● AI Supercycle, Power Grid Panic
Semiconductors are Just Getting Started: The AI Supercycle Expands to Power, Quantum, and Security
AI has triggered the semiconductor supercycle at its inception, and today’s content focuses on comprehensively summarizing the data center power crisis and energy transition investments, the practical applications of quantum computing and quantum security, the reshuffling of cybersecurity budgets, as well as real risks like shadow financing and power grid bottlenecks.
If you read it now, you’ll have a clear picture of where to invest, what to avoid, and the timing for stepping in gradually.
It naturally connects inflation and interest rate variables within the economic outlook and global stock market trends.
News at a Glance: 7 Key Headlines
- The domino effect of supply shortages created by AI: from GPU → HBM memory → advanced packaging → networking → power and cooling, tightening chain by chain.
- Memory prices and performance signal the very beginning. A semiconductor supercycle never ends in just one month.
- With surging power demand from data centers, investments in base-load power generation (nuclear/LNG), storage (BESS), and power equipment (transformers, switchgears) are booming.
- Scenarios for ultra-large data centers like OpenAI’s ‘Stargate’ are emerging, with power grids and transformers becoming real bottlenecks.
- Quantum accelerates AI: In areas such as drug discovery, materials, and optimization, hybrid quantum+AI approaches are opening commercial pathways ahead of schedule.
- Cybersecurity demands are structural due to the post-quantum (PQC) transition. Security budgets in defense and infrastructure are set to increase.
- Risks lie on two fronts: shadow financing and supply constraints in power grids and packaging. FOMO is a shortcut to losses, so respond with gradual, rule-based approaches.
The Reality of the Semiconductor Supercycle Driven by AI
- The sequence of supply shortages is clear. It started with GPUs, then HBM became constrained, and now the bottlenecks have shifted to advanced packaging (CoWoS, FOPLP), optical transceivers (800G/1.6T), and power/cooling.
- HBM is more than just a simple DRAM. It is a high value-added memory stacked using TSV to increase bandwidth, which requires simultaneous expansion of equipment and packaging capacities. This implies that the cycle may be prolonged.
- As inference gains more weight compared to training, server designs are changing. More memory capacity, CXL-based memory pooling, high-speed networking, and the adoption of liquid cooling are accelerating.
- Performance signals indicate the very early stage structurally. With a rebound in memory ASP, prolonged lead times for packaging, and increased equipment orders, profit consensus for “next year and the year after” is being raised.
- Investment points: HBM suppliers, TSV/bonding/test equipment, advanced packaging substrates (ABF), 800G optical modules, liquid cooling, data center power electronics (SiC/GaN). In this cycle, “packaging, power, and networking” within the semiconductor chain are the differentiators.
Data Center Power Crisis and Energy Transition Investment
- Power is the true cap of this cycle. Overheating emerges not from the chips but from transformers, switchgears, and transmission grid access queues. The average lead time of 2–4 years for grid connection delays the operation of new data centers.
- A scenario involving ultra-large data centers of around 260GW is being discussed in the market. Although these figures are based on estimates and assumptions, the direction is clear. Power demand is vastly exceeding expectations, and tolerance for blackouts is nearing zero.
- The realistic combination for base-load power generation is nuclear and LNG. On top of that, renewable energy (PV, wind) and large-scale ESS are added to mix costs and stability.
- Small Modular Reactors (SMRs) and long-term power purchase agreements (PPAs) for data centers are emerging as new standards. On-campus models that combine heat recovery, absorption cooling, and microgrids are on the rise.
- Key components and processes: ultra-high voltage transformers, circuit breakers, switchgears, power semiconductors (SiC/GaN), liquid cooling racks, batteries (BESS, LFP, sodium-ion), thermal management software, and DC distribution.
- Operational metrics are changing. PUE optimization, increasing electricity capacity, participation in demand response (DR), and the internalization of distributed generation become core KPIs for CAPEX and OPEX.
Quantum and Security: Next-Generation Growth Engines Enter the Main Game
- Quantum computing acts as an “AI accelerator.” In drug design (VQE/quantum simulation), battery and catalyst materials, and logistics optimization, hybrid quantum-classical workflows are increasing the number of commercial PoCs. Although it is not yet universal, it is entering a “utility phase.”
