Factors determining the competitiveness of the AI industry are expanding from model capabilities to the entire infrastructure ecosystem, including computing power procurement, storage chip supply, network connectivity, and government technology policies.Article author, source: ME News
TL;DR
- An awards ceremony unexpectedly turned into a tech industry roast. On October 8, Musk, Huang, Sergey Brin, Nadella, Su, and Michael Dell received the National Medal of Science and Technology in Washington, and reporters recorded their private conversations backstage.
- Elon Musk’s anger toward OpenAI has not subsided. He complained that OpenAI’s president, Greg Brockman, requested computing resources from him and again accused the organization of abandoning its original non-profit commitment. However, these are Musk’s personal allegations, not facts recognized by the court.
- What truly upset the tech giants was the cost of AI infrastructure. Micron’s latest quarterly gross margin reached approximately 87%, prompting Musk and Nadella to question storage chip prices and capacity expansion, highlighting the conflicting interests along the AI supply chain.
- What appears to be a lighthearted joke reflects a new phase in AI competition. Factors determining the competitiveness of the AI industry are expanding from model capabilities to the entire infrastructure ecosystem—including compute procurement, storage chip supply, network connectivity, and government technology policies.
An awards ceremony unexpectedly revealed the true nature of the tech giants.
On October 8, 2026, a highly symbolic gathering of the tech industry took place in Washington, D.C., USA. Trump awarded the National Medal of Science and the National Medal of Technology and Innovation to six leaders in the tech sector, including Elon Musk of Tesla and SpaceX, Jensen Huang of NVIDIA, Sergey Brin, co-founder of Google, Satya Nadella, CEO of Microsoft, Lisa Su, CEO of AMD, and Michael Dell, founder of Dell Technologies.
According to the official arrangement of the U.S. government, this was a formal event honoring scientific innovation and technological leadership. The six awardees represented fields such as AI chips, cloud computing, search, semiconductor design, enterprise computing, and aerospace communications—nearly encompassing the most influential sectors of today’s U.S. tech industry. However, what drew greater public attention than the political statements and ceremonial moments on stage was an informal conversation they had while waiting backstage.
Journalist Harry Lambert of Colossus documented this experience in his profile piece, “The Office of the Frontier.” According to his on-the-ground observations, the six individuals were temporarily separated from their families, assistants, and entourages due to Secret Service security arrangements and gathered together in a lecture hall at the Trump Peace Institute in Washington. With no formal agenda and no host, a group of tech giants who normally must navigate communication carefully through corporate PR, business teams, and legal counsel finally had the chance to chat face-to-face.
The conversation quickly veered away from serious tech topics. Musk complained that OpenAI executives had asked him for computing power, Brin teased the habit of adults keeping diaries, Nadella expressed dissatisfaction with capital spending by storage chip companies, Dell talked about his private jet, and Huang lamented a flight without internet. Most dramatically, Musk pulled out his phone and showed everyone a video allegedly of a Tesla Roadster flight test, joking that the footage was meant to be confidential.
Among these details, Musk remains the most outgoing and the most eager to interrupt; Huang continues to steer the conversation proactively; Su is relatively quiet; Nadella and Brin mostly play the role of responders, while Dell appears calm and restrained. Compared to their public personas at product launches, earnings calls, or media interviews, these interactions feel more personal and reveal that even tech giants don’t always engage in polite small talk.
However, viewing this gathering merely as gossip among the wealthy underestimates its informational value. The most significant aspect of this private conversation is not the entrepreneurs’ sense of humor, but the shared anxieties they inadvertently revealed: the AI industry has entered a new phase characterized by massive capital investment, severe supply chain constraints, and a complex dynamic of both competition and interdependence among companies.
Musk Slams OpenAI: Behind Personal Grievances, Computing Power Is Becoming Hard Currency
The most pointed remarks of the conversation came from Musk’s complaints about OpenAI. When Huang Renxun temporarily stepped away from the group for an interview, others began discussing AI computing supply. Musk mentioned that OpenAI’s president, Greg Brockman, had previously repeatedly asked him for computing resources during a dinner. Google co-founder Brin then said that Brockman had made similar requests to him as well. This greatly upset Musk, who directly accused Brockman of taking over the charitable organization he helped found, only to return and ask for help—his words were highly forceful.
Brockman added another perspective to the criticism, mentioning Brockman’s long-standing habit of keeping personal journals and joking that he would have burned such journals long ago. This joke is noteworthy because Brockman’s past personal records were part of the legal dispute between Musk and OpenAI. However, it is essential to distinguish between Musk’s personal opinions and judicial facts. On May 18, 2026, a U.S. federal court dismissed Musk’s claims against OpenAI and its executives on the grounds of statute of limitations. Musk stated at the time that he planned to appeal, but this ruling itself does not equate to the court having made a substantive determination on all business ethics disputes.
