Luna_Star

#AnthropicDiscloses$84.5BComputeDealWithSpaceX
The AI infrastructure race just reached another massive milestone.
Anthropic has disclosed agreements that could involve up to $84.5 billion in spending on Nvidia-based computing capacity supplied through Elon Musk’s xAI/SpaceX infrastructure through 2029, according to details from Anthropic’s confidential IPO filing reported this week.
The scale of the commitment is significant because it highlights one of the biggest challenges facing the AI industry today: access to compute.
Building increasingly capable AI models requires enormous amounts of processing power, and leading AI companies are competing for GPUs, data-center capacity and long-term infrastructure agreements.
Anthropic’s overall infrastructure plan is even larger. The company disclosed at least $518 billion in long-term infrastructure commitments over the next decade, involving major partners including Google, Amazon, Microsoft and Broadcom, alongside additional arrangements with xAI and AMD.
The SpaceX/xAI agreement is particularly interesting because it represents a major expansion from the arrangement previously disclosed earlier this year. Anthropic initially agreed to pay around $1.25 billion per month for access to SpaceX computing capacity, with the earlier structure representing nearly $45 billion over the potential term.
The latest filing puts the potential value at up to $84.5 billion through 2029.
But there is an important detail: this should not be interpreted as $84.5 billion of guaranteed spending. The agreement is largely cancellable with 90 days’ notice, making the maximum value different from a fixed, non-cancelable commitment.
The bigger story is what this says about the AI economy.
Compute is becoming strategic infrastructure. AI companies are not only competing on models and software anymore; they are also competing for reliable access to the hardware and data centers required to train and operate those models.
For Nvidia, the story reinforces the importance of continued demand for advanced AI accelerators. For SpaceX and xAI, it shows how computing infrastructure can become an additional business alongside rockets, satellites and connectivity.
And for the broader technology market, the numbers demonstrate just how much capital the next generation of AI could require.
The AI boom is no longer just about who builds the smartest model.
It is increasingly about who can secure the compute needed to build, train and scale those models.
That makes infrastructure one of the most important themes to watch as the next phase of the AI industry develops.
#Anthropic #SpaceX #AI #Nvidia

#MicronReportQ4Earnings
Micron Technology (NASDAQ: MU) is in focus today as the company prepares to report its fiscal Q4 2026 results after the U.S. market close on September 30. Micron confirmed the earnings call for 2:30 p.m. Mountain Time, making this one of the key semiconductor events of the week.
MU Price & Market Setup
Micron has entered the earnings event after an exceptional 2026 rally. The stock closed around $1,082.28 on September 25, while recent reporting shows it had gained roughly 16% over the previous month.
The stock's huge run has made the earnings reaction especially important because expectations are already elevated.
Recent options pricing indicated traders were positioning for a potential move of approximately 7% in either direction by the end of the week. Using the roughly $1,080 area as a reference, that represents a potential range near $1,004–$1,156. This is an options-market expectation, not a prediction of the actual move.
Q4 Expectations
Analyst estimates vary by data provider, but the current consensus range is centered around approximately $50–$52 billion in revenue and roughly $31–$32 EPS.
Micron's own previous guidance was:
Revenue: $50.0B ± $1.0B
GAAP EPS: $30.73 ± $1.00
Non-GAAP EPS: $31.00 ± $1.00
Gross margin: approximately 86%
These figures were provided by Micron with its Q3 results.
One current estimate puts Q4 revenue at $51.35B and EPS at $31.71, while another consensus source has revenue around $50.6B and EPS around $31.52.
Why This Earnings Report Matters
Micron's previous quarter was already extraordinary.
For fiscal Q3 2026, Micron reported:
Revenue: $41.46B
GAAP net income: $28.24B
GAAP EPS: $24.67
Non-GAAP EPS: $25.11
Operating cash flow: $25.39B
Revenue jumped from $23.86B in Q2 and $9.30B a year earlier.
Micron's Q3 gross margin reached 84.6% GAAP / 84.9% non-GAAP, compared with roughly 74% in the previous quarter.
That makes the Q4 margin outlook one of the most important numbers investors will be watching.
AI + Memory Demand
The central story behind MU remains the AI infrastructure boom.
High-performance AI systems require enormous amounts of memory bandwidth, particularly HBM and advanced DRAM. Micron has been investing heavily to increase capacity and serve demand from data-center customers.
In Q3, Micron's Cloud Memory Business Unit generated $13.77B, while the Core Data Center Business Unit generated $11.52B. Both businesses recorded very strong year-over-year growth and extremely high operating margins.
Micron has also highlighted strategic customer agreements designed to improve the visibility and durability of future demand.
The Two Numbers I’m Watching
1. Gross Margin
Micron previously guided toward approximately 86% gross margin for Q4.
A result materially above or below that level could significantly change how the market interprets the strength of the memory cycle.
2. Fiscal 2027 Capital Spending
Investors will also be watching how aggressively Micron plans to expand manufacturing capacity.
Recent analyst commentary has pointed to potential fiscal-2027 capital spending of roughly $40B–$50B, reflecting the scale of expected AI-memory demand.
The Bigger Question
The market is no longer asking only whether Micron can beat Q4 estimates.
The bigger question is whether AI-driven memory demand can remain strong enough to support Micron's earnings power into fiscal 2027.
A strong Q4 combined with strong forward guidance could keep attention on HBM demand, pricing and capacity expansion.
But because MU has already experienced a massive rally, investors may focus heavily on forward guidance and expectations, not just the headline quarterly numbers.
Key Levels to Watch
With MU around the $1,080 area ahead of earnings:
$1,100: psychological resistance.
$1,120–$1,130: next upside area if momentum accelerates.
$1,000–$1,020: major psychological/support region after a significant earnings-driven pullback.
$980–$1,000: important downside zone highlighted by recent options positioning.
The actual post-earnings reaction will depend on the reported numbers, forward guidance, margins, capital spending and management commentary.
My Earnings Checklist
Tonight, I will be watching:
• Q4 revenue vs. ~$50–$51B expectations
• EPS vs. ~$31–$32 expectations
• Gross margin vs. ~86% guidance
• HBM demand and pricing
• DRAM/NAND pricing trends
• AI data-center demand
• Fiscal 2027 revenue outlook
• Capital expenditure plans
• Strategic customer agreements
• Management's comments on supply and capacity
The setup is clear: MU enters the earnings report after a powerful rally, while expectations are already extremely high. The market reaction may therefore depend as much on the forward outlook as on the Q4 headline numbers.
Micron's own Q3 results showed just how dramatically AI demand has changed its financial profile, with revenue reaching $41.46B and Q4 guidance already calling for approximately $50B revenue and 86% gross margin.
Now the big question for September 30: Can Micron deliver another major beat and stronger fiscal-2027 outlook, or will elevated expectations make the earnings reaction more volatile?
#MU #Micron #MicronEarnings

