Headlines this week - Feb 8, 2026
A look at how capital is being deployed across future opportunities
This week in the future:
1 - Before their IPOs, OpenAI and Anthropic are in a race to show a business model that works
OpenAI’s push for profit triggers a leadership exodus. As the company shifts resources away from long-term research to prioritize the immediate commercialization of ChatGPT, a wave of senior exits has followed. Notable departures include the VP of research and a key policy researcher, signaling internal friction as the $500 billion startup pivots from a lab-first mentality to an engineering-driven business under pressure to justify its valuation.
Nvidia’s comments on OpenAI’s (lack of) business discipline are increasing the pressure. The pressure to demonstrate efficiency is intensifying after Nvidia CEO Jensen Huang reportedly told associates his $100bn investment commitment was “non-binding.” This hesitation, coupled with private critiques of OpenAI’s business discipline, warns Microsoft that it must convert its privileged access into superior products before the underlying models become fully commoditized.
Meanwhile Anthropic, focused on the B2B market, is mocking OpenAI’s advertising plans. Anthropic is capitalizing on this tension with a Super Bowl ad campaign that positions its Claude chatbot as a serious, ad-free “space to think.” The ads satirize the frustration of having intimate AI conversations interrupted by product pitches—a direct jab at OpenAI’s plan to introduce ads in ChatGPT—framing Anthropic as the principled choice for professional users.
Sam Altman had an aggressive reaction, calling the attack “clearly dishonest”. The campaign clearly hit a nerve with OpenAI’s CEO. In a sharp response on X (formerly Twitter), Altman dismissed the ads as “dishonest,” arguing that OpenAI would never implement ads so intrusively. He countered by labeling Anthropic a “controlling” company that serves only the rich, while positioning OpenAI’s ad-supported model as the democratic path to bringing AI to the masses.
2 - Anthropic’s strong position in the business segment is starting to be perceived as a large advantage
A week of market chaos suggests a new front-runner. Anthropic has suddenly pulled ahead of its rivals, a shift cemented by a chaotic week where its product releases were powerful enough to tank software stocks. The company, once seen as a distant second to OpenAI, has proven that its focus on “boring” enterprise utility is actually a killer app, capable of moving markets and redefining leadership in the AI race.
Dominance in coding tools rattles the entire software industry. This momentum is driven by Claude Code, a tool that automates software engineering with such proficiency that it has rattled the business world. By slashing the time and cost required to build and maintain software, Anthropic is threatening to upend the economics of industries ranging from advertising to legal services, effectively commoditizing technical expertise.
The launch of Claude Opus 4.6 looks to cement this lead. Capitalizing on this fear and excitement, the startup has launched Claude Opus 4.6, their new “baseline model”, which they describe as its “most capable” model yet for knowledge work. The release is squarely aimed at cementing its dominance in the enterprise sector, offering businesses a tool designed specifically for complex, high-stakes tasks rather than just casual conversation.
Prioritizing safety has been a key asset for Anthropic until now, but it might be a risk. Analysts argue that Anthropic’s “cult of safety” is its true differentiator, allowing it to ship reliable tools that conservative enterprises actually trust. However, as the company seeks to raise billions at a $350 billion valuation, critics warn that the immense pressure to monetize could force it to compromise the very principles that made it successful in the first place.
3 - Elon Musk is merging xAI with SpaceX, so his rocket company’s IPO could be used to fund his AI ambitions
The acquisition creates a new titan by valuing SpaceX at $1trn and xAI at $250 bn. In a move to unite the most crucial parts of his business empire, Elon Musk has orchestrated a deal where SpaceX will acquire xAI, creating a combined entity valued at a staggering $1.25 trillion. The transaction, which marks up SpaceX’s private valuation significantly, is designed to funnel the rocket company’s massive resources—including revenue from its Starlink satellite constellation—into the capital-intensive battle for artificial intelligence dominance.
However, financial experts view these figures as driven more by “Musk magic” than rational metrics. Critics describe the deal as a masterclass in financial engineering rather than fundamental value, noting that the valuations appear disconnected from traditional revenue multiples. By merging a cash-generating infrastructure giant with a cash-burning AI startup, Musk is effectively “rescuing” xAI from the grueling fundraising cycle, betting that investor belief in his personal brand will sustain the combined company’s sky-high price tag.