- Post-quantum cryptography (PQC) is driven by regulations. As NIST standard candidates (CRYSTALS-Kyber, Dilithium, etc.) are adopted, large enterprises, financial institutions, and the public sector are likely to mandate migration roadmaps within three to five years.
- QKD (Quantum Key Distribution) is suitable for ultra-secure lines and national backbone networks, while PQC is appropriate for extensive commercial and cloud environments. The two are complementary rather than interchangeable.
- There is a restructuring of security budgets. Defense and critical infrastructure budgets are organized into “physical security + cybersecurity + quantum security,” and along with zero trust, SBOM, and AI threat detection (MDR/XDR), the PQC transition becomes a central element of the budget.
Investment Idea Map: 6–12 Months and 12–24 Months
- 6–12 Months Focus
- Semiconductors: HBM suppliers, TSV/bonding/test equipment, CoWoS and FOPLP packaging, ABF substrates, 800G optical modules, CXL switches/controllers.
- Data Center Infrastructure: Liquid cooling, transformers and switchgears, power semiconductors (SiC/GaN), BESS, microgrid design.
- Security: Zero trust, MDR/XDR, cloud workload security, PQC libraries, HSM.
- 12–24 Months Focus
- Memory roadmap: Transition from HBM3E to HBM4, commercialization of CXL 3.x based memory pooling.
- Nuclear (SMR) value chain and the data center PPA model, renewable plus storage hybrids.
- Quantum: Advances beyond NISQ limits in error correction and logical qubits, pilot commercialization of quantum sensors (radar, geophysical).
Market Strategy: A Simple Rule to Beat FOMO
- In a rising market, “responding rather than predicting” is key. Instead of obsessing over timing, confirm trends and respond flexibly with gradual buying and rebalancing.
- Set rules. By predefining target allocation, loss limits, and intervals for additional purchases, emotional interference is minimized.
- Avoid excessive leverage. With increased volatility, the risk of margin calls escalates sharply.
- If the economic outlook shifts with changes in interest rates and inflation paths, growth stocks become more sensitive in terms of valuation. Focus on areas with performance upgrades (memory, packaging, power equipment) for defense.
System Risk Checklist: What to Monitor from Now On
- Shadow Financing and Credit Risk
- Private equity and non-bank loans, used car and consumer ABS, commercial real estate (CRE) exposures.
- Credit card delinquency rates, bank loan loss provisions, flows of deposits out of regional banks.
- Supply Chain and Policy Risks
- The speed of capacity expansion in advanced packaging, lead times for transformers and switchgears, grid connection waiting lists.
- Export control of technologies, regulations on AI and data centers, regulations on environmental and water usage.
- Energy Prices and Blackout Risks
- LNG price volatility, power supply during heatwaves and cold snaps, region-specific risks related to data centers.
The Most Important Point Not Covered Elsewhere
- The real cap of this cycle is the “power grid.” It is not the chips, but transformers, transmission, and connection permits that determine the pace of AI CAPEX.
- Packaging is the physical bottleneck. Before EUV expansion, CoWoS, FOPLP, bonding, and testing need to be scaled to release the physical products.
- The PQC transition might become a project on the scale of Y2K. An enormous total cost of ownership (TCO) will arise from certification systems, HSMs, and replacing legacy cryptographic stacks.
- The co-location of data centers and SMRs with PPA models is a complex project intertwined with finance, regulation, and thermal management. Only the early movers will achieve economies of scale.
- Liquid cooling is determined by the “component chain.” If bottlenecks occur in pumps, valves, plates, leak detection, or refrigerant materials, rack expansion will halt.
Positioning Guide: Practical Purchasing and Inspection Routine
- Monthly Check: HBM and packaging lead times, hyperscaler CAPEX guidance, transformer order backlogs, bank loan loss provisions.
- Gradual Purchase Principle: If the target allocation is 100, purchase in splits of 30-30-40; add more consistently at each adjustment.
- Sector Basket: Manage with a purpose-based allocation such as 40 in semiconductors (memory, packaging, networking), 30 in power and energy infrastructure, 20 in security and PQC, and 10 in experimental (quantum, sensors).
- Rebalancing: Maintain positions in areas with continuously upward performance, but be cautious if guidance is downgraded or lead times shorten.
< Summary >
- The AI supercycle is expanding from GPUs to include HBM, packaging, networking, power, and cooling, with memory performance recovery being an early signal.
