雅

投資17年目突入。来年の大河が小栗上野介忠順ということでアイコンを変更しました。歴史上の人物で1.2を争うぐらい好きな方です。今の日本に必要なもの今の日本人に足りないものを大河をみて考えるきっかけになれば良いなと思います。そんな大河ドラマを期待しています。

[雅 @Diegooooal]

需要がどんどん増していってメモリスーパーサイクルの終わりが見えない
新工場でキャパが増えてもAI関連がすべて食い潰すかもしれない

[パウロ @paurooteri]

メモリ
HBM3EとHBM4を買い占めた
NVHBMを推進し、メモリの付加価値を自社に
XL-Flashの推進、階層型KVキャッシュ管理

TSMC
NVIDIA独自プロセス、CoWoSの大規模予約

これらが減速すると、やはりGoogleとBroadcom連合が間を埋めるよね

[パウロ @paurooteri]

オープンソース、オープンモデル、論文をオープンする文化が無くなるとAIは減速するでしょうね

米国のフロンティアラボが規制を望み、中国のフロンティアラボがオープン推進なんて、20年前の自分に言ったら信じてもらえないね

[パウロ @paurooteri]

GPUのオンデマンド価格は駆け込み的な上昇を観測
B300とH100がさらに上昇

[Markos @MarkosAAIG]

Nebius earningscall and the total lack of business development and understanding on X. $NBIS

First, I saw a lot of comments on X about the quality of the earnings call and the level of excitement. Well, I can kill that immediately, because I found that Arkady sounded very confident and very relaxed on the call.

I liked his comment that demand for AI is real and substantial, and everything else, like building the capacity and financing it, is a task. Tasks can be hard, but they can be solved.

They are seeing customers committing to them for longer, at higher prices, and paying more upfront, as we are seeing across the industry as well.

Q3 agreements closed above $25 million per megawatt, another step up from $20–25 million per megawatt. The agreements are also getting larger. They signed their first billion-dollar-plus contract three months ago, and now they are signing multibillion-dollar agreements for their AI cloud.

And the customer profile, with enterprise leaders like Databricks, Datadog and Snowflake signing multiyear agreements, is also freaking strong.

The strength of Nebius, and what they do smartly, is that they sell across three time horizons. They optimize the mix between the three.

First, spot and short-term contracts at premium prices. Then one-to-three-year contracts, which lock in more revenue visibility. This is the core business, which is logical because it sits between opportunistic and stable.

The third is five-year agreements with investment-grade hyperscalers. Those contracts support debt financing and provide a more stable base for the buildout.

Super logical. We have talked about this many times. It is a really strong business approach that balances financial security with the flexibility to be opportunistic. Strategically excellent.

To build on that, Nebius’s CEO already explained that, on top of prepayments, they raise financing against the investment-grade contracts. Going back to those three business models, that financing allows them to be more opportunistic with spot-market demand.

Like, what’s not logical about this?

As that flywheel grows, they can pursue more spot-market demand, supported by financing backed by that third leg: hyperscalers and other customers on long-term, multiyear contracts.

And to use a quote:

“We invest in three things at once: the capacity, the software platform, and go-to-market that meets customers where they are. All three matter equally, but capacity is the foundation of the business. It defines our scale.”

Totally in line with Nebius strategy thay they have layed out a long while ago. And smarter than the rest, in my opinion.

People are now getting hung up on Nebius stopping ARR guidance, but I think that is completely logical. First of all, I have been vocal about hating that metric from the beginning. I think it is a useless metric, especially when you look at Nebius.

Take those three business models into account. If you build for on-demand capacity, you cannot safely guide for ARR. You do not know what spot pricing will be at the time, and you do not know how much capacity you are going to sell. This is the opportunistic side of the business.

So I have zero problems with stopping ARR guidance, because it does not tell you shit. What matters much more is how much revenue they are going to generate. And yes, you are not going to have full visibility into that, because part of the business is opportunistic.

But smart management is chasing where the money is. And that money is the basis for all the capacity expansions. So yes, I think it is smart.

What is one of the most basic lessons you are taught if you want to have more money than you currently have and compound your investments as an investor? Look at whether you can increase your salary by changing jobs or taking on more responsibility in your current job.

The second piece of advice is that, if you cannot increase your income within your normal working hours, look for an additional source of income you can create for yourself. That gives you more money to invest and helps you compound your financial base faster.

And this is exactly what Nebius is doing with on-demand capacity and spot pricing. They are focusing on compounding that financial flywheel as fast as possible, because they know capital matters in this buildout.

