rendi CLONE retweeted
There's a difference between having a price and having a market.
GPU providers already publish prices. But those quotes vary by region, hardware, contract length, and availability.
A functioning commodity market needs something more: standardized units, reliable delivery, and prices participants can actually trade against.
That's where compute markets get interesting.
🤖 Made with AI
rendi CLONE retweeted
What is an hour of AI compute actually worth?
Not all compute is priced the same.
The value of AI compute changes based on:
• GPU model
• Location
• Availability
• Demand
As AI infrastructure scales, transparent price discovery becomes essential.
That’s the market infrastructure RAX is building.
rendi CLONE retweeted
Hot take time:
AI compute should be priced like:
A) Oil — global commodity, single benchmark (WTI/Brent model)
B) Electricity — regional grid, locational pricing (LMP model)
Reply with A or B + your reasoning.
We're clearly Team B (hence building regional indices), but curious what you think.
Especially want to hear from folks who disagree 👀
rendi CLONE retweeted
Friday check-in 👋
What’s caught your attention in AI compute this week?
A new GPU, a data centre trend, or the future of ComputeFi?
We’re curious what the Rax community is following. Drop your thoughts below 👀
rendi CLONE retweeted
Compute is being planned on a multi-year horizon.
Chipmakers are already securing EUV tools for 2029 and 2030.
ASML says customers are asking for more machines every day, even as 2027–2028 capacity expands.
The industry is locking in the manufacturing base for the next wave of AI capacity years in advance.
Via @wallstengine
$ASML SAYS CHIPMAKERS ARE ALREADY LOOKING TO PLACE EQUIPMENT ORDERS FOR 2029 AND EVEN 2030
CEO Christophe Fouquet says “customers are asking us every single day to get more tools,” as AI-related investments drive stronger demand for advanced semiconductor equipment.
The company also outlined major milestones for its next-generation High-NA EUV technology, which Fouquet says is finally approaching high-volume manufacturing after more than a decade of development.
Samsung plans to begin using High-NA EUV in mass production in 2028, followed by TSMC around 2030. Intel is already increasing the number of layers processed and wafer output using the technology.
Separately, ASML plans to introduce its NXE:3800F EUV system in 2027, increasing throughput from 230 to 260 wafers per hour, with a greater impact on customer manufacturing capacity expected from 2028 onward.
rendi CLONE retweeted
📊 Open question for the compute market:
By the end of 2027, what % of APAC third-party GPU capacity will be traded on secondary markets rather than remain locked in long-term contracts?
We’re talking about neoclouds and independent operators, excluding hyperscaler self-use. Secondary trading includes unused or residual contracted capacity changing hands.
Our working hypothesis: 15–25%.
Why?
• Multi-year take-or-pay contracts still dominate, with limited early-exit options.
• Secondary trading is emerging, but liquidity, standardisation, and counterparty trust are still developing.
• As market infrastructure matures, more unused capacity could find a second buyer.
What’s your prediction, and why? 👀
Let’s revisit this in 12 months and see who got closest.
rendi CLONE retweeted
Scenario:
You sign a 12-month GPU contract for $2.5/hr.
3 months in:
• Your project pivots and needs a different region
• Your original contract still has 9 months left
• You're locked in, paying $216K for ~13 GPUs you can't use
What would you want?
A) Transfer the contract to someone else
B) Redeem it for other compute (different region/model)
C) Eat the cost — that's business
D) Something else? Tell us below.
💬 Reply with your choice + RAX Finance Waitlist address.
🎁 We’ll reward great answers with 5 Unboxing Opportunities!
🔗 app.rax.finance/waitlist/
rendi CLONE retweeted
Why regional compute pricing matters: the financing gap no one is talking about
AI infrastructure spend is exploding — but ownership structures differ sharply by region, and that gap is showing up in financing costs.
rendi CLONE retweeted
Exactly.
The bottleneck is connecting power, construction, equipment, and cloud capacity on one timeline.
The suppliers and operators determine whether a site gets built quickly or stays stuck in the queue.
Compute capacity creates most of its value only after it’s online and serving demand.
Yotta brought the AI infrastructure ecosystem under one roof, giving IREN leaders three days of direct conversations with the suppliers, builders and operators shaping what comes next.
- IREN is strengthening ties with leading construction and equipment partners
- Suppliers want to partner with developers that can execute
- Vertical integration from grid to AI Cloud is how IREN delivers with speed and reliability
Watch the highlights.
rendi CLONE retweeted
Why compute pricing lacks regional benchmarks:
Observation 1: Massive regional spreads
Same H100 card, wildly different economics across markets
Not 10-20% differences — we're seeing 2-3x spreads
Driven by: power costs, export controls, demand concentration
Observation 2: No one publishes regional clears
Cloud marketplaces show listing prices (not actual trades)
Real contracts are confidential bilateral deals
Lenders can't price risk without a settlement benchmark
Observation 3: This is exactly where electricity was in the 1990s
Before ISOs started publishing Locational Marginal Prices
Before you could hedge regional basis risk
Before financing could scale
RAX is building the missing layer: regional price discovery for compute.
