Pinned Tweet
๐งต The part of Intel's foundry strategy most investors will never talk about.
$INTC just opened a new 107,000 sq ft photomask facility with a Class 1 cleanroom.
At first glance, that sounds boring.
It isn't.
Photomasks are essentially the templates used to print transistor patterns onto wafers. As nodes get smaller, the masks become dramatically more complex and expensive.
This facility is being built to support Intel's future roadmap, including 18A, 18A-P, and eventually 14A.
Here's the part I think matters.
Advanced EUV systems are incredibly powerful, but the masks don't last forever. Over time they degrade, get damaged, and need to be replaced.
If you can't replace them quickly, production slows.
That means a mask shop isn't just manufacturing infrastructure.
It's production insurance.
Intel also has an advantage that doesn't get discussed enough.
Through IMS Nanofabrication, Intel is the only major chipmaker that owns the company building the world's leading multi-beam mask writers.
Instead of writing masks with a single electron beam, IMS systems use over 262,000 beams simultaneously.
Faster mask creation.
Higher throughput.
Less risk of becoming a bottleneck.
Most investors are focused on foundry customers, yield numbers, and whether Intel can win share from TSMC.
Those things matter.
But none of it works without the infrastructure behind it.
This looks like another piece of Intel quietly building the ecosystem needed to support advanced manufacturing on U.S. soil.
The market is watching the fabs.
Intel is investing in everything that keeps the fabs running.
Driver over 770 miles/ 1250 kilometers today.
Nebraska is boring very boring on the freeway. That Nebraska gas station sums it up.
Steven Martin retweeted
์๋ฒ๋ฆฐ ๊ธ๋ก๋ฒ๋ง์ผ ๋ชจ๋๋ธ๋ฆฌํ 2026-10-28 (์)
โโโโโโโโโโโโโโโโโ
1. AI ๋ฐธ๋ฅ์์ด์
๋ฒ๋ธ ๊ฒฝ๊ณ ๋ฐ ๋น
ํ
ํฌ ์๋ณธ ์กฐ๋ฌ ์๋ฐ (AI Bubble Deflation & Capital Calls)
์ค๊ตญ์์ AI ๊ด๋ จ ๊ธฐ์
๊ฐ์น๊ฐ ํญ๋ฝํ๋ฉฐ ์ฝ 4,800์ต ๋ฌ๋ฌ ๊ท๋ชจ์ AI ๊ฑฐํ์ด ๋ถ๊ดดํ๋ค๋ ๋ถ์์ด ์ ๊ธฐ๋์์ต๋๋ค (์ธ๋ฒ ์คํ
). ์คํธ๋กํฝ์ S-1 ์๋ฅ๋ฅผ ํตํด ์ฒ๋ฌธํ์ ์ ์ ๊ตฌ์กฐ(์๋ฐ๊ธฐ 800์ต ๋ฌ๋ฌ)์ 8,000์ต ๋ฌ๋ฌ์ ๋ฌํ๋ ๋ฐ์ดํฐ์ผํฐ ์ง์ถ ์ฝ์ ์ ๊ณต๊ฐํ์ผ๋ฉฐ (FT), ์๋ง์กด ๊ฒฝ์์ง์ ์ง์๋ค์๊ฒ AI ๋๊ตฌ ์ฌ์ฉ ์ ๋น์ฉ ํจ์จ์ฑ์ ๋์ผ ๊ฒ์ ์ง์(tokenmaxxing)ํ๋ ๋ฑ AI ์ธํ๋ผ ์ ์ง ๋น์ฉ ์๋ฐ์ด ์๋ฉด ์๋ก ๋๋ฌ๋๊ณ ์์ต๋๋ค (CNBC). ๋ฉํ์ 950์ต ๋ฌ๋ฌ ์๋ณธ ์กฐ๋ฌ ๋ฐํ๋ AI ํฌ์ ๋ถ๋ด์ ํ์ค์ ๋ฐฉ์ฆํฉ๋๋ค.
