Cloud computing

Oracle

2026-09-14 07:32:57

JPMorgan Reaffirms Overweight on Oracle as $30 Billion-Plus AI Deals Ease Revenue Concerns but Leave Funding Debate Intact

JPMorgan reiterated its Overweight rating on Oracle in a Sept. 11, 2026 report, keeping a $200 price target after the company posted stronger-than-expected F1Q27 results. Oracle reported $19.35 billion in revenue, up 30% year over year, while cloud revenue rose to $11.6 billion and IaaS climbed 121% to $7.4 billion. Remaining performance obligations reached $664 billion, up 46% from a year earlier and $26 billion from the prior quarter. The bank said new AI cloud contracts topped $30 billion and were largely structured through prepayments or bring-your-own-hardware arrangements, limiting the need for incremental cash from Oracle. Even so, the market is still focused on the company’s financing path as capital spending remains elevated. Oracle guided FY27 revenue to at least $90 billion and EPS to $8.10, both above consensus, while capex was set at $90 billion to $95 billion, with net cash capex capped below $70 billion. JPMorgan said FY27 and FY28 remain peak capex years, and management did not provide a timeline for free cash flow to turn positive. The report also flagged customer concentration, funding access, delivery execution, and margin ramp as key risks.

1020

Anthropic

2026-09-14 01:13:36

Anthropic signs $13.7 billion, six-year compute deal with Rum Group

Anthropic has signed a six-year computing agreement worth $13.7 billion with Rum Group, according to The Information, which cited people familiar with the matter. Rum Group is described as a company with roots in social media and long-standing ties to the Trump administration. Founded in 2013, Rum gained attention among conservatives after the 2020 U.S. presidential election as a YouTube alternative for smaller content creators. The company also hosts Donald Trump’s Truth Social and streams official White House broadcasts. The report said the deal is the latest in a string of cloud-computing agreements Anthropic has reached over the past year. Its partners have included major cloud providers such as Google, technology companies including SpaceX, and smaller emerging cloud firms such as Nscale. Demand for Anthropic’s Claude Code and Cowork products was cited as the driver behind the recent wave of capacity deals. According to the report, those agreements together account for at least 14,800 megawatts of compute capacity, with costs that could reach as much as $517 billion over the next decade. The Rum agreement is set to add to that total.

840

Anthropic

2026-09-12 02:32:39

Artemis compares Anthropic to early AWS as net ARR hits $65 billion and S-1 focus shifts to lock-in, margins, and power

Artemis Analytics argues that Anthropic’s edge is not limited to frontier models. In its latest research, the firm says the company is building an enterprise AI stack with the kind of lock-in that once turned Amazon Web Services from a disputed cost center into the dominant cloud platform. The report says Anthropic reached $65 billion in net annual recurring revenue in July 2026, up 15x year over year, and projects ARR of $275 billion by the end of 2027, a level it says would surpass AWS. The thesis rests on a few operating questions: whether Anthropic can keep hiring top researchers, whether it can secure enough power at acceptable cost, and whether it can maintain premium API pricing. Artemis estimates monthly net new ARR at $10 billion to $15 billion, says 90% of ARR comes from API sales, and argues investors should focus on net ARR after subtracting revenue shares paid to AWS, Gemini, and Microsoft. The report also lays out risks, including rising competition from OpenAI Codex and Astra, low switching costs across model providers, and the possibility that application-layer products reduce dependence on Anthropic. At the same time, Artemis includes extensive conflict disclosures, noting that the author may hold Anthropic exposure through an SPV, that Artemis is an Anthropic customer, and that none of its forecasts or valuation scenarios are verified by Anthropic.

960

Oracle

2026-09-11 08:44:26

Oracle FY2027 Q1 tops estimates as OCI revenue jumps 121% and RPO reaches $664 billion

Oracle’s FY2027 first-quarter results pointed to a company leaning harder into AI infrastructure. Revenue came in at $19.35 billion, up 30% year over year, while adjusted EPS rose 30% to $1.92, both ahead of market expectations cited in the source material. The standout figure was cloud infrastructure revenue, which surged 121% to $7.4 billion and helped lift total cloud revenue 62% to $11.6 billion. Oracle shares rose about 5.5% in after-hours trading after the earnings release. The report also highlighted the scale of Oracle’s buildout. The company delivered 850MW of new data center capacity during the quarter and shipped more than 300,000 GPUs to AI cloud customers since the end of the prior quarter. At the same time, remaining performance obligations, or RPO, reached a record $664 billion, with more than $30 billion in new AI cloud contracts signed in the quarter. Cash generation was strong, but spending remained heavier. Operating cash flow hit a record $23 billion, while capital expenditures were about $28.5 billion, leaving free cash flow at roughly negative $5 billion. Oracle also completed a $20 billion ATM stock issuance and raised its full-year adjusted EPS target for FY2027 to $8.10 from $8.05, while keeping its revenue goal at no less than $90 billion.

