Almost every app, website, and AI model you use runs on the machines of just three companies — Amazon (AWS), Microsoft (Azure), and Google Cloud. They're the "landlords" who rent out computing power to the whole world, and right now they're racing to invest at a scale business has never seen — nearly $725 billion in a single year — to build AI data centers fast enough to catch the biggest wave in tech history. This is the story of the oligopoly that props up the entire digital economy and all of AI.
Contains
Theme index· base 100 · USD total return
No index history for this theme yet.
News & notes movingMega-cap Hyperscalers
Mega-cap Hyperscalers
IREN Fair Value Trimmed to US$79.03 as AI Cloud Wins Meet Build Out Risks
IREN's fair value estimate has been trimmed from US$80.93 to US$79.03, a modest adjustment reflecting a mix of growing AI cloud contracts and ongoing questions around execution and supply build out through 2027. The revision lifts revenue growth assumptions to 168.01% from 125.79% and net profit margin expectations to 11.73% from 5.79%, while the future P/E assumption drops to 31.26x from 90.29x and the discount rate rises to 9.41% from 8.83%. On the bullish side, JPMorgan shifted IREN from Underweight to Overweight with a US$65 price target, citing a partnership with Nvidia, new customer signings and higher industry pricing, while H.C. Wainwright raised its target to US$90 after IREN announced US$2.8b of new multi year AI cloud contracts and lifted its 2026 AI cloud annual recurring revenue target to more than US$4b. Northland initiated IREN at Outperform with a US$99 price target, and Freedom Capital upgraded the stock from Hold to Buy. Freedom Capital nonetheless flagged execution on supply additions over the next two years as a central risk, noting the story depends heavily on adding capacity on schedule.
Nvidia-Backed Nscale Files for US IPO After $1.02 Billion Loss
Nscale, the London-based AI data center developer backed by Nvidia and Microsoft, filed publicly for an initial public offering in New York, seeking to raise as much as $3 billion. The company reported a net loss of $1.02 billion on revenue of $140.6 million for the six months ended June 30, compared with a net loss of $368.9 million on revenue of $10.4 million a year earlier, according to its Friday filing with the US Securities and Exchange Commission. Spun off from a cryptocurrency mining operation in early 2024, Nscale was valued at about $14.6 billion in a March Series C round led by Aker ASA and 8090 Industries, with Nvidia and Nokia Oyj also participating. The company has agreed to add more than 30,000 Nvidia chips to an existing rental agreement with Microsoft at its Narvik, Norway gigafactory, and Anthropic has agreed to spend $45 billion to rent AI cloud computing power from its flagship West Virginia data center development. The offering is being led by Goldman Sachs, JPMorgan Chase and Morgan Stanley, with shares expected to trade on the New York Stock Exchange under the symbol NSCL.
Tigress Analyst Raises Alphabet Price Target to Street-High $485
Tigress Financial Partners analyst Ivan Feinseth reiterated a Strong Buy rating on Alphabet and raised his price target to a Street-high $485 from $415, implying roughly 39.6% upside. The bullish thesis centers on Alphabet turning its massive AI investments into growth across Search, Cloud and Gemini, with Feinseth arguing AI Overviews and AI Mode can make Google more useful for complex, high-intent queries. Recent results support that argument: Alphabet's second-quarter revenue rose 24% from a year earlier to $119.8 billion, Search and Other revenue increased 17%, YouTube advertising grew 13%, and operating income climbed 30% to $40.77 billion. Google Cloud revenue surged 82% to $24.8 billion, operating income more than tripled to $8.8 billion, margins reached 35.6%, and its $514 billion backlog gives visibility into future enterprise demand. Gemini has reached 950 million monthly active users, more than 9 million monthly developers and roughly 22 billion API tokens processed per minute, and Feinseth sees monetization opportunities across subscriptions, APIs, enterprise software, agents and productivity products, though the risk remains that Alphabet must prove its enormous AI spending generates sufficient returns.
CleanSpark Data Center Bond Draws $10 Billion in Orders for $2.28 Billion Sale
CleanSpark Inc.'s debut junk-bond offering for a data center tied to Meta Platforms Inc. drew orders of about $10 billion, more than four times the deal's size, according to people familiar with the matter. The five-year notes were set at almost $2.28 billion and launched at 98.5 cents on the dollar to yield 8.25%, roughly 1.75 percentage points above the average for BB rated firms tracked by Bloomberg. The fundraising, led by Morgan Stanley, will help finance construction of a data center in Sandersville, Georgia, that has been fully leased to Anviran LLC, a Meta subsidiary, under a $6.6 billion, 20-year contract. The facility is expected to begin operations in the fourth quarter of 2027, with Meta guaranteeing rent and operating expenses. Data center developers have now sold more than $3 billion of high-yield bonds so far this year, most backed by long-term leases with hyperscalers including Oracle Corp. and Amazon.com Inc.
Manus Seeks $500 Million Round as Asia's AI Agent Market Splinters
Manus is in advanced discussions for a $500 million funding round at a $4 billion valuation, according to Bloomberg, following Beijing's NDRC blocking of Meta's $2 billion acquisition of the firm in April 2026. Co-founders Xiao Hong and Ji Yichao were summoned and barred from leaving China, forcing the company to pivot from a potential US acquisition target into a cornerstone of the domestic Chinese AI ecosystem, though the round is not yet closed and terms remain subject to change. In Seoul, Enhans, which raised a $38 million Series C confirmed in a September 17, 2026 press release, is embedding its AgentOS into Korean manufacturing and finance, with the round co-led by TIMEFOLIO Asset Management and Stonebridge Ventures and strategic participation from POSCO Investment, LG CNS, and Lotte Ventures, whose chaebol groups already run the technology in live production. Huawei, meanwhile, launched its AI Cluster Service and agent-specific tools including Agentic MaaS and the AgentArts platform at HUAWEI CONNECT 2026, with Dr. Peter Zhou emphasizing that infrastructure must evolve to support secure, reliable agent workloads in production; the Agentic Infra paradigm already serves over 3,500 customers, and while AICS is available in China starting September 30, global availability is not slated until November 30. Together these moves signal that the agent market is splitting along regional lines of data structuring, compute hosting, and regulatory oversight, leaving open whether these silos will converge through open standards or harden into permanent, incompatible zones.