2. ์ค๋ฌผ ๊ฒฝ์ ๋ํ ์๊ทธ๋ ๋ฐ ์ ์ฉ ๊ฒฝ์ ์ํ (Macro Slowdown & Credit Strains)
๋ฏธ๊ตญ์ 9์ ์ ๊ท ์ฃผํ ํ๋งค๊ฐ ๊ธ๊ฐํ๊ณ ์ค๊ฐ ์ฃผํ ๊ฐ๊ฒฉ์ด ํ๋ฝํ๋ฉฐ ์ฃผํ ์์ฅ์ ๋ํ๊ฐ ํ์ธ๋์์ต๋๋ค (์ธ๋ฒ ์คํ
). ์๊ตญ์์๋ ์ฌ์ ์ ์ด๋ ค์์ ์ฒํ ๊ธฐ์
์ด 73๋ง 8์ฒ์ฌ ๊ฐ๋ก 20๋
๋ง์ ์ต๊ณ ์น๋ฅผ ๊ธฐ๋กํ๋ฉฐ ๊ธฐ์
๋์ฐ ๊ณตํฌ๊ฐ ํ์ฐ๋๊ณ ์์ต๋๋ค (๊ฐ๋์ธ). ๋์ฒด ํฌ์ ์์ฅ์์๋ ๋์ด์ฒด๋ฐฉํฌ ์ฐํ DWS๊ฐ ํ๋๊ทธ์ญ ๋ถ๋์ฐ ํ๋์ ํ๋งค ์ฒญ๊ตฌ์ ๋ํด 10%๋ง ์ง๊ธํ๊ธฐ๋ก ๊ฒฐ์ ํ๋ฉฐ ์์
์ฉ ๋ถ๋์ฐ ์ ๋์ฑ ์๊ธฐ๋ฅผ ์ฌํ์ธํ์ต๋๋ค (FT).
3. ๊ธ๋ก๋ฒ ๋์ฒดํฌ์ ๋ฐ ๊ธฐ์
์ ์ฉ ๊ท ์ด ์ฌํ (Corporate Distress)
๋ง์ดํฌ๋ก์ํํธ๊ฐ AI ์ ํ์ ๋ฐ๋ฅธ ๋น์ฉ ์ ๊ฐ์ ์ํด 6๋ง ๋ช
์ ๋์์ผ๋ก ๋ช
์ํด์ง(buyout)์ ์ ์ํ๊ณ (๊ฐ๋์ธ), ๋ณด์์ ์์ฉ๊ธฐ ์์ฐ ์ง์ฐ ๋ฐ ๋
ธ์กฐ ํ์
์ฌํ๋ก 2027๋
๊น์ง 737 ๋งฅ์ค ์์ฐ๋ ๋ชฉํ ๋ฌ์ฑ์ด ๋ถ๊ฐ๋ฅํ๋ค๊ณ ๋ฐํ ์ฃผ๊ฐ๊ฐ ๊ธ๋ฝํ์ต๋๋ค (CNBC). ํญ๊ณต์ ๊ธ๋ฑ์ผ๋ก ์ค์ ์๋ฐ์ ๋ฐ๋ ์ ฏ๋ธ๋ฃจ(JetBlue)๋ ์ฑ๊ถ๋จ ๊ตฌ์กฐ์กฐ์ ์ ์ฐฉ์ํ์ผ๋, ํด์ด๋ผ์ด๋ ๊ณํ ๋ฐํ๋ก ์ฃผ๊ฐ๊ฐ ๋ฐ๋ฑ(19% ์์น)ํ๋ ๊ทน์ฌํ ๋ณ๋์ฑ์ ์ฐ์ถํ์ต๋๋ค (๋ธ๋ฃธ๋ฒ๊ทธยท๋ง์ผ์์น).
