LG’s humanoid robot arrives in 2027, running Nvidia’s Isaac GR00T
LG’s humanoid robot arrives in 2027, running Nvidia’s Isaac GR00T

LG will unveil a bipedal humanoid robot in the first quarter of 2027, built on Nvidia’s Jetson Thor processor, Isaac GR00T foundation model and Halos safety system. The two signed the agreement at Nvidia’s Santa Clara headquarters on 13 August. LG will…

Nvidia halved its $250bn promise to OpenAI. Then it showed investors what it owns
Nvidia halved its $250bn promise to OpenAI. Then it showed investors what it owns

Anissa Gardizy reported the cut for The Wall Street Journal on 14 August. Under the new proposed terms Nvidia would initially backstop only half of the 10GW project, and it would decide about the rest later. A deal could be signed as soon as this weeke…

Goldman Sachs is courting investors on Nvidia’s $500bn AI-compute financing deal
Goldman Sachs is courting investors on Nvidia’s $500bn AI-compute financing deal

Goldman Sachs has secured the role every Wall Street firm wanted, and it is now working the phones. The bank is in talks with investors about Nvidia’s $500bn AI-compute financing deal, having landed a prized mandate that effectively casts it as the lea…

Nvidia’s new $500B plan is risky but brilliant, especially for aging GPUs

Nvidia has a plan to make sure its GPUs won’t lose value. It wants to convince a new crop of financiers to keep lending for AI buildouts.

Nebius revenue grew 454%. Most of the cash came from customers paying in advance
Nebius revenue grew 454%. Most of the cash came from customers paying in advance

Nebius rents out computing power. It buys Nvidia chips, puts them in data centres, and charges companies to train and run AI models on them. The trade calls this a neocloud. It is based in Amsterdam, listed on Nasdaq, and was spun out of the Russian in…

AI compute becomes a tradable commodity on 5 October. The price goes public
AI compute becomes a tradable commodity on 5 October. The price goes public

Renting an Nvidia H100 for an hour costs whatever your supplier says it costs. There is no published price. Two companies buying identical capacity can pay very different rates, and neither of them will know. That changes on 5 October. CME Group, the e…

Nvidia is building a trillion-parameter open model, and it would still be smaller than China’s
Nvidia is building a trillion-parameter open model, and it would still be smaller than China’s

Nvidia released Nemotron 3.5 Lightning on Tuesday. It is a 30 billion parameter mixture-of-experts model with three billion active at any moment. The architecture is a hybrid of Mamba-2, MoE and attention layers, with a one million token context window…

Foxconn now makes more money from AI servers than from everything else combined
Foxconn now makes more money from AI servers than from everything else combined

Foxconn crossed a line this quarter that it will not cross back. Its cloud and networking division, the one that builds AI servers, made up 51% of revenue in the three months to June. That is the first time the segment has passed half. Smart consumer e…

Nvidia calms its own credit market after Jensen Huang clarifies the $500bn financing plan
Nvidia calms its own credit market after Jensen Huang clarifies the $500bn financing plan

The chipmaker’s credit-default swaps had spiked to a record on fears of circular financing. They eased once Jensen Huang recast the plan as other people’s money, with Nvidia’s own exposure capped. After weeks in which the cost of insuring its debt clim…

IBM bets $240m on cheap, open-source inference to take on the hyperscalers
IBM bets $240m on cheap, open-source inference to take on the hyperscalers

A multi-year deal with the startup Together AI will put Nvidia Blackwell systems on IBM Cloud, and the wager is that enterprises now care more about the cost of running AI than the prestige of the model doing the running. IBM has decided that the money…