Headlines this week - Oct 26, 2025
A look at how capital is being deployed across future opportunities
This week in the future:
1 - Google announced a significant technical breakthrough in Quantum Computing. But there is still lots of work to do
Google claims a significant advance, potentially achieving “quantum advantage”. For the second time, and with a more robust claim, the company asserts its quantum computer has outperformed classical supercomputers. The experiment used Google’s superconducting Willow chip, running a (specific) algorithm 13,000 times faster than the world’s best supercomputer.
The algorithm used offers potential, but not immediate, practical application. While Google didn’t claim immediate practical use, the “Quantum Echoes” algorithm demonstrated a technique potentially applicable to simulating molecular interactions, relevant for fields like materials science and drug discovery.
Experts are “thrilled” but emphasize the long road ahead. Leading quantum expert Scott Aaronson was “thrilled” by the announcement but stressed that significant work remains before commercially useful quantum computers are realized.
The gap between current devices and fault-tolerant machines remains substantial. Other top experts like Caltech’s John Preskill concur, highlighting the “substantial gaps“ separating today’s noisy intermediate-scale quantum (NISQ) devices, like Google’s Willow, from future fault-tolerant, application-scale (FASQ) machines needed for practical advantage.
Google maintains a conservative timeline for useful quantum computers. Consistent with past statements, Google remains cautious about timelines. Earlier this year, CEO Sundar Pichai estimated that “useful” quantum computers are likely still 5-10 years away (i.e. most likely, in the 2030s).
Competitors are also running in a “communications race” to maintain investor confidence. Almost immediately after Google’s news, IBM revealed a method to run quantum error-correction algorithms using conventional AMD chips (FPGAs). While not accelerating quantum timelines directly, the announcement boosted IBM’s shares, and showed companies’ intense focus on demonstrating progress in the field.
Meanwhile, government interest and investment in Quantum Computing are heating up. The Trump Administration is paying attention. This week we learned that they are negotiating equity stakes in several Quantum Computing firms, including Rigetti, D-Wave, and IonQ.
2 - Recent events suggest an increasing systemic risk in internet infrastructure and the broader economy
A massive AWS outage this week highlighted fundamental weaknesses in internet infrastructure. Monday’s widespread Amazon Web Services outage, stemming from its critical US-EAST-1 region, is seen as revealing long-standing vulnerabilities in the internet’s core architecture, particularly the reliance on centralized systems like DNS. As we increasingly rely on these infrastructure for almost everything we do, this could be seen as a warning signal.
The incident also fueled European calls for digital sovereignty, but achieving that remains challenging. The outage intensified concerns in Europe about over-reliance on US tech giants for critical digital infrastructure. However, building truly independent European cloud capabilities faces significant hurdles, including investment scale and political coordination.
At a different level, OpenAI’s deep ties with suppliers are creating new systemic risks for the tech ecosystem. By forging multi-billion dollar deals and infrastructure partnerships with Nvidia, Microsoft, Oracle, AMD, and Broadcom, OpenAI has deeply intertwined its fate with major tech players. A potential failure at OpenAI could now have significant cascading effects across the industry.
These “circular” deals echo the risky vendor financing seen in the dot-com era. The current flurry of deals where suppliers invest heavily in their major customers (like Nvidia/OpenAI) draws parallels to the telecoms boom 25 years ago. Back then, vendor financing created interconnected risks, and when customers failed, suppliers like Lucent also suffered significant losses.
Should we prioritize resiliency over pure efficiency? Comparing the AWS outage to fragility in rare earth supply chains and past information access issues, Ben Thompson (Stratechery) argues for prioritizing resilience, even at the cost of some efficiency. Over-optimization for scale can create critical single points of failure in essential systems.
3 - How AI is disrupting the cloud: Anthropic signs a multi-billion dollar cloud deal with Google
Anthropic is committing tens of billions to use Google’s custom AI chips. Anthropic has struck a blockbuster deal, valued in the tens of billions of dollars, to access Google Cloud’s custom Tensor Processing Units (TPUs). Google will bring over a gigawatt of TPU capacity online for Anthropic next year.
The deal deepens ties between Anthropic and Google, already a major investor. This agreement builds upon Google’s existing relationship with Anthropic, having previously invested more than $3bn in the AI startup. The partnership provides Anthropic with significant computing power for its AI models.
But this move comes despite Amazon being Anthropic’s largest backer. The commitment to Google Cloud is notable given Amazon has committed roughly $8bn to Anthropic, significantly more than Google. Previously, Anthropic was expected to heavily utilize AWS and its Trainium chips, seen as key to Amazon’s AI strategy.
4 - As the electric vehicle market slows down, leading vendors are starting to shift focus to autonomous cars
Elon Musk pivots focus to autonomy amid Tesla’s EV profit decline. During Tesla’s earnings call this week, Elon Musk emphasized the coming “shockwave” from autonomous vehicles, shifting the narrative (from electric cars to full autonomous ones) as the company reported a significant drop in profits.
The context for this: Tesla’s EV profits fell sharply despite record sales. Tesla’s adjusted net income dropped by more than a quarter compared to the previous year. Factors included tariff costs, loss of emissions credit revenue, and heavy AI investments, overshadowing record vehicle deliveries.
