
What happened
Nvidia's AI Enterprise lists at $4,500 per GPU annually, and paid subscriptions add supported software, security updates, and a path from CUDA dependence to recurring profit.
Why it matters
Subscriptions could offset margin pressure, but Nvidia's reported results remain dominated by data-center demand and separate renewals must add spending beyond bundled hardware.
Summaries like this, in your inbox every morning.
Nvidia's CUDA programming tools make switching platforms costly, giving the company a captive installed base. Paid subscriptions for AI Enterprise and Mission Control are designed to convert that captivity into recurring revenue by offering supported software, security updates, and tools that help customers use expensive systems more efficiently. Mission Control schedules workloads and automates recovery, while AI Enterprise lists at $4,500 per GPU annually. However, free alternatives exist, and selected GPUs already include five-year licenses, meaning separate renewals must add customer spending beyond bundled hardware purchases to cover engineering and support costs.
That challenge matters because Nvidia's reported results remain dominated by data-center hardware demand. The company trades at 25x forward P/E, versus 29x trailing reported earnings, which looks attractive against second-quarter revenue growth of 106% and management's $108 billion third-quarter revenue outlook. Analysts expect adjusted EPS to rise from $9.31 in fiscal 2027 to $15.91 in fiscal 2028, approximately 71%. But management expects gross margin to ease from 75% to 74% next quarter, so strong sales must offset some margin pressure.
The lower multiple comes with larger commitments. Nvidia generated $69.9 billion of first-half free cash flow, while supply and capacity commitments reached $279 billion across future years. Delayed customer deployments could leave inventory consuming cash and expose margins to write-downs. June's $25 billion debt issuance adds financing costs, meaning the software subscription bet carries more weight as hardware obligations grow.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
Vinci4D provides chip-design simulation software that trains on physics equations instead of customer data, an…
SEMI President and CEO Ajit Manocha told DIGITIMES that the semiconductor industry reached the US$1 trillion r…

CoreWeave fell more than 25% this week, its debt yield briefly rose above 16% on Thursday, and SK Hynix droppe…

Anthropic told investors this week that it is making 85% gross margins, a figure the columnist calls a cash co…

Modal Labs, an AI infrastructure provider, raised $750 million in a Series D led by Accel at a $15.75 billion…

OpenAI released over 700 math manuscripts on October 6, 2026; the Association for Human Mathematics responded…