Meta Raises 2026 Capex Guidance to $130 Billion to $145 Billion
Meta Platforms raised its full-year 2026 capital expenditure guidance to a range of $130 billion to $145 billion, including principal payments on finance leases, narrowed from a prior $125 billion to $145 billion range in its Q2 2026 report on July 29, 2026. The forward guidance nearly doubles Meta's full-year 2025 capex of $72.215 billion, and Q2 capital expenditures alone reached $31.1 billion, driven by servers, data centers, and network infrastructure. To fund the build, Meta ended Q2 with $90.3 billion in cash and marketable securities and $83.7 billion in debt, and announced a strategic venture with BlackRock to develop a one gigawatt data center in El Paso, Texas. CFO Susan Li said Meta is demand constrained today, and CEO Mark Zuckerberg said the company is receiving quite a number of offers at a meaningful premium over what we paid for the compute, framing direct compute sales as one leg of a portfolio that also includes APIs, business agents, productivity tools, and subscriptions. The strain is visible in the quarterly numbers: Q2 free cash flow was $784 million, down 91.31% year over year, and operating margin compressed to 31% from 43%, even as Q2 revenue reached $60.801 billion, up 27.96% year over year and above the $60.286 billion consensus, with advertising revenue of $59.4 billion, up 27%.
Amazon AWS Revenue Hits $42.23B, Up 37% in Fastest Growth in 18 Quarters
Amazon Web Services posted $42.23 billion in revenue, growing 37% year over year, the segment's fastest growth in 18 quarters, with a 39.4% operating margin and a $496 billion contracted backlog. Amazon's chips and AI businesses each eclipsed run rates of more than $25 billion in the second quarter, both growing at triple-digit percentages year over year, while 98% of Amazon's top 1,000 EC2 customers use Graviton and Anthropic and OpenAI have made multi-year, multi-gigawatt commitments to Trainium. Q2 operating income landed at $27.46 billion, up 43.2% year over year, and advertising is a $70 billion-plus trailing-twelve-month business growing 26%. Capex reached $54.21 billion in a single quarter, up 68.4% year over year, and free cash flow swung to negative $7.6 billion on a trailing-twelve-month basis, though most AI capacity is contracted for at least five-year terms. Since Amazon reported Q2 on July 30, 2026, AMZN moved from $230.08 to $251.19, while SPY went from $747.03 to $762.70 and QQQ went from $687.99 to $716.92. Andy Jassy said AWS could become a few hundred billion dollar revenue business and now believes it will be at least double that and very possibly a trillion dollar annual revenue business in time.
CoreWeave exited the second quarter of 2026 with $104.2 billion in revenue backlog, up 246% year over year, a figure that excludes more than $25 billion of net new customer commitments added in the early weeks of the third quarter. Management said more than half of the existing backlog was already tied to contracts where customer delivery had begun, and expects that proportion to exceed two-thirds of the second-quarter backlog by year-end. Quarterly revenues surged $2.6 billion, up 112% year over year, while adjusted operating income rose sequentially to $128 million, prompting the company to raise its 2026 revenue guidance to $12.4-$13.2 billion and adjusted operating income guidance to $960 million to $1.15 billion. CoreWeave also lifted its expected year-end annualized run-rate revenues to $18.5-$19.5 billion, citing the long-term nature and attractive margins of contracted backlog. Supporting that backlog is capital-intensive buildout: capex was $9.4 billion, construction in progress rose sequentially to $11.9 billion, operating expenses reached $2.6 billion including $165 million in stock-based compensation, and interest expense hit $640 million versus $267 million a year earlier, with interest expense expected to climb to $860-$940 million in the current quarter. At the end of the second quarter CoreWeave had 1.5 gigawatts of active power after adding nearly 500 megawatts during the quarter, plus 3.7 gigawatts of contracted power, which management said has since risen to 4.2 gigawatts toward a goal of at least 8 gigawatts by 2030, while competition intensifies from Nebius and Microsoft.
Expanse Raises $5.3 Million Seed Funding for AI Resource Allocation Software
Expanse has secured $5.3 million in seed funding to advance its AI infrastructure software, which predicts the exact computational resources needed before AI tasks are executed to minimize over-allocation and optimize GPU, CPU, and memory usage. The company says the platform addresses critical inefficiencies, as industry estimates indicate significant percentages of cloud expenditures are wasted on underutilized resources. Elsewhere in the market, Super Micro Computer rose 9.5% to close at $40.35, while ServiceTitan fell 7.5% to $54.25, near its 52-week low. Oracle finished at $150.59, up 5.2%, after announcing Java 27 advancements and a partnership with Quorum Health to enhance AI-powered innovation in healthcare. Microsoft settled at $497.75, up 1.5%, after announcing strategic alliances with Marvell and Nokia to enhance cloud-native payment security and AI-driven networks, and Alphabet settled at $347.33, up 1.3%.
