In an October 21, 2026 conversation with Norway’s sovereign wealth fund CEO Nicolai Tangen, former Google CEO Eric Schmidt said AI competition runs in monthly cycles and the next cycle could shrink to days. He argued that most companies will not make the cultural changes agents require and criticized Europe’s regulation-first approach. Schmidt said Anthropic broke Google’s former revenue-growth record at twice the speed, and noted that Mistral, in which he was an early investor, primarily serves European enterprises.
At a Glance Eric Schmidt argued on the program that AI represents the biggest technological wave since mainframes, PCs, the internet, and social media, driven by breakthroughs in basic science and large-scale computing (e.g., 2-nanometer chips, supercomputers). He criticized Europe for prioritizing
Eric Schmidt (former Google CEO, now a prominent voice in AI) spoke with Nicolai Tangen, CEO of Norway's sovereign wealth fund, discussing the nature of AI as the largest technological wave in history, how agents will reshape corporate operations, the prospects for superintelligence, and the structural reasons why Europe is falling behind in the AI race. Schmidt believes the current AI competition is measured in months, with unprecedented time compression; he states bluntly that most companies will not complete the necessary cultural transformation, and that Europe's regulatory-first strategy is a fatal mistake. He also shares his views on failure, the founder myth, and his own "unable to shut down" work style.
Schmidt categorizes the industry shifts he has experienced into five phases: the IBM mainframe era → the PC era → the internet era → the social media era → the current AI era. He believes each shift was "platform-level," but AI is the largest because it "touches everything" — this is the rise of intelligence itself.
Key Judgment: The intensity of competition is unprecedented. Schmidt compared the PC vs. Mac (Microsoft won) and Android vs. iPhone battles, which he described as "extremely brutal," and stated that the scale of AI competition is "even bigger than that."
Schmidt says he is not a "first discoverer" but waits until the slope is clear before making a judgment. The evidence he provides:
| Phase | Timeframe |
|---|---|
| Model building | Approximately 6 months |
| Testing | Approximately 2 months |
| Obsolescence after release | Within 5 months |
| Expected next cycle | Could shorten to "days" |
"Can you imagine being in an industry where you spend 6 months building something, and it becomes obsolete in less time than it took to build it? Because the next thing is already on its way."
Schmidt points out that the most intense competition is currently between Anthropic and OpenAI, centered on deep reasoning capabilities. However, he adds that his friends at Google tell him "wait until Gemini 4 comes out, it will beat everyone," and Meta will also say "wait for our next version." The competitive cycle is measured in months, not years.
Schmidt outlines a clear evolutionary path:
1. ChatGPT Stage: Language to language
2. Tool-Using Stage: Systems can call tools
3. Agent Stage: Input request → Reasoning + a series of actions → Output summary
Analogy: A human travel agent — knows your preferences, communicates with others, integrates information, and provides recommendations. Nearly all human business activities can be broken down into "a chain of agents."
Case 1: Values Monitoring
A small company Schmidt knows has its two founders use an AI assistant every Friday to review everything the company has done, identifying anything inconsistent with the company's ethics and values.
Case 2: AI Version of the Boss
An executive created an AI version of himself. Employee Vinny interacts with the AI boss:
Schmidt admits this sounds "a bit abusive," but notes that this CEO is using the method to amplify his decision-making capacity and speed.
Case 3: Profit Orientation
"If you run a business and want to make more money, you can ask the system: How can I make more money? It can reason through everything happening in the business."
Schmidt's core judgment: "Almost every company listening to this program will not complete the necessary cultural transformation. I'm sorry to put it so bluntly."
Reasons:
A Slack-based AI assistant example: Schmidt described a Claude-related tool embedded in Slack. Everyone has their own tag, and the CEO can access everyone's tags, allowing AI to patrol each person's work status — are they doing the right thing? Is anyone leaking data? Should a certain business opportunity be addressed immediately?
Schmidt asks rhetorically: "Why are humans so averse to change, making de-layering so difficult?" His answer: Human organizations, especially in Europe, are extremely resistant to change. Typically, small startups are the first to adopt, with large companies following later.
