Tag: gpus

‘Compute Equals Revenues’: Nvidia Needs Jensen Huang’s New Catchphrase to Be True

[ad_1]

Nvidia reported earnings on Wednesday, and as expected, the numbers were good. Really good. The company gets more than 91% of its sales from its data center unit, which generated revenue of $193,737 billion, up 68% year-over-year.

“We have now scaled our data center business by nearly 13x since the emergence of ChatGPT in fiscal 2023,” Nvidia CFO Colette Kress said in the company’s earnings call on Wednesday.

While very impressive, the number is not all that surprising given that global AI spending is expected to reach $2.5 trillion this year, and Nvidia’s largest customers, the major AI hyperscalers Amazon, Alphabet, Meta, and Microsoft, all reported record capex figures earlier this month.

The hyperscalers also made eyewatering financial commitments for 2026 totaling nearly $700 billion, which came to the dismay of many investors who have been growing wary of AI spending.

Earlier this month, Evercore analysts warned that the huge capex could turn the hyperscalers’ cash flow negative.

And despite the record after record multibillion-dollar commitment made to scale AI infrastructure and grow the technology’s adoption across the American economy, the results are yet to fully materialize. A Goldman Sachs analyst recently said that AI contributed “basically zero” to U.S. GDP in 2025.

Nvidia CEO Jensen Huang spent most of his time in the investor call trying to justify that capex growth.

“I am confident in their cash flow growing, and the reason for that is very simple: we have now seen the inflection of agentic AI and the usefulness of agents across the world in enterprises everywhere,” Huang said.

AI adoption by enterprises beyond the tech world, and whether these companies actually see real productivity gains and revenue returns from AI integration, is really important to Nvidia, because that’s a major thing that the AI industry is currently lacking to quell worries over an AI bubble.

A recent survey found that despite 70% of firms employing AI, over 80% reported no impact on employment or productivity.

Last week, OpenAI COO Brad Lightcap told TechCrunch that his company had “not really seen enterprise AI penetrate enterprise business process.”

Some experts believe that Anthropic’s Claude Cowork unveiled earlier this month is going to be a turning point in AI’s penetration into the workforce, so much so that they believe it will lead to a mass extinction-level event for software companies, and maybe even white-collar work. Huang gave a special shout-out to Claude Cowork in the call as well.

Huang also had a technical explanation to justify the capex commitments.

“In this new world of AI, compute equals revenues,” Huang said, a phrase that he repeated many times throughout the call. Huang argues that tokens, aka the chunks of data that AI models process, are the most important part of a new AI economy. The more tokens a model uses, the more computing power and time it requires. So, as models are getting more complex, the demand for computing is also going up “exponentially,” Huang said. He argued that the capex commitments will go towards building this compute capacity, which will thus power higher-level models and translate to revenue.

“The amount of token generation capability that the world needs is a lot, more than $700 billion, and I’m fairly confident that we’re going to continue to generate tokens…fundamentally because every single company depends on software, every software will depend on AI, and so every company will produce tokens,” Huang said. “If the new software requires tokens to be generated and the tokens are monetized, then it stands to reason that their data center build-out directly drives their revenues.”

Huang’s justifications may not have immediately convinced the market. Even though shares rose at first in response to the report, after the call, gains eventually pulled back to less than 1%. That’s despite revenue that exceeded market expectations.

OpenAI and China are still blind spots

Throughout the call, Huang also tried to address rumors of a falling out with OpenAI, first spurred after a $100 billion Nvidia investment announced back in September 2025 reportedly failed to progress beyond the early stages after months. Then, two back-to-back reports claimed that Huang was privately criticizing OpenAI’s business approach while OpenAI was unhappy with the inference speed of Nvidia’s chips.

In the call on Wednesday, Huang repeatedly praised the AI giant’s offerings, but revealed that the investment was still not finalized.

“We continue to work with OpenAI toward a partnership agreement, and believe we are close,” Huang said on the call. The filing also refuses to give any assurance that “a transaction will be completed.”

Another piece of uncertainty weighing on Nvidia is China. The company shared that, as of this month, the Trump administration has finally allowed it to start shipping small amounts of its H200 chips to China, where it once held 95% of the market share before Trump first banned the chipmaker’s sales to China, sparking a saga of dizzying trade tit-for-tat between the two global superpowers. But executives still don’t know if the imports will be allowed in, and are not factoring it into the revenue they expect this year.

