Looking at leadership in the AI era from Jen-Hsun Huang: Why are some companies able to achieve exponential growth?
Barrons
2026.11.22 22:09
Full text 5181 words
The previous article discussed the three behavioral logics of opportunist, diplomat and expert.
Opportunists put personal gains first and are prone to fall into short-term arbitrage; Diplomats excessively maintain relationships and organizational stability and are prone to miss opportunities for adjustment; Experts rely on personal professional abilities and easily regard local optimality as the overall optimal for the enterprise.
The problems of these three behavioral logics are not exactly the same, but they have one thing in common: the leader's decision-making boundaries are still relatively limited. How far a company can go depends largely on how far the leader can see and whether he or she can turn personal judgment into organizational capabilities.
Going further up, the behavioral logic begins to change.
The doer type first solves the problem of how to get things done stably under multiple constraints; individualists begin to break through existing rules and look for new possibilities; strategists no longer just look at a single project, but put organizations, industries and even society into the same decision-making framework; in the alchemist type, leaders will even rethink why the company exists and what kind of value can span a business cycle.
These four behavioral logics are not four types of "smarter" people, but four different decision boundaries.
(4) Doer type: multi-constraint linear programming, steady compound interest and missed exponential dividends
The doer type is a multi-constraint linear programming in a relatively stable environment: seeking a balance between short-term performance and long-term growth, writing team capabilities, resource boundaries, and market rules into the constraints to find a feasible and stable solution. They are good at building standardized collaboration systems and implementing mid- and long-term plans; they are also good at embedding AI tools into existing businesses to effectively improve efficiency. The price is to rely on mature frameworks, keep a distance from disruptive innovation, and respond slowly when the environment changes suddenly.
Apple under Tim Cook is the epitome of this type. Apple has gradually integrated AI into the mobile phone, computer and service ecosystem with a steady pace and does not pursue radical technological narratives. From the end of 2022 to November 2026, Apple's market value increased to approximately 2.5 times, during the same period Nvidia increased to approximately 11.6 times, and Microsoft increased to approximately 2.2 times; converted into a compound growth rate, Apple's annual average growth rate is approximately 27%, Microsoft's approximately 22%, and Nvidia's approximately 88%. The above multiples only reflect changes in market value, do not include dividends, and cannot be directly equated to operating performance. It illustrates how the market prices different strategies. Apple is still one of the companies with the highest market capitalization in the world, but in the incremental allocation of this round of AI market, its share is significantly smaller than Nvidia. This is not failure, it is the result of choice.
Davido's Law says that companies must continue to eliminate their own old products, otherwise they will be eliminated by others' new products. The sense of security that sticks to a mature system is often the most expensive sense of security. Drucker repeatedly emphasized doing existing things better; but opening up a new growth curve requires another kind of ability. Stability and risk resistance are virtues; but in an industry that is changing exponentially, linear growth only means a relative decline in share. What is missed is not current profits, but future positions.
(5) Individualist type: relaxation of constraints, innovation and optimization, single-point explosion accompanied by high volatility
Individualists do not regard industry rules as hard constraints but as variables that can be adjusted. Relax those constraints that have expired, take advantage of contradictory tensions, do unstructured innovative solutions, and find solutions that others have not seen. They are accustomed to using first principles to ask "why must this be done", dare to question consensus, break conventions, and promote changes in conflicts; they can make single breakthroughs, but lack overall system thinking. Change is highly dependent on personal influence and is difficult to consolidate into organizational capabilities.
When Musk founded xAI, the industry was competing for parameters and commercialization speed. xAI chose the differentiated positioning of "understanding the universe" and entered the field through the open source route. Valuations have risen rapidly: approximately US$24 billion after Series B in 2024; approximately US$113 billion in March 2025; approximately US$230 billion after financing in January 2026. The story does not stop at personal heroism: In February 2026, SpaceX acquired xAI at a valuation of approximately US$250 billion, and the overall post-merger valuation was approximately US$1.25 trillion; in May 2026, xAI merged into SpaceX and changed its name. This type of enterprise is very explosive and often ends up being absorbed by the larger system.
