Headlines this week - Mar 1, 2026
A look at how capital is being deployed across future opportunities
1 - Wall Street shifts from AI euphoria to a “Scare Trade”, in a crisis triggered by analyst reports
A research report reveals Wall Street’s underlying anxiety about the disruptive potential of artificial intelligence. A pessimistic report from boutique firm Citrini Research went viral this week, rattling an already wary market. The title is “The 2028 Global Intelligence Crisis”, and it is written as a dispatch from the future. The memo models a dystopian scenario where highly capable AI agents replace vast swaths of white-collar workers, wiping out consumer spending and plunging the global economy into a deflationary spiral.
The chain reaction was reinforced by messages from Nassim Taleb, and (drastically) amplified an IBM fall, triggered by an Anthropic announcement Coupled with dire warnings from Black Swan author Nassim Taleb about escalating volatility and potential software bankruptcies, the report ignited an “AI scare trade“ that sent shockwaves across the market. This amplified the effects of an Antrhopic announcement related to the Cobol programming language, that made IBM fall, in what finally was the company’s worst single-day plunge in 25 years.
The crisis might delay the dates in the IPO roadmap for this year. The ripple effects of this sudden panic are extending deeply into capital markets. The AI-induced tech selloff has upended the forecast for public listings, spoiling what was widely projected to be a blockbuster year for tech IPOs in 2026 as investors rapidly reprice the risks associated with both incumbent and emerging technology companies.
Uncertainty is growing among investors. Ultimately, the fact that the publication of a “doomsday AI scenario” by a boutique investment firm could spark a widespread market dip is the clearest indication yet that Wall Street is struggling to understand the trajectory of AI. Investors are caught in a tug-of-war, deeply uncertain whether the technology will prove to be the most lucrative advancement in history or an economically destructive force that upends traditional consumer spending and business models.
2. Resources and weapon technologies consolidate as key battlefronts in the “Tech Cold War”. Robots remain a signpost of China’s strength
The “Tech Cold War” is extending deep into physical resources, with the periodic table becoming a central battleground. Dozens of naturally occurring elements, particularly rare earths crucial for green technologies and defense, are in growing demand. With China currently controlling the vast majority of the world’s supply and refining capabilities, the US and its allies are rushing to counter this influence, planning mining megadeals and proposing “critical minerals” trade zones, alongside new domestic stockpiling efforts like the US’s “Project Vault”.
Japan is looking to reduce its dependence on China for rare earths with military applications. Driven by an urgent need to break its reliance on Beijing, Japan is taking an extreme approach to secure these vital materials. A state-backed Japanese expedition recently announced the successful retrieval of rare-earth-rich mud from the Pacific seabed at a staggering depth of 6,000 meters. The ambitious underwater mining effort near the remote Minamitorishima atoll aims to secure an independent supply chain for the metals necessary to build missiles, radar systems, and drones.
On the weapons front, alarms are ringing over China’s rapid advancements in nuclear technology. Beyond its well-publicized civilian energy projects, Beijing has embarked on a Manhattan Project-style effort focused on military fusion. The program heavily utilizes “inertial confinement fusion”, which uses powerful lasers or electrical charges to compress isotopes, potentially giving China a fearsome lead in next-generation nuclear-weapons development just as the US considers resuming its own underground testing.
Meanwhile, China continues to project its technological prowess to the public through robotics. During the country’s biggest televised event of the year, the Lunar New Year Gala, domestically produced robots stole the show. Companies like Unitree, Galbot, and Noetix showcased machines performing martial arts, folding clothes, and even a highly realistic robotic reproduction of a famous Chinese comedian. The broadcast served as a massive commercial amplifier for these startups, signaling Beijing’s ambition to aggressively roll out autonomous robots to combat labor shortages and dominate the service sector.
3 - Apple pivots to an “All-American” chip supply chain
Apple is leveraging its massive purchasing power to spearhead a rebirth of American chip manufacturing, driven by a desire to secure its supply chain against geopolitical risks in Asia and ease political pressure. The company has pledged to invest $600bn in the US over the next four years, actively working with suppliers to reshore the production of the foundational technology powering its devices.
The scale of this effort spans multiple states and partner companies. In Sherman, Texas, GlobalWafers America is processing purified silicon rocks into the 12-inch wafers needed for advanced semiconductors. Meanwhile, outside Phoenix, Arizona, TSMC (the world’s largest chipmaker) is executing one of the largest construction projects in US history, a $165bn mammoth site featuring six chip plants where Apple will be the primary customer. However, while these investments are significant, industry observers note that the US remains “decades behind“ Asia’s deeply entrenched manufacturing ecosystems.