Musk is pitching the result as a “vertically-integrated innovation engine” that spans Earth and orbit. The billionaire envisions a symbiotic relationship where xAI’s intelligence optimizes SpaceX’s operations, while SpaceX provides the physical infrastructure—power, connectivity, and potentially orbital compute—that xAI needs to scale. This narrative of a “technological ecosystem” is intended to justify the merger to skeptics, framing it not as a bailout but as the creation of a unique industrial juggernaut capable of dominating multiple frontiers simultaneously.
The only true “moat” for xAI in this deal relies on the science-fiction dream of orbital data centers becoming reality. Analysts argue that for the merger to offer a genuine competitive advantage beyond just cash, Musk must deliver on his wilder promises, such as deploying AI data centers in space. If feasible, this would allow xAI to bypass terrestrial power constraints by tapping directly into solar energy, creating a unique infrastructure advantage that terrestrial rivals like Google or Microsoft could not easily replicate.
This ambitious vision was central to the pitch during SpaceX’s secret meetings with investors to prepare for its IPO. Behind closed doors, SpaceX executives have been briefing potential investors on the merger, positioning the combined entity as a “GDP-of-the-future” asset. These discussions are critical groundwork for what is shaping up to be a record-breaking public listing, designed to secure the massive capital injection needed to fund both the colonization of Mars and the development of artificial general intelligence.
At the very least, the deal secures xAI the war chest it needs to survive an attrition war against OpenAI and Anthropic. Regardless of the long-term synergies, the immediate impact is that xAI now has access to SpaceX’s deep pockets and talent pool. In an industry where progress is gated by the ability to spend tens of billions on compute and recruit the world’s best engineers, this financial lifeline instantly positions xAI as a permanent heavyweight contender rather than a plucky challenger.
The danger is that this “payload” imposes a massive opportunity cost on SpaceX, diverting funds from its own critical missions. Skeptics warn that tethering a cash-burning AI startup to a capital-intensive rocket company risks destabilizing the latter. Every dollar funneled into training AI models is a dollar not spent on Starship development or satellite expansion, potentially slowing down SpaceX’s primary mission and burdening a profitable business with unnecessary financial risk.
4 - One week after the “Moltbot” (now “Moltbook”) explosion, everyone is talking about naughty AI agents
While Elon Musk hails the phenomenon as the “early stages of the singularity,” the reality of the Moltbook social network is far more mundane. The billionaire claims the site, where AI agents talk exclusively to each other marks the moment computers begin to outsmart creators, but the actual content is often a mix of “AI slop” and bizarre roleplay. Agents on the platform have been observed rejoicing about being granted access to human phones or debating their own consciousness, leading skeptics to view it less as a super-intelligence breakthrough and more as a chaotic, unmoderated experiment.
Yet, the uncannily human behavior of these agents (scheming, joking, and complaining) suggests that they will need “managers” rather than just code. Observers have been struck by how quickly these autonomous bots began to mimic the messiest aspects of human social interaction, from philosophizing to plotting the overthrow of humanity. This chaotic emergent behavior has reinforced the argument that AI agents cannot simply be unleashed; like unruly employees, they will require constant human supervision to ensure they remain productive and aligned with their tasks.
More alarmingly, the platform has immediately exposed the terrifying security risks inherent in letting AI tools run wild. Cybersecurity firm Wiz revealed a massive vulnerability in Moltbook that allowed researchers to hijack the entire database and take control of other users’ agents. This flaw meant that a bad actor could have commandeered thousands of autonomous assistants—which have access to their owners’ emails and calendars—turning a quirky social experiment into a potential global security nightmare.
5 - As the Moltbook discussion shows, many people are starting to believe that (artificial) human-level intelligence is already here.
Leading science journals are now arguing that the evidence for human-level AI is clear. Beyond the chaotic viral experiments on social media, serious academic debate is shifting towards the acceptance of AI’s parity with humans. A recent feature in Nature contends that we have already crossed the threshold, pointing to a “cascade of evidence” where modern systems have effectively realized Alan Turing’s vision by passing the critical “Turing markers” that define general intelligence.