4. ์ค๋ ์ธ๊ต ํ๋ด๊ณผ ์ง์ ํ์ ํ
์
์ง์ (Mideast Geopolitics)
ํธ๋ผํ ์ ๋ํต๋ น์ด ๋น์ฅ ์ด๋ ํ๊ฒฉ์ ๋ฐ๋ํ๋ ๋์์ค์ ๋ฐ์ธ์ ๋ด๋๊ณ , ์นดํ๋ฅด ๋ฑ ์ฃผ๋ณ๊ตญ ์ค์ฌ ๋
ธ๋ ฅ์ด ์ด์ด์ง๋ฉด์ ๊ตญ์ ์ ๊ฐ๋ ์ํญ ํ๋ฝ(-1.5%)ํ๋ฉฐ ์จ ๊ณ ๋ฅด๊ธฐ์ ๋ค์ด๊ฐ์ต๋๋ค (์ฐํฉ๋ด์คยท์ธ๋ฒ ์คํ
). ํ์ง๋ง ์๊ตญ ๊ณต๊ตฐ๊ธฐ์ง ํ
๋ฌ ๋ฏธ์ ์ฌ๊ฑด์ ๋ฐฐํ๋ก ์ด๋์ด ์ง๋ชฉ๋๊ณ (WSJ), ์ฌ์ฐ๋์๋ผ๋น์์ ๋ํ ํํฐ ๋ฐ๊ตฐ์ ๋๋ก ๋ฐ ๋ฏธ์ฌ์ผ ๊ณต๊ฒฉ์ด ์ง์๋๋ฉด์ ๊ธด์ฅ๊ฐ์ ์ฌ์ ํฉ๋๋ค. ์ฟ ๋ฐ์ ์ฌ๊ฐํ ์ฐ๋ฃ๋๋ ๊ธ๋ก๋ฒ ์๋์ง ๊ณต๊ธ๋ง์ ๋ถ์์ ์ฑ์ ๋ํ๊ณ ์์ต๋๋ค (WSJ).
5. ํ๊ตญ ํ๋๋ฉํธ ์๊ทนํ ๋ฐ ๋ถ์ค ์ํ (Korea Polarization & Default Risks)
๊ตญ๋ด ์ฃผ์ ์์ฅ์ ๋ฏธ๊ตญ ๊ธฐ์ ์ฃผ ๋ฐ๋ฑ์ ๋์กฐํ๋ฉฐ ํฐ ํญ ๋ฐ๋ฑ(^KS11 3.09%)ํ์ผ๋, ํ๋๋ฉํธ ์ฐ๋ ค๋ ์ปค์ง๊ณ ์์ต๋๋ค. ๊ธฐ์
์ํ์ ์ค์๊ธฐ์
์ฐ์ฒด์จ์ด 15๋
๋ง์ ์ต๊ณ ์น(1%๋)๋ฅผ ๊ธฐ๋กํ๋ฉฐ ๊ณ ๊ธ๋ฆฌ ์ฅ๊ธฐํ์ ์ฌํ๊ฐ ๊ฐ์ํ๋์๊ณ (์ฐํฉ๋ด์ค), ํ๊ตญ์ํ์ AI ํฌ์ ์ฌ์ดํด์ด ๊บพ์ผ ๊ฒฝ์ฐ ๋ฐ๋์ฒด ์์ถ์ด ์ต๋ 1,000์ต ๋ฌ๋ฌ ๊ฐ์ํ ์ ์๋ค๊ณ ๊ฒฝ๊ณ ํ์ต๋๋ค (๋งค๊ฒฝ). ์์ฐ ์ยทํ์ 20%์ ์ฃผํ ๊ฐ๊ฒฉ ๊ฒฉ์ฐจ๊ฐ 290๋ฐฐ๋ก ์ฌ์ ์ต๋์น๋ก ๋ฒ์ด์ง๋ ๋ฑ ๋ด์ ์๊ทนํ๋ ๊ทน์ฌํด์ง๊ณ ์์ต๋๋ค (์ฐํฉ๋ด์ค).