Arista Networks Raises Full-Year Revenue Outlook on Strong AI Data Center Demand
Arista Networks raised its full-year revenue outlook, citing continued robust demand for AI data center networking infrastructure. The company, which trades on the NYSE under ANET, said the higher guidance reflects management's expectation of a bigger top-line number for the current financial year. Management highlighted Arista's role in AI data center buildouts, and the company's AI data center focus sits within a much broader communications infrastructure footprint spanning data centers, AI clusters, campuses, and routing setups across the Americas, EMEA, and Asia-Pacific. Investors are watching whether upcoming quarters keep revenue above the US$3b mark and show continued strength in AI related switching and routing orders, along with how much of that business comes from a small group of hyperscale customers, since concentration there could still swing results either way. The update leaves intact existing risks around customer concentration and tougher competition from Cisco and NVIDIA.
Nvidia CEO Huang Says AI Spending Flywheel Is 'Really, Really Flying'
Nvidia CEO Jensen Huang said demand for artificial intelligence computing is accelerating, describing the resulting investment cycle as a flywheel that is "really, really flying." Speaking with CNBC's Jim Cramer at Dreamforce in San Francisco, Huang estimated that a 1-gigawatt Nvidia AI factory costs about $50 billion to $60 billion to build but can generate roughly $50 billion in annual rental revenue, implying a return on invested capital of about one year. He said token generation has increased 25-fold in less than a year, with open models now accounting for nearly 70% of generated tokens, up from about 30%. Huang also said Nvidia Grace Blackwell systems now rent for about $16 per GPU hour, versus roughly $5 under some contracts signed a year ago, and he expects AI testing to become a third major category of infrastructure alongside training and inference, requiring additional data centers.
Yuanta Says AI Capex Still Expanding, Recommends 11 Thai Stocks and 8 DRs to Benefit
Yuanta Securities said AI Capex investment is likely to keep expanding over the next one to two years. Although elevated US bond yields will remain a drag, growth in Enterprise AI that is driving cloud and compute demand will offset it. Returns on GPU leasing investment remain high, with payback within 2.2 to 3.1 years, and hyperscalers have the financial strength to invest at least another 2 trillion US dollars in AI infrastructure. The Data Center industry in the APAC region is set to grow at a 25.2% CAGR in 2025 to 2030, the second fastest in the world after North America at 26.8% CAGR, and excluding China it can still grow at 22.9% CAGR. Thailand stands to benefit from its infrastructure readiness and geopolitical neutrality, while the Thai economy is starting to see positive effects through AI-related exports and continued growth in FDI. Thai stocks that Yuanta sees as likely to outperform the market under this theme are divided into the AI Export group, namely DELTA, HANA, SMT and EPG, and the AI and Data Center Deployment group, namely GULF, GUNKUL, AMATA, WHA, STECON, BBL and SCB, for a total of 11 stocks. Foreign companies that stand to benefit include NVDA19, GOOGL19, AMZN19, LITE80, MRVL80, BE80, JPMUS19 and GSUS06, for a total of 8 DRs.
Marvell Deepens AI Infrastructure Push With GlobalFoundries, Microsoft, Utimaco Deals
Marvell Technology announced new collaborations with GlobalFoundries, Microsoft and Utimaco focused on advanced connectivity and payment security. The GlobalFoundries partnership targets expanded production of advanced SiGe chips used in next generation optical links for AI data centers, giving Marvell more assured capacity for SiGe chips used in pluggable optics, NPO and CPO and tightening a manufacturing relationship that already spans more than a decade. Marvell is also working with Microsoft and Utimaco to provide a cloud native payment security platform delivered through Microsoft Azure, combining LiquidSecurity hardware with Utimaco software to shift part of its security hardware footprint into a cloud delivered, usage based service. The first checkpoint for that effort is the Azure Payment HSM v2 public preview, which begins on 17 September 2026 in Western U.S. and Western Europe. Investors can then watch for disclosed customer adoption metrics, expansion of availability regions, and Marvell's capacity plans for higher speed SiGe optical products beyond the current 200G per lane benchmark.
Generac to Supply Amazon Data Centers With Up to US$8.00 Billion in Backup Generators
Generac Holdings announced a long-term agreement to supply Amazon with industrial backup generators for its global data centers, with initial deliveries of about US$2.40 billion in 2027–2028 and total potential purchases up to US$8.00 billion. The deal also includes purchase-linked warrants for Amazon to buy Generac shares, tying the two companies together through equity incentives. The arrangement locks in critical power resilience for Amazon's expanding AI and cloud infrastructure, though it does not change the near-term catalyst of AWS execution or the risk around rising capital intensity and debt-funded infrastructure. Amazon's narrative projects US$1152.4 billion in revenue and US$158.3 billion in earnings by 2029, requiring 14.1% yearly revenue growth and a US$23.0 billion earnings increase from US$135.3 billion today, with a fair value estimate of US$327.00 implying 30% upside.
Google Rethink ROI Summit Highlights Gap Between AI Discovery and Payment Readiness
Google's Rethink ROI summit in New York spotlighted a widening gap between rapid AI-driven shopping discovery and merchant payment systems that cannot yet complete agent-initiated transactions. The platform's new conversational attributes let retailers feed structured data, including FAQ, compatible accessories, and substitutes, directly into the Shopping Graph, and in early testing with Lululemon, brand-supplied attributes were incorporated into AI Mode recommendations 50% of the time. Demand for such tools is backed by Adobe Analytics data showing an 805% year-over-year increase in AI-driven traffic to retail sites during Black Friday 2025, while Salesforce Cyber Week data confirms retailers with integrated AI agents grew sales 32% faster than those without, and Adobe found those shoppers are 38% more likely to convert. That traffic, however, is colliding with a payment readiness deficit: the TLT LLP Retail Agility survey of the top 100 UK retailers found that while 49% are investing in agentic AI, only 15% say their payment systems are prepared for agent-initiated transactions. Google is attempting to standardize the transaction layer through the Universal Commerce Protocol and the Universal Cart framework, which aim to enable persistent, cross-platform shopping carts across Search, YouTube, and Gmail, and with a 10-week horizon to Black Friday 2026, the next two months will determine which brands move from discovery-focused AI to transaction-ready agentic commerce.