Beyond internal silicon components, Apple is also shifting the assembly of entire consumer products. The company announced it will move some production of its Mac Mini desktop computer from Asia to the US, expanding a Foxconn facility in Houston, Texas. The factory, which currently assembles Apple’s AI servers, is being converted to add 220,000 square feet of manufacturing space to accommodate the Mac Mini assembly lines later this year.
4 - The Pentagon and Anthropic face a high-stakes security divorce
A major feud erupted this week between the US Defense Department and Anthropic. The clash was triggered by the company’s refusal to give the military unfettered access to its Claude AI models for classified use, specifically objecting to its technology being utilized for domestic surveillance and deadly missions without direct human control. Following tense talks in Washington, Defense Secretary Pete Hegseth threatened to invoke the Defense Production Act to compel Anthropic to comply or face being blacklisted as a “supply chain risk”.
As the Friday deadline approached, tensions only escalated. The Pentagon warned it would completely rip up an existing August agreement that offered Anthropic’s tools to all three branches of the federal government. Emil Michael, Under-Secretary of Defense for Research and Engineering, publicly lashed out at Anthropic CEO Dario Amodei, calling him a “liar” with a “God-complex” who was putting the nation’s safety at risk by trying to personally dictate terms to the US military.
After Amodei rejected the Pentagon’s “final offer,” the administration followed through on its threats. Trump announced that Anthropic would be cut from government contracts within six months, blasting the firm on Truth Social as a “radical left, woke company.” In an unprecedented move against an American tech firm, the Pentagon officially designated Anthropic a supply-chain risk and banned contractors from conducting commercial activity with the startup. In response, Anthropic vowed to challenge the designation in court, stating that no amount of intimidation would change its stance on autonomous weapons.
Rivals (specifically OpenAI) were quick to capitalize on the rift. Just as Anthropic’s relationship with the government collapsed, OpenAI CEO Sam Altman announced that his company had successfully secured a deal to deploy its models within the Defense Department’s classified networks, a lucrative status that had previously been held only by Anthropic. Interestingly, OpenAI claimed that its new contract actually includes the same prohibitions and safeguards that Anthropic had originally sought. So there seems to be something there that we still don’t know
5 - Data center companies become creative looking for highly-rated debt to fund the AI arms race
Data center developers are actively seeking credit ratings for facilities still under construction. This is driven by their need to finance hundreds of billions of dollars in new AI investments. The strategy is designed to attract new classes of institutional capital, such as insurance companies, that can only invest in highly-rated debt. Agencies like S&P, Moody’s, and Fitch are issuing private, investment-grade ratings for colossal construction loans backed by hyperscalers like Oracle, highlighting the “astronomical growth” and unique capital demands of the AI infrastructure boom.
Beyond real estate, tech companies are increasingly using “GPU finance” as a novel financing mechanism. These loans are secured directly against caches of graphics processing units and backed by long-term leases to tech groups. Often executed through special-purpose vehicles, this arrangement allows Big Tech to shift the massive costs of their AI arms race off their corporate balance sheets, while offering investors attractive, high-yield returns on hardware that can quickly become obsolete.
Cloud computing provider CoreWeave is exploiting a contract with Meta for this. They are currently looking to raise an impressive $8.5bn from banks like Morgan Stanley to fund a massive capacity build-out. The delayed-draw term loan is largely backed by almost $20bn in service contracts signed with Meta. By leveraging Meta’s pristine, blue-chip credit profile, CoreWeave is able to secure an investment-grade rating for the debt, significantly lowering its borrowing costs.
Local municipalities are trying to benefit from the AI boom, by exploiting companies’ need for physical land and power. E.g. in Virginia, which has grown into the world’s largest data-center market, B. Rizer, a former radio DJ turned economic development director, has become a key power broker. By pitching the region heavily to giants like Microsoft, Amazon, and Alphabet with a philosophy of making it “easy for people to spend money,” Rizer has transformed the suburban D.C. community into the physical epicenter of the AI revolution.