[ad_2]

Ece Yildirim

Source link

February 25, 2026
Nvidia: Reports of an Elaborate Chinese GPU Smuggling Operation Are ‘Far-fetched’

[ad_1]

If some of Nvidia’s top-shelf GPUs—the physical artifacts currently at the center of the AI craze—hypothetically fell into the wrong hands, Nvidia’s next moves would have to placate a lot of parties, from shareholders to regulators to customers to China hawks in the Senate like Tom Cotton.

And a new report does say smuggled GPUs are now being used illegally by the Chinese company Deepseek, which, for someone like Cotton, would be like the One Ring being smuggled directly to Sauron. But for what it’s worth, Nvidia calls the details of the report “far-fetched.”

According to one of the tech news site The Information’s anonymously-sourced scoops, the Chinese AI company Deepseek is somehow training its latest models on Nvidia’s latest GPUs—ones built on the Blackwell architecture, pretty much the most in-demand pieces of technology in the universe. If that were true, one problem for Nvidia would be that giving companies in China access to the most advanced GPUs would be a violation of stringently enforced export rules—even after Trump moved to loosen restrictions earlier this week.

But don’t worry, China hawks. According to a company statement viewed by Yahoo Finance, the folks at Nvidia “haven’t seen any substantiation or received tips of ‘phantom data centers’ constructed to deceive us and our OEM partners, then deconstructed, smuggled and reconstructed somewhere else.”

Phew. That’s a very specific denial that really zeroes in on the details of the story, but it’s good to know that (deep breath) fake data centers created for the purpose of deceiving Nvidia or its unwitting suppliers or customers, which are dismantled, smuggled, and rebuilt somewhere in China, is something Nvidia hasn’t seen substantiated reports of, or received tips about.

“While such smuggling seems far-fetched, we pursue any tip we receive,” the Nvidia representative added, per CNBC.

And it’s true. It totally does sound farfetched if it’s not really happening. If it’s happening, the word for it is “ingenious.” In fact, it’s downright Now-You-See-Me-esque.

According to reports in May from this year, the lower-end prices of a single Blackwell GPU ranged from $6,500 to $8,000. That being the case, can you imaging the black market price? Such prices are a big part of why Nvidia is one of the rare AI companies that seem to consistently haul in money instead of just burning it, and are also why Nvidia bulls say the company is about to be worth $6 trillion.

And nothing hammers home the reasoning for an absolutely insane price tag on a piece of silicon quite like a cinematic (alleged) smuggling operation.

[ad_2]

Mike Pearl

Source link

December 10, 2025
Nvidia could be primed to be the next AWS | TechCrunch

[ad_1]

Nvidia and Amazon Web Services, the lucrative cloud arm of Amazon, have a surprising amount in common. For starters, their core businesses emerged from a happy accident. For AWS, it was realizing that it could sell the internal services — storage, compute and memory — that it had created for itself in-house. For Nvidia, it was the fact that the GPU, created for gaming purposes, was also well suited to processing AI workloads.

That eventually led to some explosively growing revenue in recent quarters. Nvidia’s revenue has been growing at triple digits, moving from $7.1 billion in Q1 2024 to $22.1 billion Q4 2024. That’s a pretty amazing trajectory, although the vast majority of that growth was in the company’s data center business.

While Amazon never experienced that kind of intense growth spurt, it has consistently been a big revenue driver for the e-commerce giant, and both companies have experienced first market advantage. Over the years, though, Microsoft and Google have joined the market creating the Big Three cloud vendors, and it is expected that other chip makers will eventually begin to gain meaningful market share, too, even as the revenue pie continues to grow over the next several years.

Both companies were clearly in the right place at the right time. As web apps and mobile began emerging around 2010, the cloud provided the on-demand resources. Enterprises soon began to see the value of moving workloads or building applications in the cloud, rather than running their own data centers. Similarly, as AI took off over the last decade, and large language models more recently, it coincided with the explosion in the use of GPUs to process these workloads.

Over the years, AWS has grown into a tremendously profitable business, currently on a run rate close to $100 billion, one that even separate from Amazon would be a highly successful company. But AWS growth has begun to slow down, even as Nvidia’s takes off. It’s partly the law of large numbers, something that will eventually affect Nvidia, too.

The question is whether Nvidia can sustain that growth to become a long-term revenue powerhouse like AWS has become for Amazon. If the GPU market begins to tighten, Nvidia does have other businesses, but as this chart shows, these are much smaller revenue generators that are growing much more slowly than the GPU data center business currently is.

Image Credits: Nvidia

The short-term financial outlook

As the above chart notes, Nvida’s revenue growth has been astronomical in recent quarters. And according to both Nvidia and Wall Street analysts, it’s set to continue.