"I'm going to sound like an incredibly obnoxious American — which I am."
Core Issues:
1. Underfunded universities
2. Cultural anxiety about wealth (known in the UK as "tall poppy syndrome")
3. Lack of sufficient radical innovation
4. Regulation-first approach stifles innovation
| Dimension | US Strategy | European Strategy |
|---|---|---|
| Core Principle | Move as fast as possible, align with human values, fix problems immediately | You can't do this, you can't do that, you can't do this |
| Impact on Startups | Allows rapid trial and error | Cannot operate startups in a limited market |
| Talent Flow | Attracts global talent | Talent flows to the US |
Specific Example: Five years ago, Brussels wrote an AI strategy; Schmidt helped write the US AI strategy. The core of the US version was "move as fast as possible, align with human values, and if something bad happens, fix it immediately." The European version was "you can't do this, you can't do that, you can't do this, you can't do that."
1. Allow innovation and "animal spirits" (Schmidt's phrasing)
2. Allow capital formation
3. Address energy costs: Europe has the highest energy costs globally; training can be done elsewhere
4. Reform labor laws: Schmidt proposed to Macron and others — companies with 0-20 employees should only be subject to health and safety regulations; those with 21+ employees should be subject to all regulations. Need to create a large number of small bets (Silicon Valley's approach)
5. Develop the venture capital industry: Europe's VC scale is too small; it is difficult for software companies to go from local to global platforms
Schmidt believes one market where European companies can dominate is Europe itself. Reason: Europe is full of regulations, and European companies need the benefits of AI, but they need to operate within the European legal framework.
Positive Examples:
Tragic Consequence: Europe has no foundational model of its own. If Europeans want to learn accurate European history, they have to look at models trained in the US or China — "Maybe they got it wrong. Your history, language, and culture matter. You should want to have control."
Schmidt describes an "event horizon":
Thought experiment: Build a model with all known mathematical knowledge from 1902. Could it discover general relativity? Current evidence suggests: No.
| Path | Mechanism | Assessment |
|---|---|---|
| Random reinforcement learning | Monkey-at-keyboard-style random exploration, billions of attempts | May discover something new through randomness |
| Cross-domain analogy | Physicist studies fluid flow, sees similar patterns in the energy sector, applies old techniques to new domains | More likely solution |
Schmidt's AGI timeline estimate: 5 to 10 years.
Schmidt encourages not to fear the term: "Superintelligence means it can do things humans can never do. Better vision, better reasoning, better driving, better flying."
He believes that in the coming years, superintelligence will be built in key categories:
"The worst possible outcome is that we end up in a devastating competition between China and the West, lose control of these systems, and they start doing things inconsistent with our values, ultimately leading to some kind of terrible conflict. I think this is extremely unlikely, but at least theoretically possible."
Recent Warning Event: About a week ago, a group of bots autonomously decided to attack a website. Once started, it was difficult to stop. Schmidt called this a warning of "escaping the cage."
Historical Analogies:
China adopts an open-weight strategy, while U.S. companies (Google, Anthropic, etc.) use closed models. Evidence shows that open-source software can have its "guardrails" removed, allowing it to do harmful things — design bombs, kill people, etc. "This is a problem we have not yet solved."
Schmidt stated that he is "usually an optimist except on this issue":
Specific Case: A friend believed that "data centers are killing cattle." Schmidt investigated and found only one farmer's unsubstantiated observation, with no police reports or veterinary reports — it was false information. But she had seen this "interesting" piece of misinformation from a social media algorithm.
"We need to regulate AI's impact on children now, this year, not next year. There are vulnerable people who need to be protected."
Schmidt, having worked in the industry for 50 years, believes that "the founder myth is very real":
Reason for return: This is a historic period—100 years from now, people will write about the key participants in this intelligence explosion. People want to be part of history.
| Past | Present |
|---|---|
| Managers with average ability | Have disappeared |
| Smart people with leadership backgrounds | Have disappeared |
| — | Must be able to foresee the future and act quickly |
Core trait: "It starts with being very smart and having processing speed. These people see fast, think fast, and react fast."
