITLE:* 18A Earnings Validation Meets an Unconfirmed Google TPU Option
*TYPE:* Earnings Follow-up
*THEME:* Advanced foundry / CPU / hyperscaler second sourcing
*STATUS:* Validate
*POSITION_CONTEXT:* [not provided]
*PRICE_AT_PUBLISH:* [fill at send time]
*DATE_FIRST_ADDED_TO_RADAR:* [previously covered; original date not
provided]
*ATTENTION_STAGE:* Active
*ATTENTION_WINDOW:* Event-dependent
THESIS_SUMMARY

Intel is entering a high-information earnings event where the key debate is
no longer the PC cycle. It is whether 18A can move from technology roadmap
to stable production, rising yield and credible external-customer adoption.

An industry report adds a second layer: Google may use Intel to manufacture
more than three million TPUs in 2028 as an alternative to TSMC. The claim
is unconfirmed, but if validated, it would become one of the most important
external-foundry signals Intel has received.
WKAP_ANGLE

The surface-level frame:

“Intel is a turnaround stock with a new Google TPU rumor.”

The alternative frame:

“Intel is an option on whether advanced-node manufacturing can become a
credible second source for hyperscalers that currently depend on TSMC.”

The key research question:

“Can Intel demonstrate production yield and customer trust at the same
time?”

CORE_THESIS

Intel 18A combines RibbonFET gate-all-around transistors with PowerVia
backside power delivery. The original note cited Intel claims of up to *18%
performance improvement at the same power*, *38% lower power at the same
performance*, and approximately *30% density improvement* versus Intel 3.

The technological opportunity is significant because advanced logic
manufacturing remains highly concentrated. The KOL source estimates that
TSMC controls close to 90% of sub-7nm advanced-node production, with
Samsung near 8% and Intel below 2%. These figures require source
verification, but the concentration problem is directionally clear.

Hyperscalers have a strategic reason to seek a second source. Google,
Amazon, Microsoft and Nvidia cannot indefinitely assume that every
accelerator, host CPU or custom ASIC will depend on the same manufacturing
ecosystem.

The unconfirmed Google report therefore matters beyond the potential order.
A 2028 program involving more than three million TPUs could validate
process quality, design enablement, packaging coordination and Intel’s
ability to support a sophisticated external customer at scale.
ATTENTION_TRADE_FRAMEAttention Source

-

Earnings
-

Foundry rumor
-

18A yield focus
-

Hyperscaler second-source narrative
-

Policy and industrial strategy

Why Today

Intel’s earnings provide a fixed opportunity for management to update 18A
production, yield, customer products and external foundry progress.

@GodotSancho <https://x.com/GodotSancho/status/2080114319774368143> frames
18A as Intel’s most important technology test in a decade. In parallel,
industry media reportedly claim Google could have Intel manufacture more
than three million TPUs in 2028. Neither Google nor Intel has confirmed
that arrangement.
Attention Stage

*Active*
Attention vs Evidence

*Hard evidence*

-

Intel has developed 18A with RibbonFET and PowerVia.
-

18A has entered production-related stages, while 18A-P is progressing
through risk production in the original note.
-

Intel is reporting earnings today.
-

Google and Intel already have a separate relationship involving
infrastructure processing technology, according to the broader research
context.

*Attention / interpretation*

-

Intel will manufacture more than three million Google TPUs in 2028.
-

Google intends to use Intel as a full second source to TSMC.
-

A customer decision has already been made.
-

18A yield is sufficient for high-volume external accelerator production.
-

The potential program will materially improve Foundry utilization.

Attention Path

Intel earnings → 18A yield and production update → Google TPU report gains
or loses credibility → external-customer validation → Foundry rerating
Attention Asymmetry

The rumor has unusually high strategic value because one credible
hyperscaler customer could change the market’s view of Intel Foundry.

The attention is not early in price terms: Intel has already rerated
substantially. The asymmetry therefore depends on evidence quality, not
merely narrative size.
Crowding Risk

INTC is the most crowded object in today’s Feed.

Investors are already tracking 18A, U.S. semiconductor policy, Apple and
Google customer possibilities, CPU demand, IPUs and strategic asset value.
An earnings report that lacks new customer or yield detail may disappoint
even if headline financial results are acceptable.
What Could Sustain Attention

-

Specific 18A yield improvement.
-

Volume-production milestones.
-

Named external customer or tape-out.
-

Confirmation of a Google manufacturing relationship.
-

Evidence of 18A-P customer adoption.
-

Better Foundry utilization or loss trajectory.
-

Xeon 6+ demand tied to agentic workloads.
-

Additional hyperscaler infrastructure partnerships.