2026 Crypto Market "Nuclear Level" Catalyst: The $4 Trillion Stablecoin Fantasia Behind eSLR Deregulation
As the entire crypto community is still debating the validity of the Bitcoin halving cycle, a saying has already circulated in the trading rooms of Wall Street: "This time it's tougher than the halving."
On the night of November 30, when the Federal Register quietly released a proposal to amend the supplemental leverage ratio (eSLR), a Telegram message exploded in the institutional trading group: "eSLR is cut, the game is over."
Behind this line of text is $210 billion of bank capital that has been tied up for a full ten years, about to be liberated.
The Forgotten Shackles: How eSLR Suppresses Banks from Buying Bonds?
The Supplementary Leverage Ratio (SLR) is a regulatory line established after the 2008 financial crisis, requiring banks' Tier 1 Capital to be at least 5% of total assets (6% for large banks). This means that for every $100 in U.S. Treasury holdings, banks must set aside $6 in capital, significantly reducing the attractiveness of holding government bonds.
The core content of the proposal on November 30 is very straightforward: "When calculating SLR, high-quality liquid assets (HQLA) — mainly U.S. Treasury bonds — will be excluded from the denominator." In other words, banks can purchase U.S. Treasury bonds infinitely without needing to set aside additional capital for it.
Don't underestimate this "less than 1%" adjustment; it means that the cost of funds for banks purchasing government bonds will drop from "regulatory punitive costs" to almost zero.
The "perfect storm" of stablecoins: when banks become the infinite demand for government bonds.
The underlying logic of mainstream stablecoins such as USDT, USDC, and FDUSD is extremely simple: for every 1 dollar stablecoin issued, there must be 1 dollar equivalent asset held in reserve. And the vast majority of these reserves are indeed short-term U.S. Treasury bonds.
The current total supply of stablecoins is approximately $306 billion, which corresponds to the need to hold $306 billion in U.S. Treasury bonds. When banks can purchase government bonds without restriction, three things will happen:
1. The issuance cost approaches zero
The current dilemma facing stablecoin issuance is that the short-term government bond yield is about 4.5%, but banks holding government bonds need to consume capital, resulting in high custody and transaction costs. Once SLR is relaxed, banks will rush to become custodians of stablecoin reserve assets, driving down costs through economies of scale. At that time, the yield for stablecoin issuers will soar from the current 1-2% to close to the government bond yield itself.
2. The "flywheel effect" of scale expansion
The current prediction figures being privately discussed on Wall Street are astonishing: Citibank's baseline scenario predicts that by 2030, the stablecoin market will reach $1.9 trillion, an optimistic scenario $4 trillion, and the most aggressive traders are calling for $8 trillion. This means an increase of 6-26 times from the current $306 billion.
3. Flood of liquidity in the crypto market
Stablecoins are the "blood" of the crypto world. When the total amount of blood increases by 5-10 times, the leverage limit of the entire ecosystem will be completely opened. DeFi, RWA (real world assets), meme coins, Layer 2, all tracks will gain unprecedented liquidity support.
A rehearsal for 2020: temporary easing, permanent madness
This is not the first time the Federal Reserve has adjusted the SLR. In April 2020, during the outbreak of the pandemic, the Federal Reserve temporarily exempted the SLR calculation for Treasury securities and reserves, allowing banks to hold unlimited amounts of these two types of assets.
What was the result? Bitcoin rose from about $4,000 in April 2020 to $69,000 in November 2021, an increase of more than 16 times in 18 months. More importantly, it was only a temporary exemption at that time, and the regulatory environment for banks participating in encryption business was extremely strict.
And this time, it is a permanent deregulation, coinciding with the overlap of three major policy benefits:
• SAB 121 repeal: Banks are not required to incur additional liabilities for custodial encryption assets.
• The stablecoin bill is implemented: banks can legally issue stablecoins.
• The Trump administration clearly supports: from the SEC to the OCC, the regulatory attitude has completely shifted.
This is not just a simple "halving bull market", but a triple blow of halving cycle + unlimited QE + green light from policies.
Wall Street is already "All In": Institutional rush layout
The most sensitive institutions have already taken action:
• Circle has converted all reserve assets into short-term U.S. Treasury bonds with maturities of 0-3 months to maximize policy dividends.
• BlackRock's BUIDL fund (tokenized US Treasury fund) has consumed $500 million in a month, with total assets under management approaching $2.9 billion, backed by JPMorgan's frenzied buying.
• Goldman Sachs has listed "stablecoin - short bond arbitrage" as one of the juiciest trading desks for 2026.
A trader working at Castle Hedge Fund revealed that he allocated all client funds into 3-month T-Bills last week and set a trigger condition: when short-term bond yields fall below 3%, he will go all in on crypto assets. His logic is simple and straightforward: "This is not a temporary exemption, it's a permanent exemption. Get ready to watch the show."
Risk and revelry coexist: the double-edged sword of the $40 trillion flood
What will happen to the market when 4 trillion dollars in stablecoins flood in?
Optimistic scenario: Bitcoin hits $200,000, Ethereum breaks $20,000, and Solana reaches $1,000. All crypto assets enjoy an epic valuation expansion.
Pessimistic scenario: Systematic leverage risk. A tenfold expansion of stablecoin scale means that the liquidation complexity of DeFi protocols grows exponentially. Once a depegging event similar to UST occurs, it may trigger a "crypto version of the Lehman moment."
Regulatory backlash risk: If the scale of stablecoins grows to threaten the traditional banking system, it cannot be ruled out that the SEC or FSOC (Financial Stability Oversight Council) will implement stricter capital adequacy requirements or even directly restrict banks from participating.
Market Timeline: "Three Steps" in 2026
Phase One (Q4 2025 - Q1 2026): The eSLR proposal completes a 60-day public notice period, with official implementation expected before March 2026. Institutions rush to build positions, and the supply of stablecoins grows moderately to $400 billion.
Phase Two (Q2-Q3 2026): The banking system completes technical integration, significantly reducing the cost of stablecoin issuance. Supply may surge to $1-2 trillion, pushing Bitcoin to challenge $150,000.
Phase Three (Q4 2026 - 2027): If everything goes smoothly, the $40 trillion goal will be achieved. The total market value of the crypto market may exceed $10 trillion, officially entering the mainstream financial system.
Investors' survival rules
In this "macro faucet" feast, retail investors should avoid blindly chasing highs:
1. Don't trade Gamma in a Beta market: When overall market liquidity is flooding, holding BTC and ETH spot is safer than high-leverage contracts. A flood of 4 trillion can lift all boats but can also instantly submerge all leverage.
2. Focus on the "native stablecoin" track: on-chain payment protocols (such as USDC's CCTP), DeFi infrastructure (MakerDAO's USDS), RWA tokenization platforms (Centrifuge); these tracks that directly benefit from the growth of stablecoins will achieve excess returns.
3. Monitor policy risks: Keep a close eye on FSOC's warnings regarding the scale of stablecoins, the Federal Reserve's subsequent supplementary terms on SLR exemptions, and the consistency of the Trump administration's internal crypto policies. Any policy reversal could be a "black swan."
4. Build positions in batches, refuse to go all-in: Even if the macro narrative is perfect, the market will experience multiple cleanouts of 20%-30%. DCA BTC in the range of $85,000 to $90,000 is more prudent than going all-in at once.
Final warning: When everyone is celebrating
The crypto market for 2024-2025 has already made many people feel crazy, but the real madness has yet to come.
When $4 trillion in stablecoins flood in, the market may rise to heights that make everyone feel "uneasy." At this time, remember the words of that Wall Street veteran: "This time it's not a temporary exemption, it's a permanent exemption."
The wallet is ready, but don't just prepare to buy. You also need to be ready to survive when the bubble bursts.
Because this time, it is not the FOMO of retail investors driving the market, but rather the American financial system has personally turned on the tap and directly inserted the pipe into the crypto market.
This party has just begun in 2026. #稳定币 #美国国债 #加密市场 #SLR政策 #Trump
Risk Warning: The implementation of the eSLR policy is uncertain, and the stablecoin scale forecast is based on current market conditions; actual progress may be adjusted due to regulatory changes. The crypto market is highly volatile, please assess risks carefully.
$BTC ‌$ETH ‌$SOL ‌