970

MarsBit

2026-09-10 04:44:09

MarsBit piece argues modern IT is still stacked on a fragile von Neumann foundation

MarsBit has published a long-form commentary that uses a single diagram to dissect the structure of modern IT infrastructure, arguing that the industry keeps adding new layers while leaving its oldest bottlenecks largely untouched. The article centers on the von Neumann bottleneck, describing it as the thin support column under today’s computing stack: CPUs keep getting faster, but memory and bus bandwidth remain limiting factors, and growing software complexity only increases data movement between memory and processors. The commentary walks from electricity and semiconductor fabs up through operating systems, high-bandwidth memory, CPUs and GPUs, C projects, JVM-based software, web tooling, and end-user applications. Its core claim is that each layer may look well engineered on its own, yet the full stack is less a unified design than an accumulation of temporary fixes, compatibility compromises, and maintenance patches. It also frames the current AI boom as a source of additional strain rather than a fix for the base layer. According to the article, AI training and inference sharply increase power demand, GPU usage, and data retrieval pressure, while AI-generated code and content can introduce quality and security problems into open-source and web ecosystems. The author extends that criticism to large tech companies, which are portrayed as repeatedly disrupting existing frameworks and standards, and to independent developers, whose fragmented projects can add flexibility but also disorder and maintenance risk.

910

Google Cloud

2026-09-10 02:45:53

Google Cloud says AI server payback is under two years, and under one year with in-house TPUs

Google Cloud Chief Executive Officer Thomas Kurian said the company’s investment in AI infrastructure is backed by commercial demand rather than speculative buildout, arguing that the overall payback period for AI servers is less than two years. Speaking at the 2026 Goldman Sachs Communacopia + Technology Conference, Kurian said the figure is cut in half when Google uses its own Tensor Processing Units, implying a payback period of under one year for TPU-based systems. He also said most infrastructure contract value comes from five-year commitments, which improves revenue visibility and lowers recovery risk on heavy compute spending. Kurian described Google Cloud’s AI stack as spanning in-house chips, Gemini models, data platforms, security products and enterprise applications. He said the company’s annualized revenue is approaching $100 billion, new customer acquisition has more than doubled from a year earlier, and deals above $100 million have more than doubled on both a sequential and annual basis. Google Cloud now has more than 17 product lines generating over $1 billion in annual revenue, while Gemini Enterprise has been adopted by 90% of the Fortune 100, according to Kurian.

920

AMD

2026-09-08 03:44:27

AMD Launches AI Personal Supercomputer to Challenge Nvidia and Cloud Dependency

AMD has unveiled a personal supercomputer called 'Threadripper Halo Station,' designed for AI-driven high-intensity computing tasks. The device aims to reshape the local AI computing landscape, reduce reliance on cloud computing, and challenge Nvidia's dominance in high-performance workstations.

870

Bitcoin minin

2026-09-03 21:31:05

Hyperscale Shuts Michigan Bitcoin Mine for AI Deal Worth More Than $1.2 Billion

Hyperscale Data has halted Bitcoin mining at its Michigan data center to free up power and infrastructure for an artificial intelligence customer, according to a company announcement released Wednesday. The agreement covers 20 megawatts for an unnamed California-based provider of AI cloud computing services and carries an initial term of 10 years. If the customer exercises both five-year extension options, the contract could generate more than $1.2 billion over 20 years. The company said the customer also has an option to add 32 megawatts within the first two years. If that expansion is taken and both extensions are exercised, total contract revenue could exceed $3 billion. CEO William Horne said the immediate shutdown of mining operations allows the company to direct the facility’s power, infrastructure, and other resources toward preparing the site for the customer. Hyperscale plans to sell its mining servers and said it expects gains from those sales. It did not disclose when AI operations will begin, and said the shutdown applies to the Michigan site. The move comes as more miners explore AI-related conversions. Decrypt noted that VanEck’s Matthew Sigel argued in March that miners could benefit by repurposing infrastructure for AI demand, while IREN’s results last month showed both the revenue potential and the cost of that shift, including a $450.4 million asset write-down tied mostly to retired mining equipment.

940