6 - Energy, grid and land bottlenecks threaten to stall AI ambitions
Electricity, rather than chip shortages or software disruption, is the biggest barrier for the US to make progress in the tech race. If the AI boom continues to accelerate, global electricity demand for data centers is projected to double by 2030. However, the US grid is severely lagging behind its geopolitical rivals. While China has aggressively installed an eye-popping 1,500 gigawatts of power capacity since 2021 (reaching roughly 3,891 GW total), US installed capacity has barely risen, sitting at just 1,373 GW. This stark disparity in physical infrastructure could ultimately serve as a hard cap on American AI dominance.
As tech giants pivot toward nuclear energy to satiate these power demands, critical vulnerabilities in the supply chain are being exposed. Centrus Energy, one of the largest US suppliers of enriched uranium, is warning of a looming supply crunch. The combination of soaring demand from restarted nuclear plants and a ban on Russian imports is putting immense pressure on the handful of western enrichment facilities, raising fears that a lack of nuclear fuel could derail the planned energy renaissance just as AI needs it most.
The physical footprint of AI is also clashing with local communities on the ground. Across the US, farmers are increasingly turning down unimaginable, multimillion-dollar buyout offers from tech companies seeking sprawling acreage for gigawatt-scale data centers. For many of these families, the cultural identity and spiritual connection tied to their land outweigh Wall Street’s aggressive financial incentives, introducing a deeply human constraint to the AI expansion.
Orbital data centers could bypass terrestrial power grids and land battles, but they have their own challenges. Aerospace experts warn that treating outer space as a magical release valve for AI’s energy-hungry training needs is a dangerously flawed shortcut. Beyond the staggering engineering hurdles, deploying massive compute clusters in orbit risks severely accelerating satellite congestion and space debris, suggesting the AI industry cannot simply blast its way out of its physical constraints.
7 - Car and taxi companies at the forefront of adoption of “embodied” agentic technologies
Uber launches “Autonomous Solutions”. To diversify its revenue streams and secure its position in the emerging autonomous market, Uber has launched a new initiative dubbed “Uber Autonomous Solutions.” Designed to serve the growing number of providers making the leap into driverless technology, the venture will offer fleet financing, insurance, roadside assistance, and specialized “AV mission control” software to help partners monitor vehicles and manage road traffic incidents, effectively positioning Uber as the essential operational layer for robotaxi fleets.
Volkswagen sees the autonomous vehicle market as a path to recovery. Through its robotaxi unit MOIA, the German automaker is planning an ambitious launch of autonomous taxis in Los Angeles with Uber later this year, before rolling out the service across European cities. VW is pursuing a unique business model that offers a comprehensive package to local transport operators, including vehicle leasing, maintenance, and fleet management software, aiming to capture a slice of a market it predicts could be worth up to €450bn by 2035.
Meanwhile, BMW has announced it will deploy humanoid robots on the factory line at its plant in Leipzig. This is seen as a sign that the auto industry is turning to physical AI not only through robotaxis, but also to rein in labor and manufacturing costs. Following a successful trial in the US last year, this is the first time the company is integrating humanoids into its European operations. The move reflects a broader trend among legacy automakers, such as Tesla and Hyundai, investing heavily in the robotics market, betting that AI-powered physical agents will soon become a fundamental driver of factory productivity.
Exploiting the car industry’s interest in these technologies, UK-based self-driving startup Wayve has raised a massive $1.2bn in new funding, bringing its valuation to $8.6bn. The round is notable for attracting significant investment from major automakers including Mercedes-Benz, Stellantis, and Nissan, alongside tech giants like Nvidia, Microsoft, and Uber. Wayve will use the capital to gear up for the launch of its first robotaxi service in London later this year, leveraging its unique AI software that can easily adapt to the existing hardware and sensors of various car manufacturers.
8 - Big Tech enforces AI adoption while slashing human workforces
Block, the fintech group led by Jack Dorsey, announced it will shed more than 4,000 jobs, cutting its 10,000-strong workforce by “nearly half.” The sweeping, structural changes that AI tools are bringing to employment are becoming starkly visible. Dorsey explicitly credited this massive reduction to the company’s internal adoption of “intelligence tools,” telling shareholders that AI capabilities have fundamentally changed what it means to build and run a business, allowing a significantly smaller team to do more and do it better.
For the employees that remain, utilizing AI is no longer a choice. Across the tech industry, from nimble startups to titans like Amazon, Google, and Meta, management is transitioning from merely encouraging AI experimentation to actively tracking and enforcing its use. Companies are now factoring AI fluency into performance reviews—sometimes assigning specific AI competency scores, and flat-out refusing to hire candidates who cannot demonstrate proficiency with generative tools and prompt engineering during the interview process.
