The day in one read

The Post-Midterm AI Correction

The November 4 midterm elections have triggered a sharp repricing of the artificial intelligence sector, as markets grapple with the fiscal and regulatory implications of a Democratic-controlled Congress. The immediate aftermath saw a bifurcated market reaction: while defense contractors and semiconductor stocks sold off on fears of stricter data center expansion limits and a pause in advanced AI development, major technology platforms rallied on the prospect of reduced capital expenditure and legislative gridlock. This political shift has accelerated a broader debate regarding the sustainability of AI infrastructure spending, with investors rotating away from pure infrastructure plays toward beneficiaries of potential regulatory constraints and software services. The tension between the massive capital requirements of the AI buildout and the emerging political will to curb it now defines the macroeconomic landscape for the industry.

Infrastructure Debt and Valuation Volatility

The financial fragility of the "Neo-Cloud" sector has come under intense scrutiny following a significant sell-off in AI-related assets, with the iShares Semiconductor ETF losing approximately 15 percent in October. CoreWeave, a specialized AI compute provider, faces severe market anxiety, with credit default swaps rising 1,200 basis points and pricing implying a bankruptcy probability of over 60 percent within five years. This distress is part of a wider valuation correction; OpenAI’s valuation was cut from $1.4 trillion to $1.25 trillion, with its latest round reduced from $30 billion to $20 billion, though only $11 billion has been committed. Analysts attribute this to price discovery and intensifying pricing pressure rather than a collapse in demand, noting that OpenAI is offering 30% discounts and users are reporting 72% cost reductions by switching to newer models like Haiku 5.5.

While OpenAI’s enterprise revenue remains robust, with Greg Brockman citing customers paying between $1 million and $100 million annually, the broader market is questioning the financial sustainability of the ecosystem. Anthropic’s GAAP revenue is projected to reach $56 billion this year, a 12x increase from last year, but growth rates are expected to slow to 2x next year. In response to these pressures, investors are reassessing risk. IREN, another Neo-Cloud, saw its stock drop 45 percent in four weeks, yet some analysts view this as a mispricing, citing its conservative financial structure and a secured pipeline of approximately 5 gigawatts of power contracts. Meanwhile, hyperscalers like Amazon report attractive payback periods of 2.5 to 3 years for AI investments, suggesting a divergence in risk profiles between specialized providers and established tech giants.

Political Headwinds and Global Expansion

The new political climate in Washington is directly impacting corporate strategy, with emerging Democratic leaders advocating for a pause on advanced AI development and a ban on artificial superintelligence. Representative Yasemin Ansari stated that the Democratic agenda includes these regulatory measures, echoing calls for severe penalties for developers of artificial superintelligence. Elon Musk responded to these statements by calling the prospect a "brave new world." This political friction is driving global expansion efforts, as companies seek to mitigate domestic regulatory risks. Google, for instance, plans to invest $15 billion in Malaysia data centers in 2027 and 2028, with campuses in Selangor and Negeri Sembilan each equipped with one gigawatt of clean power. The company also aims to train 800,000 Malaysians as part of this initiative.

In Europe, the infrastructure gap remains stark. A Bitkom study reveals that while the US adds 10 GW of data center capacity annually and China adds 6 GW, Germany adds only 0.65 GW. Germany’s private AI investment stands at $3.9 billion compared to the US’s $286 billion, and US hyperscalers hold over 90% of the German cloud market. The German government’s 2030 target of 6,000 MW is projected to be missed, with estimates suggesting only 5,471 MW will be achieved. This disparity is influencing corporate governance and strategy; Deutsche Telekom CEO Tim Höttges is lobbying Chancellor Friedrich Merz for a merger with its US subsidiary, Team Mobile, a deal valued at nearly €20 billion more than the parent company. The federal government’s 28.6% stake in Deutsche Telekom could block the transaction, highlighting the complex interplay between national sovereignty and corporate growth.