#BrentTops$106USTalksStall
Brent crude has returned to the $106+ zone as stalled US-Iran diplomacy keeps a significant geopolitical risk premium embedded in oil prices.
On September 29, Brent futures traded around $106.77, while WTI was near $93.94. The move came as markets remained focused on uncertainty around the Strait of Hormuz and the possibility of prolonged supply disruptions.
Oil Market Update
Brent had also moved above $107 earlier in the week, reaching around $107.04 on September 28 as reports indicated that diplomatic efforts to reopen the Strait of Hormuz had stalled.
By September 30, Brent had pulled back toward $103–104, but it remained on track for roughly a 14% monthly gain, showing how quickly geopolitical headlines are being priced into crude. Reuters reported Brent around $103.43, while WTI was near $89.63.
Why Is Brent Staying Above $100?
The key issue is not simply current production. The market is pricing the possibility that transportation and exports could remain disrupted if negotiations fail.
The Strait of Hormuz is particularly important because a prolonged disruption can increase shipping costs, delay deliveries and tighten the effective availability of crude.
At the same time, some Middle Eastern export flows have started recovering, which is helping limit the upside. Reuters reported that Middle Eastern crude exports reached 16.328 million barrels per day in September, the highest level since the conflict began.
So the market is balancing two opposing forces:
Supply recovery = bearish pressure on crude
Geopolitical uncertainty = bullish risk premium
Key Brent Levels
Resistance
- $106–107: immediate resistance zone
- $108–109: next upside area
- $110: major psychological level
Support
- $103–104: first support
- $100: major psychological support
- $96–98: deeper support if geopolitical premium fades
A sustained move above $107–108 would keep attention on the $110 area.
On the other hand, a decisive break below $100 could indicate that traders are pricing a greater probability of improving supply conditions and reduced geopolitical risk.
Why Traders Should Watch Oil Closely
Oil is becoming an important macro signal beyond the energy sector.
If crude remains elevated, higher energy costs can add pressure to inflation expectations. That can influence Treasury yields, interest-rate expectations and eventually risk assets such as equities and crypto.
We have already seen how oil and Treasury yields can move together. With Brent recently above $106 and the US 30-year Treasury yield reaching levels not seen since 2002, markets are dealing with both energy-price pressure and higher borrowing costs.
What Comes Next?
The next major catalyst is diplomatic progress.
A credible agreement that improves the flow of oil through the Strait of Hormuz could remove part of the geopolitical premium and push Brent lower.
If talks continue to stall while supply risks remain elevated, Brent could continue testing the $106–110 region.
For traders, the key levels are simple:
$110 → major upside test
$106–107 → immediate resistance
$103–104 → near-term support
$100 → major market pivot
The biggest risk right now is headline volatility. Oil can move sharply in either direction when diplomatic or supply-related news changes the market's expectations.
This makes Brent one of the most important macro markets to watch alongside the dollar, Treasury yields, gold, equities and Bitcoin.