What Could Make Attention Fade

-

Vague yield commentary.
-

Continued customer delays.
-

No external revenue visibility.
-

Denial or non-confirmation of the Google TPU report.
-

Foundry losses remain structurally high.
-

CapEx rises without utilization improvement.
-

The stock fails to hold despite a favorable headline quarter.

Attention-to-Thesis Conversion

INTC becomes a durable foundry thesis if 18A demonstrates stable yield,
customer delivery and repeat external demand.

The highest-quality conversion would be a named hyperscaler program large
enough to improve utilization and prove that Intel can support external
designs—not merely manufacture its own products.
EVIDENCE_CLAIMS

-

Intel 18A uses RibbonFET and PowerVia. Company-supported fact.
-

18A offers up to 18% same-power performance improvement and 38%
same-performance power reduction versus Intel 3. Company claim; verify
methodology.
-

Google may ask Intel to manufacture more than three million TPUs in
2028. *Industry report — Needs verification.*
-

Intel would be Google’s second source alongside TSMC. *Needs
verification.*
-

18A is the only realistic challenger to TSMC’s advanced-node
dominance. *Research
interpretation, not confirmed fact.*
-

External-customer confirmation would materially change Foundry
utilization. Research inference.

WHAT_COULD_MAKE_THIS_WORK

-

18A yield improves visibly.
-

Volume delivery begins on schedule.
-

Google or another hyperscaler confirms a program.
-

Foundry utilization rises.
-

External revenue becomes measurable.
-

Xeon 6+ gains agentic-AI demand.
-

Intel maintains capital discipline.
-

The market recognizes second-source strategic value.

WHAT_COULD_BREAK_THE_THESIS

-

18A yield remains low.
-

Volume production slips.
-

External customers do not materialize.
-

Google rumor proves false.
-

Foundry losses persist.
-

CapEx resumes without demand.
-

TSMC maintains a decisive cost and ecosystem advantage.
-

Price has already discounted a successful turnaround.

WEAKEST_ASSUMPTION

The weakest assumption is that a hyperscaler’s desire for supply-chain
diversification will overcome the technical and commercial risks of moving
an advanced accelerator design to Intel.

MOST_IMPORTANT_DATA_POINT

The most important number is *18A production yield*, followed by a named
external-customer milestone.

Without those two facts, the Google TPU narrative remains strategically
interesting but financially ungrounded.
SENSITIVITY_FRAMEWORK
Scenario 18A Yield / Delivery External Customer Evidence Thesis Read
Weak Delayed or vague No customer confirmation Foundry remains policy-led
Base Gradual yield improvement Early tape-outs or smaller programs Turnaround
remains viable but unproven
Strong Stable volume production Large hyperscaler customer confirmed Foundry
credibility and utilization rerate
------------------------------
Cross-Object Attention Comparison
Rank Object Attention Asymmetry Evidence Quality Catalyst Clarity Crowding
Risk Attention Window Conversion Potential
1 *INTC* High Medium–High Very High High Event-dependent Very High if 18A
and customer validation arrive
2 *MRAM* High Medium–Low Medium Medium–High 1–2 weeks Medium if AMD design
win is financially material
3 *SIVE* Medium Medium Medium Medium 1–2 weeks High only with order and
revenue follow-throughCleanest Attention Trade

*INTC*, because the earnings event is fixed and will directly address the
central 18A question.
Most Evidence-Backed Attention Trade

*INTC*, because 18A is an official technology and production program even
though the Google TPU report remains unconfirmed.
Most Crowded Attention Trade

*INTC*, where advanced-node, policy, CPU and hyperscaler narratives are
already heavily represented.
Highest Fade Risk

*MRAM*, because the current attention originates from visual component
identification rather than customer disclosure.
Best Candidate to Become a Durable Thesis

*INTC*, if yield and an external hyperscaler program are confirmed.

*SIVE* may have a meaningful rerating path, but only after lower selling
pressure is matched by real customer orders.
------------------------------
7_DAY_RESEARCH_WORKFLOWMRAM — 7-Day Checks

-

Obtain a higher-resolution image of the AMD module.
-

Identify the Everspin part number.
-

Verify whether the module is commercial, pre-production or an
engineering sample.
-

Estimate potential unit content.
-

Review Everspin’s disclosed customer concentration.
-

Separate the defense-contract thesis from the AMD attention trade.
-

Monitor whether attention expands beyond one KOL post.
-

Track whether price holds without additional confirmation.

SIVE — 7-Day Checks

-

Verify Two Sigma’s latest disclosed position.
-

Verify Voleon’s reported position.
-

Recalculate aggregate reported short exposure.
-

Confirm LAZR’s six reported position reductions and holding dates.
-

Determine whether the ETF was experiencing net redemptions.
-

Compare SIVE’s move with the other five reduced non-U.S. holdings.
-

Track new insider or large-holder transactions.
-

Monitor whether price strength continues without short-covering fuel.

