Nvidia Breaks Record After Record
Nvidia Breaks Record After Record
One of the most discussed questions in the artificial intelligence industry in recent months was: Is this growth really sustainable? Big tech companies are spending hundreds of billions of dollars on data centers, AI labs are demanding more processing power with each new model, and Nvidia stands at the center of almost all of this transformation. It was expected that such a fast-growing market would slow down at some point.
Nvidia's latest financial results say the opposite.
The company announced that it earned $96.2 billion in revenue in the second quarter results announced on August 26, 2026. This figure means an increase of 18 percent compared to the previous quarter and 106 percent compared to the same period last year. Even more striking, quarterly revenue for the data center business alone reached $89 billion. In other words, most of Nvidia's current economic weight no longer comes from classical graphics cards, but from its artificial intelligence infrastructure.
It's not just the record revenue that makes the story interesting here. Nvidia management is making clear that it thinks the growth is not a temporary wave. The company expects revenue of $108 billion for the third quarter. According to Reuters, Nvidia also expects its revenues to grow by approximately 70 percent in the next fiscal year, which ends in January 2028. That rate is well above Wall Street's expectation of about 44 percent and is an unusually high indicator of long-term confidence for a company like Nvidia, which typically only gives near-term forecasts.
Nvidia Is No Longer a Graphics Card Company
Many users still know Nvidia for its GeForce graphics cards. The gaming industry continues to be an important part of the company's brand memory. However, when we look at the financial table, we see that today's Nvidia has turned into a completely different company.
Data center revenue in the second quarter was $89 billion, an increase of 117 percent compared to the same period last year. With total revenue at $96.2 billion, the overwhelming majority of Nvidia's revenue now comes from its AI computing infrastructure.
This transformation is quite remarkable in the history of technology. The company, which once produced GPUs for gamers, has now turned into a platform company that provides basic hardware to the infrastructure of the world's largest AI laboratories, cloud providers and technology companies.
GPU doesn't just produce images anymore.
The model trains.
The model works.
The robot controls.
He does scientific simulation.
It produces videos.
It feeds the systems that write code.
And at the heart of it all is processing power.
The most striking statement used by Jensen Huang while explaining the latest results summarizes this transformation: He says that AI is now doing “useful work” and the tokens it produces create economic value. In his words, “compute is revenue”, that is, processing power now directly means income.
This sentence actually explains Nvidia's entire investment thesis for the future in a single line.
If 89 Billion of 96 Billion Dollars Comes from Data Centers, Something Has Changed
A few years ago, AI GPUs were considered niche products of the technology industry. Today, data center systems are Nvidia's main business.
ChatGPT-like models are not the only ones behind this change.
Nvidia's customer base is gradually expanding. Major cloud companies, frontier AI labs, startups, government-backed AI infrastructures, industrial companies and robotics firms are investing in the same computing resources.
According to Reuters, by 2026, major technology companies' total spending on AI infrastructure is expected to exceed $730 billion. This figure is more than double the previous year.
That's why it's becoming increasingly difficult to explain Nvidia's growth solely by "AI hype."
There are real data centers out there.
There is real energy consumption.
There are real servers.
They have real network infrastructures.
And companies are trying to build billions of dollars worth of services on these infrastructures.
Therefore, a significant part of the AI discussion has shifted to physical infrastructure rather than software.
What is Vera Rubin and Why is She So Important?
Nvidia's next big growth engine is the Vera Rubin platform.
Vera Rubin is not a single GPU name. A new generation AI platform that considers CPU, GPU, networking, storage and system architecture together. Nvidia is positioning the system specifically for agentic AI and its growing AI factories.
The company announced that Vera Rubin started full production in May 2026. The platform is planned to be deployed on major infrastructure providers such as CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure and Nebius. According to Nvidia's official statement, the system aims to deliver up to 10 times higher agent throughput in large-scale agentic AI workloads compared to the previous Grace Blackwell platform.
In its second quarter results, Nvidia reaffirmed that Vera Rubin systems are rapidly entering production and that working rack systems are available at partners.
The point here is that Nvidia isn't just trying to sell faster GPUs.
Redesigning the entire data center.
What is the Concept of AI Factory?
Recently, Nvidia has been increasingly using the term AI factory instead of the classic "data center".
The task of the classic data center was to store data and run applications.
The logic of the AI factory is slightly different.
Data is coming in.
The model handles this.
And the system produces tokens, images, videos, codes, decisions or other digital output.
