Keki Sensei 🎂 ❤

I love cake and crypto 🍰🪙

[Keki Sensei 🎂 ❤ @keki_master]

1/ I’ve been an $INTC skeptic for a long time. Especially during Gelsinger’s tenure.

Intel was stuck in the mud. While $AMD was leaping ahead. All Intel’s CPU lineups were disappointing. Chips were power-hungry. gen 14-15 had stability issues, etc.

[Keki Sensei 🎂 ❤ @keki_master]

The more time pass, the more I really think Jim Cramer is an inverse signaller for Hedge funds circle. Intentionally propped up the retail for the exit liquidity.

If you watch Wallstreet 1987, it's the same thing as "Bluehorse shoe loves Anacott steel"

[Keki Sensei 🎂 ❤ @keki_master]

At this rate it should be renamed to Strait of Schrodinger

[Open Source Intel @Osint613]

Iranian Foreign Ministry spokesperson: an agreement with Oman on a new route does not mean the Strait of Hormuz is opening or closing.

[Keki Sensei 🎂 ❤ @keki_master]

Yesterday, Jim Cramer said to sell all hyperscaler and you won't regret it.

Today they are all up like 20-30%

You can't make this shit up 🤣

[Keki Sensei 🎂 ❤ @keki_master]

@BellaItsa แต่ละตัวขึ้น 5-10 เท่า บางตัวขึ้น 20 เท่า พี่ยังเอาจนพอร์ตแตกได้ หัวจะปวด 555

[Keki Sensei 🎂 ❤ @keki_master]

$GOOG literally missed the EPS estimate (ex. SpaceX gain) and raised capex into negative FCF while revenue beat was small. Why do people keep saying it's a stellar earning?

(It's not a bad earning per se, but it's not a "surprise")

[P Equity Research 📰 @pequityresearch]

Deutsche Bank: Chip Blip

> Leading-Edge Logic (<=2nm): This sector faces the most significant supply constraints. All categories (Accelerator, CPU, and Smartphone chipset) remain well below 100% throughout the entire forecast period. The lowest point is observed in 2027, with the Smartphone chipset and Accelerator categories hovering between 72% and 75%.

> Advanced Nodes (5/4/3nm): Supply for these nodes is also under pressure, consistently trending below 100% (ranging from 71% to 82%). There is a slight projected improvement in supply for CPUs and Accelerators toward the end of 2027.

> Mature Nodes (28/22nm and >28nm): These sectors generally show better supply stability, with values mostly at or above 100%, indicating that supply is better positioned to meet current demand compared to the advanced nodes.

Memory (DRAM & NAND):

> DRAM: This segment shows a significant and persistent supply shortfall, with values ranging between 57% and 68%. HBM (High Bandwidth Memory) remains consistently constrained throughout 2026 and 2027.

> NAND: Supply conditions for NAND are relatively tighter than DRAM but show more stability, with figures ranging from 77% to 87%.

We are short in compute for the entirety of 2027!

$NVDA $AMD $INTC $ARM $GOOGL $AMZN $META

[Keki Sensei 🎂 ❤ @keki_master]

@michaeljburry I'll follow you. You are legend. Nobody could have done what you did.

The most successful substack seller and sales funneling of all time.

[Keki Sensei 🎂 ❤ @keki_master]

@awakenowzone LEAPS is actually a risk management strategy. You risk only portion of capital for the same exposure with small time decay

But yeah, if you are full port deep OTM LEAPS then that's definitely degen

[John Intel @intelfabs]

Tech: $15.6bn inflow, record 3 -week inflow

Paper hands!

[Keki Sensei 🎂 ❤ @keki_master]

@Mojo_flyin @Intel_Foundry @intel it's hard to believe the entire world advanced chip production is relying on a single Taiwan company with very high risk of geopolitical incident. US can't afford to let this continue anymore and they have only one player - $INTC

[Keki Sensei 🎂 ❤ @keki_master]

6/And that's why I think $INTC Q2 will be a huge surprise

Also if an idiot like me can come up with this numbers base on confirmed fact, I bet the streets already knew. This could probably why we're seeing so much hit piece to suppress the price before earning

What do you think

[Keki Sensei 🎂 ❤ @keki_master]

