Bezos, Schmidt, and Powell Jobs: How can the three AI investment philosophies bet on three very different futures?
Author: Shenzhao TechFlow
Original title: Bezos, Schmidt, and Powell Jobs: Three AI Investment Philosophies of Silicon Valley's Old Money
They are investing in three completely different futures.
On November 17, 2025, 61-year-old Jeff Bezos became the CEO of a company again. He returned to the trading position for the first time since stepping down from Amazon in 2021. The new company, Project Prometheus, has an initial capital of 6.2 billion US dollars, focusing on “physical AI” and targeting the manufacturing industry.
Seven months ago, 70-year-old Eric Schmidt owned a rocket company called Relativity Space and became CEO himself. He didn't explain why this age is coming to an end; maybe “every day matters in the AI era” is already the default answer.
In June of the same year, Jobs' widow, Lauren Powell Jobs, gave a rare public interview. She sits next to Jony Ive and talks about the prototype she saw at io. It was an “AI device” that was taken over by OpenAI for $6.4 billion. It had no screen and was said to look like a player hanging around the neck. Commenting on the prototype, she said, “It's incredible to watch an idea become a reality.”
Three people, three postures. But they all bet at the same casino.
Over the past three years, Silicon Valley's top purses have been doing almost the same thing: putting money from family offices, venture capital, and charitable foundations into AI. Schmidt, Bezos, and Powell Jobs are just three of the most prominent. But if you take a close look at their target list, you'll find that this isn't the same game; they're voting for three completely different futures.
Schmidt: Think of AI as the next Cold War
Schmidt's family office, Hillspire, has invested in over 22 AI companies since 2019, totaling over $5 billion, according to data quoted by Wikipedia and The AI Insider. The list includes Anthropic, SandboxAQ (a quantum+AI company spun off from Alphabet), Inworld AI, Holistic AI, and Altera. These are the targets that “people in the industry” will list.
But what really revealed his undertones was another list.
White Stork: A company that makes AI drones in Ukraine. Rebellion Defense: Defense AI. Istari: simulation simulation. Swift Beat: military software. This is a family office that uses AI as the next generation of military equipment.
Schmidt has chaired the Defense Innovation Council since 2016 and co-led the National Artificial Intelligence Security Council from 2019 to 2021. He's a player who treats AI policy, defense procurement, and energy infrastructure as the same thing. In January 2024, “Forbes” revealed that he simultaneously launched the White Stork drone project in the US and Ukraine, using the Ukrainian battlefield as a “laboratory for AI weapons.”
Then there's infrastructure.
In January 2026, he co-founded a company called Bolt Data & Energy with Texas Pacific Land and became the chairman himself. This company doesn't rent computer rooms or buy grid electricity. It has to build its own natural gas power plant in the wilderness of West Texas and pour electricity directly into the data center. The plan is to first reach 1 gigawatt, and eventually to 10 gigawatts, which is equivalent to the electricity consumption of 7 million households. Texas Pacific Land paid $50 million, plus priority water supply rights. In response to an email to Fortune, Schmidt said, “The biggest bottleneck facing AI is not algorithms; it is energy.”
In March of the same year, he took control of Relativity Space. The company is making a reusable rocket called Terran R, with the goal of impacting SpaceX's monopoly on medium- and low-orbit launches. The order was $2.9 billion as of that time.
Put these together, and the logic is very clear.
Schmidt doesn't believe in “investing in a basket of big model companies.” He believes AI will ultimately win or lose on three things, namely computing power (data centers and electricity), delivery (rockets, satellites, drones), and policy (defense committee and congressional hearings). He also voted for the model company. After DeepSeek came out, he also publicly wrote an article calling for the US to increase investment in open source, but that was just a pawn on his board, not all.
His response to DeepSeek was very telling. After DeepSeek was released in early 2025, Schmitt wrote an article in the Washington Post saying it was “a turning point in the global AI competition.” His prescriptions were not a retreat, but a step up, including more open source, more Stargate-like infrastructure, and more sharing of training methods between model labs.
In other words, he sees AI as an endurance race between countries, and he himself is already standing on the side of the racetrack and is also a member of the organizing committee. He came out as the CEO of Relativity at age 70. Outsiders looked like a mess. He himself explained that Kissinger lived until the age of 100, and “a period of major change required responsibility, not withdrawal.”
Bezos: Full Stack Control Freak
Bezos's style of play was completely different from Schmidt's.
As of 2026, Bezos has deployed more than $19 billion in AI, according to data from TechCrunch, The Information, and the Bezos Earth Fund, according to StartupHub. That number is still rising.
If you take it apart, it's mainly divided into three parts.
The first piece is Anthropic. Amazon began investing $8 billion in September 2023, and promised an additional $25 billion in April 2026. Anthropic runs on AWS and uses Amazon's Trainium chip. This ties Amazon's cloud infrastructure, Bezos' model layer bets, and Anthropic's research and production capacity into a triangle, not just financial investments. When Anthropic's valuation soared to over $60 billion, Amazon had eaten the biggest piece of outside cake.
The second piece is Bezos Expeditions' scatter investment. Bezos Expeditions is raising a multi-billion dollar AI special fund to upgrade the “Bezos Personal Angel” into an “institutional investor.” Among them, it invested in Perplexity, an AI search company, which rose from a valuation of $520 million in January 2024 to 20 billion in September 2025.
The third block is Project Prometheus.
In November 2025, Bezos and Vik Bajaj, a former Google X executive, jointly announced the establishment of this company, with a starting capital of 6.2 billion US dollars, with nearly 100 employees. The team members were drawn from OpenAI, DeepMind, and Meta. The founding consultants included Ashish Vaswani and Jakob Uszkoreit, the two authors of the 2017 “Attention Is All You Need” . The company aims to use AI in manufacturing, including automobiles, spacecraft, and chips.
