Elon Musk now literally holds the off switch on Anthropic's $2 trillion IPO.
Anthropic's filing shows it has agreed to pay Elon's AI company up to $120.5 billion for computing power. And starting early next year, either side can walk away with 90 days' notice.
The deal is with SpaceXAI, the SpaceX unit that used to be xAI, and it rents Anthropic the Nvidia chips that run Claude.
In May, the price was $1.25 billion a month for one data center in Memphis. Reuters reported the confidential draft of this filing put the total at $84.5 billion. The public version filed today says $120.5 billion.
Once all the capacity arrives in December, the bill jumps: Anthropic expects to pay Musk about $3.1 billion every month.
Anthropic reportedly brought in $16.8 billion last quarter, which works out to about $5.6 billion a month. So more than HALF of what Anthropic earns today is about to go to one supplier.
Now here's where it gets really insane:
The lock-in on those contracts ends in January and March 2027. After that, either side can cancel with 90 days' notice. Fortune reported in July that the clause gives SpaceX a way to take its chips back if Grok suddenly needs more.
Grok competes directly with Claude.
And in February, Elon publicly called Anthropic's AI evil.
So the company selling itself to the public as the safe AI lab is about to run a huge share of its product on hardware owned by its loudest critic. That critic also owns a direct competitor, and he can start the 90-day clock in early 2027.
The filing does warn that some of Anthropic's compute suppliers build their own AI models, and that warning reads very differently once the supplier is Elon.
And Elon needs this deal just as badly: SpaceX's CFO has promised $100 billion in annual revenue by the end of this year. At full speed, Anthropic's rent comes to about $9.3 billion a quarter. That's more than the roughly $7.6 billion SpaceX's entire AI business brought in last quarter.
So 2 of the biggest IPOs in history are now holding each other up, and both of them have a 90-day exit.
Anthropic has a real argument here. Computing power is the scarcest thing in AI right now, and Amazon's cloud capacity for 2027 is reportedly largely sold out. The filing even admits Anthropic badly underestimated how much compute it needed in the past. Grabbing every chip on the market is the rational move, and the same clause lets Anthropic leave if cheaper capacity shows up.
But public buyers deserve 2 things before the listing: Anthropic should publish how much of Claude runs on SpaceX hardware and what happens to its customers if that notice ever arrives. And it should get the 90-day exit removed for the full length of the contract, the same way its other big compute bills are locked in.
Until then, anyone buying ANT at $2 trillion has to answer one question for themselves:
Who do they trust to keep Claude running in 2027, Anthropic or Elon Musk?
🚨🚨🚨 Chainlink CRE fees may be much higher than we initially estimated 🚨🚨🚨
👀 We were assuming that there's only 1 orchestration fee in this DTCC transaction, but it looks like there's actually 4 separate orchestration fees, plus a CCIP fee too 👀
𝐓𝐡𝐚𝐭'𝐬 𝟓 𝐬𝐞𝐩𝐚𝐫𝐚𝐭𝐞 𝐟𝐞𝐞𝐬 𝐩𝐚𝐢𝐝 𝐢𝐧 𝐋𝐈𝐍𝐊 𝐢𝐧 𝐭𝐡𝐢𝐬 𝐨𝐧𝐞 𝐭𝐫𝐚𝐧𝐬𝐚𝐜𝐭𝐢𝐨𝐧. 🔥🔥🔥
The CRE orchestration fee isn't charged once per transaction. It's charged on every workflow step a user signs (see attached images - blurry because they're screenshots from the live demo).
In the DTCC demo at Sibos, every wallet signature shows "Orchestration fee 0.01 LINK":
13:30 Create financing contract
14:48 Deposit $10m tUSD
17:22 Request $5m loan against MSFT
29:44 Repay loan
That's 4 separate Orchestration fees, paid in LINK.
On top of that there's a separate CCIP fee (20:15) for bridging the collateral to the DTCC AppChain.
Repo + securities lending is $20T+ a day.
Even a tiny share, with multiple LINK fees per transaction, adds up fast.
The scaling is clear: more steps in a workflow means more fees.
$LINK
Chainlink keeps quietly shipping the actual rails for tokenized finance—CCIP Vault Adapters already live across 80+ chains with Aave, Lombard, and others on board—while $LINK does its classic slow-motion retest around $13 after the post-CCIP 2.0 fade from the mid-teens. Whales are active, Binance longs still outnumber shorts, and the higher-timeframe structure remains intact even as short-term selling pressure and neutral technicals dominate the tape. In a market that often prices narrative faster than infrastructure, the oracles are still building while the token plays hard to get.
BREAKING: The US economy has now experienced 78 consecutive months of expansion, the 6th-longest business cycle since 1854.