- The surge in data center power demand is driving energy transition investments, where transformers, power electronics, SMRs, and BESS are the key beneficiaries.
- Quantum serves as an accelerator for AI, while cybersecurity is growing structurally through the PQC transition. Defense and infrastructure budgets will increase their share.
- Risks include shadow financing and bottlenecks in power grids and packaging. In a rising market, control FOMO through gradual, rule-based approaches rather than predictions.
- Investment points include HBM, advanced packaging, optical networking, data center power equipment, liquid cooling, PQC, and the SMR-PPA model.
[Related Articles…]
- The Next Winner in the Semiconductor Supercycle: HBM4 and Advanced Packaging
- Data Center Power Crisis and Energy Transition: A Practical Roadmap for SMRs, BESS, and PPAs
*Source: [ 경제 읽어주는 남자(김광석TV) ]
– 반도체는 이제 시작에 불과, 양자·에너지·보안까지 판이 커진다. 슈퍼사이클의 초호황의 서막 | 경읽남과 토론합시다 | 이선엽 대표 3편
● Stablecoin Tsunami, Remittance Revolution
10% of Mexico-US Remittances Are Moving to Stablecoins: A One-Stop Overview of B2B Payment Innovation, Fee Reduction, and Exchange Rate and Regulatory Risks
Today’s Key Points at a Glance
In the field in Mexico, about 10% of US-Mexico remittances are already processed as stablecoins.
This provides a numerical comparison of how various industries and workflows are changing in B2B corporate payments, as well as how fees and settlement times are being reduced.
It includes the impact of macro variables such as exchange rates, interest rates, and inflation, and regulatory risks on adoption speed, plus a checklist for companies preparing to implement this immediately.
It offers a separate section that delves into practical points rarely covered elsewhere, such as “financial/settlement design, on-chain FX hedging, accounting/internal controls, ERP integration, and liquidity pool management.”
On-Site News: Stablecoin Adoption Accelerates in the Mexico-US Corridor
Mexico, which shares a border with the United States, witnesses substantial and frequent remittance and trade flows.
According to discussions at a local conference, about 10% of remittances between the US and Mexico are processed using stablecoins.
Both institutions and fintech companies plan to gradually increase this percentage, as the cost and speed advantages compared to cash-based remittances are evident.
Due to structural changes such as a surge in remote work, increased trade from nearshoring, and diversification of small-scale suppliers, the demand for digital payment infrastructure in Mexico is growing rapidly.
Why Is the Mexico-US Corridor Evolving So Early?
In 2023, Mexico received record-breaking overseas remittances of approximately 63 billion dollars (according to Banxico’s data).
The strong labor and trade links between the US and Mexico, along with frequent remittance occurrences, create rapid network effects.
With supply chains being reorganized due to nearshoring, the number of B2B payments from small and medium-sized manufacturers has surged, and they are seeking methods that are cheaper and settle instantly compared to cards or telegraphic transfers.
Changes Occurring in B2B Payments
Payment for parts and raw material supplies: US headquarters → Mexican subcontractors settled in USDC/USDT, with local settlement in MXN on the next business day.
Cross-border salary/freelancer payments: Transition from bi-monthly to weekly or on-demand payments, allowing more flexibility in labor cost management.
Handling returns, rebates, and limit payments: Process small, frequent transactions via on-chain transfers to reduce bank fees and waiting times.
Trade finance and pre-discounting: Tokenize invoices for early cash conversion (discounting) and automate settlement and netting.
Cost and Speed Comparison: Traditional Remittances vs. Stablecoins
Fees: Traditional cross-border transfers typically charge 2-6% (varying significantly by corridor and method).
Stablecoins often incur network fees under 0.1%, and combined on/off ramps result in rates ranging from 0.3 to 1.0%.
Settlement Speed: The conventional method generally requires T+1 to T+3.
Stablecoins achieve payment confirmation within 1-10 minutes, and even including conversion to local currency (MXN), the process is completed within hours.
Transparency: On-chain tracking and automated reporting significantly enhance the visibility for procurement and finance teams.
Macro Variables: The Impact of Exchange Rates, Interest Rates, and Inflation
Exchange rates: During a strong dollar phase, the timing of converting MXN on the receiving end determines the performance.
Interest Rates: As stablecoin reserves are commonly invested in US Treasuries, the level of US interest rates affects ecosystem liquidity and incentives.
