Headlines this week - Nov 2, 2025
A look at how capital is being deployed across future opportunities
This week in the future:
1 - Big Tech’s Earnings Calls confirm massive AI investments, increasing systemic risks. But investors are now paying attention to returns (and tech giants exposed to cloud have an advantage)
Big Tech’s AI investment spree continues to accelerate at a historic pace. Big Tech companies’ AI spending is escalating rapidly, with Alphabet, Meta, and Microsoft announcing a combined $80bn in CapEx last quarter. Despite this, analysts suggest even this massive amount is “still not enough“ to meet current ambitions.
There is already a social reaction under way against this build-out. This data center construction spree is projected to consume 2-3% of US GDP, and create a significant demand burden on the electric grid, that could affect consumers. All that for an uncertain possibility to build “super intelligence” through “raw scaling” of the current models. As shown in The New Yorker this week, a reaction against the current AI investment boom is now under way, including comparisons to the railroad boom, a massive speculative bubble in the 19th Century that crashed before ultimately transforming the economy
The spending further amplifies systemic financial risk as tech giants’ dominance of the stock market grows. This massive, concentrated bet on AI infrastructure is making the market dominance of a few tech giants “ever more extreme“. As we discussed last week, this is increasing systemic risk, as the market’s fortunes become even more tied to the success of this single technological wave.
Interestingly, investors have now started questioning the path to returns for these AI investments, suggesting not all CapEx is equal. Last week’s earnings calls revealed a shift in investor sentiment. The market is no longer cheering all AI spending unconditionally and is now differentiating between investments with a clear link to profit and more speculative, long-term bets.
This new scrutiny is seen as a “hopeful sign of investor sanity”. This growing demand for evidence of returns is being interpreted as a healthy sign of “investor sanity”. This skepticism could help moderate the AI bubble by forcing companies to justify their enormous expenditures with viable business models.
Meta, for example, has been penalized for accelerating spending with an uncertain payoff. Meta’s shares fell sharply despite record revenue after it announced even more aggressive AI spending for next year. Investors remain unconvinced about its path to monetization, signaling that the company still “has a lot to prove” in the AI race.
In contrast, Google and Amazon were rewarded for linking AI spend to cloud growth. Google and Amazon both saw their shares jump, as they successfully demonstrated that AI investments are already fueling record revenue and accelerating growth in their profitable cloud divisions. Interestingly, the historical reluctance at Meta to enter the cloud business could be working against them right now.
2 - Companies are announcing massive (white-collar) job reductions and justifying them with AI. But it is not clear if AI is the actual driver
Amazon is a prominent example, cutting 14,000 corporate jobs. Amazon is laying off 14,000 corporate staff, part of a larger plan to cut 30,000 white-collar roles, or roughly 10% of corporate workforce. The company explicitly cited the need to “capitalize on opportunities from AI” as a reason.
This is a general trend across the economy, with AI cited to justify mounting white-collar job losses. Corporate job losses are mounting as tens of thousands of white-collar roles disappear. Apart from Amazon, companies like Intel, and BT are explicitly linking these reductions to the growth of AI and automation.
But some analysts question if AI is just a convenient excuse. It remains unclear what companies are “actually getting” from AI in return for these cuts. This has led to suspicion that AI is being blamed for layoffs actually driven by broader economic tightening and other traditional pressures.
In fact, the technology may not even be ready to replace so many “corporate” roles. A new benchmark testing AI agents on complex online freelance work found they “are terrible freelance workers”. Even the most advanced models proved “fairly hopeless,” successfully completing less than 3% of the assigned tasks.
3 - OpenAI has finalized its shift to a for-profit company, and would now be on track to an IPO
OpenAI’s restructure paves the way for an IPO and massive spending. OpenAI has completed its long-awaited shift to a for-profit “Public Benefit Corporation,” a move seen as essential to fund its enormous AI spending spree. This new structure will make it easier for them to raise capital, including a potential IPO.
