Yesterday I appeared on CNBC’s Closing Bell Overtime to discuss NVIDIA’s earnings. You can watch an edited video of this conversation via the link below, or read on for an AI-generated, human-edited overview and summary of the complete discussion.

Link to edited segment.

Overview

NVIDIA remains the bellwether for the AI trade despite stock movement, supported by strong revenue growth, high GPU demand, and the pervasiveness of its architecture. Concerns over gross margins, massive capital expenditures, and potential custom chips from competitors like OpenAI are balanced by profitability among frontier model developers and expansion into industrial robotics. Additionally, software partnerships, such as the strategic link between Salesforce and Anthropic, highlight ongoing integration across the enterprise ecosystem while mitigating concerns about data privacy and the broader impact of AI automation on traditional software models.

Outline

NVIDIA Earnings and Market Position

• Mike Santoli of CNBC opens the discussion by noting that NVIDIA stock is down slightly after its earnings release despite doubling hyperscaler revenue and maintaining full steam on AI infrastructure buildout.
• Santoli asks Patrick Moorhead about NVIDIA’s ongoing role from an investment perspective as the market looks at other ways to express its views on AI.
• Moorhead explains that NVIDIA remains the bellwether because tenfold more software companies use its GPUs compared to any competitor.
• Moorhead notes that despite a revenue increase of 106% and a quadruple beat, the stock reacted to missed buy-side whisper numbers and gross margin questions.
• Moorhead points out that frontier model companies like Anthropic and OpenAI are showing profitability, which should relieve market debates about whether anyone is making money in AI.

Long-Term AI Investments and Industry Drivers

• Santoli suggests that current operating profitability for AI companies still requires massive future revenue or capital raising to cover compute commitments.
• Moorhead responds that revenue will not lock into gear unless capex commitments and AI factory buildouts continue as planned.
• Moorhead highlights the industrial sector and especially industrial IoT as a major long-term growth driver for AI that goes beyond standard datacenter demand.
• Santoli mentions that everyday news flows, such as OpenAI developing custom chips or Chinese models leading in token consumption, create static around NVIDIA’s growth sustainability.

Competitive Landscape and Hardware Supply

• Moorhead addresses OpenAI’s chip named Jalapeño, arguing that custom chips from model makers and hyperscalers are expected — but hard to scale meaningfully against NVIDIA.
• Moorhead notes that rolling a single custom chip costs $300 million to $500 million, making it difficult for rivals to take significant market share from NVIDIA in a rapidly expanding market.
• Moorhead emphasizes that NVIDIA further secures its dominance by buying up the majority of high bandwidth memory, starving smaller rivals even when supported by Broadcom or Marvell.
• Santoli agrees that NVIDIA’s strong balance sheet provides a distinct advantage over competitors for the coming years.

Enterprise Software and the Salesforce–Anthropic Partnership

• Santoli asks Moorhead for his thoughts on the strategic partnership between Salesforce and Anthropic and what it could signify for enterprise software.
• Moorhead states the partnership benefits Anthropic CEO Dario Amodei by countering criticism regarding customer data privacy through Salesforce’s trusted brand.
• Moorhead cites the framing of the deal as a user interface enhancement that helps clients navigate Salesforce more effectively.
• Moorhead concludes that while discussions around AI disrupting SaaS will persist, this collaboration helps mitigate existential fears in the market.
• Santoli closes the interview by quoting Dario Amodei’s remark about not wanting to destroy anyone and thanks Moorhead for his analysis.

Companies Mentioned

• NVIDIA
• Salesforce
• Anthropic
• OpenAI
• Broadcom
• Marvell

Keyword Summary

NVIDIA, AI, GPUs, hyperscaler, capex, frontier models, custom chips, high bandwidth memory, enterprise software, Salesforce