Almost simultaneously, GM says they are also doubling down on autonomous driving features. Following its recent scale-back of EV plans, General Motors is now promoting advanced driver-assistance features. The company announced plans for “hands-free, eyes-off” highway driving capabilities in its Cadillac Escalade IQ, targeted for 2028.
5 - Is experience dead? AI was supposed to be the junior-job killer, but it could also destroy learning / experience economies
AI simulations threaten to neutralize experience-based competitive advantages. According to a WSJ opinion piece this week, complex business processes can now be simulated in detail, allowing less experienced companies to “learn” rapidly and eroding the traditional learning curve advantage held by established players. This could shift competition towards adaptability and speed.
At the same time, rapid AI evolution demands continuous training, devaluing past expertise. The fast pace of AI development requires constant employee retraining to adapt to new tools and workflows. Consequently, deep experience with older processes becomes less valuable compared to the ability to quickly learn and apply new AI-driven methods.
Meta’s AI layoffs may illustrate this shift in valuing talent. Recent job cuts in Meta’s AI division, affecting “legacy” teams while the company continues aggressive, high-cost hiring for new AI roles, could exemplify this trend. It suggests a potential de-prioritization of incumbent experience in favor of fresh talent aligned with new AI paradigms.
6 - Are the missions to the Moon and Mars at risk? NASA faces friction with SpaceX
NASA is seeking alternative Moon landers, claiming that SpaceX faces delays. Amid concerns that SpaceX’s Starship development is behind schedule for the Artemis program, NASA is soliciting proposals for alternative lunar landers. The agency aims to ensure redundancy and competition for future missions.
As a result, Elon Musk has publicly clashed with NASA leadership. Following NASA’s expressed concerns and search for alternatives, Elon Musk publicly criticized acting NASA Administrator Sean Duffy, questioning his qualifications and the agency’s direction. This public spat highlights the growing tension between the agency and its key contractor.
(Broader) internal NASA issues and potential over-ambition could jeopardize the Moon race against China The friction with SpaceX occurs amid broader turmoil at NASA, including leadership gaps and budget pressures, potentially delaying the Artemis timeline. Some argue NASA’s complex Artemis architecture is overly ambitious compared to China’s more straightforward lunar plans.
Debate also emerges on the complexity and the ultimate goal of a mission to Mars. Beyond the Moon, the long-term vision for Mars is also being questioned. Some experts argue against permanent colonization, suggesting the focus should instead be on scientific discovery and resource utilization that benefits Earth.
7 - Nucleus launches Origin, an open-source model that promises to “predict human longevity on embryo”
Nucleus Genomics has released open-weight AI models for In-Vitro Fertilization (IVF) and genetic optimization. Nucleus, a startup we’ve been following here, has launched Origin, a family of nine open-weight AI models, claiming benchmark performance in predicting health risks and longevity from embryo DNA. The models were trained on millions of diverse genetic profiles.
The models aim to inform embryo selection by forecasting disease risk. Origin predicts risks for conditions like Alzheimer’s and cancer across ancestries, potentially allowing IVF couples to reduce disease likelihood in offspring. Nucleus states selection among five embryos could halve the risk for common diseases.
8 - Amazon evolving into a robot-first operational model
Visiting a modern Amazon fulfillment center can be an “unsettling” glimpse into the future. Highly automated facilities feature robots managing vast inventories, moving merchandise, and directing human workers on where items should go. An investor described his experience at the WSJ this week, and said it was like observing a machine orchestrating human activity.
Amazon plans to replace hundreds of thousands more jobs with robots. Internal documents reviewed by The New York Times reveal Amazon’s ambition to replace potentially more than half a million warehouse jobs globally with robotic systems over time. This automation push aims to increase efficiency and reduce labor costs.
Remaining human workers will increasingly take direction from AI and robots. Amazon is also testing new AI tools and robotic systems, including Digit humanoid robots, to further automate tasks. This points towards a future where algorithms and robots manage workflows, essentially telling human employees what tasks to perform (not so different from what the WSJ article describes).
9 - Medicine emerging as a “killer” use case for chatbots
OpenEvidence, a “ChatGPT for medicine,” raised $200m amid soaring demand. OpenEvidence secured $200 million funding, valuing the AI startup at $3.5bn. This reflects massive user growth among clinicians and rising patient use of chatbots for self-diagnosis and second opinions. Companies like Microsoft, and other startups, are also trying to capture the opportunity.
10 - Like Google DeepMind and others, Anthropic wants to contribute to accelerate drug discovery with AI
Anthropic is tailoring Claude for life sciences and drug discovery. Following rivals like DeepMind, Anthropic is adapting its Claude chatbot to integrate with tools used by scientists. The aim is to accelerate research by assisting with tasks like data analysis and literature review in the life sciences sector.
The AI will connect with biological databases and lab software. Claude is being integrated with specialized platforms, including lab management systems, genomic analysis tools, and biomedical databases. Early users like Novo Nordisk and Sanofi report significant time savings on tasks such as clinical study documentation.