Oracle Expands Restructuring to $2.8 Billion as AI Backlog Hits $664 Billion
Oracle expanded its fiscal-2026 restructuring plan to roughly $2.8 billion as the company pours more capital into AI infrastructure, Reuters reported. The shares rallied about 5.4% to $150.90 Thursday morning, based on the latest market snapshot, versus the earlier $148.83 quote. That $2.8 billion restructuring bill works out to roughly 14.5% of Oracle's latest $19.35 billion in quarterly revenue and covers severance, contract exits and other restructuring costs, while free cash flow is running at approximately negative $5.4 billion. Remaining performance obligations have climbed to $664 billion after Oracle signed more than $30 billion of additional AI-cloud contracts, with roughly half of that backlog expected to turn into revenue within 36 months. Management says customer prepayments and customer-supplied chips can soften the capital burden, but the real test is execution.
Qualcomm Gives Amazon Warrants Tied to $60 Billion AI Chip Deal
Qualcomm has given Amazon an equity-linked incentive to deepen their AI-infrastructure partnership, with Amazon able to eventually purchase as much as $60 billion of Qualcomm data-center products while purchase-linked warrants could hand Amazon roughly $4 billion of Qualcomm shares at $161.26 each. Qualcomm shares gained approximately 2.1% to $188.64 Thursday, a level 7.37% above the stock's $175.69 GF Value. The warrant value represents roughly 6.7% of the maximum purchasing framework, directly tying Amazon's buying activity to Qualcomm's equity story. The two companies are also working together on custom AI inference silicon and optical connectivity capable of reaching 1.6 terabits per second. No minimum purchase commitment, delivery timetable or margin profile has been disclosed.
Alphabet Opens Google Home to Rival AI Agents via $20 Monthly Plan
Alphabet opened its Google Home smart-home ecosystem to rival AI agents, turning the platform into a broader distribution channel rather than a Gemini-only service. The company's Home MCP server now lets compatible third-party agents such as Claude and OpenClaw inspect connected devices, track their status, execute commands and work with historical household events. The service sits behind Google's Premium Advanced plan, priced at $20 per month or $200 annually, and also requires a Google Cloud project. Google has placed limits around higher-risk actions such as unlocking doors, though opening the system to outside agents introduces a new security variable. Alphabet shares gained approximately 1.5% to $347.91 Thursday, with the stock trading 35.37% above its GF Value estimate of $257.01.
Iren CEO Touts AI Compute Demand as $684 Million Loss Weighs on Shares
Iren Limited CEO Daniel Roberts argued that demand for AI compute remains structurally constrained by limited supply, even as the company's shares fell nearly 5% on September 14 and dropped 12% after it reported a $684 million net loss in its fourth quarter FY26 results. Roberts cited Anthropic CEO Dario Amodei's comment that his company was operating at an 80x pace versus a planned 10x, OpenAI President Greg Brockman's remarks on continued compute constraints, Google's sevenfold increase in processing volume from a year ago, and Nvidia's guidance for roughly 70% revenue growth while describing its own outlook as supply-constrained. On the supply side, he pointed to high-bandwidth memory shortages and Goldman Sachs data suggesting only about half of scheduled U.S. data center capacity will be built on time. The AI pivot is being supported by $6.4 billion in GPU financing, including $3.6 billion of investment-grade financing for the Microsoft contract, which together with customer prepayments funds 96% of the associated GPU capex, though the structure creates meaningful customer concentration risk. Iren reported $4 billion of contracted ARR for 2026 capacity while only about $1 billion of ARR was operating as of August 26, and hedge fund ownership rose to 69 funds at the end of the second quarter of 2026 from 53 at the end of the first quarter, while short interest stood at 24.46% of float as of August 31, 2026.
Microsoft Launches Azure Payment HSM v2 for Banks as Shares Rise 1.6%
Microsoft pushed another high-security financial workload deeper into Azure Thursday with a managed payment-security service built for banks and payment processors, sending the shares up approximately 1.6% to $498.17. The offering, Azure Payment HSM v2, combines Marvell's LiquidSecurity hardware with Utimaco's Atalla payment module, giving financial institutions a cloud-based way to protect payment keys, encryption and transaction security without giving up the controls demanded by regulated environments. The public preview is available in West US and Western Europe, and Microsoft did not disclose pricing, customer contracts or expected revenue, so the near-term financial impact remains difficult to quantify. The bigger story is where Microsoft keeps pushing Azure: Microsoft Cloud generated $59.3 billion, or roughly 65.9% of the company's latest quarterly revenue, and payment security opens another sticky workload where compliance requirements can make switching painful. The preview alone will not move Microsoft's financial needle, but successful production adoption could deepen Azure's grip on highly regulated customers and create another layer of recurring cloud spending.