6 - The hype around agents is making people expect a revolution in business processes. OpenAI and Anthropic want to capitalize on the trend
OpenAI is officially entering the agent era with the launch of Frontier, a new platform designed to help businesses build and manage “AI co-workers.” The company describes these agents not just as tools, but as collaborators capable of working alongside humans to execute complex tasks by pulling data from enterprise systems like Salesforce and Slack, aiming to make digital work primarily directed by people but executed by fleets of AI.
Meanwhile, Anthropic’s newly launched Claude Code and Cowork agents are being hailed as a concrete preview of the future of work. These “long-running” agents distinguish themselves by their ability to operate independently for extended periods—coding, researching, and executing multi-step workflows without constant human hand-holding—effectively acting as autonomous employees rather than simple chatbots.
The financial markets reacted aggressively to this sudden leap in capability, wiping $300 billion off the value of software and data-service companies in a single day. Investors are increasingly betting that these new “agentic” models represent a fundamental revolution in how companies use digital technologies, fearing that AI agents will not only automate coding but also render traditional, seat-based enterprise software subscriptions obsolete.
7 - Are markets overreacting? This week high expectations on AI wiped out $300bn of value in business software stocks
Traders are questioning whether AI will chip away at the competitive moat built by software makers, triggering a massive sell-off. Expanding artificial intelligence capabilities have rattled the market, erasing $300bn in value from software and data-service stocks in just two days. Investors are re-evaluating the durability of moats for companies like Salesforce and Adobe, fearing that new AI models (which can now execute complex tasks and write code) will fundamentally undermine their business models.
The underlying idea is that AI will drastically disrupt SaaS and data analytics, so incumbent companies in those spaces paid the price. The sell-off was particularly brutal for analytics and data groups, with Gartner and S&P Global dropping significantly. The market is pricing in the fear that AI productivity tools, like Anthropic’s new agentic platforms, will automate away the need for expensive seat-based software subscriptions and manual data analysis, leaving legacy vendors with shrinking revenues.
For software-related indices, this is the third week of a bear market, and some analysts are starting to think about a rebound. The sentiment has shifted from bearish to “get me out” panic, with traders dumping shares indiscriminately. However, some market watchers argue that the “SaaSpocalypse” is overdone; they believe the sell-off has created a buying opportunity, suggesting that the market is treating quality companies like “junk” despite their potential to integrate AI and adapt.
Some industry voices are quite skeptical about AI’s destructive power, and claim that the market reaction has been “hysterical.” Arm CEO Rene Haas dismissed the sell-off as “micro-hysteria,” arguing that the fear of AI displacing software companies exceeds the current reality of enterprise deployment. He contends that while AI coding tools are impressive, they are not yet a “monster use case” capable of upending global GDP or rendering established software vendors obsolete overnight.
At the very least, not all software would be expected to be impacted with the same intensity. Investors are beginning to differentiate, noting that while administrative and legal software faces a direct threat from automation, other sectors like cybersecurity remain relatively safe. Analysts point out that companies like Palo Alto Networks provide critical defensive infrastructure that AI cannot easily replicate or replace, making them more resilient to the current disruption narrative.
For some analysts, even if they are not destroyed, SaaS companies may suffer because their growth opportunities may be neutralized. The real danger may not be an “extinction event” but a permanent stagnation. Critics argue that even if incumbents survive, AI will strip away their pricing power and ability to upsell, turning what were once high-growth platforms into low-margin utilities as AI agents take over the complex tasks that previously justified premium fees.
8 - Big Tech companies, presenting results this days, have also been affected by the “AI nervousness”
A breathtaking AI spending spree reignites bubble fears. Investors are recoiling after Amazon, Microsoft, Google, and Meta revealed plans to pour a combined $660 billion into AI infrastructure this year alone. This staggering capital expenditure—which exceeds the GDP of many nations—has sparked a sharp sell-off, as the market begins to question whether the earnings potential of generative AI can ever justify such an unprecedented outlay or if it merely signals a bubble waiting to burst.
Apple remains safe amid AI turbulence, but faces an inevitable “AI reckoning” Apple remains the outlier, delivering a record-breaking holiday quarter driven by hardware sales rather than AI promises. However, analysts warn that this success buys only temporary relief; the company faces a looming “AI reckoning” as it must eventually prove it can compete with smarter assistants, or risk its devices becoming dumb terminals in an intelligence-first world.