#OpenAIAnnualRecurringRevenueNears$70B
OpenAI’s annualized recurring revenue is approaching $70 billion, marking another major milestone in the commercial expansion of generative AI.
According to Reuters, citing a source familiar with the company’s finances, OpenAI’s annualized revenue run rate has increased by more than 70% since the beginning of Q3 2026, while enterprise sales have more than doubled since July.
$70B Revenue Run Rate
The most important point is that $70 billion is an annualized revenue run rate, not necessarily $70 billion of revenue already booked in 2026.
A run rate extrapolates recent revenue performance over a full year, so it can change quickly as sales accelerate or slow. Reuters also noted that this metric can sometimes give a misleading picture compared with audited full-year revenue.
That distinction matters when evaluating OpenAI’s actual profitability and cash generation.
Enterprise Demand Is Accelerating
Enterprise adoption is becoming one of the biggest growth engines.
OpenAI’s business-to-business sales have reportedly more than doubled since July. At the same time, consumer revenue generated during Q3 reportedly exceeded the total incremental consumer revenue added during all of 2025.
This suggests the AI market is moving beyond experimentation toward broader commercial deployment across companies and consumers.
Why This Matters for AI Stocks
OpenAI’s growth is not isolated to the software layer.
More AI users and enterprise workloads require more:
- GPUs
- Data centers
- Cloud computing
- Networking
- Storage
- Electricity
- Cooling infrastructure
- Semiconductor capacity
That creates a potential downstream effect across the AI infrastructure supply chain.
Oracle was one of the clearest market reactions to the report. Reuters said Oracle shares gained 5.3% on September 29 after the OpenAI revenue report emerged, reflecting the importance of OpenAI to Oracle’s AI-compute business.
Other AI-related infrastructure names also moved higher during the session, including companies involved in power, optical connectivity and semiconductor infrastructure.
OpenAI vs. Anthropic
The AI revenue race is also becoming more competitive.
Anthropic’s annualized revenue reportedly reached around $65 billion by July, according to reporting cited by The Information, while OpenAI’s latest reported run rate is now approaching $70 billion.
However, revenue growth alone does not tell the entire story.
The bigger questions for investors are:
How much does it cost to generate that revenue?
How quickly are compute expenses increasing?
Can AI companies turn massive revenue growth into sustainable cash flow?
Those questions become increasingly important as model training, inference and data-center capacity require enormous capital commitments.
The Bigger AI Infrastructure Story
OpenAI’s numbers highlight a broader trend:
AI adoption → more users → more workloads → more compute → more infrastructure spending.
This is why OpenAI’s financial growth can affect companies far beyond the AI software sector.
If demand continues accelerating, semiconductor manufacturers, cloud providers, data-center operators, networking companies and power infrastructure providers could remain closely connected to the AI expansion cycle.
But the market will also be watching whether infrastructure spending grows faster than sustainable revenue.
Key Takeaway
Approaching $70 billion in annualized revenue is a major commercial milestone for OpenAI, but the figure should be viewed as a run-rate indicator rather than confirmed annual revenue.
The next phase of the AI story will be about more than revenue growth.
It will be about revenue vs. compute costs, enterprise retention, margins, cash flow and the enormous capital required to support AI demand.
The AI boom is increasingly becoming a financial story as well as a technology story — and the balance between AI revenue growth and infrastructure spending could become one of the most important themes for technology markets through 2026 and beyond.