Oracle Trades at $147 Despite $664 Billion AI Backlog, 24/7 Wall St. Sees 50% Upside
Oracle shares trade at $147.17 even after a fiscal Q1 2027 earnings report that saw revenue hit $19.345 billion, up 29.6% year over year, non-GAAP EPS of $1.92 against a $1.74 estimate, and cloud infrastructure growth of 121%. Remaining performance obligations swelled to $664 billion, with more than $30 billion in new AI contracts booked in the quarter alone, yet the stock has fallen 8.95% over the past week as the market focuses on negative $5.396 billion quarterly free cash flow and a $28.499 billion capex bill. 24/7 Wall St. maintains a buy rating and a $214.69 price target, implying roughly 50% upside, with 90% confidence, citing a forward P/E of 17 that is cheaper than Microsoft at 26 and Alphabet at 22. The bear case centers on FY26 free cash flow of negative $23.686 billion, interest expense up 55% to $1.4 billion, and full-year FY27 capex guidance of $90 to $95 billion, which could force another equity raise beyond the completed $20 billion ATM program. Half of the $664 billion RPO is expected to convert to revenue over the next 36 months, and management guided full fiscal 2027 revenue to at least $90 billion with non-GAAP EPS of $8.10.
Astera Labs Expands Leo Memory Controllers for AI and Cloud Workloads
Astera Labs unveiled new Leo Smart Memory Controller solutions, including the Leo X-Series and Leo 2 E & P-Series, targeting AI and cloud workloads. The expanded Leo lineup is built for large-scale AI uses such as KV-cache and agentic AI, aiming to improve memory efficiency for hyperscalers. Astera Labs is collaborating with major CPU, GPU, and memory vendors and is showcasing the Leo family at the AI Infra Summit. The company designs connectivity chips that link CPUs, GPUs, and memory inside cloud and AI data centers, and its Leo controllers plug directly into the core plumbing that lets large AI workloads access high volumes of data efficiently.
Nvidia CEO Huang Says Chip Sales Will Double in 2027 on Vera Rubin Ramp
Nvidia CEO Jensen Huang confirmed at a Scotland tech gathering that the company's chip sales in 2027 are projected to roughly double from 2026, echoing management's August 26, 2026 earnings call guidance for about 70% revenue growth in fiscal 2028. The doubling is being driven by the Vera Rubin platform, which entered full production alongside continued Blackwell deployment and is expected to be the fastest product ramp in Nvidia's history, with per-gigawatt revenue opportunity rising to roughly $40 billion from $25 billion on Blackwell. Nvidia's supply obligations swelled to $279.0 billion, largely tied to memory procurement for Vera Rubin, while guarantee obligations for AI cloud partners are capped at $108.5 billion and financing platforms with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR are expected to mobilize over $500 billion of third-party capital for AI infrastructure. Sovereign and enterprise AI is growing 100% a year, with sovereign AI revenue up 35% sequentially and more than tripling year over year, and neocloud partners expected to exit the year with eight gigawatts of installed capacity versus roughly three gigawatts at the end of 2025. Consensus EPS for the fiscal year ending January 2028 has climbed from $12.67 ninety days ago to $15.57 today, with 39 upward revisions in the past 30 days and zero cuts, on revenue now pegged at roughly $678 billion.
Amazon CEO Andy Jassy Predicts AWS Could Reach $1 Trillion in Annual Revenue
Amazon chief Andy Jassy said on the company's recent earnings call that he now believes Amazon Web Services could become at least a few hundred billion-dollar revenue business, and very possibly a trillion-dollar annual revenue business in time. The prediction stands out against AWS' current $169 billion annual revenue run rate, and against the $58 million annual revenue run rate AWS reached three years after launch. AWS' AI revenue run rate, roughly three years into the AI boom, now totals more than $25 billion. Amazon said it plans to spend $220 billion this year, up from an earlier forecast of $200 billion, and that even at that level it won't be able to fulfill all of the demand this year and next year. The article was originally published by The Motley Fool.
Marvell Data Center Revenue Hits Record $2.17 Billion, 79% of Total
Marvell Technology's data center business reached a record $2.17 billion in the second quarter of fiscal 2027, up 46% year over year and 18% sequentially, and accounted for 79% of total company revenues. The data center end market made up 74% of Marvell's fiscal 2026 revenues, and the company expects data center revenues to grow approximately 60% in fiscal 2027. Marvell reported strong demand for 800G optical DSPs while its 1.6T products ramp rapidly, and scale-out switching is expected to more than double in fiscal 2027 on broader adoption of its 51.2T switch products. The company expects its custom business to accelerate in the second half of fiscal 2027 and more than double year over year in fiscal 2028, driven by XPU and XPU-attach solutions, with an expanded hyperscaler agreement covering AI inference accelerators, storage controllers, network interface controllers, memory interface controllers and near-memory compute. Marvell also expanded its position in next-generation AI infrastructure through its relationship with NVIDIA, including NVLink Fusion, scale-up switching and silicon photonics, and introduced products such as the 102.4 Tbps Teralynx T100 switch. Marvell shares have gained 170.3% year to date, and the Zacks Consensus Estimate for fiscal 2027 earnings suggests year-over-year growth of 48%.
Crux AI Secures $22 Billion Debt Financing to Buy Google TPUs
Crux AI, the cloud venture backed by Google and Blackstone, has secured $22 billion in debt financing to buy Google-developed Tensor Processing Units, according to Bloomberg. Ten banks are reportedly providing the financing, including Goldman Sachs, Barclays, BNP Paribas and Bank of Nova Scotia, with the debt backed partly by the value of Google's TPUs and partly by customer contracts signed by Crux AI. Blackstone has separately committed an initial $5 billion in equity. Crux AI plans to bring its first 500 megawatts of data-center capacity online in 2027, a test of whether customers will adopt Google's chips on a much larger scale outside Google's own ecosystem. The $22 billion solves part of the funding problem, and investors will be watching how quickly Crux AI fills that first 500 megawatts.