1/ My Estimation on $INTC Q2 earning

Confirmed by CFO
*18A yield ahead of plan. 7-8% improvement/month
*Fastest ramp in 5 years
*6-7x vol Panther Lake vs Q1
*DCAI double digits growth vs Q1
*Increased Intel 3+7 wafers input
*Q2 will have full benefit of double price hike in Q1

[ImNotHarsh | 📈💸 @imnotharsh]

$INTC Q2 2026 Earnings Preview - 7/23, After Market Close

I said it multiple times before and I will say it again: Intel’s balance sheet and earnings will flip on a dime, thanks to Intel Foundry. And Wall Street still isn’t fully ready. The re-pricing is not complete for the near-term roadmap (this year alone), where Intel’s FCF trajectory and P/E multiple will turn on its own axis.

Current Consensus Estimates:

- Revenue ~$14.4 - 14.7B (Street avg ~$14.4 - 14.68B)
Non-GAAP EPS ~$0.19 - 0.21
- Company guide from Q1:
--> Revenue $13.8 - 14.8B (mid ~$14.3B), non-GAAP GM 39%, non-GAAP EPS $0.20, GAAP EPS $0.08.

Q1 actual was a massive beat ($13.6B rev, $0.29 non-GAAP EPS vs much lower Street). Stock has ended up re-rating to the ~$110 mark, with market cap ~$550–565B. The bar is higher, but the setup for another positive surprise is real.
__________________________________________________________

What matters most for a positive reaction:
1. Beat and a sequential growth with improving supply.
- CPU demand remains strong, even with their pricing power. Increased cost seemingly has not brought demand downward on the trajectory. The constraint has been supply, not demand. As 18A ramps (Panther Lake as we know is already in high-volume production and shipping), more product hits the market to satisfy the clear demand they have stated is not constrained. DCAI sequential growth (guided double-digit in Q1) and better CCG supply should drive the top line. Margins should expand as the mix improves and early 18A ramp costs normalize.

2. Capex commentary is the potential rocket fuel (get it? because $SPCX Terafab. funny? no? okay, let's move on)
- Watch this closely. If management raises capex or reallocates more toward tools / capacity (specifically because of customer roadmaps), for advanced packaging, glass substrates, High-NA EUV for 14A, or EMIB-T expansion, then that is bullish. It signals real external demand and volume commitments. Capex was guided roughly flat-to-slightly-down overall with higher tool spend. Any upward revision tied to hyperscaler / foundry wins will be read as validation, not waste. “We’re investing more because the customers are here and the lines need to ramp”, to me is a strong catalyst.

3. Foundry turnaround becoming material
- Foundry, as we know, is the multi-year asymmetric upside. Advanced packaging (EMIB / EMIB-T) is quietly turning into a multi-billion rocket ship (same joke again, let's move on). TSM CoWoS remains sold out deep into 2027. Intel’s CFO David Zinsner has upgraded external packaging commentary to “billions per year” once the first lines fully ramp and also noted strong engagement and deals close to closing.
__________________________________________________________

I have used this example before, but I find it apt for the Earnings preview itself. Think about the packaging roadmap like this, in the barebones state:
- 1 meaningful deal (e.g., $GOOGL Google TPU packaging or early $TSLA Tesla Terafab-related work) = $1–3B annual run-rate by end of this year / into 2027. Shows up in H2 numbers and proves the model.
- 2 deals (Google + $AMZN Amazon Trainium or similar) =
run-rate in 2027–2028. U.S./Vietnam/Malaysia capacity expansions make this executable. EMIB-T is built for the huge multi-die AI chips everyone wants.
- 3+ deals (add $AAPL Apple silicon packaging, Meta, or others) = $10–20B+ annual packaging revenue by 2028–2030. Wafer wins typically follow packaging wins. This becomes the on-ramp to full Foundry scale.

Intel doesn’t need every deal to win big, and does not need to beat TSM right now. That's not what this is about. Even partial execution while TSM is capacity-constrained and geopolitical risk remains is a game-changer. Foundry losses should continue narrowing as utilization and external mix improve. Positive commentary on external packaging revenue, customer wins, or yield progress on EMIB-T (already showing strong high yields in recent technical updates) will move the needle hard.
__________________________________________________________

3 core thesis points that need more talking about:
1. The CPU shortage is real and Intel confirmed the issue is supply, not demand
- Immense demand exists across client and especially data center/agentic AI workloads. When supply constraints ease further (Q2 and H2 with 18A/Panther Lake ramp), volumes and margins improve. Xeon remains the cost-effective, high-TCO host CPU of choice for many hyperscalers.