Enterprise Adoption and Strategic Shifts

As the hype cycle matures, the focus is shifting from technical proof-of-concepts to core business strategy and operational efficiency. At Figma, CFO Praveer Malani describes a transition from manual SQL-based data reconciliation to AI-driven automation, allowing finance teams to move from retrospective reporting to forward-looking strategic decision-making. Malani notes that 95% of the challenge in automating tasks like travel and expense reconciliation is getting data into the correct format, not the AI model itself. He warns against "slop," where AI generates excessive documentation without human verification, and predicts that if business complexity remained static, Figma would not need the same number of finance employees in 12 to 18 months.

This strategic shift is evident across the industry. Canva is cutting 22% of its workforce, totaling 1,271 employees, due to a growth slowdown and ballooning AI costs. Co-founder and CEO Melanie Perkins stated that the company must change to address slowing growth, noting that success in its first chapter does not guarantee future success. Canva has reduced its revenue growth forecast from 30% to 20% and is developing its own AI models to reduce reliance on third-party providers. Similarly, Dropbox cut 9% of its staff, and its CTO has departed. These moves reflect a broader trend where companies are consolidating teams and reducing subspecialization to manage costs and improve efficiency in an environment of intense competition from general-purpose chatbots and AI startups.

Safety, Sovereignty, and the Road to AGI

The trajectory of AI safety and recursive self-improvement (RSI) remains a critical concern for industry leaders. Demis Hassabis, co-founder of Google DeepMind, warned that the next phase of closed-loop RSI without human review carries significant risk, potentially leading to an "intelligence explosion" resulting in an alien intelligence difficult for humans to understand or control. To mitigate these risks, Hassabis advocated for accelerated cyber defense, hardened sandboxes, and robust monitoring systems. He proposed an industry-run standards body with government oversight, similar to financial auditors, and suggested that leading labs should red-team each other. On the topic of consciousness, Hassabis argued that intelligence and consciousness are double dissociable and advised against building conscious entities until society has debated the moral implications and neuroscience better understands the neural correlates of consciousness.

Sovereignty is also a key theme, with Hassabis predicting six to eight frontier AI offerings globally, with open-source models lagging only six months to a year behind. He advised countries to avoid building frontier models themselves, instead focusing on downstream applications and collaborating with multinational companies. In the US, the political landscape is influencing these discussions, with some lawmakers calling for a pause on advanced AI development. Meanwhile, in the enterprise sector, Thomas Wollmann of Merantix Momentum argues that AI strategy and business strategy are now inseparable, requiring organizations to evaluate whether their models remain defensible against AI disruption. He emphasizes that 80-90% of successful transformation is human-centric, citing legacy process restructuring as a greater challenge than technical implementation.

Also Notable

In the autonomous vehicle sector, Waymo will officially split from Uber in the Austin and Atlanta markets starting in January 2028, offering rides directly through its own smartphone app. Co-CEO Takidra Mawakana stated that the rider experience via Uber has been inconsistent because riders cannot determine if they will receive a robo-taxi or a human-driven car. In the education sector, Google faced backlash after announcing that schools using its K-12 software will lose access to Search unless they opt into using AI tools. Senator Bernie Sanders accused Google of extortion, while former FTC chair Lina Khan condemned the decision as anti-competitive. Google responded by stating that AI products are off by default and that schools can enable Gemini, Gemini Notebook, and Search individually.

In the crypto space, a $6 billion Binance hack, executed via a fake job interview containing malicious code, did not severely impact markets as funds were recovered. Circle reported net inflows of over $2 billion to Binance in 24 hours post-hack. Circle’s USDC stablecoin volume stands at $70 billion, down from $75 billion a year ago, while Tether shows similar stagnation. In the pharma sector, Novo Nordisk reported revenue guidance of 2% growth to 2% decline, an improvement from prior expectations, driving a 5% stock increase. Orion, a Finnish pharma partner to Bayer on the cancer drug Nubeka, reported €280 million in Q3 revenue from the drug, over half its total, and projects Nubeka could exceed €1 billion annually. The company’s stock is up 55% this year, valuing the firm at €15 billion.