Nvidia sees this almost like industrial production.
Electricity + data + GPU → intelligence.
This perspective makes it easy to understand why Nvidia directly links processing power to revenue.
If AI agents are doing research, providing customer service, writing code, creating advertisements or managing the production line on behalf of the company, the computational capacity used is no longer just an IT cost.
Production input.
And Nvidia is positioning itself as the machine maker of this new economy.
Vera CPU is also a part of this strategy
The “Vera” in the name Vera Rubin is not a coincidence.
Nvidia also announced the Vera CPU in May 2026. The company describes this processor as a new generation CPU designed especially for AI agents, reinforcement learning and data processing. According to Nvidia, Vera can complete tasks up to 1.8 times faster compared to classic x86 processors in certain workloads.
More importantly, companies planning to use Vera include Anthropic, OpenAI, and other major AI labs. System manufacturers such as Dell, HPE, Lenovo and Supermicro are also preparing Vera-based infrastructures.
This shows that Nvidia no longer wants to remain strong only in the GPU field.
CPU.
GPU.
Networking.
Storage.
Software.
AI framework.
He wants to present them all within the same system.
In other words, the company no longer sells chips.
It is trying to sell its entire AI infrastructure.
Why Did Nvidia Come to Such a Strong Position?
Software plays as much of a role here as hardware.
One of Nvidia's biggest advantages is the CUDA ecosystem.
For years, researchers, developers, and companies have been developing GPU-based computing applications on CUDA. Therefore, the company is not competing only with the physical chip.
It competes with a huge software and developer habit that has been built over the years.
This transforms Nvidia from a classic hardware manufacturer into a platform company.
A competitor might make a faster or cheaper AI chip.
But getting the customer to change the entire software infrastructure is much more difficult.
Nvidia is trying to create a similar effect on the AI compute side that Apple's ecosystem advantage has in the mobile world.
But Rivals Are Not Standing Idle
While Nvidia's growth has been phenomenal, the market is not completely unchallenged.
AMD is trying to expand in the AI accelerator market.
Intel is developing its own AI solutions.
Google uses TPU.
Amazon is developing its own Trainium and Inferentia chips.
Microsoft is working on Maia.
Meta develops custom AI hardware.
Almost all major technology companies are investing in special silicon projects that can reduce their dependence on Nvidia.
Reuters also states that investors especially watch this custom-chip trend as one of the long-term risks for Nvidia.
But for now, the market is growing so fast that the emergence of alternative chips is expanding Nvidia's total AI compute capacity rather than reducing its sales.
So the pie is expanding.
China Still a Great Uncertainty
One of the important risks facing Nvidia is China.
US export restrictions on advanced AI chips are affecting the company's data center operations in China. Nvidia specifically states that it does not include data center compute revenue from China in its $108 billion revenue estimate for the third quarter.
This is important.
Because the company expects record growth even without taking China revenue into account.
On the other hand, the limitation of a giant market like China in the long term may create serious opportunity costs for Nvidia.
Therefore, the company's future is shaped not only by technology but also by geopolitics.
Growth Comes at a Cost: Memory and Energy
The growth of AI data centers isn't just driving demand for GPUs.
HBM memory.
Electricity.
Cooling.
Fiber connections.
Networking equipment.
Data center land.
There is an explosion of demand across the entire infrastructure chain.
For this reason, reports emerged that Nvidia customers may receive price increases of more than 15 percent in next-generation systems. One of the important reasons for these increases is the rise in high-performance memory prices.
In other words, scaling artificial intelligence is not just a "let's produce more chips" problem.
The entire physical infrastructure needs to grow together.
And perhaps this is where one of the AI industry's long-term frontiers will emerge.
Compute can be found.
What about electricity?
AI Bubble?
This question still remains.
Nvidia's financial results show that AI demand is indeed there right now. There's no doubt that companies are ordering billions of dollars' worth of systems and data centers are expanding rapidly.
But there is another question:
Will AI products built on this infrastructure be able to return the invested money?
The answer on Nvidia's side is quite optimistic for now.
Jensen Huang says AI tokens are now “productive and profitable”. So models are not just experimental technology; It produces real economic jobs.
But it may be too early to say the same for the entire market.
Not every AI startup will be successful.
Not every data center investment will provide the same return.
It would not make economic sense for each company to train its own model.
Therefore, Nvidia's growth does not prove that all companies in the AI sector will be successful.
Rather it shows:
The gold rush is still going on and Nvidia is the company selling shovels.