In its recent earnings report covering the fourth quarter of its fiscal 2024 (the three months ending January 31, 2024), Nvidia told its investors that it anticipates $24 billion worth of revenue in its current quarter (Q1 FY25). Compared to its year-ago first quarter, Nvidia expects to post growth of around 234%.

That is simply not a number we often see from mature public companies. However, given the company’s massive revenue ramp in recent quarters, its growth rate is expected to decline. From a 22% revenue gain from the third to fourth quarter of its recently concluded fiscal year, Nvidia anticipates a more modest 8.6% growth rate from the final quarter of its fiscal 2024 to the first of its fiscal 2025. Certainly, on a year-over-year comparison and not a look back at just three months, Nvidia’s growth rate remains incredible for the current period. But there are other growth declines on the horizon.

For example, analysts expect Nvidia to generate $110.5 billion worth of revenue in its current fiscal year, up just over 81% from its year-ago results. That’s dramatically lower than the 126% gain it posted in its recently concluded fiscal 2024.

To which we ask: So what? For at least the next several quarters, Nvidia is expected to continue scaling its revenue past the $100 billion annual run rate mark, impressive for a company that in its year-ago period today saw total revenues of just $7.19 billion.

In short, analysts, and to a more modest degree Nvidia, see huge buckets of growth ahead for the company, even if some of the eye-popping revenue growth figures will slow this calendar year. It’s unclear what happens on a slightly longer timeframe.

Momentum ahead

It seems that AI could be the gift that keeps on giving for Nvidia for the next several years, even as more competition from AMD, Intel and other chipmakers begins to emerge. Much like AWS, Nvidia will face stiffer competition eventually, but it controls so much of the market right now, it can afford to cede some.

Looking at it purely at the chip level, not at boards or other adjacencies, IDC shows Nvidia firmly in control:

Image Credits: IDC

If you look at the board level with these market share numbers from Jon Peddie Research (JPR), a firm that tracks the GPU market, while Nvidia still dominates, AMD is coming on stronger:

Image Credits: Jon Peddie Research

C Robert Dow, an analyst at JPR, says some of these fluctuations have to do with when new products are introduced. “AMD gains percentage points here and there depending on cycles in the market — when new cards are introduced — and inventory levels, but Nvidia has been in a dominant position for years, and that will continue,” Dow told TechCrunch.

Shane Rau, an IDC analyst who follows the silicon market, also expects the dominance to continue, even as trends shift and change. “There are trends and countertrends, the markets in which Nvidia participates are big and getting bigger, and growth will continue, at least for another five years,” Rau said.

Part of the reason for that is Nvidia is selling more than just the chip itself. “They’ll sell you boards, systems, software, services and time on one of their own supercomputers. So any of those markets are big and growing and Nvidia is attached to all of them,” he said.

But not everyone sees Nvidia as an unstoppable force. David Linthicum, a longtime cloud consultant and author, says that you don’t always need GPUs, and companies are beginning to realize that. “They say they need GPUs. I look at it, do some of the back of the envelope math, and they don’t need them. CPUs are perfectly fine,” he said.

As this happens, he thinks Nvidia will begin to slow down and competition will loosen its stronghold on the market. “I think that we’re going to see Nvidia morph into a weaker player over the next couple of years. And we’re going to see that because there’s too many substitutes that are being built out there.”

Rau says other vendors will also benefit as companies expand AI use cases with Nvidia products. “What I think you’ll see going forward is growing markets that’ll create tailwinds for Nvidia. But then there’ll be other companies that also follow in those tailwinds that will benefit from AI particularly.”

It’s also possible that some disruptive force will come into play and that would be a positive outcome to keep one company from becoming too dominant. “You almost hope disruption will happen because that’s the way markets and capitalism work best, right? Someone gets an early lead, other suppliers follow, the market grows. You get established players, who are eventually disrupted by a better way to do the same thing within their market or within adjacent markets that are crossing into theirs,” Rau said.

In fact, we are beginning to see that happening at Amazon as Microsoft gains ground via its relationship with OpenAI and Amazon is forced to play catch-up when it comes to AI. Whatever happens to Nvidia in the long run, it’s firmly in the driver’s seat right now, making money hand over fist, dominating a growing market and having just about everything going its way. But that doesn’t mean it will always be this way or that there won’t be more competitive pressure down the road.

[ad_2]

Ron Miller

Source link

March 24, 2024
The chip industry can’t keep up with the A.I. revolution

[ad_1]

Everyone is talking about chips again, thanks to A.I and a rosy forecast from Nvidia. The news drove investors to flock to A.I.-related stocks to the tune of $300 billion in added value last month.

