Nvidia on Wednesday issued a quarterly revenue forecast that exceeds Wall Street expectations, but the outlook did little to quiet investor questions about the company’s long-term position in the artificial intelligence chip market. The company projected third-quarter revenue of $108.0 billion, plus or minus 2%, versus analysts' average estimate of $104.19 billion based on LSEG data.
Shares of the world's most valuable company slid 1.2% in extended trading following the announcement. As of the close the prior trading day, Nvidia's year-to-date gain stood at 14.2% - markedly lower than the performance of rivals AMD and Intel, both of which have more than doubled so far this year.
In its outlook, Nvidia said it did not assume any data center chip sales from China. The company’s quarterly results and guidance are widely watched because its graphics processors power many of the advanced AI models and major data centers around the globe, making Nvidia a bellwether for AI infrastructure demand.
Industry spending patterns underlie some of the tension around Nvidia’s guidance. Several major technology firms, including Microsoft and Meta, recently reiterated expectations that Big Tech will allocate more than $730 billion to AI infrastructure this year, a sharp increase from last year’s roughly $400 billion outlay. Yet an increasing share of that planned spending is being directed to in-house chip development, as customers pursue custom silicon intended to reduce dependence on Nvidia’s high-cost, supply-constrained processors.
Competition in the inference segment - the part of AI that handles running models and answering queries - is intensifying. Nvidia’s GPUs, which are heavily used for training and inference, now face greater pressure from central processors and specialized chips that are optimized for inference tasks. Rivals including Intel and AMD are explicitly targeting that market. Several Chinese companies, among them Baidu, are also producing proprietary chips aimed at inference workloads.
At the same time, scrutiny of Nvidia’s role in the broader financing ecosystem for AI infrastructure has increased. The company agreed to guarantee some deals as part of a new arrangement with six major financial institutions that aims to raise more than $500 billion for AI infrastructure, drawing attention to its expanding influence across technology and finance.
Nvidia has taken actions to broaden its architecture in response to competitive pressures. In March, the company unveiled a new central processor and an associated AI system built around technology licensed from inference-focused startup Groq, under a $17 billion agreement that pairs Groq's chips with Nvidia's upcoming Vera Rubin platform.
Corporate endorsements of Nvidia’s hardware have continued: earlier this month SpaceX's chief executive said the rocket firm would rely exclusively on the company’s processors. Nvidia itself has quantified the long-term opportunity for its AI chips, saying the addressable revenue could exceed $1 trillion through 2027 - roughly double the $500 billion opportunity it previously cited through 2026 for its Blackwell and Rubin families of chips.
For investors and market participants, the latest guidance offers a mixed signal: robust near-term demand implied by the above-consensus revenue target, but persistent questions about competitive dynamics, geographic assumptions and the implications of Nvidia's expanding role in financing that large-scale infrastructure buildout.