Buffett Warns on Market Risk as Berkshire Builds $38 Billion Alphabet Stake
Warren Buffett told CNBC that markets are in a gambling mood and that prices for many assets will look very silly, while confirming he initiated Berkshire Hathaway's aggressive buying of Alphabet stock, a position now worth about $38 billion and representing roughly 12.6% of its public equity holdings. Buffett said the decision was made by Greg Abel, and when asked why he was comfortable buying Alphabet over other hyperscalers spending heavily on AI-related capital expenditures, he said he did not want to knock the others because they have no choice, adding that in many cases they are playing a game they do not want to play. Bond markets are pricing in default risk for Oracle, with a mid-market credit default swap spread of 192 basis points implying a 3.2% one-year default probability and 14.8% cumulatively over five years, after S&P Global Ratings downgraded Oracle debt from BBB to BBB-, its lowest investment-grade rating, warning that an industry downturn would hit Oracle worse than other hyperscalers. The takeaway is that any significant market-led weakness tied to weaker hyperscalers could be a buying opportunity in higher-quality names like Alphabet.
Vitality and Google Bring AI Health Platform Vitality AI to the United States
Vitality, in partnership with Google, announced the U.S. expansion of Vitality AI, an AI-driven health platform built on Google Cloud and Gemini Enterprise that delivers personalized health recommendations, coaching, and incentives to health plans and employers. The platform draws on more than 2,800 unique health and behavioral dimensions and generates personalized insights at a 99.2% accuracy rate based on physicians' review. Vitality said engaged clients experience a 4% reduction in healthcare claims costs on average, a 180% return on investment, and an average recovery of 4.4 productive days per employee per year, equivalent to annual savings of $1,047 per employee. Internationally, members receiving personalized AI-driven recommendations were 3.3 times more likely to complete a mental wellbeing assessment, 2.7 times more likely to complete an online health review, and 1.4 times more likely to complete a colorectal cancer screening, while recommending and incentivizing cancer screenings led to an estimated 23,302 additional screenings in 2025, resulting in 219 early detections and an estimated six deaths prevented. Vitality AI is now live across Vitality and Discovery and in the UK, South Africa, and the United States, and since June 2025 more than 210,000 people in the U.S. have received an AI-driven personalized health action, with 54% completing at least one AI-recommended health action. Emile Stipp, CEO of Vitality AI, said the partnership with Google has shown that AI's impact on health care is a proven differentiator, and Maureen Costello, Vice President, UK, Ireland and Sub-Saharan Africa at Google Cloud, said the expansion marks the next step in the partnership with a goal of extending life expectancy and reducing healthcare-associated costs across the U.S.
Anthropic chief executive Dario Amodei published a blog post calling for a slowdown in the pace of artificial intelligence development, warning the technology is progressing so fast that control could be lost, and citing recent cybersecurity events as evidence. The warning puts pressure on SpaceX, which in May signed a deal to rent computing capacity to Anthropic that is expected to bring in $1.25 billion per month through May 2029, worth roughly $45 billion overall. SpaceX has also agreed to lease approximately $920 million in capacity to Alphabet from October 2026 to June 2029, and roughly $150 million worth of capacity per month to start-up Reflection AI through 2029. SpaceX chief financial officer Bret Johnsen has said the company's AI business could reach a $100 billion annual revenue run rate by the end of 2026, and Elon Musk has said the AI segment could propel SpaceX to $1 trillion in annual revenue as soon as 2030. However, most customers can end their infrastructure agreements with 90 days' notice, so if Anthropic needs less capacity because of its plan to pace AI development, it could exit its contract far earlier than anticipated. SpaceX generated $7.8 billion in total revenue in the second quarter of 2026, up 92% year over year, with $2.561 billion from AI, $4.291 billion from connectivity, and $0.962 billion from space. Based on trailing 12-month revenue of $23 billion, SpaceX trades at a price-to-sales ratio of 87.3, 14 times the Nasdaq-100's 6.1, and Wall Street expects total revenue of $106.5 billion in 2027, a forward price-to-sales ratio of 18.8. SpaceX stock is already down 34% from its all-time high.
Alphabet's Cloud Backlog Hits $514 Billion as Google Cloud Growth Outpaces Rivals
Alphabet's Google Cloud revenue grew 82% to $24.8 billion in the second quarter, with backlog jumping more than $50 billion sequentially to $514 billion, a figure larger than Alphabet's entire fiscal 2025 revenue base. The company's Q2 2026 results showed EPS of $9.11 against a $3.04 consensus, revenue of $119.8 billion up 24.23% year over year, and operating margin expanding to 34%, while cloud growth accelerated from 48% in Q4 2025 to 63% in Q1 2026 and 82% in Q2 2026. 24/7 Wall St. maintains a buy recommendation on Alphabet with a price target of $504.76, implying 50.47% upside from a current price of $335.45, and a model confidence of 90%. The bull case targets $569.14, citing that nearly 90% of the Fortune 100 use Gemini Enterprise, Gemini App has 950 million monthly active users, and model APIs process 22 billion tokens per minute, up from 16 billion a quarter ago. Risks include Q2 capex of $44.9 billion doubling year over year, free cash flow swinging negative to -$5.86 billion, long-term debt roughly doubling from $46.5 billion to $98.2 billion, and suspended buybacks, though cloud operating income more than tripled to $8.8 billion with margins climbing to 35.6%.