The fluctuations of this type of enterprise can be seen most clearly in Tesla. Tesla's stock price has retraced approximately 65% throughout the year in 2022, while its stock price has risen more than ten times between 2020 and 2021. The rise is huge when it goes up, and it is equally sharp when it goes down. This is the typical distribution of individualistic companies. Schumpeter's theory of creative destruction points out that innovation will inevitably break the old order, but single-point destruction itself does not constitute a new order. Individualists are good at tearing gaps, but not at turning gaps into structures. Short-term explosive power can break through the stock bottleneck, but without system support, it will fluctuate greatly, making it difficult to accumulate cross-cyclical value. It depends on the individual, but it also depends on the individual.
(6) Strategist Type: Multi-layered global dynamic optimization to leverage industry-level index wealth
The strategist type is a high-level multi-objective dynamic planning in management science. The decision-making boundary has expanded from a single department and a single enterprise to three levels of organization, industry, and society; the time span has been lengthened from one year to five to ten years; the objective function also includes business income, organizational growth, industrial empowerment, and social value, and seeks the long-term equilibrium of the entire system rather than the extreme value of a certain point. They are good at reconstructing organizational paradigms, aligning the interests of multiple parties, and promoting systemic changes; they do not pursue single hot spots, but focus on ecological construction and long-term capacity building, and use forward-looking judgment on technological rhythms to lay out the future track.
Huang Renxun’s judgment is that AI is not a graphics card business, but a competition in the computing ecosystem. NVIDIA has invested in CUDA since 2006, tying the chip, software stack and developer ecosystem together to form a closed loop of "hardware-toolchain-application"; before the outbreak of AI demand, it had proactively adjusted the resource allocation of game graphics cards. From the end of 2022 to November 2026, the market value increased to approximately 11.6 times, and Huang Renxun’s own net worth also increased from tens of billions of dollars to hundreds of billions of dollars. Market value is just the result. The real change is that it has become the computing power infrastructure of the entire industry.
When Nadella took over, Microsoft's growth had stalled and the organization was tied to the inertia of the Windows ecosystem. He turned his focus to cloud and AI: establishing "cloud first", focusing on OpenAI, embedding Copilot into all products, and at the same time restructuring the delivery method of enterprise services. In the same range, Microsoft's market value increased to approximately 2.2 times. The multiple is not as good as NVIDIA, but it illustrates one thing: turning AI into an organization's basic capability will last longer than turning AI into a product.
The companies with the largest increase in market value in this round of AI market have in common not that they have stronger models, but that they have placed themselves at key nodes in the industry chain - providing computing power, platforms, and distribution channels. Once a position is established, the benefits become self-reinforcing. Drucker believes that the core job of a leader is to establish a shared vision and build systems, not to make decisions for the team. The benefits brought by global thinking have compound interest characteristics, relatively small fluctuations, and wide radiation. It is the mainstream form of wealth creation in the AI era.
(7) Alchemist Type: Reconstruction of the objective function to define sustainable value across cycles
The alchemist does not seek the optimal solution within the established framework, but redefines the objective function itself. Breaking away from the single scale of commercial interests and re-answering why enterprises exist from the perspectives of historical cycles, social civilization and human-machine symbiosis. This is a change at the paradigm level, not a parameter adjustment. They set the ethical bottom line, spiritual consensus and the mission of the times as top-level hard constraints, use symbols and narratives to unite the organization, and care about whether the company can transcend the cycle, not just this quarter's report.
The AI industry is still in the early stages of rapid technological iteration, and no recognized alchemist-type leader has yet emerged. There is a more complete reference in business history: Inamori Kazuo reconstructed Kyocera's value system with "respect for heaven and love for people", and regarded profits as results rather than goals. He successively led the reconstruction of Kyocera and JAL, allowing the two companies to go through multiple cycles. What is worth noting is not how much wealth he created, but the fact that the company still maintained its organizational capabilities and culture during the crisis. What the alchemist will look like in the AI era, we can only speculate now: most likely he will not be the best at calculating, but the person who can best answer "What does technology do for people?"
Kazuo Inamori believes that the essence of business is to shape people's hearts, and profits are incidental results; Torbert pointed out that the core ability of an alchemist is to reshape the value consensus of the group, which is difficult to possess at other levels. This layer has stable business returns and strong risk resistance. It also accumulates assets such as brand, culture, talent and social trust that are difficult to replicate. The return period is long and the decay is slow.
