Jensen Huang’s Latest Article: AI’s “Five-Layer Cake”
TechFlow Selected TechFlow Selected
Jensen Huang’s Latest Article: AI’s “Five-Layer Cake”
The true significance of AI lies not merely in smarter software, but in an infrastructure revolution on a scale comparable to electricity and the internet.
Navigating Web3 tides with focused insightsBy: Jensen Huang
Translated by Peggy, BlockBeats
Artificial intelligence is one of the most powerful forces shaping our world today. It is not merely a clever application or a single model—it is infrastructure, as essential as electricity and the internet.
AI runs on real hardware, real energy, and real economic systems. It transforms raw materials into scalable “intelligence.” Every company will use it; every country will build it.
To understand why AI unfolds in this way, it helps to return to first principles—to examine the fundamental shifts occurring across the computing landscape.
From “Pre-Built Software” to “Real-Time Generated Intelligence”
For most of computing history, software has been “pre-built.” Humans first describe an algorithm, and computers execute instructions accordingly. Data must be meticulously structured, stored in tables, and retrieved via precise queries. SQL is indispensable precisely because it enables this entire system to function.
AI breaks this paradigm.
For the first time, we have computers capable of understanding unstructured information—capable of seeing images, reading text, listening to sound, and comprehending meaning; capable of reasoning about context and intent. More importantly, they can generate intelligence in real time.
Every response is newly generated. Every answer depends on the context you provide. This is no longer software retrieving pre-existing instructions from a database—it is software performing real-time inference and generating intelligence on demand.
Because intelligence is generated in real time, the entire computational technology stack supporting it must also be reinvented.
AI as Infrastructure
Viewed through an industrial lens, AI can be decomposed into a five-layer stack.
Energy
The foundational layer is energy.
Real-time intelligence generation requires real-time electricity generation. Each token produced corresponds to electrons moving, heat being managed, and energy being converted into compute capacity.
Beneath this layer, there is no abstraction. Energy is the first principle of AI infrastructure—and the fundamental constraint determining how much intelligence a system can produce.
Chips
Above energy sit chips. These processors are designed to convert energy into compute capacity with exceptional efficiency and at massive scale.
AI workloads demand enormous parallel compute capability, high-bandwidth memory, and ultra-fast interconnects. Progress at the chip layer determines how quickly AI can scale—and how cheaply “intelligence” ultimately becomes.
Infrastructure
Above chips lies infrastructure: land, power delivery systems, thermal management, civil engineering, networking systems, and orchestration software that integrates tens of thousands of processors into a single machine.
These systems are, in essence, AI factories. They are not designed for storing information—but for manufacturing intelligence.
Models
Above infrastructure sit models. AI models can understand diverse domains: language, biology, chemistry, physics, finance, medicine—and reality itself.
Language models represent only one category. Some of the most transformative work is unfolding in areas such as protein AI, chemical AI, physics simulation, robotics, and autonomous systems.
Applications
At the top sits the applications layer—the place where real economic value is created. Examples include drug discovery platforms, industrial robots, legal copilots, and autonomous vehicles.
An autonomous vehicle is essentially an “AI application carried by a machine”; a humanoid robot is an “AI application carried by a body.” The underlying technology stack is identical—the final form simply differs.
Thus, AI’s five-layer structure is: Energy → Chips → Infrastructure → Models → Applications. Every successful application pulls downward across all layers—right down to the power plant supplying its electricity.
An Infrastructure Buildout Still in Its Early Stages
We are only just beginning this buildout. Current investment totals only several hundred billion dollars—while future infrastructure requirements will reach multiple trillions of dollars.
Globally, we are witnessing the construction of chip fabs, computer assembly plants, and AI factories—built at unprecedented scale. This is becoming one of the largest infrastructure projects in human history.
Labor Demand in the AI Era
The labor force required to support this buildout is immense.
AI factories require electricians, plumbers, pipefitters, structural steel workers, network technicians, equipment installers, and operations personnel.
These are highly skilled, well-compensated roles—and currently in acute shortage. Participating in this transformation does not necessarily require a Ph.D. in computer science.
Meanwhile, AI is boosting productivity across the knowledge economy. Take radiology, for example. AI is already assisting in medical image interpretation—yet demand for radiologists continues to grow.
This is not contradictory.
A radiologist’s true responsibility is patient care; image reading is only one component of that work. As AI takes over more repetitive tasks, physicians can devote more time to judgment, communication, and treatment.
Hospitals become more efficient, enabling them to serve more patients—and thus requiring more staff. Productivity creates capacity; capacity drives growth.
What Changed Over the Past Year?
Over the past year, AI crossed a critical threshold.
Models have become sufficiently capable to deliver real-world impact at scale.
- Significant improvement in reasoning ability
- Marked reduction in hallucinations
- Dramatically enhanced grounding in the real world
For the first time, AI-based applications are generating tangible economic value.
Clear product-market fit has emerged in fields including drug development, logistics, customer service, software development, and manufacturing.
These applications are strongly pulling demand across the entire underlying technology stack.
The Role of Open-Source Models
Open-source models play a pivotal role. The vast majority of AI models worldwide are freely available. Researchers, startups, enterprises—and even entire nations—rely on open-source models to compete in advanced AI.
When open-source models reach the technological frontier, they do more than reshape software—they activate demand across the entire technology stack.
DeepSeek-R1 exemplifies this. By making a powerful reasoning model widely accessible, it accelerated rapid growth at the application layer—and simultaneously increased demand for training compute, infrastructure, chips, and energy.
What Does This Mean?
When you view AI as infrastructure, everything becomes clear. AI may have begun with Transformers and large language models—but it is far more than that.
It is an industrial-scale transformation reshaping:
- How energy is produced and consumed
- How factories are built
- How work is organized
- How economic growth occurs
AI factories are built because intelligence can now be generated in real time. Chips are redesigned because efficiency dictates the speed of AI scaling. Energy is central because it determines the maximum amount of intelligence a system can produce. Applications explode because models have finally crossed the threshold of “scale-ready usability.”
Each layer reinforces the others.
That is why this buildout is so massive, why it impacts so many industries simultaneously, and why it cannot be confined to any single country or domain.
Every company will use AI.
Every country will build AI.
We remain in the early stages.
Vast amounts of infrastructure remain unbuilt, vast numbers of workers remain untrained, and vast opportunities remain unrealized.
Yet the direction is unmistakably clear.
Artificial intelligence is becoming foundational infrastructure for the modern world.
And the choices we make today—how fast we build, how broadly we participate, and how responsibly we deploy—will determine what kind of era this ultimately becomes.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
@nvidia