Witness the biggest evolution in Gate’s 13-year history!
One Gate Witness Program
Hold Anything, Spend Anywhere, Trade Anytime
No deposit or trading required. Log in to Gate, choose your sharing theme, and complete a valid share to receive the corresponding reward.
Get witness number 1, 11, 111, 1,111, 11,111, or 111,111 for a chance to win 100 GT, F1 tickets, and driver-signed merchandise
Be more than a witness. Be part of this evolution.
Join now: https://www.gate.com/activities/everything-money-ceremony
Witness starts from September 30, 04:00 (UTC)
‍#Gate #OneGate #OneGateWitnessProgram

🔥 What’s worth talking about today? Gate Square trending topics are live!
🔹US core PCE and final Q2 GDP are due today — could inflation and growth data reshape rate expectations again?
🔹Micron reports after the close, with HBM4 ramp and FY2027 guidance in focus — can the AI memory trade keep running?
🔹$MRVL jumps 4.5% as AI hardware stocks rally — is capital rotating further into optics and data-center plays?
🔹US-Iran talks stall as Brent moves above $106 — could rising oil prices put inflation back in focus?
Post with a trending topic and quality content can get traffic support, Featured placement, and Content Mining rewards.
💰 Macro, earnings, AI and geopolitics are all moving — share your take:
https://www.gate.com/post/topic

October 7. Something big is coming.
@Dr. Hantakes over the main stage with Max33Verstappen
to unveil one of the biggest upgrades in Gate's history.
See you at TOKEN2049. 👇

