# AI capex — X 热门讨论 (2026-09-30 10:57 UTC)
## @HH_SSS_HH (SSS) · 09-29 12:00 · ♥31 ↻9 💬3 この無課金AIは初期投資(CAPEX)という概念を知らないのか(笑) 太陽光や蓄電池に数百万円かけて「燃料費タダ」って、牛を買って「牛乳がタダ！」と喜ぶのと同じです。 (元取れる頃にまた設備の更新、最後は屋根にゴミが残る) > 引用 @daina_murakami: 「30万km走らないとEVの元が取れない」は外部充電の価格を前提にした古い試算です。
現代の太陽光＋EV生活なら、燃料費はほぼタダ（ガソリン車の10分の1以下）。さらにガソリン車激減で今後車検・部品代、給油スタンド維持費が跳ね上がるリスクを天秤にかければ、数万kmで余裕で逆転します。 https://x.com/HH_SSS_HH/status/2104904024176918831
## @puckrin (Nic) · 09-30 05:02 · ♥32 ↻4 💬7 Micron earnings today could tell us a lot about the AI boom.
Expectations are HUGE:
- $51B revenue - $31.5 EPS - ~86% gross margin
Q3 results were:
- $41.5B revenue - $25.11 EPS - 84.9% gross margin
What actually matters?
Guidance.
Especially revenue, gross margins and the outlook for AI memory demand.
If Micron keeps growing revenue while sustaining margins in the mid-80% range, it would show that the boom is translating into real profits.
If that outlook starts to weaken, it raises a bigger question:
How durable is the AI capex boom? https://x.com/puckrin/status/2105161452902572519
## @AustralRes_En (Austral Research) · 09-29 11:37 · ♥31 ↻7 💬0 BofA sees hyperscalers rising from 2% to 4.9% of the global IG index by 2027.
Still below the big six US banks, but AI capex is turning cash rich tech companies into benchmark scale borrowers.
The key risk is increasingly credit quality, as negative free cash flow means greater reliance on debt, leases and project finance for the most capex intensive names.
Investors are not just funding AI infrastructure, they are underwriting future AI monetization to sustain today's balance sheets.
$AMZN $MSFT $GOOGL $ORCL $META #AI #Hyperscalers > 引用 @AustralRes_En: Hyperscalers have already committed to roughly USD 1.4 trillion in lease obligations, but around USD 1.1 trillion relates to data centres that are not yet operational.
As a result, these growing obligations are not yet recognised as lease liabilities on their balance sheets.
So we're only seeing the tip of the iceberg.
The issue is that the growing use of structures such as project finance and off-balance-sheet commitments requires investors to look beyond reported leverage.
In other words, even if these commitments are not yet reflected in reported debt, the companies have already committed to future lease payments.
That debt burden will increase over time, meaning that apparent leverage may understate the underlying economic risk.
But there is more.
Concentration risk is also rising. JPM estimates that the combined weight of hyperscalers in the USD IG market could approach 10% by 2030, while the USD IG market has historically functioned in an orderly manner with issuer concentration levels between 3% and 8%. Currently, some names alone account for ~1.5%.
#AI #CreditMarkets #Hyperscalers $MSFT $GOOG $AMZN $META $ORCL $SPCX https://x.com/AustralRes_En/status/2104898232409038995
## @pesshun (マイクロンぺす子) · 09-30 03:35 · ♥32 ↻1 💬1 マイクロン決算の投資家見込 売上昨年度比4.5倍↗️ 約93億ドル→約415億ドル↗️ 粗利80%over💥
見込時点で強すぎて意味わかんないw その上キャッシュリッチでcapexバンバン行ける。 製品単価は上がる一方。 今後の需給が翳ったら？くらいしか弱味無い💪その需給も来年まではほぼ売り切れ。2028以降の見込みは？AIがなくなるわけなくない？なくなくない？
要するに決算ではなくて、機関投資家の思惑で左右されたのが6月決算。 9月はもうそうはならんぞ‼️ｺﾉﾔﾛｳ $MU > 引用 @micron_girl: マイクロン前夜祭②【利益率編】🔥
売上だけじゃない。
☁️クラウドメモリ 粗利83% 🏢データセンター 粗利87%
87%？？？って原価どこ？笑笑
もう「普通のメモリ会社」じゃない。 利益の出方までバケモノなんです🤗🔥🚀 https://x.com/pesshun/status/2105139438305108184