This is well above the long-term average of ~49 months and the median of ~38 months.
By comparison, the longest period without a recession was between 2009 and 2020, at 128 months.
All 5 of the most recent US expansions rank among the 7 longest economic growth periods since business cycle dating began in 1854.
Unconventional monetary policy, historically large budget deficits, alongside the AI investment boom, appear to be extending the duration of business cycles.
The data says the US economy is remarkably strong.
The biggest disconnect with investing in Chainlink is seeing all of these mammoth integrations with Swift, DTCC, etc., and thinking "ok that's great, but how is the LINK token involved?"
Well now that's settled:
Chainlink’s overall platform is the Chainlink Runtime Environment (CRE). These institutions are integrating CRE.
All transactions orchestrated through CRE have an "orchestration fee".
In addition to that, Chainlink services within CRE also require a fee (requirements for data, interoperability, compliance, privacy, etc.). The more complex a transaction, the more of the Chainlink platform is required, and the higher the fees.
The attached images are the evidence of this. They show example orchestration fees required when using Chainlink CRE.
Now imagine how this scales:
Swift is connecting thousands of banks to blockchains through CRE.
DTCC is connecting multiple private bank chains and public chains to their Collateral Appchain through CRE.
Bottomline is connecting 600+ banks to blockchains through CRE.
Swift alone sent 60 million messages per day in 2025. 15 billion messages in total 2025. Chainlink is what makes it possible to turn ISO 20022 messages into blockchain transactions.
AI Agents will also consume LINK through CRE with Chainlink's new product Chainlink for Agents, which lets AI agents act through CRE.
$LINK
Introducing Ondo Intelligent Portfolios, the first three portfolios powered by BlackRock.
Ondo Intelligent Portfolios introduces a new onchain product category: curated investment portfolios delivered as single onchain transferable tokens.
The first three portfolios are based on portfolio strategies developed by BlackRock for Ondo, marking the first time eligible onchain investors can access exposure to such strategies through a single token.
1. BLKHIon: Ondo High Income Powered by BlackRock
2. BLKDIGon: Ondo Diversified Growth Powered by BlackRock
3. BLKGRWon: Ondo High Growth Powered by BlackRock
Diversified, professionally constructed strategies have historically required brokerage accounts and traditional fund structures. Now, delivered as peer-to-peer transferable tokens from Ondo, these onchain portfolios become accessible to eligible non-US investors in permitted jurisdictions through the wallets, exchanges, and DeFi applications they already use.
“Tokenization creates new ways for portfolio strategies to be delivered through digital infrastructure. Diversified portfolio strategies can be incorporated into tokenized investment products, enabling eligible investors to access diversified allocations through a single instrument. It shows how established portfolio construction approaches can be delivered through new channels and technologies.” - Lisa O’Connor, Global Head of the Model Portfolio Solutions team and Co-CIO for Global Solutions within the Multi-Asset Strategies group at BlackRock
Ondo Intelligent Portfolios can unlock novel capabilities:
→ Programmatic rebalancing
→ Full composability with DeFi
→ Complete transparency onchain
→ Multiple asset classes in a single token
This is just the start for Ondo Intelligent Portfolios. The infrastructure is now in place for leading financial institutions to bring their asset allocation expertise onchain.
$LINK People are underestimating what the Infosys connection actually means.
Chainlink doesn't need to sell its stack to every bank one by one.
Infosys already builds and maintains critical infrastructure for financial institutions around the world, and Chainlink is now listed among its global alliance partners.
That's the important part.
CCIP. CRE. Compliance. Data. Proof of Reserve.
If Chainlink becomes part of the toolkit companies like Infosys use when bringing financial institutions onchain, Chainlink gets something much more valuable than another logo:
A distribution layer into the existing financial system.
The institutions don't all have to find Chainlink.
Their technology providers can bring Chainlink to them.
That's how infrastructure standards scale.
chainlink bought back $15m in LINK over 90 days with zero token unlocks during the same window. reserve holds 4.5m LINK with $49.5m cumulative inflows. SVR feeds generated $12.4m in 2026 from recaptured liquidation value alone. CCIP just passed wormhole in transfer volume across 70+ chains. DTCC collateral appchain targeting Q4 2026 deployment for automated margin call settlement, currently processing $2.6 quadrillion annually. 52 banks committed to pangea for atomic FX settlement. CME launched 24/7 LINK futures. two spot ETFs live. 8,000 new wallets added in 5 days. token is 85% below ATH. the buyback engine is running, the unlock pressure paused, and the biggest infrastructure deployment in crypto history is 5 months out. price doesn't reflect the progress and that's either a gift or a trap. stop trading start believing.