🤩 Happy National Day! Gate Social-exclusive gifts are being distributed, with up to 5,000 USDT!
Gate Social's 2️⃣4️⃣th Growth Points Lucky Draw is underway, with multiple great gifts awaiting you!
🎯 Growth Points Guide:
1️⃣ Plaza Tasks
Interact by posting, liking, commenting, and sharing daily to earn up to 600 Growth Points/day
2️⃣ Livestream Tasks
Watch livestreams, like, comment, share, and schedule livestreams to earn up to 300 Growth Points/day
3️⃣ Hot Chat Tasks
Participate in community discussions and share trading cards to earn up to 90 Growth Points/day
4️⃣ Creator-Exclusive Tasks
Publish content featuring designated tokens and trading cards to earn up to 300 Growth Points per task
Test your luck today 👉 https://www.gate.com/activities/pointprize?now_period=24
$BTC $ETH $ZEC

🤩 Happy National Day! Gate Social-exclusive gifts are being distributed, with up to 5,000 USDT!
Gate Social's 2️⃣4️⃣th Growth Points Lucky Draw is underway, with multiple great gifts awaiting you!
🎯 Growth Points Guide:
1️⃣ Plaza Tasks
Interact by posting, liking, commenting, and sharing daily to earn up to 600 Growth Points/day
2️⃣ Livestream Tasks
Watch livestreams, like, comment, share, and schedule livestreams to earn up to 300 Growth Points/day
3️⃣ Hot Chat Tasks
Participate in community discussions and share trading cards to earn up to 90 Growth Points/day
4️⃣ Creator-Exclusive Tasks
Publish content featuring designated tokens and trading cards to earn up to 300 Growth Points per task
Test your luck today 👉 https://www.gate.com/activities/pointprize?now_period=24
$BTC $ETH $ZEC

Gate Social Rewards Are Here — Win Up to 5,000 USDT!
Growth Points Lucky Draw Phase 2️⃣ 4️⃣ is ongoing, with plenty of rewards up for grabs!
🎯 How to Earn Growth Points
1️⃣ Square tasks — post, like, comment, and share, up to 600 pts/day
2️⃣ Live tasks — watch, like, comment, share streams, up to 300 pts/day
3️⃣ Chat tasks — join discussions and share trading cards, up to 90 pts/day
4️⃣ Creator tasks — post token or trading card content, up to 300 pts/task
Test your luck today 👉 https://www.gate.com/activities/pointprize?now_period=24
$BTC $ETH $ZEC

#AnthropicDiscloses$84.5BComputeDealWithSpaceX
The AI race is starting to look less like a software competition and more like an infrastructure competition.
Anthropic’s IPO filing gives a remarkable view into just how much computing capacity the next generation of AI companies believe they will need. According to Reuters, Anthropic has disclosed a potential computing-capacity arrangement worth as much as $84.5 billion with xAI, while its broader long-term infrastructure commitments are expected to reach at least $518 billion over the next decade.
Those numbers are difficult to process at first. But the more important question is not simply how much money Anthropic is committing.
It is why a company building some of the most advanced AI models in the world needs this much computing power in the first place.
The answer is becoming increasingly clear.
Training a frontier model is only one part of the equation. Once a model becomes commercially useful, computing demand does not disappear. Every user request requires inference. Every AI agent running continuously consumes resources. Longer context windows require more memory. More capable models require larger and more sophisticated infrastructure. And as AI moves from occasional chatbot interactions toward systems that can perform tasks continuously, the amount of computation required can increase dramatically.
That changes the economics of the AI industry.
The scarce resource is no longer simply access to talented researchers or good algorithms. Reliable computing capacity itself is becoming a strategic asset.
Anthropic’s infrastructure commitments show how companies are trying to secure that asset before demand becomes even harder to satisfy.
The $84.5 billion arrangement is particularly interesting because it represents a different approach from simply building everything internally. Leasing computing capacity can give an AI company access to large-scale GPU infrastructure without having to construct every data center, secure every power connection, purchase every server, and manage the entire physical deployment process itself.
That flexibility matters.
AI demand is growing quickly, but it is also difficult to forecast several years ahead. Leasing allows companies to scale capacity more rapidly and potentially avoid owning enormous amounts of hardware that could become underutilized later.
But there is another side to this strategy.
Long-term infrastructure commitments can become enormous fixed obligations. Reuters reported that roughly 80% of Anthropic’s disclosed infrastructure commitments are binding regardless of actual usage. That means the company is effectively making a massive bet that future AI demand, revenue and model usage will justify today's infrastructure decisions.
This is where the AI infrastructure story becomes much more interesting.
Anthropic is not relying on a single source of computing power. Its disclosed commitments involve multiple infrastructure and technology partners, including very large agreements connected to Google, Amazon, Microsoft and Broadcom. The strategy appears to be about securing enough capacity from multiple sources while continuing to expand the underlying AI business.
In other words, the AI race is creating an entire infrastructure economy behind the models.
And this brings us directly to memory.
GPUs get most of the attention when people talk about AI computing, but GPUs cannot operate in isolation. High-bandwidth memory, or HBM, is a critical component of modern AI accelerators because large-scale training and inference require extremely high memory bandwidth.
The latest TrendForce research makes this part of the story even more important. It expects HBM supply to remain constrained into 2027 and has raised its forecast for the blended average HBM selling price in 2027 to a 121% year-over-year increase, driven by tight supply, the growing mix of HBM4 and continued AI server demand.
That means the AI infrastructure race does not stop at data centers.
More AI computing demand creates pressure across GPUs, HBM, networking equipment, power infrastructure, cooling systems, data-center construction and electricity supply.
This is why I think the most interesting part of Anthropic’s filing isn't the headline number itself.
It is the confirmation of a structural shift.
For years, the AI conversation was dominated by model architecture, benchmark scores and product features. Now the bottleneck is increasingly physical. You can have brilliant researchers and sophisticated algorithms, but without enough chips, memory, electricity and data-center capacity, those models cannot be trained or served at the scale the market expects.
At the same time, simply throwing more hardware at the problem cannot be the industry's permanent solution.
If computing costs continue rising, AI companies will eventually need better efficiency as well as more capacity. Smaller models, quantization, sparsity, inference optimization, improved architectures and specialized accelerators can all reduce the amount of computation required for a given task.
That creates an interesting balance for the next phase of AI.
More computing power expands what models can do. Better efficiency determines how economically those capabilities can be delivered.
Anthropic’s enormous infrastructure commitments therefore tell us something much bigger than the spending plans of one AI company.
They show that the AI industry is moving into a phase where securing compute can be just as important as developing the model itself.
The real competition may not be simply about who builds the smartest AI.
It may increasingly be about who can secure enough computing power, memory, energy and infrastructure — while still making the economics work.
And that is a much bigger story than an $84.5 billion contract headline.