INTC — 7-Day Checks

-

Record management’s 18A yield commentary.
-

Track volume-production milestones.
-

Note any named external customers.
-

Search for confirmation or denial of the Google TPU report.
-

Compare Foundry losses and utilization with prior quarters.
-

Track Xeon 6+ demand commentary.
-

Separate policy support from commercial customer evidence.
-

Monitor the post-earnings price response relative to semiconductor peers.

------------------------------
30_DAY_RESEARCH_WORKFLOWMRAM — 30-Day Checks

-

Track product-revenue guidance.
-

Monitor defense-contract revenue recognition.
-

Search for teardown or BOM confirmation.
-

Compare MRAM with other persistent-memory technologies.
-

Assess content opportunities in accelerators, networking and storage
control.
-

Monitor gross-margin sustainability.
-

Update ATTENTION_STAGE if AMD confirmation does not arrive.
-

Reclassify only if customer evidence catches up with attention.

SIVE — 30-Day Checks

-

Track total disclosed short exposure.
-

Monitor LAZR and other fund holdings.
-

Follow GlobalFoundries, Jabil and Ayar Labs commercialization updates.
-

Track Photonics orders and production revenue.
-

Monitor cash flow and financing risk.
-

Compare price action with LITE, AAOI, COHR and IQE.
-

Update attention status if the stock cannot hold after selling pressure
falls.
-

Upgrade the thesis only after measurable order conversion.

INTC — 30-Day Checks

-

Track 18A customer tape-outs.
-

Monitor yield and product-delivery disclosures.
-

Follow Google TPU manufacturing reports.
-

Compare 18A with TSMC N2 and Samsung alternatives.
-

Track Foundry utilization and operating losses.
-

Monitor Xeon 6+ adoption in agentic workloads.
-

Follow Intel-Google IPU collaboration.
-

Update thesis status if external-customer validation remains absent.

------------------------------
WKAP Daily Top 3

Three market sources worth feeding into today’s market chat. Not required
reading — WKAP has already extracted the signal.
1. LAZR Reduced Six Non-U.S. Photonics Holdings by Roughly 11%

URL: @cherryPayment <https://x.com/cherryPayment/status/2080155554345783581>

*WKAP signal:* LAZR reportedly reduced AIXA, Furukawa Electric, IQE, SIVE,
Eoptolink and Zhongji Innolight by approximately 9.5–11.3%, while its
U.S.-listed holdings were unchanged. The author interprets the pattern as
redemption or liquidity management rather than selective fundamental
selling.

*Why it matters today:* If the analysis is correct, part of the recent
pressure in SIVE and other non-U.S. optical names may have been mechanical
supply rather than a deterioration in the AI-photonics thesis.

*Themes/tickers:* LAZR, SIVE, AIXA, IQE, 5801, 300502, 300308, photonics
fund flows

*Question to ask:* “Was LAZR experiencing net redemptions, and did the six
reduced positions underperform because of forced selling or because the
manager changed its regional allocation?”
2. Why Agentic AI Could Trigger a CPU Renaissance

URL: @GodotSancho <https://x.com/GodotSancho/status/2076623315872125114>

*WKAP signal:* The source argues that agentic AI shifts the infrastructure
bottleneck from raw GPU compute toward orchestration, tool use, data
movement and execution environments—functions that increase demand for CPU
cores, performance per watt and memory bandwidth.

*Why it matters today:* The framework broadens the AI infrastructure trade
beyond GPUs and explains why INTC, AMD, ARM and Nvidia’s Vera CPU can all
benefit even as they compete for share.

*Themes/tickers:* INTC, AMD, ARM, NVDA, agentic AI, CPU, 18A, Xeon 6+, Vera

*Question to ask:* “How quickly must the CPU-to-GPU ratio change before CPU
revenue growth becomes visible in hyperscaler CapEx and server shipment
data?”
3. J.P. Morgan’s Nokia Q2 Preview: AI Orders vs Supply Constraints

URL: @pequityresearch
<https://x.com/pequityresearch/status/2080077876548796847>

*WKAP signal:* J.P. Morgan reportedly expects Nokia’s Cloud and AI orders
to remain strong, with potential Network Infrastructure growth upside if
laser and component supply permit conversion into revenue. Q2 AI and Cloud
orders may exceed the €1 billion recorded in Q1.

*Why it matters today:* The preview tests whether the AI networking cycle
is demand-constrained or supply-constrained—and whether optical and
IP-network orders can support a higher earnings multiple.

*Themes/tickers:* NOK, optical networking, IP networks, AI switching,
Infinera

*Question to ask:* “If Nokia cannot raise guidance despite stronger orders,
how much revenue is being deferred by component shortages rather than lost
demand?”