Inflation: The disparity in prices between Mexico and the US influences trade prices and the magnitude of wage remittances, driving demand for on-demand payments.
Supply Chain Reorganization: With nearshoring increasing the number of transactions and partners, stablecoins are highly suitable for small and frequent cross-border payments.
Global Economic Cycle: In economic downturns, the need for fee reduction acts as an adoption catalyst, while in recovery periods, speed and scalability needs drive adoption.
Regulatory, Compliance, and Risk Checks
KYC/AML: It is crucial to verify the identity of on/off ramp operators and comply with the travel rule.
Issuer Risk: Manage risks related to de-pegging or delayed redemption by ensuring reserve transparency and auditing.
Accounting and Taxation: In some jurisdictions, stablecoins may be classified as digital assets rather than equivalents of cash, so establish evaluation and disclosure policies in advance.
Market and Liquidity: Risks include network congestion on specific chains, sudden fee spikes, and a scarcity of local off-ramps.
Policy Environment: As guidelines for stablecoins in both the US and Mexico are evolving, it is safer to work with partners who hold licenses.
‘Core Internal Secret’ Points Rarely Discussed
On-chain FX Hedging Design: Hedge the gap between holding USDC and MXN settlement using NDF (non-deliverable forwards) or through spot/swap transactions to reduce exchange rate slippage.
Treasury Operation Model: Separate pending receivables and payables by on-chain wallet and periodically net them, reducing the number of transfers by 30-50% and cutting network and off-ramp fees.
Liquidity Pool Optimization: Pre-secure MXN inventory from off-ramp partners in anticipation of end-of-month settlement peaks to prevent sudden spread increases.
ERP and Settlement Integration: Map wallet addresses and supplier IDs one-to-one, and automatically attach on-chain transaction hashes to journal entries to enhance audit traceability.
Internal Controls (SOX/Internal Approvals): Use multi-signature and policy-based wallets to align payment approval flows with existing approval lines, addressing concerns that “crypto assets cannot be controlled.”
An Execution Guide for Companies (30-Day Roadmap)
Week 1: Distinguish between payment corridors and amounts, select the chain (considering speed, fees, and ecosystem), and send RFPs to 2-3 potential on/off ramp candidates.
Week 2: Test KYC/AML and sanction filter integrations, and simulate small, high-frequency transactions in a sandbox environment.
Week 3: Integrate the exchange rate API and on-chain payment trigger, and implement an automatic wait and split transfer logic when the target spread is exceeded.
Week 4: Finalize accounting policies (evaluation and disclosure), design multi-signature authority, run pilot transactions (20 for sales and 20 for purchases), and prepare a cost/speed report.
From an Investment Perspective: Who Benefits
Payment Infrastructure: Providers of cross-border on/off ramps, payment APIs, and chain analysis (compliance) solutions.
Latin American Fintech: Networks for remittances and local payments, and logistics/trade finance platforms that benefit from nearshoring.
Banks and Securities Firms: Diversifying fee and deposit income through stablecoin custody, reserve investment in Treasuries, and institutional on-ramps.
Risks Remain: Uncertainty in regulation, de-pegging events, and a narrowing spread in revenue models increase valuation volatility.
Tracking Adoption Speed with Data and Indicators
Monitor monthly remittance trends from Banxico alongside the USD/MXN exchange rate volatility between the US and Mexico.
Track on-chain wallet volumes (Mexico-US corridor), average transaction fees, and off-ramp spreads to gauge actual adoption speed.
By checking chain congestion and settlement delay rates during end-of-month/quarter peaks, bottlenecks in B2B payments can be identified early.
6-12 Month Scenarios
Base Scenario: The stablecoin percentage in B2B small, frequent payments expands to the 15-25% range, and average remittance costs continue to decline.
Positive Scenario: With clear guidelines in the US and Mexico and expanded bank on-ramps, trade finance and salary payments shift decisively.
Negative Scenario: In the event of regulatory shocks or major de-pegging incidents, a temporary retreat may occur, leading to a reorganization centered on key players.
Conclusion: Why This Is Important Now
With increased trading volumes due to nearshoring, companies demand cost reduction, speed, and transparency simultaneously.
Stablecoins uniquely address all three simultaneously, and the Mexico-US corridor is at the forefront of this change.