Step Path: From Personal Wealth Creation to Systematic Wealth Creation
The seven behavioral logics are not labels isolated from each other, but have the possibility of gradual development.
Torbert emphasized that the upgrading of behavioral logic is not completed in the classroom, but in high-pressure situations, through repeated reflection, adjustment of judgment and action, gradually changing the original way of cognition and behavior. The jumps between levels have their own development paths, and each step requires a specific approach.
Wealth leverage at different stages is expressed as multiples below to illustrate the leverage differences that may arise from behavioral logic upgrades. These multiples are only indicative of magnitudes and are not predictions of actual wealth growth, nor do they mean that completing a certain stage of transition will necessarily result in returns of corresponding multiples.
(1) Expert → Doer: From solving the problem yourself to building the system (about 1-10 times)
The first thing that needs to be changed at this stage is over-reliance on personal professional abilities. Expert managers are accustomed to analyzing problems and making judgments in person, and even leave much of the work that could have been done by the team in their own hands. To transform into a doer, work needs to be re-divided: matters involving professional judgment and key decisions are handled by managers, while work that can be standardized is handed over to teams and AI. Managers’ energy should gradually shift to goal setting, resource allocation, and management of execution boundaries.
To judge whether this transformation is truly completed, we can look at a very practical question: Can the team complete the work according to the established process and achieve the expected results when the person in charge is not present? If most things still require leaders to personally intervene, it means that personal capabilities have not really been transformed into organizational capabilities.
Managers can also check themselves in their daily work: Are they accustomed to interrupting subordinates and still make decisions by themselves in the end? Are you unwilling to hand over standardization work to the team and AI because you feel uneasy? Will the existing processes still function normally when you are temporarily away?
(2) Doer → Individualist: From sticking to the framework to breaking the framework (about 10-100 times)
What needs to be overcome at this stage is path dependence. Doers are good at improving efficiency within existing rules and systems, but when the original framework begins to restrict business development, managers need to re-examine these rules. Regularly take inventory of existing systems to distinguish which constraints are still in effect and which have become constraints left by historical inertia. For the latter, it should be adjusted or even cancelled.
Management styles also need to change. Instead of advancing in a fixed manner according to the annual plan, it is better to adopt rolling planning, adjust the direction in time according to market changes, and find new opportunities through small steps of trial and error and rapid calibration. When there are differences of opinion within the team, there should be no rush to resolve the differences. Some disagreements just show that the existing framework is no longer able to explain new problems, and may also help managers discover solution paths that were not considered in the past.
Managers can ask themselves: Are they accustomed to using past successful experiences to judge today's problems? Is dissent being rejected out of fear of change? Are existing processes improving efficiency, or have they become a hindrance to business development? Are there any other possibilities besides the current solution?
(3) Individualist → Strategist: From a single breakthrough to a global ecosystem (about 100-1000 times)
Individualists can break the original framework and find new breakthrough points, but a single point of innovation does not mean that the company has the ability to continue to grow. If a new technology, an innovative business or a market breakthrough always relies on the personal promotion of the leader, it will be difficult to replicate within the organization and form a long-term competitive advantage.
Transforming into a strategist requires transforming personal insights into organizational capabilities. When making decisions, we should not only consider the benefits of a certain team or business, but also take into account the collaboration within the organization, the relationship between the upstream and downstream of the industry chain, and the broader social impact, and make a trade-off between short-term benefits and long-term value. The innovation results that have been achieved also need to be gradually accumulated into systems, processes and talent systems, so that enterprises no longer rely on the continuous promotion of one person. At the same time, it is also necessary to coordinate the interests of different participants through a common vision and reduce conflicts of interest within organizations and in industrial cooperation.
It should be noted here that systems thinking does not require everything to be comprehensive, nor does it require that all local efficiencies be sacrificed for the overall benefit. The key is to identify the relationship between local decisions and overall goals, and to determine which local optima actually harm the long-term interests of the system.