The Real Paradox: Nvidia Wins as AI Companies Compete
OpenAI is developing a new model.
Anthropic is developing a larger model.
Google is expanding Gemini.
Meta invests in Llama.
New laboratories are being established.
All of these companies compete with each other.
But they all have common needs:
More compute.
Therefore, Nvidia's market may grow as competition among AI model companies increases.
Competing labs are using more GPUs to outdo each other.
New benchmark.
New model.
New agent.
New video model.
New reasoning system.
All of them create infrastructure needs.
This business model puts Nvidia at a very interesting point in the AI race.
Which model is the winner can sometimes be secondary for Nvidia.
The important thing is that the race continues.
“Agentic AI” Era with Vera Rubin
It is no coincidence that the new platform was designed specifically for agentic AI.
Today's chatbots mostly answer questions when asked.
Agentic AI, on the other hand, takes on longer missions.
He's investigating.
It uses the web.
The code is running.
It creates a file.
It checks for errors.
He tries again.
It communicates with other systems.
Such tasks consume much more tokens and computing power than a simple chatbot response.
This is pretty good news for Nvidia.
As AI becomes more capable, the need to compute per user does not have to decrease.
On the contrary, it may increase.
Having an AI agent work on your behalf for half an hour requires much more action than answering a single question.
The economic side of Nvidia's "AI is turning into useful work" statement is exactly here.
What Does This Mean for the Creative Industry?
For Voldi Creative, what makes Nvidia's financial results interesting is that they show how fast the infrastructure behind the creative industry is growing, rather than stock market numbers.
Many tools we use today are based on GPU infrastructure.
Generative video.
Image generation.
AI voice.
3D production.
Upscaling.
Generative fill.
Motion.
AI coding.
There are huge computing systems behind these.
While we type a prompt and wait a few seconds, hundreds or thousands of GPUs are running elsewhere.
So AI “cheaper” creative production is a bit misleading.
The cost for the user decreases.
But on the infrastructure side, extraordinarily large capital investments are being made.
Nvidia's $96 billion quarter is one of the clearest indicators of this.
Why Are AI Tools Keep Getting Better?
One reason for this is model development.
But the other reason is compute.
When more computational resources are given to a video model, higher resolution, longer duration, better physics or more consistent characters are possible.
Image generation is the same way.
LLMs are the same way.
Therefore, the rapid increase in quality we see on the creative side is not just a “better algorithm”.
There are also more GPUs.
Therefore, Nvidia's financial growth indirectly tells us where creative software is heading.
If AI video produces 10 seconds today, it may produce longer tomorrow.
If character consistency is a problem today, it may decrease after a few generations.
If high resolution is expensive today, it can be standardized as the infrastructure grows.
Nvidia is Now an Indicator of the Technology Sector
For a while, Intel's financial results were monitored to understand the health of the PC industry.
Then Apple's iPhone sales showed the state of the mobile economy.
Today, Nvidia's results have turned into a similar indicator to understand the appetite of the AI industry.
If the company expects a third quarter revenue of $108 billion, it is not only Nvidia's success that is behind this.
Google has an investment.
Microsoft has an investment.
Amazon has an investment.
AI laboratories have investments.
States have investments.
There are new data centers.
So Nvidia's balance sheet is a summary of how much money the world spends on artificial intelligence.
And for now there is no slowing down
It has been frequently discussed that AI investments may slow down throughout 2026, major tech companies may begin to cut spending, and GPU demand may have peaked.
Nvidia's latest results weaken this expectation, at least in the short and medium term.
Quarterly revenue of $96.2 billion.
$89 billion in data center revenue.
108 billion dollars is expected for the next quarter.
And a growth forecast of around 70 percent for the next fiscal year.
Moreover, Vera Rubin is just coming into play on a large scale.
These figures show that the AI race is not over.
Perhaps more importantly, it shows that the race has entered a new phase.
In the first stage, everyone was making models.
Now, industrial-scale infrastructure is being established for these models.
And Nvidia is in the middle of this transformation.
Maybe years later, when we look at the technology history of the 2020s, we will remember Nvidia not only as "the incredibly successful company selling AI chips."
We will see it as one of the companies that establish the infrastructure of a larger transformation.
Because the most important raw material of the artificial intelligence economy today may be data.
But what turns it into economic value is still another source:
computing power.
And right now, there is no other company in the world that monetizes this as well as Nvidia.
Nur Oğuz