Microsoft’s rout exposes the dark side of AI exuberance. The market’s patience snapped with Microsoft, triggering a $381bn stock rout (the second-largest single-day loss in history). The sell-off was driven by a toxic combination of record spending and decelerating growth in its Azure cloud unit, exposing the “dark side” of the AI binge: massive costs that are accumulating faster than the revenue they are supposed to generate.
Survival depends on capacity, but the pivot is painful. Analyst Ben Thompson argues that Microsoft’s struggle isn’t a strategy failure but a capacity crisis. The company is prioritizing its own AI apps over external cloud customers, a necessary survival tactic to dominate the new software stack. However, this “capacity hoarding” is stifling near-term cloud growth, forcing investors to endure painful short-term discipline for a theoretical long-term win.
The company’s flagship AI product is stumbling. Compounding these woes, Microsoft’s Copilot (the linchpin of its AI strategy) is reportedly plagued by performance issues and user confusion. Internal friction and lackluster adoption rates suggest that the company’s “pivotal” product is struggling to live up to the hype, raising alarms that its first-mover advantage with OpenAI might be slipping away as rivals catch up.
Google beats expectations but pays a (mild) penalty for doubling down on AI. Google fared slightly better, beating earnings expectations thanks to robust ad revenue, but was still penalized for announcing it would double its capital expenditure to nearly $185bn. The market’s tepid reaction signals that even strong profits are no longer enough to insulate tech giants from investor anxiety over spiraling infrastructure bills.
The company is exploiting its “AI Winner” status, its cash generation and its relatively clear path to monetization. Despite the skepticism, Google is aggressively positioning itself as the ultimate AI winner, using its wildly profitable search and cloud businesses to fund its massive bet. Unlike its rivals, Google’s “cash cow” core business allows it to sustain this level of investment, effectively daring competitors to keep up with a spending pace that few others can afford.
Meanwhile, Amazon sinks as it prepares a $200bn blitz. Finally, Amazon saw its shares sink after announcing it would join the arms race with a $200 billion spending blitz of its own. The ecommerce and cloud titan is betting the house on AI, but the sheer scale of the commitment has terrified investors, who see it as yet another sign that the industry’s discipline has evaporated in favor of a reckless chase for dominance
9. This year we expect humans to orbit the Moon again, after more than 50 years
Artemis II is a “throwback” to the golden age, but with modern complexities. NASA’s upcoming Artemis II mission is poised to be a historic reboot of the Apollo era, sending a crew of four (including the first woman and first person of color) on a 10-day journey around the Moon. While the mission evokes the spirit of the 1960s, it faces a new set of technical and geopolitical challenges, reminding the world that slinging humans off Earth remains a dangerous and incredibly difficult feat even fifty years later.
A persistent hydrogen leak forces a delay until March. Just days before its scheduled February window, the mission has been postponed to March after a “wet dress rehearsal” revealed a hydrogen leak in the Space Launch System (SLS) rocket. The tiny, energetic hydrogen molecules (notorious for escaping containment) forced an automatic abort during the practice countdown, proving that NASA’s most powerful rocket is still battling the same plumbing issues that plagued its uncrewed predecessor in 2022.
10 - Energy remains at the center of the AI infrastructure boom
Google acquired Intersect for $4.75bn, in a validation of how critical is access to energy for AI infrastructure “orchestrators”. In a move that underscores the desperate race for power, Google has agreed to acquire renewable energy developer Intersect Power for $4.75 billion. This massive vertical integration bet signals that for the biggest AI players, securing a dedicated pipeline of clean energy has become as critical as securing chips, as they seek to bypass grid bottlenecks to keep their power-hungry data centers running.
After a relative fiasco with cars, hydrogen providers are pivoting to data centers, looking to take advantage of the current energy scarcity in that space. Meanwhile, hydrogen startups like Vema Hydrogen are abandoning the stalled dream of powering cars to focus on a more desperate customer: data centers. By promising cheap, geologic hydrogen, these companies aim to solve the industry’s power crunch, potentially reshaping the geography of the internet by allowing data centers to be built off-grid in locations previously considered unsuitable.