Nvidia reported earnings on Wednesday, and as anticipated, the numbers had been good. Actually good. The corporate will get greater than 91% of its gross sales from its information heart unit, which generated income of $193,737 billion, up 68% year-over-year.
“We’ve got now scaled our information heart enterprise by practically 13x since the emergence of ChatGPT in fiscal 2023,” Nvidia CFO Colette Kress stated in the firm’s earnings name on Wednesday.
Whereas very spectacular, the quantity is not all that stunning provided that international AI spending is anticipated to attain $2.5 trillion this yr, and Nvidia’s largest clients, the main AI hyperscalers Amazon, Alphabet, Meta, and Microsoft, all reported file capex figures earlier this month.
The hyperscalers additionally made eyewatering monetary commitments for 2026 totaling nearly $700 billion, which got here to the dismay of many buyers who’ve been rising cautious of AI spending.
Earlier this month, Evercore analysts warned that the big capex may flip the hyperscalers’ money movement adverse.
And regardless of the file after file multibillion-dollar dedication made to scale AI infrastructure and develop the expertise’s adoption throughout the American economic system, the outcomes are but to totally materialize. A Goldman Sachs analyst lately stated that AI contributed “basically zero” to U.S. GDP in 2025.
Nvidia CEO Jensen Huang spent most of his time in the investor name making an attempt to justify that capex development.
“I’m assured of their money movement rising, and the cause for that is quite simple: we’ve got now seen the inflection of agentic AI and the usefulness of brokers throughout the world in enterprises all over the place,” Huang stated.
AI adoption by enterprises past the tech world, and whether or not these corporations truly see actual productiveness features and income returns from AI integration, is actually vital to Nvidia, as a result of that’s a serious factor that the AI trade is at the moment missing to quell worries over an AI bubble.
A latest survey discovered that regardless of 70% of corporations using AI, over 80% reported no impression on employment or productiveness.
Final week, OpenAI COO Brad Lightcap instructed TechCrunch that his firm had “not actually seen enterprise AI penetrate enterprise enterprise course of.”
Some specialists consider that Anthropic’s Claude Cowork unveiled earlier this month is going to be a turning level in AI’s penetration into the workforce, a lot in order that they consider it’s going to lead to a mass extinction-level event for software program corporations, and possibly even white-collar work. Huang gave a particular shout-out to Claude Cowork in the name as properly.
Huang additionally had a technical rationalization to justify the capex commitments.
“On this new world of AI, compute equals revenues,” Huang stated, a phrase that he repeated many occasions all through the name. Huang argues that tokens, aka the chunks of information that AI fashions course of, are the most vital a part of a brand new AI economic system. The extra tokens a mannequin makes use of, the extra computing energy and time it requires. So, as fashions are getting extra advanced, the demand for computing is additionally going up “exponentially,” Huang stated. He argued that the capex commitments will go in direction of constructing this compute capability, which can thus energy higher-level fashions and translate to income.
“The quantity of token era functionality that the world wants is lots, greater than $700 billion, and I’m pretty assured that we’re going to proceed to generate tokens…basically as a result of each single firm relies upon on software program, each software program will rely on AI, and so each firm will produce tokens,” Huang stated. “If the new software program requires tokens to be generated and the tokens are monetized, then it stands to cause that their information heart build-out immediately drives their revenues.”
Huang’s justifications could not have instantly satisfied the market. Despite the fact that shares rose at first in response to the report, after the name, features finally pulled again to lower than 1%. That’s regardless of income that exceeded market expectations.
OpenAI and China are nonetheless blind spots
All through the name, Huang additionally tried to tackle rumors of a falling out with OpenAI, first spurred after a $100 billion Nvidia funding introduced again in September 2025 reportedly failed to progress past the early levels after months. Then, two back-to-back experiences claimed that Huang was privately criticizing OpenAI’s enterprise strategy whereas OpenAI was sad with the inference velocity of Nvidia’s chips.
In the name on Wednesday, Huang repeatedly praised the AI large’s choices, however revealed that the funding was nonetheless not finalized.
“We proceed to work with OpenAI towards a partnership settlement, and consider we are shut,” Huang stated on the name. The submitting additionally refuses to give any assurance that “a transaction might be accomplished.”
One other piece of uncertainty weighing on Nvidia is China. The corporate shared that, as of this month, the Trump administration has lastly allowed it to begin delivery small quantities of its H200 chips to China, the place it as soon as held 95% of the market share before Trump first banned the chipmaker’s gross sales to China, sparking a saga of dizzying commerce tit-for-tat between the two international superpowers. However executives nonetheless don’t know if the imports might be allowed in, and are not factoring it into the income they anticipate this yr.
Disclaimer: This article is sourced from external platforms. OverBeta has not independently verified the information. Readers are advised to verify details before relying on them.
