NVIDIA's Rubin Chip Secures First Asian Orders as AI Financing Model Draws Market Scrutiny
Nashnova编辑部
India's AM Intelligence has ordered 9,000 Vera Rubin rack-scale systems, making it one of Asia's earliest adopters; meanwhile Nvidia's financing guarantees for AI clients have crossed $500 billion, putting its emerging 'AI central bank' role under market scrutiny.
Who placed this Asian order — and why does it matter?
AM Intelligence, a subsidiary of Indian renewable-energy group Greenko, has ordered 9,000 Vera Rubin rack-scale systems as part of an $8 billion, 1-gigawatt compute buildout.
Servers will go live next year in Hyderabad, running trillion-parameter AI models and next-generation agentic AI applications for clients across India, the US, Finland, and Malaysia.
This means → Rubin hasn't formally shipped yet, but Asia is already locking in capacity. The first batch of compute has already been bought by an undisclosed US client.
Why can an energy company sell AI compute?
Greenko founder Mahesh Kolli says clients are "desperately searching" for compute, and low-cost renewable energy is AM Intelligence's core edge.
In plain terms = the biggest cost in a data center is electricity. Whoever has the cheapest power can undercut on compute pricing — and this company generates its own.
The group plans to fund expansion through a debt-equity mix, targeting 5 gigawatts of data-center capacity across India and Europe by 2030.
What does Rubin mean for Nvidia's earnings?
Nvidia reports quarterly results on August 26. Analysts expect Q2 revenue to nearly double year-over-year to $92.18 billion — the fastest growth in seven quarters — with data-center sales projected to more than triple.
The market's central question: can Nvidia smoothly migrate clients from Blackwell — its current-generation AI chip — to Vera Rubin? Shipments are expected to begin this fall.
Morgan Stanley estimates Rubin could contribute roughly $9 billion in revenue in the fiscal quarter ending October, and expects Nvidia to frame the chip as "dramatically improving AI-factory economics."
This means → Rubin's ramp pace directly shapes the second-half growth curve. How smoothly the migration goes will be the single biggest topic on the earnings call.
Where is the risk in the 'AI central bank' model?
This month Nvidia helped arrange $500 billion in AI-infrastructure financing from six major US financial institutions. Last week it agreed to guarantee up to $105 billion in 20-year leases for an OpenAI data center in Ohio.
Zacks chief strategist Brian Mulberry put it bluntly: this makes Nvidia "a central bank of AI" — the risk is total, undiversified exposure to AI, and the model works only if AI-tool adoption keeps growing.
In plain terms = Nvidia isn't just selling chips anymore — it's helping clients borrow money and backstopping their leases. If AI demand growth slows, those guarantees could turn into real cash losses.
How did Jensen Huang respond — and what does the market think?
Huang argued the Ohio data-center guarantee is not revolving financing: OpenAI will pay its own rent, and Nvidia is simply providing long-term backing for the facilities that will house its chips.
Nvidia stock is up 11.8% year-to-date but has lagged key rivals, briefly ceding the title of world's most valuable company to Apple last month.
This reflects a shift in how the market prices Nvidia — it's no longer just about chip sales, but about the scale of contingent liabilities (potential obligations that become real if the guarantees are triggered). Rubin's ramp trajectory and how these guarantees are disclosed will be the key verification points on this earnings call.
Content is for reference only, not financial advice.