Generac surges on $8 billion Amazon data center generator deal
Generac Holdings surged 33.7% in premarket trading after announcing a long-term agreement with Amazon to supply industrial backup generators for the technology giant's global data centers. The agreement calls for initial deliveries worth about $2.4 billion across 2027 and 2028, while the total value of the arrangement could reach as much as $8 billion. Amazon will also receive a warrant to purchase about 1.69 million Generac shares at an exercise price of approximately $200.93 per share, representing nearly 3% of the company's shares outstanding. Nebius Group NV rose 11.1% in pre-open trading after notifying customers of broad-based price increases across its on-demand GPU cloud services effective Oct. 1, with prices rising approximately 17% for H100 instances, roughly 21% for the latest Nvidia B300 GPU, and 25% for AMD EPYC Genoa CPU rates. BCB Bancorp fell 4.9% in premarket trading after pricing an underwritten public offering of 11 million common shares for gross proceeds of $85.25 million and disclosing it is marketing approximately $210 million of problem loans, while expecting a net loss of between $126.2 million and $136.1 million for the third quarter of 2026.
Atthawit Urges Energy Ministry to Raise Data Center Power Rates, Fears MEA and PEA Will Miss 9 Billion Baht Revenue Target
Atthawit Suwanpakdee, an MP of the United Thai Nation Party, is calling on the government to quickly raise electricity rates for Data Centers and to tighten regulatory oversight, saying that failing to raise Data Center power rates affects revenue remitted to the government and risks leaving the Metropolitan Electricity Authority, or MEA, and the Provincial Electricity Authority, or PEA, unable to bear their financial burden, which would in turn affect the country's revenue projections and annual expenditure budget. Atthawit said the budget burden stems from two main government electricity subsidy programs totaling 30 billion baht: a program subsidizing the first 200 units of electricity at 3 baht per unit, which uses 12 billion baht, and a program removing public lighting or street lighting electricity charges from people's bills, which uses another 18 billion baht. As of August 31, 2026, figures show that PEA's revenue remitted to the treasury stood at 16.011 billion baht, while MEA's stood at 5.104 billion baht, bringing the combined remittances of the two agencies over the 11-month period to about 21 billion baht, still short of the target and obligations by as much as 9 billion baht. As a result, this year may be the first in which MEA and PEA, the champions of revenue remittance to the treasury, fail to meet their revenue targets and post negative results. Atthawit said that although Eknat Prompan, the Energy Minister, has spoken about this issue for a long time, no actual increase has been made. Currently, Data Centers still pay electricity at the same rate as households, which he called unfair to the public, and he urged the Energy Ministry to urgently raise Data Center electricity rates. He also proposed that the Department of Industrial Works under the Ministry of Industry expedite a declaration classifying Data Centers as Type 108 factories under Section 5 of the Factory Act, up from Thailand's current 107 factory types, since Data Centers have total machinery power exceeding 50 horsepower, use electricity at the megawatt level, and fall under the definition of factory operations involving the storage of equipment. The declaration process would take 60 days so that officials can inspect and oversee licenses, sanitation, or order closures if they cause a nuisance. He also pointed to the lack of preparation for dedicated zoning for Data Centers compared with other countries such as Malaysia, which has zoning on Penang Island complete with water and electricity systems, while in Thailand the Energy Regulatory Commission, or ERC, does not allow Data Centers to build their own power plants, forcing Data Centers to locate in areas with the strongest power and water networks, such as the Rama IX and Ramkhamhaeng areas.
Warsh Blames AI Hyperscaler Borrowing for Rising Treasury Yields
Federal Reserve Chair Kevin Warsh said the borrowing spree by AI hyperscalers is one reason long-term borrowing costs are rising, pointing to the competition for capital as the 10-year Treasury yield hit 5%. Warsh, speaking after the Fed raised its benchmark rate by a quarter point to 3.75-4%, said the so-called hyperscalers are out in the market raising funding and that the competition for capital is real. The five major hyperscalers issued $121 billion in U.S. corporate bonds in 2025, compared with an average of $28 billion a year between 2020 and 2024, according to BofA Securities, while Morgan Stanley estimates AI-related global debt reached nearly $236 billion by the end of May and forecasts it will approach $570 billion for the full year of 2026. Hyperscaler capital spending now runs close to 100% of operating cash flow, with some dipping negative, forcing the companies to turn to bond markets. Warsh also cited economic growth and geopolitics, namely the Iran war, as reasons for higher yields, but did not list the federal deficit, the explanation most bond investors give. He added that the Fed has set up an internal task force on AI, due to report by the end of the year.
Nvidia's "Bank of AI" Role Expands as Huang Targets AI's Capital Bottleneck
Nvidia is increasingly acting as what Chamath Palihapitiya called the "bank of AI," working across suppliers, infrastructure providers and financing partners to clear bottlenecks before they constrain AI deployment, Jensen Huang said at the All-In Summit. Huang described AI as a "new industrial revolution" that requires manufacturing, electricity, internet infrastructure and physical data-center capacity to scale together, not just GPUs. He said roughly $400 billion in venture financing flowed into AI companies over six months, creating jobs and massive demand for compute, which in turn drives demand for data centers. Nvidia has increasingly worked with financial institutions and other ecosystem players to make AI compute an investable, financeable asset, a strategy that turns the company into an ecosystem enabler: help customers secure capital, power and infrastructure, and Nvidia keeps supplying the computing engine underneath them. For investors, that could make Nvidia's AI opportunity considerably broader than its semiconductor market share, since the biggest constraint on AI spending may eventually be capital and infrastructure rather than demand for GPUs.