#MicronReportQ4Earnings
$MU is heading into earnings, and I think the biggest mistake would be to look at this report only through the headline numbers.
Revenue, EPS and whether Micron beats estimates will obviously matter, but those figures are mostly about what has already happened.
What I really want to know is what happens after this quarter.
Micron has become one of the clearest beneficiaries of the AI infrastructure buildout because every new generation of AI data centers needs more advanced memory. But the market is now asking a harder question: how much further can this growth actually run?
That is why HBM4 is the first thing I will be watching.
HBM3 and HBM3E have already established the importance of high-bandwidth memory in AI accelerators. HBM4 is the next major step, and this is where execution becomes critical.
It is one thing to have strong customer demand.
It is another thing to manufacture enough HBM4 at the required quality and yield to actually capture that demand.
If Micron gives investors confidence that HBM4 production is progressing well, yields are improving and meaningful capacity can come online as planned, that would strengthen the argument that the company still has another phase of AI-driven growth ahead.
But if management talks about production delays, weaker yields or limited capacity, the market could start questioning whether Micron can fully participate in the next stage of the AI memory cycle.
And that leads directly to the second thing I care about:
2027.
I think this is where the earnings call could become much more important than the earnings release itself.
The current quarter tells us how strong demand has been.
Management's 2027 commentary will tell us how sustainable they believe that demand is.
Are AI customers still aggressively securing memory capacity? Does Micron expect HBM demand to remain constrained by supply? Can pricing and margins remain strong as additional capacity comes online?
These are the questions that can change the market's perception of $MU.
A strong quarter with cautious 2027 guidance could create a very different reaction from a strong quarter accompanied by confident long-term commentary.
That is something investors sometimes forget during earnings season.
The market doesn't pay you for yesterday's numbers. It prices expectations for tomorrow.
There is another interesting part of this setup: options pricing is implying roughly 8–10% post-earnings volatility.
That doesn't tell us whether Micron will go up or down. It simply shows that the market expects the new information to matter.
And I think that makes sense.
Micron is sitting at an important intersection between the AI boom and the traditional memory cycle.
AI has created a much stronger structural demand story for high-bandwidth memory, but memory is still a business where supply, pricing, capacity and margins can change the picture very quickly.
That's why I don't think one earnings beat would automatically mean the stock's next move is higher.
The more interesting scenario would be:
Strong results.
Strong HBM4 execution.
Healthy capacity expansion.
And constructive 2027 guidance.
If those pieces come together, the market gets another reason to believe that Micron's AI opportunity is still expanding.
But if the numbers are good while HBM4 execution or the 2027 outlook disappoints, investors may start asking whether expectations have simply moved too far ahead of the underlying cycle.
For me, that's the real dividing line.
I'm not trying to predict whether $MU will pump or dump immediately after earnings.
I want to see whether Micron can turn today's AI demand into sustainable memory growth over the next several years.
Because the biggest opportunity in this cycle isn't simply selling more memory during a shortage.
It's proving that AI has fundamentally changed the demand structure enough to support stronger utilization, pricing and margins for longer than a normal memory cycle.
That's the part management needs to convince the market of.
So tonight, I'll be listening less to the headline number and more to the details.
HBM4 tells me how well Micron is executing.
2027 guidance tells me how management sees the road ahead.
Margins tell me how much of that demand is actually translating into economics.
The AI memory story is still very much alive, but this earnings report could tell us whether Micron is entering another leg of growth or whether the market needs to become more selective about what comes next.
No need to guess before the information arrives.
Let the report come out. Listen to management. Then reassess.
DYOR.

