
During his participation in Foxconn’s Keynote at Computex 2025Jensen Huang, CEO of Nvidia, made a statement that must move the foundations of the semiconductor industry: according to him, Moore’s Law no longer applies to the current technological evolutionas did it in January.
Instead comes a new logic – which he called informally (and boldly, let’s face it Lei de Huangbased on structural advances that go beyond the miniaturization of transistors.
The speech came in response to a question about the speed with which AI chips have evolved. Huang explained that performance gains today are no longer determined by reducing the size of the nodes, but by Complementary packaging technologies and interconnectionallowing to scale Datacenter computing with results previously considered unattainable.
Why Moore’s law is no longer enough
Formulated in 1965 by Gordon Moore, co -founder of Intel, the so -called Moore Law noted that the number of transistors in a chip would double every two years, which would imply proportional growth in performance. For decades, this trend has sustained advances in the information technology sector.
“The shrinkage of the knot no longer delivers the gains it delivered. Today we got much more connecting different platforms on a large scale than just reducing transistors”
However, according to Huang, this metric has lost practical validity. He states that the physical limits of miniaturization were increasingly evidentand the gains by area became marginally smaller.
This observation is accompanied by a context: Nvidia has accelerated the rhythm of releases.
While industry used to operate in annual or biennial cycles, the company already works with a Calendar of launches every six monthsand according to the executive, there are concrete plans to reduce this interval to three months soon.
The new performance engines
The NVIDIA CEO attributed the recent jumps to two technical pillars. The first is the use of Advanced packaging technologieslike the CoWoS (Chip-on-Wafer-on-Substrate) from TSMC. Such an architecture allows for greater component density and large -scale thermal optimization.
The second pillar is the NVLinkNVIDIA proprietary technology that allows Connect multiple chips as if they were a single cohesive system. With this, the company can extract performance superior to the summar processors in interconnected racks, especially in artificial intelligence applications.
Today, performance does not grow only with silicon. It grows with systems. With interconnections. With architectures designed to operate on scale
Jensen HuangNVIDIA CEO
The statement suggests that the company is redefining the innovation parameters, and that next generations of chips will not only depend on the evolution of lithographybut from a broader and more systemic approach.

Frequency of releases and went to the center of the strategy
Huang’s speech also signals a Nvidia’s strategic repositioning in the AI sector. Products such as Vera Rubin and Blackwell Ultra, separated by just six months, show a new cadence: a pace that challenges traditional hardware development cycles.
With the continuous expansion artificial intelligence market, the company seeks to maintain relevance by delivering new products to increasingly short windows. The statement that performance can double in just one quarter points to a aggressive model of incremental evolutionsomething rarely seen before on an industrial scale.
Also read:
What to expect from now on
Jensen Huang’s lines consolidate Nvidia’s vision for the future: more than following the evolution of technology, the company wants to lead a new stage, redefining how the sector measures performance and progress.
This “Huang Law”, still informal, represents a key turn in contemporary technological discourse. If before the focus was on increasingly smaller transistors, now the focus is on ability to design and integrate complete systems with intelligence, modularity and scalability.
The impact of this goes beyond the laboratories: it reaches directly as data centers, companies and AI developers will plan their next decades.
If Moore’s law was northern computing in the 20th century, it is possible that Huang’s law will become the guide to the next generations.
Fonte: Taiwan Economic Daily

Join the Adrenaline offers group
Check out the main offers of hardware, components and other electronics we find over the internet. Video card, motherboard, RAM and everything you need to set up your PC. By participating in our group, you receive daily promotions and have early access to discount coupons.
Enter the group and enjoy the promotions
Source: https://www.adrenaline.com.br/hardware/ceo-nvidia-afirma-lei-de-moore-morreu-lei-huang/