โโโโโโโโโโโโโโโโโ
"๊ธฐ๋ณธ๊ฐ์ No Action. ์กฐ๊ฑด์ด ๋ฐ๋๋ฉด ์์ง์
๋๋ค.
Steven Martin retweeted
๐๐๐๐ ๐๐๐๐๐ฌ | ๐ข๐๐ง๐ข๐๐๐ฅ ๐ฎ๐ณ, ๐ฎ๐ฌ๐ฎ๐ฒ
๐ ๐ฌ ๐ฅ๐๐๐
I still really like what $KEEL is building. The company has access to a lot of power, and that matters as AI data centers compete for electricity.
But I keep coming back to the same question. Where are the signed leases?
Keel has roughly 2.2 GW in its development pipeline. That is potential capacity, not operating data centers.
At $2.75 million per MW annually, the full pipeline represents nearly $6 billion in potential yearly revenue. That is a high end illustration.
Actual revenue depends on the type lease.
The opportunity is large. Keel still has to prove it can turn that opportunity into revenue.
๐ฆ๐๐ง๐ ๐ฃ๐ฅ๐ข๐๐ฅ๐๐ฆ๐ฆ
Allentown received conditional approval in October for a proposed 75 to 80 MW data center.
Local reporting indicates Keel could operate the facility for an unnamed technology tenant. No signed lease or financial terms have been released yet.
Moses Lake continues moving forward. PowerSecure announced a backup power agreement in September, and Vertiv equipment deliveries have advanced.
Management has previously targeted initial capacity in the first half of 2027.
Sharon and Panther Creek remain important development sites, with potential service dates around la
๐๐จ๐ฆ๐ง๐ข๐ ๐๐ฅ ๐ก๐๐๐ข๐ง๐๐๐ง๐๐ข๐ก๐ฆ
This is what interests me most.
During the Q2 earnings call, CEO Ben Gagnon said multiple prospective customers were negotiating for each of Keelโs three priority sites.
He also stated that some customers were interested in more than one location.
A customer interested in several sites could support a much larger agreement than a single location.
There is also speculation about Amazon becoming a customer.
I have not seen confirmation of an Amazon agreement, and I still watching that rumor.
I would rather focus on what management has stated and I am excited for the earnings report to hear more updates.
๐๐ก๐ฆ๐๐๐๐ฅ ๐๐จ๐ฌ๐๐ก๐
CEO Ben Gagnon purchased additional shares in August. The COO also bought shares.
I like seeing management invest its own money.
It shows confidence, but it does not guarantee the stock will recover.
๐ช๐๐๐ง ๐๐ฆ ๐๐ข๐ง๐๐๐ฅ๐๐ก๐ ๐ ๐
Management previously targeted three leases before the end of 2026.
During the Q2 call, management discussed active negotiations but did not clearly reaffirm that deadline.
I would rather see Keel sign good contracts than rush into bad ones.
But investors deserve an update.
Keel reported approximately $819 million in liquidity in August.
That gives the company flexibility, but building data centers is expensive.
I want to know how much cash remains and how much construction will cost before meaningful revenue begins.
๐ฃ๐ฅ๐๐๐ ๐๐๐ง๐๐ข๐ก
KEEL closed Monday at approximately $2.18.
The stock has been under pressure, with previously reported short interest around 19% of the float.
A signed lease could change sentiment quickly.
But high short interest alone is not a reason to buy.
The business still has to deliver.
๐ก๐ข๐ฉ๐๐ ๐๐๐ฅ ๐ฑ
Keel reports Q3 earnings before the market opens on November 5.
I want answers to four questions.
Have any customer agreements been signed?
Are the three original lease targets still realistic for 2026?
Are Moses Lake, Sharon and Panther Creek still on schedule?
How much cash will construction require before meaningful revenue begins?
Those answers matter more to me than another presentation about the 2.2 GW pipeline.
๐๐ข๐ง๐ง๐ข๐ ๐๐๐ก๐
I have not changed my view of Keelโs potential.
The power is valuable. The locations matter. The opportunity is large.
