Just finished watching the latest episode of the All-In Podcast. Jensen Huang resigned abruptly in six days, and Anthropic is going public next Thursday. The complete timeline of Jensen Huang's resignation: This event unfolded so quickly, completely overturning Nvidia's 33-year-old power structure within six days. On November 6th (Friday, right after the midterm elections), The New York Times, citing a leaked intelligence memo, claimed that Alice Wang, an executive at the Singapore-based data center company Megaspeed, closely associated with Huang, was suspected of working for Chinese intelligence and assisting in the transfer of approximately $7.5 billion worth of chips to Chinese entities since 2023 (short-selling firm Culper Research had actually named this company six months earlier). On November 8th (Sunday), Huang sent an email to the entire company admitting to an extramarital affair and accepting an investigation by the board. Subsequently, The New York Times also broke the news of his affair with a direct subordinate and another person. On November 12th (Thursday), Jensen Huang, who had led Nvidia for 33 years, officially resigned. CFO Colette Kress became interim CEO, and Nvidia's stock price plummeted 15% in a week, wiping out approximately $1 trillion to $1.5 trillion in market capitalization. Why was the board so eager to force Huang to resign? Friedberg and Chamath's analysis suggests that this wasn't simply a moral judgment, but rather a consequence of the stringent requirements of fiduciary duty and executive liability insurance clauses. When signing quarterly financial statements, executives of publicly traded companies must provide written confirmation to auditors that there are "no undisclosed conflicts of interest with subordinates or significant business partners." If this confirmation contains false statements, it directly triggers a breach of contract, leaving the board completely without insurance protection. Therefore, the board was, to some extent, forced to make a painful decision to protect itself, rather than entirely acting out of judgment against Huang's personal conduct. The four hosts' in-depth questioning: "These are two separate actions." The hosts unanimously agreed that this was not an isolated sex scandal. Sacks characterized it as two separate actions—the first being Alice Wang and Megaspeed's alleged involvement in chip smuggling, which should be thoroughly investigated. The second was someone risking imprisonment to illegally leak classified intelligence memos to the media with the aim of bringing down Huang Renxun. These two events are completely different in nature and should not be confused. The program also raised several rather sharp questions. The first question is why there wasn't a pre-arranged defensive briefing? Sacks compared this to the case of Congressman Eric Swalwell and Fang Fang, a suspected spy. According to standard operating procedures in the National Security Department, if there is suspicion that a key U.S. figure has contact with spies, a warning is usually issued first, requiring them to cut off contact, not allowing the situation to develop for over a year before suddenly detonating it through the media. The second question is the reversal in the espionage characterization. A Democratic congressman, who is about to become chairman of the House Intelligence Committee, revealed to Politico after receiving a classified briefing that Alice... Wang may not be a Chinese government agent, and Huang, as the CEO of a private company, does not possess government secrets. Furthermore, the hardware entity cannot reverse engineer through personal connections. Therefore, the logical chain of this accusation is quite fragile. The third point of contention is the real motive: remote chip shutdown and centralized power. Several hosts pointed out that Washington hawks and some forces advocating centralized AI control have been pushing for a "Chip Security Act," requiring all high-end AI chips to have built-in GPS tracking and remote government shutdown capabilities. Huang is a staunch defender of open-source ecosystems and the globalization of computing power. After removing Huang, his successor, in order to protect himself, may be forced to compromise with Congress and implant hardware backdoors. This is the true core motive of this power struggle. The fourth point of contention is even more ironic. China is also refusing to buy US chips. Sacks revealed a reality selectively ignored by Washington hawks: the Chinese government itself is also blocking Nvidia chips (including the H20 and the already approved H200), aiming to force domestic companies to use domestically produced chips like Huawei Ascend (such as CloudMatrix). (384-card supernode) to support domestic alternatives. So essentially, it's a joint effort by Chinese and American hawks to drive Nvidia out of China. If Jensen Huang really falls due to lobbying for sales in China, the outcome would be quite ironic. The fifth point of contention is the economic and industrial impact, known as the risk of Nvidia Bank shutting down. Chamath warned that the entire flywheel of the AI economy has heavily relied on Nvidia Bank, meaning that Jensen Huang, leveraging his personal authority, provided hundreds of billions to trillions of dollars in equity investment and credit guarantees to startups and new cloud data centers. If a traditional professional manager takes over, risk control will inevitably be tightened, chip iteration speed will be slowed, and Nvidia's past advantage of one generation per year will be lost. Coupled with the continued obstruction of construction by a large number of anti-data center NGOs, US GDP growth could lose 50 to 150 basis points, or even more. Anthropic is going public next Thursday, with a valuation significantly lower than previously rumored. The latter half of the program focused on Nasdaq... Anthropic, which is going public, is scheduled to list next Thursday (November 19th). Regarding the offering size, it plans to issue 825 million shares at $70 to $78 per share, targeting a valuation of approximately $1.4 trillion (a decrease from the previously rumored $2 trillion). Including private placements, the total fundraising is expected to be around $90 billion, giving it a post-IPO cash reserve of $230 billion to $300 billion. It has already been oversubscribed six times. Thanks to the strong performance of the enterprise market, Anthropic's annualized revenue surged to $81 billion in October, and its third-quarter revenue reached $17 billion, significantly surpassing OpenAI's $10 billion during the same period. Some Wall Street institutions, in the most optimistic scenario, have given a forward valuation forecast of $4.2 trillion by 2028. Jason points out that Anthropic is adopting a super-bundling strategy reminiscent of Microsoft's past practices. Recently, it has been releasing a series of Claude... Meetings (comparing to Zoom), whiteboard and spreadsheet tools (comparing to Excel), Claude Legal, and personal finance tools—all bundled together in a $20 to $50 monthly subscription—are systematically eroding the competitive moat of traditional SaaS and Microsoft Office. Chamath adds that another key reason for the market's enthusiasm for Anthropic lies in its recursive self-improvement capabilities, touted to investors during roadshows. This means the computational cost per unit of intelligence is continuously decreasing, and the model itself can achieve self-iterative acceleration.