Article title: 奥特曼 empathizes with 黄仁勋, "heroes have flaws too," but does he support him returning as 英伟达 CEO…
Body:
奥特曼 empathizes with 黄仁勋, "heroes have flaws too," but does he support him returning as 英伟达 CEO…

Mengyao reporting from Waifeisi
Quantum位 | WeChat account QbitAI
The fallout from 黄仁勋's departure continues to simmer, and 奥特曼 has finally spoken publicly.
On November 16 local time, at a Silicon Valley summit hosted by Semafor, OpenAI CEO Sam 奥特曼 publicly addressed the fallout from 黄仁勋's departure for the first time.
奥特曼 even made no secret of his admiration for 黄仁勋 on stage: "黄仁勋 has always been a hero in my mind."

Regarding the fallout surrounding 黄仁勋's departure and the future of the entire AI industry, 奥特曼 also shared a more complete view of his own:
- 奥特曼 acknowledged that heroes are not perfect.
- 黄仁勋 has always been a hero in 奥特曼's mind, and it is hard to imagine the AI revolution without him.
- 黄仁勋's departure will not stop the entire AI industry from continuing to move forward.
- OpenAI and 英伟达 established a mutually supportive relationship early on, with both sides betting on each other very early.
- As for whether 黄仁勋 should return to 英伟达, 奥特曼 believes he does not have enough information to make a judgment.
Three years ago, 奥特曼 himself went through a board ouster that shook Silicon Valley.
Now it is 黄仁勋's turn, and he clearly understands the feeling better than most.
奥特曼 speaks for the first time about 黄仁勋's departure: heroes have flaws too
On November 12, founder 黄仁勋, who had led 英伟达 for 33 years, suddenly announced he was stepping down as CEO and from the board.
Once the news broke, 英伟达's stock price fell by about 15% at one point within a week, and more than $570 billion in market value evaporated.
Subsequently, multiple Silicon Valley executives including 埃隆·马斯克 signed an open letter, and 180,000 people signed a petition calling on 英伟达's board to reinstate 黄仁勋.

However, many outside observers noticed the wording of this open letter supporting Huang—
it was identical to the open letter in 2023 when OpenAI employees pressured the board to demand 奥特曼's return.
At the Semafor summit, host Reed Albergotti put this question directly to 奥特曼.
After all, in November three years ago, 奥特曼 was also suddenly ousted by OpenAI's board, and four days later staged a dramatic return of the king.
In response, 奥特曼 admitted that he has "strong personal empathy" for 黄仁勋's current situation.
He recalled that 英伟达 bet on OpenAI early on, and OpenAI likewise bet on 黄仁勋.
In 奥特曼's view, without 黄仁勋 and the 英伟达 he built with his own hands, it is hard to imagine what today's AI revolution would look like.

But feelings are feelings, and business is business.
When asked whether 英伟达's business commitments still carry as much weight as before after 黄仁勋's departure, 奥特曼's answer was very pragmatic.
Over the past few years, 英伟达 has not only been OpenAI's largest chip supplier, but also one of its most important backers.
Not long ago, 英伟达 had just agreed to provide up to $105 billion in financial guarantees for OpenAI's data center project in Ohio.
Statistics show that 英伟达's total investments and commitments in the AI ecosystem have exceeded $100 billion, covering a range of core players including CoreWeave, Nebius, and Anthropic.
There is widespread concern that once CFO 科莱特·克雷斯 takes over as interim CEO, 英伟达 may pull back, affecting liquidity across the entire AI industry.
奥特曼 does not agree with this pessimistic view.
He pointed out that 英伟达 is now a massive giant company, and everyone still very much trusts its ability to guarantee.

As for whether the entire AI industry will lose momentum after losing 黄仁勋, 奥特曼 is quite optimistic.
He believes that AI market demand is now strong, AI's value has been validated, more and more people are starting to actually use AI, and both industry revenue and market demand are growing strongly.
Therefore, even if 黄仁勋 leaves, the entire AI industry can continue to move forward.
奥特曼 also emphasized that 黄仁勋, through his extraordinary ambition, played an extremely special and valuable role in driving the AI industry forward.

