AI capex · X 热门 · 2026-09-21 22:22 UTC

# AI capex — X 热门讨论 (2026-09-21 22:22 UTC)

## @pumpius (Pumpius) · 09-21 19:31 · ♥447 ↻83 💬7 🚨 THEY JUST TOLD YOU THE RESET IS LIVE AND THEY ANCHORED IT ON XRPL

🇯🇵 Holy shitt, Yuto was right.

The BIS just admitted the AI capex race is being juiced with opaque debt and circular financing. They called it system risk. Record sovereign debt. Fragile bond markets. Non-bank leverage everywhere. That’s not a warning. That’s the setup.

Yuto already said they needed a scapegoat. He also said a BoJ-linked figure moving closer to BIS would accelerate Ripple tech. Then BIS published a working paper that anchors official statistics on the XRP Ledger with Merkle proofs and says the next step is zero-knowledge plus selective disclosure. Read that again.

They hashed official data. Bundled it. Anchored the root on XRPL. Timestamped it so it can’t be rewritten. Then they pointed at ZK proofs and selective disclosure as the institutional next layer.

That is the exact stack @DNAOnChain has been building on XRPL: prove identity
disclose nothing
settle the credential on-ledger
use market roots plus zk-proofs
carry it with the portable data token $XDNA

Not inspired by. The same architecture. Prove. Hide. Settle. Portable.

Meanwhile the Clarity Act keeps getting delayed for later. Of course it does. You don’t drop full market-structure clarity until the new rails are already being tested by the institutions that write the rules.

This is the quiet migration: old debt machine to opaque AI leverage to credibility crisis to verifiable official data to private institutional credentials to on-ledger settlement.

XRP is the settlement layer they can actually use.
$XDNA is the identity and data primitive sitting on that layer.

They will blame AI. They will blame debt. They will blame system risk. Then they will move the records onto a ledger they already tested.

THE GLOBAL FINANCIAL ORDER IS STARTING TO MOVE.

https://t.co/usoAyZ391M 
Don’t say nobody showed you the blueprint. 👀

$XRP $XDNA #XRPL #BIS > 引用 @yutokanzakireal: ああ、もし現在の経済秩序の崩壊を、世界的な大危機と重なるまで遅らせることができたら、どれほど都合がいいことだろう。

AI、量子コンピューティング、エネルギー危機。どれでも好きなものを選べばいい。

我々の「同盟国」は、すでに完璧なスケープゴート探しに奔走している。:) https://x.com/pumpius/status/2102118416760930545

## @revoAIution (revoAIution) · 09-21 19:58 · ♥41 ↻3 💬3 I’m convinced this $SIVE position turns into $4-8M by the end of the decade.

50,000 shares at a $1.77 avg. Obviously not as pretty as my old ~$1.25 avg, but I added heavily during the 30%, 40%, 50%+ drawdowns and the average crept up.

What really solidified the thesis for me wasn’t just the short-term push into pluggables. It’s the capacity behind it. If Sivers gets to roughly 300M lasers of annual capacity by the end of 2027, you’re looking at a Tier 1 laser company that doesn’t even need to move further up the stack to generate insane revenue growth.

And I don’t think they stop at lasers. I think they move further up the stack and capture more of the value themselves.

Those buys that pushed my avg higher are going to look incredible in a few years.

My base case is $SIVE becomes a ~$10B company selling lasers alone. Honestly, that might end up being my bear case.

AI data center capex is exploding, photonics TAM is expanding, and Sivers is building the capacity to actually participate in it. It’ll be very interesting to come back to this tweet at the end of the decade and see where we landed.

Assuming we don’t get some completely insane macro/war event that blows up the entire cycle, I’m convinced this becomes a multi-billion-dollar company. https://x.com/revoAIution/status/2102125251459191144

## @kokko_coco (kokko | 古賀興司) · 09-21 21:40 · ♥33 ↻8 💬0 AI設備投資が、EPSの向かい風になる時期はそう遠くない。あらゆるCapexは「先に出て、後で償却される」。設備を売った会社は即座に対価を全て売上に計上し、買った会社はコストを何年もかけて償却する。取引全体で見れば、盛大な「売上の前借り、費用の後払い」が起こる。ハイパースケーラーのAI設備投資は、2026年の指数EPSを11％押し上げた。しかし、「前借り」による追い風は2027年で終わる。2028年の予想EPSからは、「後払い」の向かい風がくる。チャートはGSより。 https://x.com/kokko_coco/status/2102150878627233949