Leopold Aschenbrenner's Situational Awareness fund is heavily reliant on the real AI bottleneck: power & data center infrastructure to bring massive GPU clusters online
Here are his top holdings:
$BE: 22.79%
$SNDK: 18.79%
$CRWV: 14.42%
$IREN: 10.40%
$CORZ: 10.09%
$APLD: 8.30%
$RIOT: 3.69%
$CLSK: 2.71%
Thesis in one sentence: The AI race won't be won by who has the best model, it will be won by who can bring compute online the fastest and at scale.
AI power demand is becoming one of the biggest AI Bottleneck of this decade.
Goldman Sachs predicts that power demand could surge 200% by 2030.
Power Generation
$VST — Vistra Corp.
$CEG — Constellation Energy
$NEE — NextEra Energy
$D — Dominion Energy
$GEV — GE Vernova
Nuclear
$OKLO — Oklo Inc.
$NNE — Nano Nuclear Energy
Grid, Electrification
$ETN — Eaton Corp.
$POWL — Powell Industries
$HUBB — Hubbell Inc.
$ATKR — Atkore Inc.
Data Center Infrastructure
$VRT — Vertiv Holdings
$FIX — Comfort Systems USA
$EME — EMCOR Group
$PWR — Quanta Services
$MTZ — MasTec
Fuel Cells / Hydrogen
$BE — Bloom Energy
$PLUG — Plug Power
$FCEL — FuelCell Energy
$BLDP — Ballard Power Systems
$HYLN — Hyliion Holdings
$GNRC — Generac Holdings
Energy Storage
$FLNC — Fluence Energy
$EOSE — Eos Energy Enterprises
Solar Energy
$TE — T1 Energy
Bookmark for research and invest in the best ones!
There is a near-perfect correlation between US oil prices and US CPI inflation, as shown in our below analysis.
Oil prices have averaged near $100/barrel since March 6th, or 79 days.
The longer this persists, the more inflation we will see.
Asset owners are the only winners.
A message for my $TSLA friends eyeing the SpaceX IPO $SPCX.
I traded $TSLA for years. I know the community. I know the excitement when Elon takes something public. But before you chase @SpaceX at $1.75 trillion, read the S-1 carefully.
SpaceX doesn't need your money.
They raised at $800B in private tenders six months ago. They could raise $50B privately tomorrow with a phone call. This IPO isn't about raising capital. It's about giving insiders liquidity.
95% of @SpaceX shares are held by insiders. Only 5% will be publicly traded. Insiders hold $1.66 trillion in paper wealth they currently can't sell. The IPO changes that.
And they've structured it so insiders can sell BEFORE the standard 180-day lock-up expires. @SpaceX built in early release provisions -- after the first earnings report, insiders can sell up to 20% of their shares.
They're also reserving 30% of IPO shares for retail. Ask yourself -- when has Wall Street ever given retail the best seats in the house unless retail was the product?
100x revenue.
$4.9B net loss.
xAI burning $6.4B a year while @Starlink subsidizes it.
This isn't 2020 Tesla at 20x revenue with a clear path to profitability. This is a different risk profile.
Now here's the part I want you to actually consider.
SpaceX's S-1 sizes their satellite-to-phone business (Starlink Mobile) at a $740 billion TAM. Their Connectivity segment does $11.4B at 63% EBITDA margins. Those numbers are real and impressive.
But buried in the S-1, @SpaceX names their D2D competitor: $ASTS .
@AST_SpaceMobile $40 billion market cap.
Not $1.75 trillion. $40 billion.
Here's what $40B buys you:
98.9 Mbps proven from space to unmodified phones (SpaceX does 3-5 Mbps)
The only low-band D2D spectrum access on Earth (indoor coverage SpaceX can't match)
All three US carriers forming a joint venture around ASTS technology
Google invested $358M
their largest public equity holding
AT&T, Verizon, Vodafone as equity investors
$3.5B cash, $1.2B contracted backlog
3,900 patents, custom ASIC in production
Three satellites launching on a Falcon 9 next month
60 carrier partners covering 3 billion subscribers
@SpaceX at $1.75T is pricing perfection across rockets, satellites, AI, and Mars. One miss and it corrects hard.
$ASTS at $40B is pricing uncertainty in a $740B market where the technology is already proven and the carriers have already chosen sides.
The Tesla community knows what it feels like to find a mispriced stock before the world catches on. $TSLA at $30 pre-split wasn't obvious to anyone except the people who did the work.
$ASTS at $106 in a $740B market with 33x faster speeds than SpaceX D2D, a carrier JV, and institutional discovery just beginning -- that's the same kind of setup.
So before you throw money at a $1.75T IPO where insiders are building exit ramps, maybe look at the $40B competitor they named in their own filing.
Not financial advice. Just math.