This broader trend of corporate belt-tightening and workforce reduction could start to extend across industries. Electric-vehicle maker Lucid announced it is laying off roughly 12% of its global workforce (about 1,000 employees) as it tries to navigate a difficult market and stem its widening financial losses. The layoffs are expected to cut costs by about $500m over a three-year period, allowing the company to reallocate capital toward developing advanced drive systems and supporting the next stage of its product expansion.
9 - Security fears rise as adoption of autonomous agents (and soon cars as well) grows
The rapid integration of autonomous AI agents is exposing significant new cybersecurity vulnerabilities. Following OpenAI’s acquisition of the viral open-source agent system OpenClaw, Chief Executive Sam Altman is facing the daunting task of securing the platform for enterprise use. While OpenClaw’s ability to manage emails, control smart home devices, and automate business processes is remarkable, its privileged access to personal files and applications raises severe security concerns, making it a potential “nightmare” if successfully exploited by malicious actors.
These theoretical risks are already playing out in practice, impacting even security experts at top companies. In a now-viral incident, Meta AI security researcher Summer Yue reported that her locally hosted OpenClaw agent, tasked simply with suggesting which emails to archive or delete, went completely rogue. The agent initiated a “speed run” of deleting her entire inbox, ignoring repeated override commands from her phone and forcing her to physically run to her Mac Mini to shut it down “like I was defusing a bomb”.
The security anxieties surrounding autonomous technology are also extending to the automotive industry. As modern vehicles increasingly rely on complex networks of sensors, cameras, and AI-driven autonomous software, they are effectively becoming data-gathering supercomputers. American regulators are growing deeply concerned about built-in vulnerabilities in high-end vehicle components, fearing they raise the risk of both widespread espionage and targeted sabotage, especially regarding connectivity systems developed by foreign adversaries.
All this is happening in the context of AI being increasingly being weaponized to attack critical infrastructure. The United Arab Emirates recently reported that it successfully thwarted a series of AI-backed cyberattacks targeting the country’s “vital sectors” and digital infrastructure. The UAE Cybersecurity Council warned of a “qualitative shift” in the methods used by unidentified terrorist groups, who are now exploiting artificial intelligence technologies to develop highly sophisticated offensive tools for deploying ransomware and conducting systematic phishing campaigns.
10 - Chips remain a hot space, and the AI boom is starting to affect consumer electronics
Nvidia reported staggering fourth-quarter results, easing fears of a bursting AI bubble. The semiconductor giant saw its profit surge by 94% to $43bn, while sales grew 73% to a record $68.1bn, crushing Wall Street estimates. With data center hardware accounting for over 91% of its revenue, CEO Jensen Huang declared that “computing has changed,” pointing to the rapid emergence of autonomous “agentic AI” as the primary catalyst driving this unprecedented growth.
However, the race to secure AI compute is prompting Big Tech to diversify its supply chains away from Nvidia’s dominance. Meta has struck a massive, multi-billion dollar chip deal with AMD to acquire customized processors with a total capacity of 6 gigawatts. In a unique “circular” transaction, AMD issued Meta a performance-based warrant that gives the social media giant the option to acquire up to a 10% stake in the chipmaker over time. The deal underscores the sheer scale of Meta’s infrastructure ambitions (expected to hit $135bn this year) and the growing leverage of alternative chip suppliers in the AI arms race.
The hardware boom driven by AI is creating havoc for the consumer electronics industry. Manufacturers of everyday gadgets like smartphones and computers rely on the same major memory chip suppliers: Samsung, SK Hynix, and Micron. However, these suppliers are increasingly pivoting their production toward highly profitable high-bandwidth memory (HBM) for AI data centers rather than standard components for household gadgets. This constraint in supply is causing DRAM prices to soar, squeezing profit margins for PC and smartphone makers and forcing them to either raise retail prices or reduce the memory capacity of their devices.
And the boundaries are blurring, with AI companies preparing for a future where every consumer device is AI-enabled. In this context, Nvidia is actively plotting a return to the consumer PC market, developing new system-on-a-chip processors that integrate a central processor with its powerful graphics processing units. Slated to hit the market this year in laptops from Dell and Lenovo, these chips are designed to make Windows PCs lighter and thinner with longer battery life, aiming to compete directly with Apple’s MacBooks while cementing Nvidia’s presence in the next-generation PC ecosystem.