By managing macro variables such as exchange rates, interest rates, inflation, and regulatory risks, companies can be the first to capture the excess benefits of B2B payment innovation.
< Summary >
About 10% of US-Mexico remittances are processed via stablecoins, as B2B payments undergo a rapid transition.
With fees ranging from 0.3 to 1.0% and settlements completed within hours, stablecoins clearly outperform traditional methods in terms of cost and speed.
By managing risks related to exchange rates, interest rates, inflation, and regulations, immediate efficiency gains can be achieved in nearshored supply chains.
The practical essentials are on-chain FX hedging, treasury netting, ERP/internal controls integration, and off-ramp liquidity management.
[Related Articles…]
Checklist for Exchange Rate Risks in Emerging Markets Amid a Strong Dollar in 2025
Key Points in Supply Chain Reorganization from Mexico Nearshoring
*Source: [ Jun’s economy lab ]
– 멕시코는 이미 이것으로 돈 보낸다
● AI Cash Crunch, Musk Pay Battle, Power Market Panic
OpenAI Funding Signal, Tesla Musk Compensation Plan Shareholders Meeting, Vistra Energy Earnings Miss — Why AI, Electricity, and Governance Are Shaking the Market Right Away
Core Issues Covered in Today’s Article
The OpenAI CFO’s mention of ‘fundraising’ and why it directly relates to AI infrastructure, the electricity market, and the interest rate/dollar environment are summarized.
The impact of Tesla’s Elon Musk compensation plan shareholders meeting on corporate valuation, governance premium, and tech stock valuation is explained in detail.
An analysis of why Vistra Energy’s earnings miss reveals the “electricity betting” risks of the AI era is provided.
A separate look is given at the most crucial link behind the news headlines that “no one talked about.”
From an investor perspective, a one-page sector-by-sector play strategy and risk checklist is summarized.
Issue 1) OpenAI CFO’s Hint at Fundraising — Meaning and Ripple Effects
According to reports, OpenAI’s Chief Financial Officer (CFO) hinted at the need for additional fundraising.
This signals that the expansion of large-scale model training and inference infrastructure, data center electricity and cooling costs, and semiconductor supply bottlenecks remain severe.
The current interest rate level and strong dollar have resulted in an environment where the required return on growth capital is higher, making funding structurally more expensive.
Therefore, aside from traditional equity issuance, OpenAI is likely to combine structured capital such as cloud prepaid credits, long-term power purchase agreements (PPAs), supply chain financing, and vendor financing.
Consequently, the pace of model performance improvement depends on the simultaneous securing of the “trio of computation, power, and cash.”
For semiconductor suppliers, long-term purchase agreements for GPUs, HBM memory, and networking equipment are increasing, raising the risk of dependency on specific vendors.
In the electricity market, the increasing AI load is expected to raise the base demand, leading to a reassessment of wholesale power prices and capacity values.
Issue 2) Tesla Musk Compensation Plan Shareholders Meeting — The Intersection of Governance and Valuation
Tesla’s upcoming shareholders meeting regarding Elon Musk’s compensation plan has come into renewed focus.
The outcome of the compensation plan has significance on two main axes.
First, the governance premium.
The approval of the compensation plan enhances leadership stability, while also introducing variables such as potential dilution and regulatory/litigation risks.
Second, business execution.
Resource allocation for new growth areas such as FSD, robotaxi, energy storage (ESS), and AI computing could be linked to the outcome of the compensation plan.
Across the U.S. stock market, valuation multiples for large tech stocks are highly sensitive to inflation, interest rates, and regulatory frames.
Therefore, the shareholders meeting outcome can be seen as an event that realigns tech stock premiums and expectations for electric vehicles and robotics.
Issue 3) Vistra Energy Earnings Miss — A ‘Reality Check’ on the AI Electricity Theme
News has emerged that Vistra Energy reported earnings below expectations.
The key points are twofold.
First, even as power stocks have overheated in the short term due to AI demand, traditional risks such as hedge positions, maintenance schedules, and weather variations still drive earnings guidance.
Second, even though data center electricity demand is on a long-term upward trend, short-term profits can fluctuate due to spread volatility and mark-to-market effects from accounting.
Investors must continue to monitor conventional indicators such as the ‘electricity price cycle’ and ‘generation mix/maintenance calendar.’