In fact, the conflict between OpenAI and Musk extends far beyond internal disagreements among the founding team. In 2015, Musk co-founded OpenAI; in 2018, he left the organization; since then, as ChatGPT drove the commercialization of generative AI, he re-entered the large model competition through xAI. The dispute between the two sides has gradually expanded from issues of organizational mission and control to multiple areas including talent, funding, and computing resources. For Musk, OpenAI is both an institution he once helped create and a major competitor he now must confront.
More notably, Brockman sought computing support from both Musk and Brin, revealing that resource relationships in the AI industry are far less clearly defined than outsiders might assume. Competitors may also be potential infrastructure partners—companies compete in the model market yet still need to secure suppliers capable of providing computational resources. When computing power becomes a critical constraint for model training and product operations, those who can consistently access low-cost, large-scale computing resources are more likely to gain the upper hand in the next round of competition.
NVIDIA’s earnings report provides compelling evidence. Data from its second fiscal quarter of 2027, released in August 2026, showed quarterly revenue of $96.2 billion, a 106% year-over-year increase, with data center revenue reaching $89 billion, up 117% year over year. Such remarkable growth demonstrates that global AI capital expenditures are still rapidly translating into revenue for infrastructure providers. At this stage, chips and computing resources are not merely technological tools—they are key sources of corporate competitiveness and bargaining power.
Micron's 87% gross margin has left tech giants collectively dissatisfied.
Compared to the personal grievances sparked by OpenAI, Musk and Nadella’s discussion about Micron more directly reveals the current conflicts of interest in the AI supply chain.
When discussing memory chip supply, Musk expressed shock at Micron’s recent quarterly gross margin of approximately 87%, stating he had never seen such an extreme supply chain price shock. Nadella also believed that Micron had not invested enough to expand its production capacity. Although their complaints were made in informal settings, they directly highlight the real challenges facing AI infrastructure investment: demand-side parties continue to increase budgets, while key component suppliers are reaping extremely high profits amid tight supply conditions.
The data is indeed noteworthy. Micron’s fourth fiscal quarter results for fiscal year 2026, reported as of September 30, 2026, showed quarterly revenue of $54.229 billion, with a GAAP gross margin of 86.8% and a non-GAAP gross margin of 87.0%, significantly higher than the 45.7% recorded in the same period last year. Annual revenue reached $133.188 billion, with a GAAP net profit of approximately $84.969 billion. Even considering the backdrop of explosive demand for AI storage, such profit performance clearly indicates a significant shift in the supply and demand dynamics of the memory chip market.
However, equating high gross margins simply to Micron deliberately restricting capacity is not accurate. Financial reports show that Micron’s net capital expenditure for fiscal year 2026 reached $27.37 billion, indicating it has not refrained from expansion. The issue is that semiconductor manufacturing involves factory construction, advanced processes, equipment delivery, and product certification—even if a company is willing to increase investment, it cannot immediately convert capital spending into additional supply. Therefore, Musk and Nadella’s comments are better understood as expressing frustration over the speed at which suppliers are expanding capacity, rather than as a factual assessment of Micron’s business practices.
Microsoft’s own financial results also reflect this pressure. The first-quarter earnings report for fiscal year 2027, released on October 28, 2026, showed quarterly revenue of $92.3 billion and a 48% year-over-year increase in Azure revenue. Meanwhile, the company’s capital expenditures for the quarter reached $45.2 billion, with approximately three-quarters allocated to short-cycle assets primarily consisting of CPUs and GPUs, further impacted by rising memory and storage costs. Microsoft’s management expects capital expenditures for calendar year 2026 to total approximately $170 billion. These figures demonstrate that the growing demand for AI infrastructure is not merely an entrepreneur’s anecdotal concern—it has become a critically significant expense line item in the financial statements of major tech companies.
A concerning situation has emerged: the more aggressively AI infrastructure is invested in, the more pronounced supply constraints in critical areas may become, and rising costs will further raise the capital threshold for the next round of construction. Chip manufacturers, cloud service providers, and model developers are collectively driving the expansion of the AI industry, but profits and risks are not evenly distributed. For investors, assessing the health of the AI industry in the future cannot rely solely on user growth at model companies—it must also include monitoring chip supply, capital expenditure efficiency, and whether new computing power ultimately generates sufficient commercial returns.
Jensen Huang complains that the plane has no internet; it seems like a joke, but it’s highly representative.
