Headlines this week - Nov 23, 2025

A look at how capital is being deployed across future opportunities

1 - AI investments becoming a credit event, as even Big Tech companies are running out of cash

• Aggressive spending in AI is eroding Big Tech’s cash flows. The intensity of AI investments is “making Big Tech weaker” by pressuring balance sheets and depressing free cash flows for companies like Microsoft and Amazon, forcing investors to re-evaluate these firms as capital-intensive builders rather than just software giants.

• So even Amazon is now starting to use debt to fund its AI race. The tech giant is raising $12bn in its first bond sale in three years, to fund data centers, joining Alphabet and Meta in shifting from cash to debt to finance the AI build-out. Amazon’s capital expenditures rose 61% to $34.2bn in the third quarter alone.

• The massive AI build-out is becoming a credit market event. Big Tech’s aggressive “debt binge” to fund infrastructure is raising risks, with analysts warning that the flood of new bonds (expected to hit record issuance levels) could reshape the corporate debt market and alter investor risk calculations.

• Private capital looking for opportunities will share the burden. Beyond corporate debt, Funds like Brookfield are also pouring capital into AI projects. Brookfield is raising a specific $10bn AI infrastructure fund backed by Nvidia and the Kuwait Investment Authority, with plans to build and acquire up to $100bn worth of data centers and power assets.

2 - “AI Exuberance” continues: Investor skepticism clashes with a renewed political and corporate push for dominance

• This week we learned that many global fund managers believe the AI boom has gone too far. A Bank of America survey finds a record net 20% of investors believe companies are overspending on infrastructure, with half explicitly calling AI a bubble and identifying it as a major tail risk for markets.

• Meanwhile, big public investments could be on the way, with Trump preparing a new “Manhattan Project” for AI. The incoming administration plans to unveil a “Genesis Mission“ to treat AI development as a national emergency, utilizing executive orders to bypass state regulations and accelerate industry growth through federal deregulation.

• At the same time, FOMO seems to be a stronger force than skepticism, and valuations (and risks) keep growing. As an example, Elon Musk’s xAI is seeking $15bn at a significantly higher valuation (around $230bn, almost 2x in just 6 months). The company is looking to secure the funds needed to buy chips and build data centers to compete with OpenAI and Google.

3 - Nvidia’s record-breaking earnings fail to cure the market’s vertigo

• Nvidia’s financial performance suggested the AI boom was far from over. The chipmaker’s strong earnings report blew past expectations, signaling that fears of a slowdown were premature and that demand for AI infrastructure remains robust despite recent market volatility.

• The company’s CEO explicitly dismissed “AI bubble” concerns. With revenue jumping +62% to $57bn and beating Wall Street estimates, Jensen Huang declared the industry sees something “very different” from the skeptics, pointing to sustained demand for next-generation Blackwell chips.

• But even a blockbuster earnings report could not sustain a broader market rally. While Nvidia’s results initially sparked optimism, the momentum quickly faded as lingering concerns over high valuations and economic headwinds dragged major indices lower, proving that strong AI fundamentals alone cannot prop up a wobbly market.

4 - The Next AI Frontier: Moving beyond LLMs toward “World Models” and physical intelligence

• It is confirmed: AI pioneer Yann LeCun is leaving Meta to launch a startup focused on “world models.” The “godfather of AI” is now expected to found a startup to build systems that reason and understand the physical environment, moving away from Large Language Models, which he argues have reached their limits, in pursuit of true machine “superintelligence”.

• This is being perceived as a “quiet revolution” that would be brewing behind the current AI hype cycle. While Nvidia dominates headlines, experts are pivoting toward architectures that mimic human learning and spatial intelligence, raising questions about whether Big Tech’s massive capital expenditures on current LLMs could eventually become “stranded assets“ if the technology shifts.