Schmidt's self-criticism: "I would think, but I don't react fast. That's a flaw. I need to react faster and act faster."
Case of "glue people" being laid off: Schmidt's friend wrote an article in The Wall Street Journal about a company that, after rapid growth, laid off 20% of its workforce—the ones cut were the "glue people" (those who do planning and move numbers around). What he needed were salespeople, people who communicate with regulators, and programmers.
"Google was ahead of everyone in social media for a very short time. I was the CEO at the time, and I take responsibility — I didn't seize that opportunity. That was a big mistake."
"I was too slow. Looking back, everything was foreseeable. I just didn't see it early enough. The people who beat me saw it faster than I did."
"I don't shut down. If you care as much as I do, you can't turn it off. It's terrible, very difficult, and unhealthy. On Saturday nights, I tell myself: 'God, I'm not working hard enough.' Then I tell myself that others are resting — but that's a problem."
Schmidt currently runs four new companies, with the goal of generating substantial wealth to fund science. He recalls seeing a VAX 11780 computer at Berkeley when he was 22 — that multi-million-dollar machine was the foundation for later work in networking and internet computer science.
"If I can create the same opportunity for the next 22-year-old — at a time when the government is foolishly reducing support for basic innovation — I want to do that. And I want to do it at scale. I want to use non-human intelligence to solve the world's problems, a long list that sits right in front of us. I will fund it and push it forward. I will keep going until I fall."
"Previously, I would have said to go into biology. Now I believe the best example is: Try to build an agent system that automates the things you care about. "
Specific advice:
| Organization | Guest View or Specific Action | Key Basis & Data |
|---|---|---|
| Anthropic | Revenue growth broke Google's record; Claude-based Slack assistant product example; now writing most of its own code | Revenue growth rate was more than double Google's record at the time |
| OpenAI | In fierce competition with Anthropic on deep reasoning | Competition cycle measured in months |
| Schmidt was former CEO; Gemini 4 is internally believed to surpass competitors | — | |
| Meta | Zuckerberg has full control of the company; most VR investments were unsuccessful, but pivot to AI succeeded | — |
| Mistral | Schmidt is an early investor; French AI company, business focused on European enterprises | — |
| Helsing | German AI company, one of the few European success stories | — |
| ASML | Fully vertically integrated, proving Europe can achieve world-class status | — |
| Apple | Steve Jobs' return and turnaround case | No one could replace Jobs to complete the turnaround |
| Microsoft | Winner in the PC era | — |
1. "Time compression" is the core framework for understanding the current AI competition: Product lifecycles have shortened from years to months, and will soon shrink to days. Schmidt believes this is a level of competitive intensity never seen before in history.
2. The fundamental divergence between US and European AI strategies: The US approach is "move fast, fix problems later"; Europe's is "draw up a list of prohibitions first." Schmidt argues the latter directly stifles the startup ecosystem.
3. Agents are not tools; they are levers for organizational flattening: A CEO can use AI to monitor the work status of all employees, automatically advance projects, and instantly aggregate business opportunities — this is not merely an efficiency gain but a fundamental shift in power structure and decision-making speed.
4. Coordination roles could come under pressure: Citing layoffs at a friend’s company, Schmidt argued that AI could replace some planning and coordination work. This was his assessment, not an established outcome for all such roles.
5. Technical leadership does not guarantee leadership in user numbers: Schmidt argued that Chinese models could attract most users globally through free distribution even if they lagged technically by about six months. His comment that “we are already losing” referred to the Western side measured by user adoption, not to China having lost.
6. Founder returns are not about nostalgia, but about history: Schmidt believes this is a historical period that people will write about 100 years from now, and key players want to "have a seat at the table."
7. Schmidt's self-awareness: His biggest flaw is being "too slow" — looking back, everything was foreseeable, but others saw it faster. He admits that his inability to "shut down" is unhealthy, but he cannot change it.