$ASTS 🛰️
cc @SawyerMerritt@unusual_whales@DanBTC916
chainlink CCIP absorbed $4b in assets from layerzero in 60 days after their vulnerability disclosure. kelp DAO moved $1.2b, lombard $800m, kraken moved their entire bridge infra. zero CCIP exploits since launch vs $2.8b stolen from competing bridges. SGX just adopted chainlink for OTC FX settlement in the world's 3rd largest FX market. LINK sits at $6.8b market cap securing $110b in value. the cross-chain infrastructure war ended and most people were looking at shitters when it happened
🦔Microsoft canceled its internal Claude Code licenses this week after token-based billing made the cost untenable, even for a company with effectively infinite cloud resources. Uber's CTO sent an internal memo warning the company burned through its entire 2026 AI budget in just four months. American AI software prices have jumped 20% to 37%, and GitHub (owned by Microsoft) is dropping flat-rate plans for usage-based billing across its products.
My Take
The AI subsidy era is ending in real time. The same company that put $13 billion into OpenAI and built the Azure infrastructure powering most of Anthropic's compute just looked at the bill from a competitor's coding tool and decided it was not worth paying. That is not a productivity failure on Anthropic's end. Token-based pricing is forcing every enterprise customer to confront the actual cost of running these models at scale, and the number turns out to be far higher than the flat-rate experiments suggested.
This ties directly to my Gemini Flash post yesterday. Anthropic, OpenAI, and Google all raised effective prices in the last six months. Enterprises that built workflows assuming AI costs would keep falling are now watching annual budgets evaporate in months. Two outcomes look likely from here. Either enterprises scale back AI usage to fit budgets, which slows the revenue ramp the labs need to justify their valuations ahead of IPOs, or the labs cut prices and absorb the losses, which makes the unit economics worse at exactly the wrong moment. Both paths land in the same place, the numbers stop working, and somebody has to take the writedown.
Hedgie🤗
8 undervalued stocks Leopold Aschenbrenner is signaling you to buy through his latest 13F reports. It is not $BE, $MU, or $NVDA, but:
• $HIVE — HIVE Digital
• $RIOT — Riot Platforms
• $TE — T1 Energy
• $APLD — Applied Digital
• $IREN — IREN Limited
• $BTDR — Bitdeer Technologies
• $KEEL — Keel Holdings
• $CLSK — CleanSpark
These stocks are all undervalued, with massive potential to create generational wealth.
Not financial advice. Do your own research.
DTCC processes $2.5 quadrillion annually and just named ondo as a tokenization partner for July production launch on Canton Network. $4.6 trillion in daily repo volume moving to atomic settlement. clearstream integration bridges european custody (€16T AUC) into the same pipe. CLARITY Act cleared senate banking 15-9 two weeks ago. october full production opens access to 3,600+ DTCC member firms. the entire thesis breaks if DTCC decides to build tokenization in-house, but the coalition they assembled (goldman, BNY mellon, moody's, microsoft) suggests they chose the partner route. ondo already runs $600m in tokenized treasuries. that number looks small until you realize the rails it sits on are about to connect to the largest settlement system on earth
kraken went exclusive chainlink CCIP for all current and future wrapped assets. $4b+ migrated from layerzero in 48 hours after the $292m kelp exploit. ZRO dropped 20%+ on the exodus. the part that compounds: every protocol that wants kraken integration now has to build on CCIP. 47% of layerzero OApps were running the same 1-of-1 DVN config that got drained. 94% admin key overlap between layerzero labs and nethermind DVNs. production multisig keys were being used for shitter trading and LP provisioning. the bridge market just split into two tiers in a single weekend and the institutional tier has one provider
$IREN : Ran some numbers on Saturday Napkin Math.
$IREN math is getting WILD 👀
Already signed:
• Microsoft = 300 MW → $9.7B
• NVIDIA = 60 MW → $3.4B
That’s:
• 360 MW contracted
• $13.1B total value
• $2.62B annualized revenue
BUT…
Childress still has:
• 390 MW UNSOLD
If future AI-cloud pricing improves:
• another ~$15B+ possible
• ~$3.8 - 4.5B/yr potential
Then add:
• Sweetwater 1 = 1.4 GW open capacity
At current AI infra pricing:
• ~$12–18B/yr possible from SW1 alone
Total potential 2027 🤯 for 750 + 1400 MW = 2.15 GW
• ~$15–22B+/yr revenue opportunity WOW.
They have total 5 GW so still has 2.8 GW, this is insane opportunity.
I swear this guy does NOT miss. Incredible.
This is getting insane... $VELO up 40% in two days since this mf'er bought, $LUNR up 48%, $NBIS up 51%, etc.
Just buy what he buys and you'll beat the markets lol.