• Investors are placing multi-billion dollar bets on equipping robots with general-purpose brains, that would look to learn from the physical world. Startup Physical Intelligence has raised $600m at a $5.6bn valuation to develop software that allows robots to learn from real-world tasks, validating the industry’s growing focus on bridging the gap between digital AI and physical action.

5. The Great Deregulation: Europe and the US race to dismantle part of the existing AI guardrails

• Trump weighs an executive order to preempt state AI laws. The Trump administration is preparing a federal initiative to override a “patchwork” of local regulations in states like California and New York, arguing that inconsistent rules hinder innovation. The order would task the Justice Department with challenging state laws deemed overly restrictive, favoring a unified, lighter federal approach to help the US win the global AI race.

• Simultaneously, the EU admits it may have botched its attempt to regulate AI. Facing heavy lobbying from tech firms and member states like France and Germany, officials plan to delay or limit key pieces of the AI Act, including narrowing the definition of “high-risk” applications. Critics warn this deregulatory “race to the bottom” signals that Europe’s complex rulebook is being dismantled to avoid stifling the region’s faltering economic growth.

• More generally, Europe is also rethinking its aggressive approach to Big Tech. Fearing that overregulation has left the continent struggling to compete with the US and China, European policymakers are drafting a “digital simplification package“ to scale back landmark rules. The initiative represents a major shift from the “Brussels effect” era, with plans to rewrite parts of the GDPR to make it easier for companies to use data for AI development.

6 - The Quantum Reality Check: Scalable machines require an industrial revolution, not just better physics

• Scalable quantum computing faces a massive manufacturing bottleneck. Nobel laureate John Martinis warns that at the current pace, useful machines are decades away, arguing the industry must pivot from “jungle of wires” prototypes to mass-producible integrated circuits to reach the necessary one million qubits.

• He already anticipated this ideas in this interview from 3 weeks ago. In the interview, he explains the fundamentals of quantum mechanics and his prize-winning research. He also discusses the current state of quantum computing, the impact of AI, and the intensifying technological race between the US and China.

7. The Nokia Reinvention: A pivot to enable “AI at the edge” backed by a massive US bet

• Nokia pledges $4bn to expand US operations in partnership with the Trump administration. The Finnish telecom equipment giant will invest $3.5bn in R&D and $500m in manufacturing across New Jersey, Texas, and Pennsylvania to optimize network infrastructure for AI and national security applications.

• The company is attempting a difficult “third reinvention” to become a software-led player. Following a recent stock bounce fueled by Nvidia’s backing, Nokia is pivoting from hardware to cloud and AI services, though analysts warn this “softwarization” strategy faces fiercer competition than its traditional duopoly with Ericsson.

8 - Google’s Gemini 3.0 Reshapes the AI Landscape: A stock-soaring comeback against OpenAI

• Alphabet shares soared to record highs following “glowing reviews” for its latest AI model. The stock jumped nearly 7% as analysts compared the new Gemini favorably to OpenAI’s GPT-5, fueling investor confidence that Google has finally delivered a superior product in the generative AI arms race.

• Google is aiming to shake up the chatbot market with this major release. By rolling out a version that meets high user expectations, the company seeks to leverage its inherent advantages of massive scale and profitability to gain a decisive edge in the intensifying competition for AI dominance.

9. Off-Planet Compute: Tech giants look to space to solve the AI energy crisis

• Energy constraints are forcing Big Tech to look off-planet. With the AI race straining Earth’s power grids, companies are exploring building data centers in outer space to bypass terrestrial capacity limits and secure the massive energy required for future models

10 - Agentic AI Security Risks: Windows 11’s new autonomous features open a new front for system threats

• Microsoft is testing “agentic” AI features capable of autonomous tasks like organizing files, but warns they introduce “novel security risks.” To mitigate threats like “cross-prompt injection,” agents will operate on separate user accounts with distinct permissions, though granting read/write access to personal folders remains a significant privacy gamble that is currently disabled by default.