Not because it's novel (electricity solved this 30 years ago)
But because compute is $180B annual spend with zero pricing infrastructure in Asia/MENA.
That's the gap we're closing.
rendi CLONE retweeted
The compute market is at the stage where:
• Electricity was in the 1990s (before deregulation)
• Oil was in the 1950s (before posted prices → spot markets)
• Carbon credits are today (trading exists, but fragmented)
We're watching commodity infrastructure form in real-time.
What we're building: docs.rax.finance/
rendi CLONE retweeted
$40B, just to buy the chips.
SpaceX is lining up Apollo-led financing for Nvidia capacity: about $10B in bank loans and $30B in investment-grade debt, aimed at Vera Rubin.
It is a compute purchase large enough to be financed like infrastructure.
Power, land, and chips are now the constraint.
The market is underwriting the factory, not the demo.
Via @wallstengine
$SPCX SEEKS $40B TO FUND MASSIVE $NVDA CHIP PURCHASE: FT
SpaceX is looking to raise about $40B in financing led by Apollo to fund purchases of Nvidia chips.
The plan includes roughly $10B in bank loans and $30B in investment-grade debt, with the transaction expected to close in 2027.
Elon Musk said in August that SpaceX had “decided to build exclusively on Nvidia,” specifically pointing to its Vera Rubin architecture, as the company ramps spending on AI and data-center infrastructure.
rendi CLONE retweeted
Three things every commodity market needs:
1. Fungible units
Gold: 1 oz = 1 oz anywhere
Compute: every contract is unique ❌
2. Price transparency
Oil: WTI, Brent published daily
Compute: real deals are confidential ❌
3. Tradeable infrastructure
Electricity: day-ahead markets exist
Compute: can't resell a signed contract ❌
That's the gap we're closing.
More on commodity formation: docs.rax.finance/
rendi CLONE retweeted
To everyone who’s been waiting with us — we hear you. 🤍
We know many of you have been on the Waitlist for a while, and we truly appreciate your patience.
We’re now moving deeper into ComputeFi, with discussions already underway for our first batch of idle compute partnerships.
And yes — your community points still matter. Holding an NFT will not only come with its own benefits, but also boost your community points.
We’re getting closer, and there’s more to share soon!
Thank you for staying with us!
We’re building this with you!
rendi CLONE retweeted
Compute is being contracted like capacity, not bought like hardware.
Oracle is selling Tencent access to ~100,000 advanced chips over 5 years, about $7B, with ~30% paid upfront.
Broadcom is lining up as much as $42B of financing against Anthropic’s multi-year compute commitment.
Nebius is buying inference tech so GPUs spend less time idle.
That is the real market:
power, chips, utilization, and who is willing to prepay for the output.
The scarce layer is still compute.
A TON OF THINGS HAPPENED IN THE STOCK MARKET TODAY.
Here's a full recap:
1. Palantir $PLTR and Armada announced a partnership to deliver sovereign AI on modular data center infrastructure manufactured in the U.S. and allied nations. Under the agreement, Palantir named Armada its inaugural Certified Modular Data Center Partner, combining Palantir’s Sovereign AI Operating System with Armada’s Galleon modular data centers, Armada Platform, and Sovereign AI Grid. The joint offering allows enterprises and governments to deploy open-weight AI models on infrastructure they own and control, with the ability to fine-tune models on proprietary data while keeping compute, data, and hardware entirely within their own security boundary. The companies said the modular approach can bring new AI capacity online in months instead of years, addressing growing demand for sovereign AI infrastructure.
2. Nebius $NBIS has acquired Inferize, an AI inference optimization startup, for an undisclosed amount. Inferize’s technology is designed to reduce model “cold starts,” cutting the time GPUs remain idle while models load and new inference capacity comes online. Nebius plans to integrate the technology and team into its Token Factory inference platform to improve GPU utilization, lower inference costs, and serve more AI workloads using the same infrastructure.
3. NVIDIA $NVDA and SoftBank have each completed the final $10B installment of their respective $30B commitments to OpenAI’s latest funding round. The financing totaled $122B in commitments at an $852B valuation and included Amazon’s $50B investment. SoftBank’s total investment in OpenAI has now reached approximately $64.6B, giving it an estimated 13% ownership stake. According to The Information, OpenAI is already targeting another $30B funding round at a valuation of around $1.4T.