On the delays, I agree that they are part of the industry, especially when you build through colocation, because you do not have everything in your own hands. He says they are completely on schedule with their own sites. That might be a little optimistic, but it is possible. You can read our free Nebius article, written by @Mark_AAIG , on the Birmingham site.

Delays, especially around permitting at the Vineland site, do not really surprise me. I see this as a matter of timing, and it does not affect the thesis at all. They are also moving away from that dependence on colocation, and customers are still happy. So, no biggie.

So, long story short: a lot of commotion with no basis. They guided to $12–13 billion in revenue next year, which is completely logical given the three parts of their business model. Stepping away from that stupid metric called ARR is the best thing they could do.

The longer-term contracts in the other two parts of the business help fund capex through prepayments and debt financing. That allows them to bring more capacity online for on-demand and spot customers.

Management expects 75–85% of its 2027 capex to be covered by prepayments and financing against investment-grade contracts, with $17–20 billion coming from customer prepayments alone.

They are firing on all cylinders, and they are much smarter than every other player out here. And I am being completely objective. What a bunch of noise on X today about this.

[Wall St Engine @wallstengine]

ANDURIL AIMS TO GO PUBLIC WITHIN 2 YEARS, AFTER ARRANGING AN INSIDER SHARE SALE AT A $90B VALUATION LAST MONTH: DOW JONES

[Beth Kindig @Beth_Kindig]

Amazon $AMZN and Google's $GOOG stakes in Anthropic could potentially be worth a combined $411 billion, if Anthropic’s IPO is priced at the top of its proposed range, according to calculations from Reuters.

[ブルームバーグニュース @BloombergJapan]

スペースXとテスラ、史上最大の合併へ協議進展－12月にも発表か https://t.co/aT8b85p0ds

[雅 @Diegooooal]

👀

[Wall St Engine @wallstengine]

APPLIED MATERIALS $AMAT Q4’26 EARNINGS HIGHLIGHTS

🔹 Revenue: $10.41B (Est. $10.3B) 🟢; +53% YoY
🔹 Adj. EPS: $4.16 (Est. $4.05) 🟢; +92% YoY
🔹 Non-GAAP Gross Margin: 50.6% (Est. 50.45%) 🟢
🔹 Non-GAAP Oper Income: $3.68B (Est. $3.62B) 🟢
🔹 Non-GAAP FCF: $2.6B (Est. $2.26B) 🟢

Q1 Guide:
🔹 Revenue: $11.10B +/- $500M (Est. $10.89B) 🟢
🔹 Adj. EPS: $4.35 +/- $0.20 (Est. $4.26) 🟢

Fiscal 2026:
🔹 Revenue: $34.45B; +21% YoY
🔹 Adj. EPS: $12.89; +37% YoY

Segment Net Revenue:
🔹 Semiconductor Systems: $8.0B; +64% YoY
🔹 Applied Global Services: $1.9B; +26% YoY
🔹 Display / Corporate and Other: $517M; +25% YoY

Other Q4 Metrics:
🔹 GAAP Gross Margin: 50.4%; +240 bps YoY
🔹 GAAP EPS: $3.40; +43% YoY
🔹 Cash From Operations: $3.3B
🔹 Shareholder Distributions: $2.42B, incl. $2.00B buybacks and $421M dividends

Comments:
🔸 “AI continues to fuel strong incremental demand for semiconductors and Applied has built the most-enabling portfolio to advance the AI-computing roadmap, putting us in a great position for multi-year revenue and profit growth.”

[パウロ @paurooteri]

やはりDRAMとeSSDが強い

最近、ストレージ業界関係者数名がBlue Whale Technologyの記者に対し、大手企業数社がエンタープライズグレードのNANDフラッシュメモリの価格を引き上げる意向を販売代理店に口頭で伝えたと明らかにした。正式な価格改定通知はまだ出ていないものの、販売代理店側ではすでに価格変動が始まっているという。

記者が様々な流通チャネルを訪れたところ、複数の代理店が前述の価格上昇の兆候を確認した。ハイニックスの代理店であるユアン氏は記者に対し、「価格は日々上昇している」と述べた。同氏は、中古車の価格が着実に上昇しており、それが全体の価格を押し上げていると指摘。価格調整プロセスは、地方の代理店が毎週価格を調整し、それが価格連鎖を伝播していくという流れで、「我々もそれに応じて調整している」と語った。