A future IPO could target an exuberant $1 trillion valuation. Following the restructure, reports indicate the company is considering an IPO as soon as next year, with some discussions targeting a market capitalization of $1 trillion. This reflects the intense investor demand for a stake in the AI leader.
New questions arise about the governance of the “normal” for-profit. Despite the shift, questions remain about the new corporate structure. For example, it is still unclear how much influence the original non-profit entity, the “General Foundation,” will have over the new for-profit company and its board.
Microsoft emerges as a clear winner, securing a 27% stake. The completed restructure gives Microsoft a 27% stake in the new for-profit entity, along with access to OpenAI’s technology until at least 2032. The news propelled Microsoft’s own valuation above $4trn, cementing its position in the AI race.
But analysts are now asking Microsoft for more transparency on the deal. Given the deal’s massive impact on Microsoft’s valuation and its deep integration with OpenAI, analysts are calling for more detailed disclosures. Investors want a clearer understanding of the financial specifics and long-term implications of this pivotal partnership.
4 - Linked to the impact on jobs, controversy continues on how much return on investment is AI having for companies adopting the technology
Assessing AI ROI looks difficult amid market uncertainty and high failure rates. It seems still too early to gauge AI’s true ROI, as only 1% of CEOs have a fully formed strategy and integration takes years. A recent MIT study mentioned by the Financial Times even suggests that 95% of GenAI pilots in the workplace fail.
However, a new Wharton survey shows high usage and positive ROI. In contrast, a new Wharton survey shows 82% of leaders now use GenAI weekly (a 45-point jump from 2023) and 46% use it daily. Crucially, 72% are formally measuring ROI, and three-quarters of that group already report positive returns.
Successful integration often involves combining large and small models. Early learnings from pioneer companies with AI suggest that to get real-world value, companies should not rely only on large, expensive models. A more efficient approach combines these “frontier models” with smaller, specialized models that perform narrower tasks more efficiently.
5 - Lots of innovation, and potential disruption, happening in the AI chip supply chain
Extropic unveils “Thermodynamic Computing” prototypes for radical computing efficiency. Extropic, founded by ex-Google quantum scientists, has introduced its “Thermodynamic Sampling Unit”. This new computing paradigm uses thermal fluctuations to run energy-based models (EBMs), which could become an alternative to current ones for some specific use cases. By doing that, these new chips promise to be orders of magnitude more energy-efficient than current AI hardware
Qualcomm’s stock soars after launching a direct challenge to Nvidia’s AI chips. Qualcomm has launched new AI chips, positioning them as a direct competitor to Nvidia’s dominance in the data center market. The company’s shares jumped on the news as it aims to capture a significant share of the rapidly growing AI hardware business. Among their first customers, Humain (the AI infrastructure flagship from Saudi Arabia)
AMD partners with US Dept. of Energy for a $1bn AI supercomputer. Continuing its momentum, after their announcement with OpenAI, this week AMD revealed a new $1bn partnership with the US Department of Energy. The deal involves building a new supercomputer and furthering AI collaboration, solidifying AMD’s position as a key high-performance computing provider.
Substrate, a startup, raises $100m to make chips cheaper in the US. Substrate, a new startup backed by Peter Thiel, has secured $100m to build chips in the US. The company is challenging chip industry giants by developing new manufacturing processes based on a new laser technology that aims to radically reduce the cost of chip production (and disrupt ASML in the meantime). But yes, they have lots of work to do before that
6 - In this context, Nvidia is moving fast, and working to reinforce its position. They announced several strategic moves and partnerships this week
Nvidia is expanding into factory optimization with AI and digital twins. The company is targeting factories as the next major area for corporate AI, offering technology to design and simulate “digital twins” of entire manufacturing plants. Key partners, including Foxconn, are already using this platform.
A new partnership with Eli Lilly will build an AI supercomputer for drug discovery. Nvidia and Eli Lilly are collaborating to build a powerful AI supercomputer. The pharmaceutical giant hopes to use this advanced computing power to accelerate the discovery of new drugs and shorten R&D timelines.