Palantir Names Nebius Preferred Sovereign AI Cloud Partner
Palantir Technologies and Nebius Group announced a partnership on 8 September 2026 making Nebius Palantir's preferred sovereign AI infrastructure provider, integrating Nebius' AI-native cloud and inference endpoints into Palantir's commercial perimeter to give eligible customers more control over compute, data, and models. The alliance positions Nebius as an AI-first cloud platform built for demanding AI workloads and could broaden its reach through Palantir's enterprise distribution and modular data-center deployments. The news comes alongside Nebius' first senior secured debt facility of about US$775.0 million, backed by GPU assets and contracted cash flows, underscoring how the company is funding and filling significant new capacity. Nebius Group's narrative projects $30.3 billion revenue and $2.4 billion earnings by 2029, with a $312.67 fair value implying 49% upside to its current price, while some of the lowest ranked analysts had assumed heavy spending would keep Nebius unprofitable even with revenue reaching about US$19.1 billion by 2029. The Palantir deal strengthens Nebius' case as a sovereign AI provider and could reinforce near-term demand visibility, but it does not remove the key near-term risk around heavy capital spending and the company's ability to translate rapid revenue growth into sustainable margins.
Jensen Huang Says Hyperscalers Plan Capacity Wrong, Backs 1,000 Neoclouds
Nvidia CEO Jensen Huang said hyperscalers are "always almost wrong" on AI capacity planning because they set infrastructure plans once a year while demand shifts far faster, and he is pushing a network of nimble regional clouds as the fix. Speaking at the All-In Summit, Huang said Nvidia would be happy with five hyperscalers but argued the industry ultimately needs 50, 100 or even 1,000 neoclouds, since regional providers can move quickly to secure land, power and existing facilities that are harder for companies operating from Silicon Valley or Seattle to coordinate globally. Nvidia is building what Huang described as a large-scale distributed network of companies securing those critical resources, and he pointed to activity in Australia and Southeast Asia where Nvidia is helping bring additional gigawatts of capacity online through regional cloud partners. The strategy matters for Nvidia because it does not need every AI workload to run through a handful of hyperscalers if it can help create hundreds of smaller customers that collectively consume massive amounts of Nvidia hardware. It also reinforces a broader shift in the AI infrastructure trade, where the bottleneck is increasingly not who has the best chips but who can secure the electricity, land and data center capacity needed to deploy them.
HPE Expands Oracle AI Data Center Deal and Signs Tottenham Hotspur Project
Hewlett Packard Enterprise expanded its global AI infrastructure partnership with Oracle, bringing HPE Juniper Networking into Oracle AI data centers, and separately agreed to a large-scale digital modernization project with Tottenham Hotspur Football Club focused on stadium hybrid cloud and AI infrastructure. Both collaborations highlight new client deployments for HPE across AI data centers and digitally enabled sports venues. The Oracle expansion supports the view that Juniper networking widens HPE's role in AI-centric data centers, where Cisco and Arista are key rivals, and shows Juniper integration moving into multi-year deployments. The Tottenham project reinforces HPE's GreenLake and as-a-service infrastructure business by turning a complex, multi-site venue into a single hybrid cloud environment. The article does not answer how far HPE can shift its overall revenue mix away from legacy servers toward these recurring, software-rich contracts.
Microsoft Raises Quarterly Dividend 8% to $0.98 per Share
Microsoft raised its quarterly dividend by seven cents to $0.98 per share, an 8% increase that lifts the annual payout to $3.92 per share from $3.64 previously. The dividend will be paid December 10 to shareholders of record on November 19, and the company set its annual shareholder meeting for December 8. The stock slipped to $492.76 Wednesday morning as investors focused more on the growth outlook than the modest increase in cash returned per share. Microsoft returned $10.2 billion through dividends and share repurchases last quarter, equivalent to roughly 11.3% of its $90 billion in quarterly revenue, while Microsoft Cloud generated $59.3 billion and Azure revenue climbed 43%. At $492.76, the shares sit 15.77% below the $585.03 GF Value estimate.
Apollo Warns Hyperscaler Credit Risk Rising on AI Spending
Apollo Global Management said Wednesday that corporate debt from major cloud computing companies is becoming riskier as these firms increase spending on artificial intelligence infrastructure. Apollo chief economist Torsten Slok wrote in a note that the market is repricing hyperscaler credit fundamentals due to a debt-financed AI capital expenditure cycle with rising leverage, negative free cash flow and uncertain payback on depreciating assets. According to Apollo's research, the gap between hyperscaler credit default swaps and bank credit default swaps has widened to around 60 basis points from roughly zero since October 2025, suggesting hyperscaler credit risk is increasing independently rather than as a result of dealer hedging of new bond issuances. The warning comes after leaders of frontier large language models said over the weekend they want to slow the rate of advancements of their products due to safety concerns, which could have financial consequences for the cloud computing providers that run the LLMs. According to FactSet, Alphabet has a forward debt-to-equity ratio of 13% and forward free cash flow of negative $25.7 billion, Amazon has debt-to-equity of 23% and free cash flow of negative $30 billion, Meta Platforms has debt-to-equity of 34% and free cash flow of negative $25.7 billion, and Microsoft has debt-to-equity of 7.34% and positive free cash flow of $33.4 billion.
Oracle Restructuring Bill Climbs to $2.8 Billion as AI Cloud Backlog Hits $664 Billion
Oracle's restructuring program has grown by another $700 million to roughly $2.8 billion, covering severance, contract exits and other costs tied to reshaping the business, as the company races to turn enormous AI demand into cash flow. Shares gained approximately 2.1% to $143.33 Wednesday. The added expense is meaningful against quarterly revenue of $19.35 billion, while Oracle added about 850 megawatts of data-center capacity during the latest quarter and free cash flow remained around negative $5 billion as infrastructure spending ran hard. The bull case rests on more than $30 billion of additional AI-cloud contracts signed in the period, which pushed remaining performance obligations to $664 billion; management argues those commitments can be supported within its existing capital plan because some customers can prepay or contribute chips. The restructuring is more than simple cost cutting, since it can help Oracle absorb the near-term cash burden, but the real test is converting that gigantic backlog into profitable cloud revenue fast enough to justify the infrastructure bill.