At the GTC developer’s conference on March 16, 2026, Jensen Huang, the CEO of Nvidia, stated on Monday that he anticipates purchase orders for two of the company’s major AI chip technologies totaling up to $1 trillion, which is twice as much as the $500 billion estimate from the previous year.
The stock bettered 1.63% to a peak of the day prior to stabilizing at the higher gains of 1.8% reflecting the investor excitement in the market despite the high levels of competition.
Record Demand Fuels Growth
Huang also highlighted high confidence orders that had already been cemented when he delivered his speech on the California stage.
The statement;
Now, I don’t know if you guys feel the same way, but $500 billion is an enormous amount of revenue.
He said.
Well, I’m here to tell you that right now where I stand a few short months after GTC DC, one year after last GTC right here where I stand, I see through 2027, at least $1 trillion.
The update consolidates Nvidia as a dominant force in AI data-center infrastructure, where its finders of the GPUs support the significant technology spend allocation.
Even though the market capitalization of the corporation is the largest in the world market, the shift away from the field of AI training to real-life inference, when models perform the tasks of users is an important strategic change.
Inference Race Heats Up
Nvidia is facing growing competition in the AI hardware market by companies like Advanced Micro Devices, Intel, and mainstream technology giants like Google and Meta Platforms, both of them also designing their own in-house silicon.
With the development of AI systems, the focus has now moved on to the mobilization of the models rather than their education, which is also known as inference. Such development requires inference workload-specific processors instead of depending solely on traditional graphics processing units.
To manage this shift in the paradigm, Nvidia has also been exploring technologies that would simultaneously lower the cost and latency, at an equal rate, without compromising high-performance.
The acquisition of Groq that is under consideration is an example of a strategic move towards increasing inference computing opportunities.
The architecture of Groq is focused on fast processing of data in terms of simplified execution pipelines, and its latency is significantly lower when compared to conventional models with GPUs.
Analysts argue that the orchestration of this technology into the system that has been present in the Nvidia ecosystem would make the inference computing to be both efficient and cost effective.
Looking Ahead
Speculation shows that in the future Nvidia may create a new generation of inference based chips that integrate the architectural ideas of Groq with the commonly used CUDA platform.
This sort of development would solidify the market presence of Nvidia because developers can maintain compatibility with already existing CUDA-based platforms, but enjoy the performance of dedicated inference engines.
In case this approach turns out to be successful, it may enable Nvidia to reserve a pivotal place in the AI computing industry despite the increasing competition and the tendency of companies to design custom silicon architecture.