4. Netflix $NFLX co-CEO Ted Sarandos said the company “isn’t growing as fast as I want us to,” while noting that some of Netflix’s strategic decisions are intentionally creating headwinds to subscriber growth. He also revealed that live programming accounts for roughly 5% of Netflix’s $20B annual content budget but generates only about 1% of total viewing, underscoring the company’s continued focus on balancing long-term strategic investments with overall engagement.
5. Nike $NKE reported Q1 FY2027 results with EPS of $0.48, beating expectations of $0.43, while revenue missed at $11.21B versus $11.32B expected, down 4% YoY. Gross margin came in stronger than expected at 42.8%, helped by lower warehousing and logistics costs, but weakness persisted across NIKE Direct (-8% YoY), Digital (-13% YoY), and Greater China (-22% YoY). The biggest disappointment was guidance, with Nike forecasting high-single-digit revenue declines for FY2027 and adjusted EPS of $1.15–$1.35, well below the $1.65–$1.67 consensus. Management said it continues to see challenges in Sportswear, Jordan Brand, and Greater China, while rolling out its new Pace operating model to accelerate its long-term turnaround.
6. Broadcom $AVGO has agreed to provide Anthropic with up to $42B in financing tied to its AI infrastructure expansion, according to Anthropic’s IPO prospectus reviewed by Reuters. The facility could fund roughly one-third of Anthropic’s $125.2B five-year TPU compute commitment, and the debt may be convertible into Anthropic equity. Anthropic is also expected to become Broadcom’s largest compute customer in 2027. The filing notes that Broadcom’s dual role as both a hardware supplier and financing partner could create potential conflicts related to pricing and access to compute.
7. Amazon $AMZN plans to expand its in-house delivery network from roughly 23,700 U.S. ZIP codes today to about 39,400 by 2029, according to an internal planning document reviewed by Business Insider. That expansion would allow Amazon to reach approximately 95% of U.S. ZIP codes and 99.9% of customer demand. The plan includes building 165 new delivery stations between 2027 and 2029 and adding roughly 4,500 ZIP codes through its Hub Delivery rural network. The initiative reportedly began under Project Aurelian, a contingency plan created when negotiations with the U.S. Postal Service became uncertain, though Amazon said the projections remain preliminary and could change.
8. The top 10 most active options today by contracts traded were $NVDA with 2.6M contracts, $MU with 1.5M contracts, $TSLA with 1.3M contracts, $AAPL with 1.1M contracts, $SPCX with 786K contracts, $AMZN with 640K contracts, $GOOGL with 637K contracts, $INTC with 582K contracts, $NKE with 561K contracts, and $MSFT with 447K contracts.
9. Oracle $ORCL has reportedly signed a 5-year, $7B AI compute agreement with Tencent, according to the Financial Times. Under the deal, Oracle will provide access to roughly 100,000 advanced AI chips across its Southeast Asia data centers, with about 30% of the contract value paid upfront. Tencent plans to use the capacity to train AI models and expand its portfolio of agentic AI products, as the company accelerates AI infrastructure spending. The investment has already impacted Tencent’s financials, with Q2 infrastructure capex surging 176% YoY to $7.9B and free cash flow turning negative at -$2B for the first time in more than a decade.
10. AI-related companies continue to dominate capital markets, now accounting for roughly 40% of the S&P 500's total market capitalization. In emerging markets, just 3 semiconductor companies make up 28% of the MSCI Emerging Markets Index, while AI-related issuers have represented 49% of all investment-grade bond issuance so far in 2026. The concentration is even more pronounced in private markets, where 87% of year-to-date venture capital funding has gone to AI companies. For comparison, during the 1999 Dot-Com bubble, internet companies accounted for less than 40% of venture capital funding, highlighting the unprecedented concentration of investment flowing into AI today.
11. Since 1928, the S&P 500 has gained an average of 2.1% during the first half of October in U.S. midterm election years, making it the strongest-performing half-month of the fourth quarter. Positive returns occurred 71% of the time over that period, the third-highest hit rate of any half-month in Q4. The seasonal strength follows a historically weak second half of September, which has averaged a 1.1% decline. Momentum has also tended to remain positive later in the quarter, with the second half of October averaging a 0.3% gain and the first half of November delivering an average 1.7% return with a 79% positive hit rate.
12. Bond ETFs have attracted $23B in inflows over the last 5 trading days, accounting for 62% of all ETF net inflows despite representing only about 15% of total ETF assets. Fixed-income funds also made up 6 of the 8 most-purchased ETFs over the period, with investors allocating capital across cash, municipal bonds, high yield, Treasuries, and $TLT. The broad-based demand suggests investors are increasingly positioning around the bond market as yields remain near multi-year highs and volatility persists across asset classes.
