On November 30, 2025, Indian entrepreneur Nikhil Kamath interviewed Elon Musk on the podcast “People by WTF”.
There was no press conference, no PPT, just an in – depth conversation about AI, currency, and the future of work.
Facing the US national debt that has exceeded $38 trillion, Musk gave a clear judgment:
The only solution is for AI and robots to boost productivity, and they must outpace inflation within three years.
When asked “which companies investors should focus on”, he named NVIDIA and Google.
But throughout the conversation, what he really focused on was not specific companies, but the path of wealth flow: from chips to platforms, from platforms to systems, and from systems to entrances.
Section 1 | NVIDIA is great, but it’s not the end
Musk took NVIDIA and Google as examples, but what he really emphasized was not the chip performance, but a deeper logic: in the early stage of AI, wealth was concentrated on the computing power supply side, and in the next stage, it will flow to platform – type companies that can build a complete AI ecosystem.
1. Chips: A necessity, not a moat
GPU computing power is the infrastructure of AI, which has made NVIDIA the company with the highest market value in the world. But in Musk’s value map, chips are just a tool layer.
What he really focuses on is the higher – level capabilities: systems, platforms, closed – loops, and interfaces. These terms do not refer to single – point technologies, but to how to connect data collection, model training, inference deployment, and scenario applications.
This complete connection ability can be understood as an “AI factory”: a complete production line from raw materials to finished products, rather than just equipment for a certain link.
2. The value of platforms: Data, scenarios, and closed – loops
When talking about Tesla, Musk said that it is the world’s leading real – world AI because it not only has models, but more importantly, data, scenarios, and feedback form a closed – loop. When talking about the X platform, he also emphasized that AI must be embedded in real user behaviors, interaction data, and decision – making scenarios.
The common point between Tesla and X is the core value of the platform:
Continuous data generation
Continuous model iteration
Natural user retention
The same logic also applies to Google. The data and user habits it has accumulated in search, maps, and YouTube are the real reasons why Musk thinks it will be quite valuable in the future.
3. Reconstruction of value distribution
From this perspective, Musk’s division is very clear:
“Chip companies provide tools, and their profits come from sales; platform companies build ecosystems, and their value comes from the network effect. The former is a business with high gross profit, and the latter is an asset with high barriers.”
This also explains why xAI started to build its own inference cluster from the beginning, rather than relying on cloud APIs.
Only by fully controlling the entire process from data to deployment can one control the value.
Section 2 | From conversation to execution: The next leap of AI
The most frequently used word in Musk’s conversation is: system.
What’s the difference between this word and “platform”? A platform controls data and users, while a system controls tasks and execution.
For example: ChatGPT is an excellent conversation tool, but it can only answer questions. What Musk wants is for AI to be able to post tweets on your behalf, track interactions, adjust strategies, and even complete a series of coherent actions.
In other words, the next – stage competition of AI does not lie in the strength of the model, but in who can build a complete task – execution system.
Take the X platform as an example:
The user says “Post a tweet about a certain topic”
The AI understands the intention, generates content, and calls the publishing interface
Tracks interaction data and gives optimization suggestions
This is not a multi – round conversation, but a goal – driven task – execution chain. What is needed is not a better chat box, but a workflow system that can schedule multiple modules and connect multiple steps.
The core capabilities of this type of system include:
Receiving target instructions rather than simple questions
Scheduling multiple AI modules and external interfaces
Continuously executing and giving status feedback
Why does Musk emphasize the integration of SpaceX, Tesla, and xAI? Because future AI is not a single – point tool, but a collaborative system across data, scenarios, and hardware. Tesla has driving data and in – vehicle hardware, SpaceX has satellite networks and space computing power, and xAI provides model capabilities. Only by integrating the three can a complete closed – loop from data collection to inference deployment be built.
This closed – loop ability is not available in the tool calls of OpenAI and Anthropic. They still stay at the level of ability output, rather than system operation.
The first – layer transfer is from chips to platforms, and the second – layer is from platforms to systems. The former controls data and users, and the latter controls tasks and closed – loops.
Whoever can make AI not only answer questions but also complete work will master the value entrance
Section 3 | WeChat++: Musk’s ambition for the entrance
What does the entrance look like? Musk’s answer is: WeChat++.
In China, WeChat is a unified platform for people to send messages, make payments, call taxis, order food, and manage finances. Most digital behaviors in life are completed in a super app. What Musk wants to do is an AI – upgraded version of this model.
X is not just a social media, but a unified entrance in the AI era.
1. From WeChat to WeChat++: The value of a unified entrance
When Musk first created X.com, his goal was to establish a “clearinghouse for financial transactions” and a “more efficient currency database”. After acquiring Twitter, he said it was an opportunity to re – examine this vision.
The current functions of X are:
Text, picture, and video posting
Secure messaging and audio – video calls
Automatic translation (connecting users of different languages)
The Grok AI assistant
Payment function will be added in the future
Integrating these functions together, X has mastered the complete behavioral context of users. AI no longer faces fragmented single – time requests, but understands a person’s complete digital life.
This is the technological implementation of what Musk calls “bringing the world together into a collective consciousness”.
2. The value of the AI entrance: Not just traffic, but control
In the traditional Internet era, the value of a platform was to aggregate traffic. But in the AI era, the value of the entrance has been upgraded to:
Context control: Mastering the user’s complete context and needs
Call authority: Deciding which AI to use and which services to call
Execution ability: Making AI not only answer, but also complete tasks on behalf of users
If X can achieve this vision, it will become:
A training ground for AI (data comes from real user behaviors)
A workbench for AI (executing tasks here rather than just having conversations)
The user’s only startup panel (no need to switch between multiple apps)
X wants to be “a place where you can do anything”, which is an inevitable requirement for realizing AI capabilities, rather than a marketing slogan.
3. The complete picture of three – layer value distribution
Looking back at these three sections, Musk’s judgment on the flow of AI wealth forms a clear three – layer structure:
First layer: Chips → Platforms
Computing power is the infrastructure, but the data ecosystem is the moat
Second layer: Platforms → Systems
Having data is not enough; one must be able to build a complete closed – loop for task execution
Third layer: Systems → Entrances
Whoever controls the user’s only interface will control the value distribution right in the AI era
NVIDIA provides tools, Google builds capabilities, OpenAI creates systems, and what Musk wants to do,
is to occupy the entrance that is closest to the user and most difficult to be replaced.
The complete path of AI wealth flow from chips, platforms, systems, to entrances is clear.
Conclusion | The end of wealth: Those who occupy the entrance
This interview answered a core question: Where will AI wealth flow in the next decade?
The answer is not in the speed of chip iteration, nor in the number of model parameters, but in who can build a complete system from data to execution and who can occupy the only entrance between users and AI.
Chips will continue to evolve, and models will continue to make breakthroughs,
but ultimately, those who master the value are those who can transform technology into an irreplaceable user interface.
📮Original link:
https://www.businessinsider.com/elon-musk-ai-fix-america-debt-crisis-inflation-2025-12?utm_source=chatgpt.com
https://www.businessinsider.com/elon-musk-robotics-money-irrelevant-currency-universal-basic-income-2025-11?utm_source=chatgpt.com
This article is from the WeChat public account “AI Deep Researcher”. Author: AI Deep Researcher. Published by 36Kr with authorization.