The quick version
- Nvidia is leaderless for now. Jensen Huang resigned as CEO, president and board member on 12 November. Chief financial officer Colette Kress is interim CEO while the board searches for a permanent successor (NVIDIA).
- Markets wobbled but did not panic. The shares fell as much as 14% on Thursday and closed down 4.3%. That still wiped more than $180 billion off the company’s value (Bloomberg).
- Rivals are not standing still. Google released its fourth AI chip system of the year this week (Yahoo Finance), and Huawei already holds 28% of China’s AI chip market (ThinkChina/Caixin).
- The real bottlenecks are physical. Memory chips cost more than five times what they did a year ago (Fortune), and power grids from Dublin to Texas are running short of capacity.
- Leadership in AI is now a system, not a company. Whoever controls supply, power, capital and the rules will matter more than any one chief executive.
What happened at Nvidia
Jensen Huang, who co-founded Nvidia in 1993, resigned as chief executive, president and board member on 12 November. Chief financial officer Colette Kress is serving as interim CEO while the board looks for a permanent successor (NVIDIA).
The board has said it had “no reason to believe that either national security or NVIDIA secrets were compromised,” and that its independent inquiry is continuing (NVIDIA). It also named Harvey C. Jones, a former Synopsys chief executive, as its first-ever chair (InvestmentNews).
Zuddies is not repeating the unconfirmed claims that have circulated about the reasons for his departure. Until the inquiry reports, the facts above are what is established.
The circumstances of his exit have dominated the headlines. The more useful question for readers is what it shows about who actually leads in AI.
Five things that decide who leads
1. Chips, and who else can make them
For years, “AI leadership” meant buying Nvidia. That is changing. On 9 November Google added two new versions of its own AI chips, the TPU 8tx for confidential workloads and the TPU 8i Max for very large tasks. Both are due next year (TechCrunch). Google’s chip unit used to release one new processor every two years. This year it has released four systems (Slashdot/Bloomberg).
In China, Huawei shipped 572,000 AI chips in the first half of 2026, taking 28% of the Chinese market against Nvidia’s 34%, according to IDC. DeepSeek, ByteDance and Meituan are among its customers (ThinkChina/Caixin). Huawei cannot get the most advanced manufacturing tools, so it links large numbers of less powerful chips together instead. It is a costly approach, but it works.
What it means: Nvidia is still the leader, but buyers now have real alternatives. A leadership crisis makes those alternatives look even more attractive.
2. Memory and manufacturing
The less glamorous parts of the supply chain are now the tightest. Global memory chip revenue jumped 369% year on year in the third quarter, to just under $200 billion, as AI data centres bought up supply (TechRadar/Counterpoint). One Taiwanese memory maker’s CEO warned that by 2027 “even money can’t buy goods” (Fortune).
Even Nvidia reportedly asked TSMC, the world’s largest chip manufacturer, to consider making advanced memory. TSMC says it remains focused on its core business, and neither company has announced a memory programme (eeNews Europe).
What it means: The companies that control manufacturing, mostly in Taiwan and South Korea, have as much say over the pace of AI as the companies that design the chips.
3. Electricity
AI runs on power, and power is running out in many places. In Ireland, data centres used 23% of metered electricity in 2025, up from 5% in 2015, and EirGrid considers the Dublin region saturated for large new connections (Projet Celsius). On 7 November about 2,000 people marched in Dublin in Ireland’s first national protest against data centres, calling for a moratorium (RTÉ).
The US faces the same strain. Developers are struggling to secure grid connections and local approval, which is delaying projects (Data Center Knowledge).
What it means: A country can have the best chips in the world and still fall behind if it cannot plug them in.
4. Money, and investor patience
The AI labs are worth more than most countries’ stock markets, but the mood is cooling. OpenAI closed $11 billion in new funding at a $1.25 trillion valuation, less than half of the $30 billion it first targeted (Reuters). Anthropic set its IPO price range this week, valuing it at up to about $1.4 trillion, well below the $2 trillion reported a few months earlier (TechCrunch).
In Ireland, some tech firms are now trying to win back workers they let go after their AI bets faltered (Business Post).
What it means: The leaders will be the companies that can keep raising money once the hype fades, and that means showing real revenue.
5. The rules, and who can buy what
Governments decide who gets the best chips. President Trump has said China should not get any Nvidia chips above the H100, and Treasury Secretary Scott Bessent says leakage “hurts our great companies in the AI race” (BBC). Yet the administration also allowed sales of the more powerful H200 chips to China in exchange for a 25% tariff (Axios).
Europe is taking a slower path. The EU’s AI Omnibus pushed the main high-risk AI rules back to December 2027 and August 2028 (EY).
What it means: Export rules and regulation can change the balance of the AI race faster than any product launch.
Why it matters for Ireland
Ireland sits at the centre of two of these five pressures. It hosts a large share of Europe’s data centres, so it feels the power and planning debate first. Enterprise Minister Peter Burke calls data centre growth a strategic opportunity for the Irish economy (StackPath/RTÉ), and protesters say it is pushing up household bills. Ireland is also home to the European headquarters of many US tech firms, so jobs here move up and down with AI spending.
What to watch next
At the time of writing (14 November), these were the next milestones. Some may have passed by the time you read this.
- Nvidia’s quarterly results. Analysts expect questions about whether its revenue forecast still holds without Huang (AFP).
- Anthropic’s IPO. The company was expected to price its shares on 18 November and begin trading the next day (Fortune).
- The APEC summit in Shenzhen. Trump and Xi were due to meet, with chip exports in the background.
- The board’s inquiry. Its findings, and the choice of a permanent CEO, will matter more than any headline.
Zuddies view
This section is opinion.
The week’s lesson is not that Nvidia is finished. It is that no single company, or person, decides who leads in AI any more. The winners will be the countries and companies that secure chips from more than one supplier, build the power to run them, and set rules that are clear and stick. Europe, and Ireland in particular, has a real say in the power and the rules. It should use that say on purpose, not by accident.
Join the discussion
Should Ireland keep building data centres to stay in the AI race, or pause until the grid and household bills can cope?
What we know and what we don’t
| What we know | What we don’t know yet |
|---|---|
| Huang resigned on 12 November. Colette Kress is interim CEO (NVIDIA). | Who will be Nvidia’s permanent CEO. |
| The board said it had no reason to believe secrets were compromised (NVIDIA). | What the independent investigation will conclude. |
| Google and Huawei are expanding their AI chip ranges (TechCrunch, ThinkChina/Caixin). | How many Nvidia customers will actually switch. |
| Memory prices and grid limits are tightening (Fortune, Projet Celsius). | Whether the shortages ease in 2027 or get worse. |
| The US allows some advanced chip sales to China, with conditions (Axios). | How export rules will change after the Trump–Xi meeting. |