Microsoft showed a version of Windows in which an AI coding model and parts of Copilot run on the computer itself instead of in the cloud, and a security tool, Microsoft Execution Containers, that limits which data an AI agent can reach. It also presented…
Finance angle The finance angle is where AI work gets paid for. A model that runs on the laptop shifts spending from cloud usage to the device: Nvidia earns from the chip, Microsoft from the hardware and Windows, and the buyer pays once instead of per query. The price puts the machine against Apple's MacBook Pro, which the report…
Read the report at Reuters via BNN Bloomberg ↗
SpaceX is looking to raise about $40 billion to pay for a large purchase of Nvidia artificial-intelligence chips, according to a Financial Times report relayed by EconoTimes. The plan is roughly $10 billion of bank loans and $30 billion of investment-grade…
Finance angle The finance angle: about $40 billion of new borrowing — $10 billion in loans and $30 billion in investment-grade debt — that flows almost entirely to Nvidia as chip revenue, with private-credit firms such as Apollo and Pimco earning the interest. SpaceX shareholders carry the debt; rental income from Anthropic and…
Read the report at EconoTimes ↗
Paris-based Mistral AI put its Large 4 model into public preview on October 6. The company describes a multimodal model with 1 trillion parameters, 49 billion of them active at a time, trained on 3,800 Nvidia Grace Blackwell GPUs in Mistral's own European…
Finance angle The finance angle is price and ownership of the hardware. $1.36 and $4.18 per million tokens is the rate European companies and public bodies can now compare with US model providers, and open weights let them run the model on their own servers instead of paying per token. The training used 3,800 Nvidia GPUs that…
Read the report at Mistral AI ↗
The Wall Street Journal reported that Lambda, a California company that rents out Nvidia-based computing for AI, is raising as much as $4 billion in a round led by Blackstone and Coatue Management, at a valuation of $14.5 billion before the new money. It is…
Finance angle The figures are $4 billion of equity against a $50 billion backlog: a 'neocloud' has to buy the chips and build the capacity before the contracted revenue arrives, so the money raised flows on to Nvidia and to data-centre and power providers. A backlog is committed orders, not revenue, and it more than tripled in…
Read the report at Seeking Alpha ↗
Foxconn, the Taiwanese contract manufacturer that assembles AI servers as well as consumer electronics, reported third-quarter revenue of T$3.03 trillion (about $95.4 billion), up 47% from a year earlier and above the LSEG SmartEstimate of T$2.83 trillion.…
Finance angle The finance angle: Foxconn's monthly sales are an early, hard number on what cloud companies are actually paying for AI hardware, weeks before the chip designers report. Revenue of T$3.03 trillion in a quarter, T$200 billion above the consensus figure, says orders for AI server racks were still being delivered at…
Read the report at GuruFocus ↗
Firmus Grid, an Australian AI data-centre operator backed by Nvidia and Blackstone, plans to allocate roughly half the shares in its initial public offering to existing shareholders, Investing.com reported, citing Bloomberg. The company is seeking up to…
Finance angle At up to A$5.5 billion the sale would exceed Medibank's 2014 float of just under A$5 billion, among the largest Australian listings. The money goes to GPU infrastructure for Firmus's first data centre in Batam, Indonesia, so much of it flows on to chip suppliers, Nvidia first among them, which is both a shareholder…
Read the report at Investing.com ↗
SoftBank Group paid the third and final $10 billion tranche of its $30 billion follow-on investment in OpenAI on October 1 through Vision Fund 2. Its total investment in the ChatGPT maker is now $64.6 billion for a stake of about 13%. The payment was…
Finance angle The $30 billion was SoftBank's share of the $122 billion round that valued OpenAI at $852 billion, alongside $50 billion committed by Amazon and $30 billion by Nvidia. OpenAI receives the cash and spends much of it on computing capacity, which is revenue for chip and cloud suppliers. SoftBank carries the financing…
Read the report at The Next Web ↗
Tencent has reportedly signed a five-year agreement, worth about $7 billion, to lease roughly 100,000 advanced AI chips housed in several Oracle data centres in Southeast Asia, with around 30% paid upfront. It is described as Tencent's largest overseas…
Finance angle The finance angle: about $7 billion over five years, roughly $2.1 billion of it upfront, would flow to Oracle's cloud business from a single customer, for hardware that Nvidia could not sell into China directly. The cost shows on Tencent's side: its second-quarter capital spending rose 176% to RMB 52.8 billion and…
Read the report at TrendForce ↗
SharonAI Holdings disclosed its first committed senior secured debt facility backed by GPUs: US$356 million at a fixed rate of 9.95% before fees, held in a special-purpose vehicle and secured on the chips and the cash flows from customer contracts. Goldman…
Finance angle The 9.95% coupon shows what private credit charges to finance AI chips for a smaller operator: well above investment-grade borrowing costs, and payable from rental income on hardware that loses value quickly. Sharon AI says it has secured over US$2.6 billion of debt and equity in ten months against a customer order…
Read the report at SEC EDGAR (SharonAI Holdings 8-K) ↗
GMI Cloud, a Mountain View-based GPU cloud provider, raised a $223 million Series B led by ARCHIV with Nvidia, KT Corporation, Trend Micro and others taking part, plus a $445 million credit facility led by CTBC. It plans to add GPU capacity in the US, Taiwan…
Finance angle Two-thirds of the $668 million is debt, the usual structure for GPU clouds: lenders finance the chips against customer contracts. GMI says contracted annual recurring revenue has passed $600 million, more than nine times its level at the end of 2025. Most of the money goes to buy GPUs, so Nvidia is both an investor…
Read the report at Pulse 2.0 ↗
Nvidia's board authorised a further $150 billion of share repurchases, the largest increase any company has announced and above Apple's $110 billion in 2024. Total buyback capacity is now $235 billion, which Nvidia plans to use through fiscal 2028. The…
Finance angle The $235 billion is cash earned from AI chip sales to cloud companies being returned to shareholders rather than spent on capacity or acquisitions. It comes while Nvidia's shares, up just over 20% this year, trail AMD (more than doubled) and Intel (more than tripled) as competition in AI chips grows, and with the…
Read the report at Reuters via Investing.com ↗
Stories are summarised in our own words from the linked reports; every one credits its publisher. Facts and exposure, never a tip: XDQ Labs is not registered with SEBI, the US SEC, FINRA or any European regulator. US and European share prices are not shown.