Market Context) Interest Rates, the Dollar, Inflation, and Tech Stocks
The possibility of persistently high interest rates and a strong dollar act as a headwind for growth stocks, but supply constraints in AI infrastructure create excess profits for some companies.
Inflation pushes up electricity rates and equipment prices, altering the profitability equations of projects.
Thus, while tech and power stocks may move together, the timing of profits and volatility are likely to differ significantly.
Investor Checklist) Sector Positioning and Strategy
AI Infrastructure Equipment & Components.
Focus on bottleneck components such as GPUs, HBM, optical networking, and power semiconductors with high earnings visibility.
Monitor announcements of long-term supply agreements, yield improvements, and new product roadmaps.
Electricity & Utilities.
Pay attention to regional operators with increasing data center loads and companies with efficient assets in nuclear and gas turbines.
However, always consider hedge gains/losses, maintenance schedules, and regulatory rate bases.
Platform & Model Companies.
Key points that reduce fundraising burdens include cloud prepaid credits, partnership terms, and whether they develop their own chips.
Electric Vehicles & Robotics.
Even after overcoming the shareholders meeting event risk, regulatory, litigation, and quarterly achievement of product roadmaps must be re-evaluated.
A Point No One Mentioned) The ‘Triple Bond’ of Electricity-Finance-AI
The key is that the growth of AI is determined not by software issues but by ‘electricity and capital costs.’
Electricity rates, capacity rates, and forward contract prices essentially determine the AI model release cycle and service margins.
Therefore, future sources of earnings surprises will be driven not by the number of parameters but by the securing of kilowatt-hours and financing rates.
OpenAI’s hint at fundraising presages the pace of LLM releases, pricing policies, and the reorganization of partner ecosystems.
Vistra Energy’s earnings miss conversely validates this logic.
If power securing is not flawless, the AI rally may not translate into profits that match the momentum.
Data Points and Monitoring Timelines
Fundraising.
Track new investment rounds of major AI companies, power PPA signings, and announcements of long-term chip supply agreements.
Governance.
Monitor the results of Tesla’s shareholders meeting, subsequent legal issues, dilution disclosures, and stock option exercise timelines.
Electricity Market.
Monitor weekly key indicators such as regional wholesale power prices, capacity market bid prices, maintenance schedules, and weather event risks.
Risks and Alternative Scenarios
Rising Interest Rates.
If U.S. long-term interest rates rise, the compression of growth stock multiples may resume.
Strong Dollar.
Companies with a high proportion of overseas sales may face both currency translation impacts and a slowdown in demand.
Policies & Regulations.
Environmental regulations on data centers, semiconductor export controls, and power rate regulations could increase volatility.
Execution Risks.
Delays in power ingress, insufficient substation capacity, and disruptions in chip supply are direct variables that can shake earnings guidance.
Quick News Summary
The OpenAI CFO has hinted at the need for additional fundraising.
This implies that AI infrastructure costs have risen in a high interest rate and strong dollar environment.
Tesla is moving ahead with a shareholders meeting regarding the Musk compensation plan.
The outcome may readjust the governance premium, dilution risks, and resource allocation for new businesses.
Vistra Energy missed earnings expectations.
This shows that despite the AI electricity theme, traditional electricity risks can still shake operating results.
Practical Insights
AI is ultimately an electricity cycle industry.
It is difficult to explain the gap between performance and stock prices without considering equipment, power, and capital together.
Rebalance your portfolio toward bottleneck components, high-quality power assets, and platforms with stable governance.
Keyword Scanner
U.S. Stock Market, Interest Rates, Inflation, Dollar, Semiconductors.
< Summary >
The OpenAI fundraising issue reveals the reality of AI infrastructure costs and power procurement.
Tesla’s Musk compensation plan shareholders meeting is an event that could change governance premiums and the pace of growth investments.
Vistra Energy’s earnings miss reminds us of the short-term volatility and traditional risks in the AI electricity theme.
The investment key is to choose assets with bottlenecks and stable governance on the triangular axis of equipment-power-capital.
[Related Articles…]
Roadmap for Re-rating Tech and Power Stocks After the Peak of Interest Rates
*Source: [ Maeil Business Newspaper ]
– 오픈AI CFO, 자금조달 문제 내비쳐?ㅣ테슬라, 머스크 보상안 주총 개최ㅣ비스트라에너지 예상치 하회 실적ㅣ홍키자의 매일뉴욕