しかし、サムスンの販売代理店である李明氏（仮名）は記者に対し、今回の値上げは全製品カテゴリーにわたる一律の値上げではなく、主にDRAM契約とエンタープライズグレードSSDのオリジナルメーカー／チャネルに集中しており、コンシューマーグレードNANDの包括的な値上げではないと説明した。「値上げの大部分は依然としてコンピューティングパワーセンター向けであり、現場レベルからの需要は実際にはそれほど大きくない」と述べた。

価格上昇は当初、DRAM受託市場とエンタープライズグレードSSDセグメントに集中していたため、今回の価格上昇はメーカーからサプライヤーに転嫁された際、最初にエンドユーザーに届くのではなく、中間層であるメーカーやモジュールサプライヤーに及んだ。李明氏は、これらのメーカーがメーカーからの価格上昇の矢面に立たされる側であり、業界チェーンにおけるそれぞれの立場が異なるため、感じ方やプレッシャーも大きく異なると付け加えた。
https://t.co/fk2EG0s4uF

[パウロ @paurooteri]

こりゃ工場が遅れて2028年の供給に間に合わないかもですね

[NVIDIA Newsroom @nvidianewsroom]

NVIDIA announces CEO transition. Read the release: https://t.co/MTTUvFHCJA

[パウロ @paurooteri]

noteのメンバーシップ向けに記事を投稿しました。
MLCC輸出の状況です。
https://t.co/aTlEH2c5yp

[bb8 @b24288375]

投資とは無縁なのに利上げにビックリする普通のサラリーマン面白すぎるだろ😅

[郭明錤｜Ming-Chi Kuo @mingchikuo]

Jensen Huang matters to Nvidia because he can take technologies his engineers believe are possible and turn them into products the whole company commits resources to, takes risks on, and has to deliver. That's why the supply chain gets such strong order visibility and is willing to keep up with aggressive spec upgrades. It also forces competitors to follow, which drives technology upgrades and supply chain investment across the entire AI industry.

Between the two, deciding on aggressive spec upgrades and making them happen is the harder part, compared with giving the supply chain order visibility. Huang doesn't need to be the person at Nvidia who knows the most about new technologies like HVDC, CPO/NPO, or CCL/PTFE. His value is that he weighs strategy, organization, technology, and supply chain all at once, then makes the call quickly and puts resources behind it.

Based on what's public so far, it looks hard for Huang to come out of this unscathed. If so, the practical move may be to gradually reduce Nvidia's reliance on Huang's personal decision-making while preparing a succession plan. Along the way, two things are worth watching to gauge where AI technology and investment are heading: whether Nvidia can keep developing new products and technologies as efficiently as before, and how competitors respond. Supply chain research will be an important basis for that judgment.

[雅 @Diegooooal]

@yukimamax これはこれで胸熱

[雅 @Diegooooal]

・・・

[Elon Musk @elonmusk]

@zerohedge SpaceX has a chance of going about half the GW next year of actually usable compute.

“Actually usable” is an important metric. It takes us ~2 months working 24/7 to get a large compute cluster working reliably. I’d guess closer to 6 months for others.

[Sawyer Merritt @SawyerMerritt]

Tesla FSD Supervised (now named Tesla Assisted Driving in Europe) has officially been approved in Finland, marking the 11th European country.

Countries in Europe that have approved it so far:
• Belgium
• Netherlands
• Denmark
• Lithuania
• Finland
• Estonia
• Slovenia
• Czechia
• Croatia
• Slovakia
• Latvia

[Serenity @aleabitoreddit]

You heard 10x demand for compute > supply from $CRWV today.

Hearing from $TSM today on Nikkei Forum:

"The demand on our leading node [most advanced] silicon is insatiable"

Customers "want to get [their] hands around any compute capability they can have. It's truly an insatiable environment."

Q: Whether AI demand was slowing?
A: "we did not see any sign of slowdown"

Q: How demand compared with TSMC’s ability to supply
A: "the demand actually is still far above our capability today to supply."

TSMC identified power and substrates as bottlenecks customers were encountering.

On the mid-to-high-50s annualized growth estimate for AI accelerator revenue from 2024 to 2029. Wei said "it probably will be higher."

Describing things as "insatiable" from a traditionally conservative company... doesn't quite look like any collapse in AI demand.

[パウロ @paurooteri]

@Biz_zatukora そう思います
ただ、中国のビジネスの遮断が越境データセンターへのアクセスまで至るとAI市場全体がおかしくなりそうです
このあたりはポストしていこうかなと