But all this optimism shouldn’t distract us from one of the chip industry’s key problems: Chips have stopped providing real jumps in computing power, right as we see an explosion of power-hungry applications like generative A.I.

Historically, the computing power of chips has doubled every two years in what became to be known as “Moore’s Law.” But we haven’t seen that jump in performance for a while. Now, a microprocessor’s performance increases by only about 10-15% each year—and the actual increase in speed for a given software application is often much smaller. And the process of rearchitecting software for these chips can be expensive and buggy.

This slowdown could not have come at a worse time. Chips are simply not able to keep up with some of the most computation-intensive applications yet seen. The size of models used for tasks like computer vision, natural language processing, and speech processing has increased by 15 times in just two years, an order of magnitude higher than the increase in computer power in chips over the same period. The most advanced machine learning models, like those that power GPT-4 and ChatGPT, have increased by 75 times, again much more than the power of the graphics processing units (GPUs) that underlie them.

The gap between what’s needed and what’s provided can only be filled by more chips. And that’s making computing expensive for everyone. It’s now so expensive to build advanced machine learning models that they are now the exclusive domain of rich, powerful corporations.

Why are chips lagging so far behind?

There are technical challenges. It’s hard to make chips smaller than they already are—transistors, at their thinnest dimension, are only a few atoms thick.

But it’s a partial explanation at best. Chips haven’t kept up with the needs of contemporary applications for quite some time—and even on the best of days, improvements in chip speed have lagged improvements in software algorithms.

A better reason is that the chip industry has not been all that innovative, especially recently. Microprocessors have worked in more or less the same way for 80 years, even as devices get smaller. We haven’t changed how we use computer memory in decades. And the GPUs that power advanced machine learning also haven’t changed much in the past 10 years.

Slowing miniaturization is exposing the lack of disruptive ideas in the industry. No chip company appears in recent lists of innovative companies. And the unchanging ranks at the top of the industry suggest an oligopoly.

Innovation needs an ecosystem where companies, typically startups, want to experiment in the hopes of a breakout success.

The chip industry doesn’t have many of those experiments.

First, the cost of experimentation is extremely high. It often takes $10-30 million just to get the first product, and another $70-100 million to scale up. These extraordinarily large sums of money discourage risk-taking, entrepreneurship, and funding. As a result, not many chip startups are formed, and the few that get funding come from teams of seasoned chip veterans. This recipe leads to incrementalism, not disruption.

Second, the gestation period for new ideas is too long. It typically takes a few years before the first product sample is created and it may take just as long again to see revenue. This long period, again, discourages both innovators and investors that typically prefer to “fail fast”.

Third, the chip industry is too consolidated, dropping from 160 companies in 2010 to 97 in 2020. A lack of buyers constrains the size of exits, further discouraging investors.

Chips attract less than 1% of total U.S. venture capital investment, despite the emergence of A.I., the Internet of Things, electric cars, and 5G.

Finally, the chip industry may not be attracting talent. Today’s STEM graduates see better prospects in industries with much faster growth (including, perhaps ironically, A.I.). This chip industry also has a branding problem—even chip industry executives agree that the sector has a weak brand. Young employees and future innovators want to tinker with software more than struggle with new hardware.

The U.S. government must use its CHIPS Act funding as a lever to make the chip industry more welcoming to innovation.

To bring down the cost and time needed for a new idea, the government should require recipients of government money to allocate money to more agile methods of hardware methodologies, open-source tools, and open standards. It should require recipients to make commonly-used hardware components widely accessible at a low cost, so that other companies can combine them with innovative components to create new hardware platforms cheaply and quickly.

Academic CHIPS Act beneficiaries should be required to modernize chip design curriculums to emphasize accessibility and impact.

The government should allocate some National Science and Technology Council funding to develop a shared, subsidized infrastructure for design and fabrication with mature, trailing technologies to reduce the cost of producing proof-of-concept hardware.

And, finally, it can encourage the passage of right-to-repair legislation to help stimulate a culture of tinkering with hardware.

The chip industry has managed to mask its struggle with disruptive innovation for some time. And as miniaturization comes to a close, or at least loses its effectiveness, it’s time to address the innovation head on. Technological progress—literally—depends on it.

Rakesh Kumar is a Professor in the Electrical and Computer Engineering department at the University of Illinois and author of Reluctant Technophiles: India’s Complicated Relationship with Technology.

The opinions expressed in Fortune.com Commentary pieces are solely the views of their authors, and do not reflect the opinions and beliefs of Fortune.

[ad_2]

Rakesh Kumar

Source link

June 12, 2023