When it comes to where AI's next wave of breakthroughs will appear, 奥特曼 set his sights on scientific research.
He revealed that over the past few months, AI has already begun to show potential for driving scientific discovery in fields such as mathematics and physics, and some exciting progress is emerging.
In his view, this is only the beginning.
As model capabilities improve further, AI is expected to participate in complex engineering tasks such as chip design, helping humanity solve technical problems that previously required large numbers of experts.
At the same time, the way work is done inside enterprises will also change:
A large amount of repetitive, process-driven work is expected to be gradually handed over to AI.
From scientific discovery to engineering design, and then to daily enterprise operations, 奥特曼 expects AI to take on increasingly complex real-world work and thereby create new economic value.
Also clarifies the memory stick controversy, responds to data center and IPO challenges
Besides talking about his old partner 黄仁勋, 奥特曼 also directly addressed several major recent public questions during the interview that day.
The first was the memory stick remark that recently angered PC gamers.
Earlier this month, global DRAM memory prices soared under pressure from AI computing demand, with prices at one point rising 5 to 7 times.
At the time, 奥特曼 said in an interview that it was quite bad to see scarce memory sticks sitting in ordinary people's personal computers, and that concentrating them in AI data centers would allow them to play a greater role.
Those remarks quickly drew mockery in overseas communities, and many PC builders even gave him the nickname Ram Altman (Memory 奥特曼).
At the summit, 奥特曼 took the opportunity to clarify, saying his previous words had been misunderstood by outsiders.
He actually hopes everyone can afford large memory and should install large memory in their own computers, and he very much supports people tinkering with hardware at home.

The second issue was the real-world problem of double counting in AI chips and the difficulty of getting data centers built.
There were previously reports that 奥特曼, when communicating with Samsung executives, mentioned that there is a bubble of double counting in chip demand on the market.
奥特曼 explained the logic behind this on stage.
He said that many companies are now developing AI chips, and every supplier is counting on selling most of its capacity to a small number of leading large-model companies such as OpenAI and Anthropic.
But from the buyer's perspective, OpenAI does not need to swallow all suppliers' chips wholesale, which leads to overlapping expectations in the upstream supply chain.
However, he stressed that the actual construction demand for data centers is still very real.
奥特曼 acknowledged that the old approach of building data centers near residential areas no longer works, and future developers must build data centers far away from people.

In addition, 奥特曼 also talked about the latest progress in recursive self-improvement (RSI) inside OpenAI.
奥特曼 revealed that OpenAI has begun using AI technology to accelerate the development of its own models.
However, he made clear that OpenAI has not yet launched fully autonomous RSI, that is, letting AI continuously improve its own capabilities without human supervision.
The reason is that at this stage it is still impossible to ensure that this process is safe enough.
As AI capabilities continue to improve, how to ensure humans always retain control will also become an important issue OpenAI needs to solve.

Finally, regarding the capital market developments that outsiders are watching, 奥特曼 also gave his latest statement.
This week, competitor Anthropic just launched a highly anticipated IPO process.
Some analysts worry that Anthropic's listing will siphon off a large amount of Wall Street capital in advance, potentially shrinking the fundraising space left for OpenAI.
奥特曼 seemed quite calm about this—
He believes the real situation in the primary market over the past year is that investor demand far exceeds the supply of high-quality assets, and he is not worried at all about a lack of future funding support.
Moreover, giving more ordinary people the opportunity to invest in OpenAI and share in the economic gains brought by AI development is one of the important reasons for pushing the company to go public.
But he also admitted that in the current period of high industry turmoil, becoming a public company does have quite a few practical drawbacks.
Therefore, although OpenAI will certainly move toward the public market at some point in the future, it is not in a hurry right now.
[1]https://www.youtube.com/watch?v=oi-E6xB4wSU&t=39s
- 黄仁勋 prepares to return to 英伟达 2026-11-16
- A dark horse emerges again among domestic large models, drawing Silicon Valley's attention! 2026-11-13
- First batch of Fable 5.2 hands-on tests is here! Netizens: Opus suddenly lost its appeal 2026-11-10
- Robots do the work, AI acts as factory manager! This giant wants to build the next generation of autonomous intelligent warehouses 2026-11-03
Related reading

奥特曼 returns to the board acquitted! The truth behind OpenAI's internal strife comes to light, investigation results released
Title: 奥特曼 empathizes with 黄仁勋: "heroes have flaws too," but supporting his return as 英伟达 CEO…

奥特曼's impassioned mini-essay hypes AGI: deep learning just works, no need to argue
Superintelligence will arrive in "a few thousand days"

ChatGPT's blockbuster upgrade! A stronger GPT-4 goes live, API pricing slashed, the applause at the launch never stopped
奥特曼 is calling on you to customize ChatGPT and make money

奥特曼's 17 year-end takeaways go viral! OpenAI president: counterintuitive but true

Just now, 奥特曼 laid out three judgments on AGI: the Scaling Law holds, there's no reason to slow investment
By 2035, every person will be able to summon capability equivalent to the sum of everyone's intelligence in 2025.

A mysterious large model floods feeds overnight, so capable it's suspected of being GPT4.5; 奥特曼 dodges the question and speaks in riddles
gpt2-chatbot, which identifies itself as a "personality v2" version of ChatGPT