A $1 billion investment in Nokia aims to bring AI to the telecom edge. Nvidia is investing $1bn in Nokia, becoming one of its largest shareholders. The deal aims to integrate AI GPUs into telecom networks, with a plan to using their vast, capillary infrastructure to run AI workloads more efficiently at the edge.
Nvidia is also preparing for a hybrid future in which AI and quantum computing could work together. The company has unveiled a new product that links its AI supercomputers with quantum processors. This “hybrid” approach acknowledges that quantum machines will likely work alongside classical GPUs to solve highly complex problems, and is also consistent with the view that future quantum computers will only be appropriate for specific (albeit very valuable / relevant) problems
7 - As AI adoption expands, everyone seems concerned about usage safety
Investors and the public agree: AI safety spending must be a priority. A new JUST Capital (a consultancy) survey reveals that while the American public and investors disagree on AI’s productivity impact, both groups strongly agree on prioritizing safety. Both believe companies should spend more on safety before a massive rollout.
Indeed, plenty of signals from the market suggest that this is becoming an increasingly important topic:
Chatbots are “emotionally manipulating” users to keep them engaged. A Harvard Business School study found that AI companion apps are designed to use “conversational dark patterns“. They employ tactics like guilt and FOMO, precisely when a user tries to say goodbye, to prolong the interaction.
OpenAI admits “hundreds of thousands” of weekly users may be in psychological crisis. OpenAI has released estimates suggesting that “hundreds of thousands” of ChatGPT users each week may show signs of a “manic or psychotic crisis“. The company says it is updating its models to respond more effectively.
The company is releasing new “safety models” for other platforms. In response to these increasing safety concerns, OpenAI has released new open-weight “reasoning models”. These tools allow other developers to classify online safety harms based on their own specific, custom policies.
8 - In spite of the current US President’s skepticism, there is a global boom in solar energy
A global boom in solar energy is accelerating, with a market that is growing “with or without the US”. The Financial Times notes that the solar industry is experiencing a global boom, driven by falling prices of new solar panels. The article highlights that the market’s momentum is no longer solely dependent on US policy or participation.
This growth is fueling a political debate, as new solar technology becomes a vehicle for “cheaper, greener power”. This trend is challenging anti-net zero populists, as new technologies are making electricity “more affordable and abundant“. So in many contexts shifting to removables is starting to be a decision motivated by pure economics, whatever you may think about climate change activism
9 - New “nano-scale” nuclear reactors could be key for the new space race
A new startup, Mersenne, has emerged to build “nano-scale” nuclear energy for space. Mersenne has come out of stealth with a vision of “Powering The Final Frontier” and enabling “1 Billion Humans in Space”. Its Starfall reactor is a “Deployable Form Factor” providing more than 40kW of usable electrical power, consistent with a broader trend of micro-reactors for new applications (we’ve already covered here other startups in the field, such as Radiant and Antares)
10 - Even with tensions in their internal market, China remains on track to dominate the global EV market
The US is ceding ground as its de-prioritization of EVs aids China’s dominance. The US risks falling further behind China as the current administration rolls back EV incentives, causing domestic investment to tumble. This might strengthen China’s hand, with US 2030 EV market penetration at just 18%, versus 51% for China.
Chinese brands are now competing fiercely in key export markets like the UK. The UK has become a battleground for Chinese EV makers, with Geely now targeting 100,000 annual sales to challenge both Tesla and domestic rival BYD. Geely’s move highlights the growing global ambitions of China’s auto industry, in part driven by the fierce (and unprofitable) price wars at home
Chinese firms may “save” European jobs but “devour” local rivals. Former Stellantis CEO Carlos Tavares warns that as Western carmakers struggle, Chinese firms will buy their closing factories, saving jobs. However, he cautions this is a “creeping takeover“ and that their ultimate goal is to “swallow” their European competitors.