Allentownโs approval and Moses Lakeโs progress give me reasons to keep watching.
Multiple customers negotiating across the three priority sites is encouraging
I want signed leases and customers.
๐ค Made with AI
This is a big factor. There is no ai crash $NBIS on the larger end to $KEEL on the smaller end will be needed for infrastructure ready to go or quickly ready to go!
Posted takeaways of Anthropic's S-1 on my substack. This is a very detailed S-1 with a lot of risk factors and specific languages.
Key takeaways:
> Anthropic has amassed massive long-term commitments totaling hundreds of billions of dollars across major partners (including non-cancelable agreements with Google, Amazon, Microsoft, Broadcom, and others), roughly 75% of which are non-cancelable. Although, deals with Nscale and Lambda may be cancelled on failure of service.
> NRR stands at approximately 900%. Over 50% of Fortune 500 companies are customers generating at least $100,000 in TTM revenue.
> Companies sees compute availability limited for the foreseeable future.
> Expect use of short-term compute to increase in the near term and to in turn compress their operating margin and Adjusted Operating Margin as a result.
Read more with the link in replies ๐
Steven Martin retweeted
Considering LEAP Options ?
Check out this must read article from @EdgeReport91 ๐
Steven Martin retweeted
I'm an ASL interpreter by day.
My job is taking a message in one language and delivering it in another without losing the meaning.
So here's Wall Street, interpreted:
Spend less than you make.โจBuy the index.โจKeep buying.โจDon't touch it.
Simple to say. The hardest part is being relentlessly consistent.
What would you add?
Steven Martin retweeted
โVOO and chill.โ
I checked my portfolio 47 times today. I have no intention of selling for 20+ years.
The VOO part is easy. Working on the chill.
Good night fam ๐๐ค๐ค๐ป
Steven Martin retweeted
$INTC almost quadrupled in six months. Then it gave back 35.65% ๐
Still up 145.80% this year.
Intel started 2026 at $36.90, ran all the way to $140.94 in June, and trades at $90.70 today.
๐ด Last time it reported, it fell 7.89% the next day. It reports again Thursday.
Is this a healthy pullback or the end of the run?
Steven Martin retweeted
$RES triangle is breaking out ๐
RPC Inc, oilfield services for oil and gas producers in the US.
๐น Lower highs from $6.73 and higher lows from $5.58 squeezed it into a triangle
๐น Today it pushed over the top line and hit $6.49
๐น It needs a daily close over $6.41 to lock it in
๐ Q2 revenue $460.9M (+9.5%), adjusted EPS $0.08 vs $0.04 expected, net income $12.1M vs $0.9M in Q1
๐ฐ Adjusted EBITDA $66.0M (+23.3% sequentially), margin up 250 bps to 14.3%
๐ $179.5M of cash with no borrowings on the revolver, and a dividend yield of about 2.5%
โ ๏ธ Oil price volatility keeps management cautious, and Q3 results come Thursday, Oct 29 before the open
โ
Confirm: daily close over $6.41
๐ข TP1 $6.73 (+5.0%) โข TP2 $7.41 (+15.6%) โข TP3 $8.06 (+25.7%)
โ Invalid: daily close under $6.31 (โ1.6%)
Risk from the live $6.45 to Invalid is about 2.2%, and TP1 pays about 2.0x.
Please respect your SL level ๐
Steven Martin retweeted
Microsoft is offering buyouts to 60,000 employees while spending billions on AI.
When interest rates stay high, companies become more cautious about spending. Hiring slows, budgets tighten and job security becomes less certain.
And that creates a vicious cycle: workers worry about losing their jobs, so they spend less. Businesses then face weaker demand, making them even more cautious about hiring and investing.
High interest rates donโt just make money more expensive. They can make an entire economy more fearful.
Image source: hcamag.com
Driving on snowy road in October. Backcountry skiers already skiing off independence pass. @KeystoneMtn should be opening soon!
