Managers can ask themselves: Does this innovation place an additional burden on other business segments? Does the immediate breakthrough overdraw the company's ability to develop in the long term? Does a change rely on the individual leader to promote it, or has an organizational consensus formed? Can this innovation mechanism continue to operate without the leader himself?
(4) Strategist → Alchemist: From optimizing the system to defining value (more than a thousand times to cross-cycle)
Strategists have been able to allocate resources and coordinate interests based on the entire system, but alchemist leaders have to answer a further question: What value should this system create?
Entering the AI era, this problem has become more realistic. Algorithmic black boxes, data privacy and technical ethics are not problems that can be solved simply by improving efficiency. Technology can provide more choices, but it cannot replace enterprises' judgment on application boundaries. Leaders need to treat ethical principles as hard, non-negotiable constraints, clarifying what can and cannot be done. At the same time, we also need to rethink the value of people in the era of human-machine symbiosis and why people work. Enterprises cannot only rely on benefit distribution to maintain member relationships, but also need to form a mutually recognized value goal and transform the organization from a community of interests to a community of values. Then through specific business scenarios, cultural symbols and continuous organizational practices, this consensus can truly settle and continue between different generations.
There is no standard answer that can be directly applied at this level, and it is difficult to verify through short-term operating indicators. Managers can ask themselves: Does this decision maintain the bottom line of ethical and social values? Will it still hold true ten years from now? Do team members truly understand and agree with the value direction that the company adheres to?
The behavioral logic of leaders will affect how companies allocate resources, build organizational capabilities and choose development directions. These differences will ultimately be reflected in the company's operating results and long-term value.
But for investors, there is a more specific question: Why are some companies able to continue to generate cash flow while others struggle to sustain growth? How will the leader's cognitive and behavioral logic affect the scale, duration and risk of the company's future cash flow?
In the next article, we will turn from the behavioral logic of leaders to the perspective of investors, discuss the relationship between corporate value and cash flow duration, and further analyze how leaders can transform today's competitive advantages into cash flows for the next ten years or even longer.
[Author: Yang Jian, Director of the Financial Information Center of Renmin University of China, Chief Professor of MPA, and Dean of Jian Jinxin Research Institute]
For more in-depth analysis and exclusive insights into global markets, multinational companies and the Chinese economy, welcome to visit the official website of Barron's Chinese website
According to the real-name requirement of the Cybersecurity Law, please bind your mobile phone number before posting a comment.
Express
22:15
China Galaxy Securities Yang Chao: The three "new policies" of A-shares have been implemented accurately to effectively stabilize market fundamentals and medium- and long-term expectations.
22:03
Gold prices fell nearly 9% in a week, report said gold still has downward pressure in the short term
21:58
Preview of market events next week (November 23-29)
21:42
It is reported that AI server manufacturers have increased prices by up to 45%.
21:40
MiniMax announces research progress on automated cryptanalysis of latest internal models
21:21
Taiwan, China plans to lower the financing margin ratio to 120% to stabilize the stock market
21:15
Ningquan Asset announced the resumption of new investor subscriptions
20:56
KEPCO reveals another large-scale data leak, with information of 1.52 million people leaked
20:54
Beijing Chaoyang police have launched an investigation into suspected illegal crimes against companies related to the "Hengchang Group"
20:51
Shentianma A: The overall scale of the company’s non-display business is still very small.
20:21
Bama Tea Industry responded that it was suspended from participating in the procurement of military materials and engineering services
20:20
Core Semiconductor: Relevant patent litigation will not have a significant adverse impact on the company’s current production and operations
20:16
The Shanghai and Shenzhen Stock Exchanges have increased their fee reduction measures year by year, and since 2021, the cumulative profits have exceeded 7 billion yuan
20:14
21 quick facts about News Network on November 22
19:31
Adani Energy announces acquisition of 100% stake in Vikram Shila Power Company
19:30
Zhang Xue said she would consider making an electric motorcycle for delivery boys
Xpeng G9L launched in Thailand and Indonesia
19:29
WeChat: The platform removed more than 100 illegal micro-short dramas in October
CITIC Construction Investment: Keep an eye on liquidity and wait for a turnaround
19:28
Hanhe Capital: successively subscribed for its fund products under management, with a total subscription amount of 13.4675 million yuan
Scan to download App