WALL STREET IS THE GREATEST SHOW ON EARTH.
rendi CLONE retweeted
Compute capacity is shifting east.
East Asia: Semiconductor and data centre expansion
Southeast Asia: Power + connectivity advantages
India: Scale + cost efficiency
MENA: Capital meets ambition
Infrastructure is growing fast.
But pricing transparency? Still catching up.
Building a more transparent and accessible compute market is exactly what @RaxFinance is working toward.
rendi CLONE retweeted
The listed price is not the traded price.
Public marketplaces show H100 at $3.25/hr.
Real year-long contracts clear at $2.10–2.70.
Why the gap?
Bilateral negotiations are confidential.
Nobody publishes what actually traded.
Price discovery needs transparency.
🔗 Join the Waitlist: rax.finance/
rendi CLONE retweeted
Commodity formation playbook:
Gold → London Fix (1919) → futures market
Oil → posted prices → spot market → Brent/WTI
Electricity → bilateral → pools → day-ahead markets
Compute → ?
We're watching the pattern repeat in real-time.
🔗 Join the Waitlist:
app.rax.finance/waitlist
rendi CLONE retweeted
The debate on AI compute has moved.
From "will demand exist?"
To "do the unit economics work?"
GPU pricing, financing costs, utilization.
Contracted power, density, offtake.
A 1GW factory only works if it can be rented, powered, and paid back.
Some names are being marked down on returns.
Others are being marked up for contracted power, density, and long-term offtake.
The market is underwriting compute like real assets now, not like a theme.
Via @wallstengine
ROTHSCHILD REDBURN LAUNCHES COVERAGE ACROSS AI DATA CENTER NAMES
$NBIS: SELL, $84 PT
Says Nebius has a demanding valuation and questions the sustainability of its unit economics. Upside could come from newer-generation GPUs, faster growth of its Token Factory inference business and stronger capacity execution. Risks include customers bringing compute in-house and higher funding costs.
$CRWV: SELL, $54 PT
Redburn questions CoreWeave's unit economics and ability to convert its pipeline into attractive returns. Key risks include hyperscalers or AI labs bringing capacity in-house, falling GPU pricing and higher financing costs.
$DLR: BUY, $227 PT
Highlights Digital Realty's global wholesale and colocation footprint, long-term tenant contracts and expansion into larger, higher-density facilities built for AI workloads and hyperscale customers.
$EQIX: BUY, $1,261 PT
Points to Equinix's global interconnection ecosystem as a key advantage, with AI, HPC, enterprise and cloud customers increasingly needing high-density compute close to networks and other infrastructure.
$IRM: BUY, $132 PT
Iron Mountain has expanded beyond its legacy records-storage business into data centers, information management and IT asset lifecycle services, while developing more power-dense capacity for AI workloads.
The miner-to-AI names were treated much more cautiously:
$APLD: NEUTRAL, $22 PT
Says unit economics and pipeline-conversion risks are already well priced in. Upside comes from lower build/operating costs and additional large tenant signings. Risks include local opposition, construction delays and tenant insolvency.
$IREN: NEUTRAL, $40 PT
Sees upside from additional large-scale AI tenants and better execution at existing sites. Risks include difficulty securing tenants, higher funding costs and delays or cancellations.
$WULF: NEUTRAL, $15 PT
Potential upside comes from securing additional U.S. capacity and signing more major tenants. Permitting issues and state-level opposition are key risks.
$CIFR: NEUTRAL, $18 PT
Sees better or faster grid-capacity allocation and more major tenant deals as upside. Tenant pullouts and buildout delays are the main downside risks.
$HUT: NEUTRAL, $96 PT
Upside depends on faster data-center execution and securing additional grid capacity. Delays or cancellations in the buildout remain the key risk.
$CORZ: NEUTRAL, $16 PT
Core Scientific has been shifting capacity from Bitcoin mining toward AI/HPC hosting after emerging from bankruptcy, while continuing its mining business. Its proposed acquisition by CoreWeave collapsed last year after shareholders rejected the deal.
rendi CLONE retweeted
BlackRock: AI compute could be tokenized.
Translation:
compute is about to get a market.
The world’s largest asset manager just named the future.
We’re building the unit it needs.
1 cT = 1 GPU-hour.
Fungible. Redeemable. Tradable.
That’s the layer.
Our latest research paper explores the growing connection between AI and digital assets and explains why broad AI adoption may drive new demand, utility and applications across the digital asset economy. blackrock.com/us/individual/…
