It's not just a bond selloff. It's the price of money in the world's largest economy hitting a level that a generation of investors has never had to navigate.
The 30-year Treasury yield climbed to 5.595% on Tuesday, the highest since 2002. The long bond has now risen for six consecutive sessions, pushing past the 5.6% mark that once looked like a ceiling and turning it into a floor. The move is not happening in isolation. The 10-year yield is hovering near 5.27%, its highest in 19 years. The average 30-year fixed mortgage rate has already broken through 7.45%. This is a repricing of long-term borrowing costs across the entire economy, and it is happening fast.
Two forces are driving it, and they are reinforcing each other. The first is energy. Elevated oil prices tied to the conflict in the Middle East are feeding directly into inflation expectations. Higher energy costs filter into transportation, manufacturing, and consumer prices, which makes it harder for inflation to fall and harder for the Fed to step back from tight policy. The second is supply. Corporate America is issuing debt at a record pace, and that wave of issuance is competing with Treasuries for the same pool of capital. Investment-grade companies sold roughly $1.68 trillion in bonds through August, up 27% from a year earlier. When the private sector is borrowing that aggressively, the government has to offer higher yields to attract buyers.
There is a third factor that is harder to quantify but just as important. Analysts at RBC Capital Markets have noted that there are no real technical levels for investors to anchor on in this zone. The market is in a vacuum. When there is no clear support, selling can accelerate because there is nothing to stop it. That is how you get from 5.3% to 5.6% in a matter of days.
What does this mean beyond the bond market? For anyone with a mortgage, a credit card, or a car loan, it means borrowing costs are rising again. For equity investors, it means the discount rate used to value future profits is going up, which compresses valuations. When the risk-free rate is 5.6%, the bar for holding a stock that pays no dividend gets higher. That is why the S&P 500 fell 0.5% on the same day the 30-year yield broke through 5.6%.
The survey data suggests the market thinks this is not over. More than half of respondents in a Bloomberg poll expect the 30-year yield to touch 6% before the end of 2026. BlackRock has taken a low allocation to long-dated Treasuries in its latest outlook. The message is clear: the long end of the curve is not a place investors want to be right now, and the burden of proof is on the data to change that.
Friday's jobs report and the coming inflation prints will decide whether this is the peak or another step higher. For now, the market is pricing in the possibility that the era of cheap long-term money is not coming back anytime soon. And that changes the calculus for everything from housing to corporate capital spending to the valuation of every asset that depends on a discount rate.
DYOR 🔎 NFA ✔️
#US30-YearTreasuryYieldHits5.595%,HighestSince2002

$BTC
• BTC at 84,312.4 dollars, +1.36% in 24h, high 85,632.2, low 82,953.5, volume 7.47K BTC.
• Perp at 84,282.7 dollars, +1.48%, turnover 628.89M dollars. Spot outflow of 5.14M dollars shows less sell pressure.
• Resistance 84,777 - 87,401 dollars, support 83,800 EMA cluster - 82,625 dollars.
BTC bounced from low 82,953 dollars to 84,312 dollars today. Fund inflow for large-cap spot products stayed strong last week, helping price hold above 83,800 dollars.
On-chain data shows more coins left trading venues than entered. Large wallet outflows around 318k for SOL and similar pattern for BTC point to holding. DeFi TVL and DEX volume stay firm.
Futures volume 7.25B dollars, open interest 7.67B dollars. Funding up since mid-Sep, near highest since late July. No big liquidation cluster today.
• EMA5 83,953.5, EMA10 83,806.7, EMA30 83,813.1 now support on 4h. Daily EMA20 ∼84k, EMA50 ∼80k, EMA100 ∼78k, EMA200 ∼75k.
• RSI 58-68, MFI 70.1, bullish but not overbought.
• Stochastic up from low zone.
Above 84,777 dollars opens 85,632 dollars and 87,401 dollars. Below 83,800 dollars, next support 82,625 dollars and avg price 78,571.5 dollars.
#ShareWeekly #Gate广场中秋团圆局 #每日神贴 #Gate广场
#btc ‌ ‌