After discussing computing power and chips, the conversation gradually turned to private jets. Huang Renxun complained that NetJets, Buffett’s private aviation service, is unsuitable for international travel. Musk then teased Dell, speculating that he must own an entire fleet of private jets. Dell replied that he owns only one, since a person can only be on one plane at a time. He then mentioned that he frequently holds video conferences on his plane using SpaceX’s Starlink, and Brin also expressed high satisfaction with his experience using Starlink.
Huang’s experience presents an interesting contrast. According to journalist accounts, he chose not to own a private jet, and during a recent Delta Airlines flight, he encountered a network outage. With the in-flight Wi-Fi unavailable, he had to quietly read approximately 7,000 emails and complained that the lack of internet connection prevented him from working normally. Musk jokingly suggested that the best way to isolate an AI would be to put it on a Delta Airlines plane, since there’s no way for it to escape.
This is, of course, merely a lighthearted anecdote and should not be interpreted as a formal assessment by tech giants on the aviation industry. However, it does reflect a broader trend: as cloud collaboration, real-time data processing, and AI tools become integrated into daily workflows, reliable internet connectivity is shifting from a convenience to a foundational productivity infrastructure. Huang’s complaint about a flight’s connectivity failure and Dell’s praise for Starlink’s connectivity both point to the same underlying need.
From a business perspective, this also explains why Musk can build competitive advantages across seemingly different industries. Tesla, SpaceX, Starlink, and xAI have distinct business models, but all are connected to computing, communication, and automation capabilities. For future applications reliant on real-time data and remote computing, network coverage and connection stability may be just as critical to product experience as computational resources.
Six tech giants sit together, behind them lies the power network of the U.S. tech industry.
If we shift our perspective from the conversation itself to the identities of the participants, the significance of this gathering runs even deeper.
Jensen Huang controls critical resources in the global AI accelerated computing market; Lisa Su leads AMD’s ongoing competition in high-performance computing; Satya Nadella oversees one of the world’s leading cloud platforms; Larry Page is closely tied to Google’s technology ecosystem; Dell represents servers and enterprise infrastructure; and Elon Musk spans aerospace, communications, automotive, and artificial intelligence. Though these six individuals come from different industries, their businesses are increasingly converging around AI infrastructure—any bottleneck in a key component could disrupt the expansion plans of other companies.
This industry-government relationship is also extending to the policy level. In March 2026, the White House announced the first members of the President’s Council of Advisors on Science and Technology, including Brin, Huang, Su, and Dell. By October, at an award ceremony, the Trump administration brought together these technology leaders to emphasize the importance of scientific and technological innovation for America’s future competitiveness. The connection between tech companies and the government is evolving beyond traditional industry policy consultations to encompass AI infrastructure, research organization models, and national technological capacity building.
Of course, close engagement between entrepreneurs and the government does not automatically imply policy bias, nor can a single award ceremony be directly interpreted as a quid pro quo. However, it is noteworthy that the U.S. government is increasingly viewing major technology companies as vital components of the nation’s innovation capacity. This raises new policy challenges: as advanced AI models, computing infrastructure, and foundational research become ever more dependent on a handful of corporate giants, balancing the encouragement of private investment, the preservation of market competition, and the support of public scientific research will remain a long-term challenge.
In this sense, the casual conversation among the six giants offers a window into the structure of the U.S. tech industry. They share common interests yet face clear competition; they want the AI market to continue growing but are unwilling to bear the rising costs across other industry segments; they benefit commercially from technological advancements but must also confront the increased demands on capital and resources that come with technological expansion.
What truly matters is not what these giants have said in strong terms.
Looking back at this behind-the-scenes conversation in Washington, the most widely shared moments were clearly Musk’s sharp criticisms of OpenAI and Huang’s frustration over the plane’s lack of internet—details that made these top tech executives seem more human and revealed the complex personal dynamics between them.
But the greater value of this conversation lies in how it reveals the real pressures behind the AI industry's boom: compute competition continues to intensify, supply of storage chips remains tight, massive capital expenditures continually test corporate cash flow and return on investment, and the U.S. government is readjusting the relationship between technology policy and industrial development. Even tech giants with vast resources cannot fully escape these constraints.
My assessment is that the AI industry is shifting from solely competing on model technical capabilities to also competing on financial strength, control over infrastructure, and commercial monetization efficiency. Over the past few years, the market has focused on who can train the most advanced models; in the coming years, who can consistently access computing resources at reasonable costs and turn expensive infrastructure into tangible revenue will be equally important.
Elon Musk can complain about OpenAI, Satya Nadella can question Micron, and Jensen Huang can criticize airline network quality—but these seemingly unrelated grievances all point to one truth: technological progress does not eliminate resource constraints; it merely shifts their location. For the AI industry, which is undergoing rapid expansion, this may be more worthy of long-term attention than any feud among tech giants.