2. Rising demand for cost-effective, margin-improving silicon is real
- Hyperscalers and enterprises continue looking for TCO improvements without sacrificing too much performance. Intel’s Xeon and platform approach delivers this better than pure “more FLOPS at any cost” solutions. Not every customer will pay Nvidia / $AMD premium pricing indefinitely. Business is business and ROI matters. Intel is positioned for the more rational, scaled AI buildout.

3. The Foundry bottleneck against the hedge of a monopoly with limited capacity
- $TSM cannot deliver everything, everywhere, all at once. Intel Foundry may never be more than a strong #2 in leading-edge wafers, and it doesn’t need to be. Execute on packaging first (the near-term multi-billion opportunity) and the rest follows. Customers are there. We all know they have them. It’s only a matter of time until this is more fully validated in the numbers and guidance.
__________________________________________________________

A beat on revenue / EPS, with any positive Capex raise tied to customer demand, and Foundry / packaging that shows external revenue momentum or narrowed losses means we have another strong reaction and continued re-rating. The core products business is already generating solid profits that fund the Foundry build. FCF is expected to turn positive for the full year. The market has barely re-priced the recovery story. It has not fully priced the packaging ship or the multi-year Foundry optionality if even half the pipeline converts.

Q2 is the next proof point of this growth story we get to watch play out in real time. We must watch Capex language and packaging commentary above all.
__________________________________________________________

Look, Intel is working with the whole Mag 7. Don't forget that. Intel partners with every major neocloud, and these are major collaborations. Intel is staked by the movest valuable company on the planet $NVDA. It is staked by $SFTBY Softbank. It is staked by the US Government $DJT. It is being backed by Elon Musk's Tesla / SpaceXAI. Let us not lose sight of the fact that the most powerful people and institutions on the planet have Intel's back. And Intel, themselves, have the superteam to execute on the backing. It is led by the most brilliant minds on the Executive team and one of the best CEOs in this space (no pun intended. the joke that never stopped).

This is America's most valuable asset. It is the greatest company in the world, and somehow only a small % of people even understand what Intel is doing. The future of memory, the future of robotics, the future of Quantum Computing - these are all verticals that Intel has its grasps on. They have been researching and building on these areas for years. So, by the time it comes to fruition, Intel will be way ahead of everyone else. Except its investors. Only we will be there.

A generational earnings is coming soon. The whole narrative is about to flip around. Now? Maybe. But, this next year looks prime.

LONG. LIVE. INTEL.

[Keki Sensei 🎂 ❤ @keki_master]

5/ Based on all facts above, my Q2 estimated revenue is - $16.5-$17B.

Base Q1 rev = 13.6B
+2.15B Panther Lake growth
+0.51B DCAI growth
+1.77B price hike effect
-1B Inventory benefits

=$17.03B

(Adv packaging/ASIC growth/Terafab fees not included since I have no data)

[Keki Sensei 🎂 ❤ @keki_master]

AI bubble this, AI bubble that. This could probably be the lowest P/E bubble I've seen.

And this doesn't even accounted for more relaxed accouting rules during 2000 that inflated the "E" in P/E

[Keki Sensei 🎂 ❤ @keki_master]

@Mojo_flyin my small concern is, they will have to book around 10B loss for US gov escrow share on GAAP basis due to share price rised from 40 to 140. (they booked around 1B in Q1 from 30s to 40s)

it's basically just stupid accounting rule and I hope market would ignore ugly GAAP EPS lol

[Keki Sensei 🎂 ❤ @keki_master]

@ObservingMindTh ผมเห็นต่างครับ internet ตอนนั้นเงินไหลจาก user เข้าบริษัทช้ากว่า AI มาก อย่าง Anthropic ปีเดียว runrate จะ 50B แล้ว เทียบ fwd PE จะเห็นภาพชัดครับ

นี่ยังไม่นับ accounting rule change ที่เข้มขึ้นหลังจาก dotcom ที่ทำให้ P/E สมัยก่อนต่ำกว่าความเป็นจริงจากการแต่งงบประมาณ 10-15% ด้วย