Why manufacturing? Because it fits right in with Bezos' other businesses. Amazon has the Kuiper satellite constellation, and the first customers after the manufacturing AI was launched were in his own home.
Musk called Project Prometheus a “copycat” on X.
But structurally, this isn't plagiarism.
Bezos holds the model layer through Anthropic, the application layer through Perplexity and Figure, and the computing power layer through Amazon. Now we're reinventing Prometheus, incorporating AI into manufacturing, and putting the “physical world execution layer” next. This is a full-stack style of play. From training chips to deployment to the factory floor, each layer has its own card.
About 10 days after launch, Project Prometheus quietly acquired a company called General Agents. This company makes a “computer agent”, an AI agent that can directly operate the entire computer. WIRED later revealed that it only took four days from the beginning of this acquisition to complete delivery.
Donely's CEO Harsha Abegunasekara commented, “What General Agents really cracked was speed; Ace runs almost instantaneously on your computer.” His company was a competitor to General Agents.
From angel investing to forming a special fund to personally becoming CEO, it only took Bezos 18 months. He's actually building a bigger system than Amazon.
Powell Jobs: The low-key faction
If you look at these three people together, Powell Jobs is the least like an “AI investor.”
Her family office, Emerson Collective, has invested in at least 9 AI-related startups since 2022 and participated in a total of over $1 billion in funding rounds, according to CNBC, citing private wealth data platform Fintrx. This number is not on the same scale as Schmidt or Bezos.
But what's interesting is the list itself.
Proximie: Remote Surgery Connectivity Platform; Atropos Health: Clinical Data AI; Formation Bio: AI Pharmaceuticals' Curipod: Norwegian AI teaching tool; Mistral: France's big model company, Europe's sole player against OpenAI.
There are no national defenses, data centers, and rockets.
Emerson Collective's official website clearly describes the investment direction: education, energy and environment, digital health, fintech, media. AI is just one tool interspersed with these topics. She holds a majority stake in “The Atlantic Monthly” and is very skilled in Colombian-style “soft power” investments.
But the one she actually hit was not on the same line as the previous ones.
After Jony Ive left Apple in 2019, Powell Jobs voted for his design company LoveFrom through Emerson Collective. Ive later said in an interview with the Financial Times, “If it weren't for Laurene, there wouldn't be LoveFrom.” A few years later, Ive founded another hardware company called io, which specializes in AI devices, and Powell Jobs followed suit. In May 2025, OpenAI bought io in a $6.4 billion all-stock deal, making Ive a billionaire on the books. Emerson Collective followed suit.
Another investment was also critical; Emerson Collective was one of Mistral AI's early investors. At the time, the French company was still the only one left on the big model in Europe.
Putting these together, her AI bet is focused on two directions, either “using AI to solve specific human problems, or “reshaping the way humans and machines interact” (io's device, Ive's design).
VC Sheet described Emerson Collective in an assessment: “A deliberately vague LLC that puts venture capital, philanthropy, policy advocacy, art, and media ownership under one roof and can use any of the most effective tools of grants, policy lobbying, or investment.”
Philosophically, she is closer to the older Dongan Family Office. Influence is more important than return, long term is more important than short term, and microphone is more important than spotlight.
Three investment philosophies
Put the three lists together, and you'll see three sets of judgments about the future of AI.
Schmidt is betting on national competition and infrastructure bottlenecks. In his world, AI will ultimately be determined by “who has the most electricity, who has the fastest rocket, and who has the most powerful drone.” The model is just an entrance ticket; the real moat is in the physical layer. So he personally ran to be the head of Relativity and Bolt. What he wanted was not reward, but control.
Bezos is betting on the spread of applications at the level of the industrial revolution. He believes AI will eventually infiltrate every machine tool, every aircraft, and every satellite like electricity. So he locked the model layer through Amazon, the manufacturing layer through Prometheus, and embedded the consumer application layer through Expeditions. What he is betting on is not whether a certain company can win, but whether the “whole” structure can win.
Powell Jobs was betting on something else. In the end, she can't stand the current model of human-computer interaction when gambling. In interviews with the Financial Times, she and Ive repeatedly emphasized “humanity deserves better.” The reason behind her investment in io, LoveFrom, medical AI, and education AI is the same judgment, that is, the biggest market in the next ten years is to “fix the side effects of the Internet in the previous decade.”
Three sets of judgments, three styles of play.
Which one is right? No one knows. Schmidt probably overestimated the weight of geopolitics in the AI economy. Bezos probably underestimated the capital consumption of the “full stack” asset-heavy model. A typical example is that Prometheus hasn't shipped yet, and it is already rumored that it will finance another 10 billion dollars. Powell Jobs, on the other hand, has to face an even more embarrassing problem: io's devices will not be mass-produced until 2027, and OpenAI's own financial model has been repeatedly questioned by the market.
But one thing is certain. When the winners of the previous generation of the internet collectively turned their family funds to AI, this was no longer a small outlet for a certain racetrack. Bolt has raised 150 million dollars in start-up capital, and Anthropic alone ate the $33 billion promised by Amazon. Capital flows of this magnitude will themselves shape the industrial geography of the next ten years.
As to who will have the last laugh, we'll have to wait until 2030 to look back. Before that, all three old players were still on the table, and their chips were still being added.
Twitter:https://twitter.com/BitpushNewsCN
Compare the TG exchange group:https://t.me/BitPushCommunity
Compare TG subscriptions:https://t.me/bitpush
