The day in one read

The AI Valuation Reckoning

The global markets experienced a violent correction on November 5, as the Nasdaq 100 fell 6.4%, its worst single-day performance since March 2020, erasing $2.3 trillion in market value. The sell-off was driven by a confluence of political anxiety following the U.S. midterm elections and deepening skepticism regarding the profitability of artificial intelligence infrastructure. The Philadelphia Semiconductor Index dropped over 12%, with Nvidia losing 9.3% of its value, or $450 billion, while memory stocks like Seagate and Western Digital fell as much as 22%. This volatility marks a sharp reversal from the post-election optimism seen earlier in the week, where the Dow had jumped over 1,500 points. Investors are now grappling with the reality that the massive capital expenditure required for AI data centers may not yield the immediate returns previously projected, leading to a significant rotation out of high-multiple tech assets.

Sovereign AI and Infrastructure Strategy

In response to geopolitical fragmentation, major technology players are accelerating the buildout of sovereign AI infrastructure, prioritizing energy independence and data sovereignty over pure efficiency. Mistral, Europe’s most valuable AI company at a valuation exceeding 21 billion euros, released the weights for its ML4 model, a 1-trillion-parameter system trained entirely in France on 10 megawatts of self-owned power. Mistral’s VP of Science, Pierre Storck, argued that relying on internal reinforcement learning rather than external distillation is essential for technological independence, comparing access to AI tokens to access to electricity. This strategy aligns with broader European efforts to reduce dependence on U.S. cloud providers. In Germany, the federal government unveiled a 2030 AI strategy aiming to quadruple data center capacity from 0.5 gigawatts to 2 gigawatts, a target that critics note remains significantly below the United States’ current 8 gigawatt capacity. Meanwhile, NEAR Protocol is positioning its blockchain-native architecture as a solution for privacy-preserving AI, launching uncensored models for cybersecurity defense and using formal verification to address the "security Armageddon" posed by AI-driven vulnerability discovery.

The Political Economy of AI Policy

The 2026 U.S. midterm elections have shifted the political landscape, with Democrats flipping seven Senate seats and winning the Minnesota governorship by a historic margin. This outcome has introduced new regulatory risks for the tech sector, particularly regarding data center expansion. OpenAI’s Chief Global Affairs Officer, Chris Lehane, has framed the 2028 election as a defining moment for AI policy, invoking a New Deal metaphor to argue for state capacity to manage economic disruption. However, critics like Kate Brennan of the AI Now Institute contend that these proposals are corporate advocacy that lacks mechanisms for accountability, noting OpenAI’s previous opposition to safety regulations like California’s SB 1047. In Minnesota, Governor-elect Amy Klobuchar has pledged to ban non-disclosure agreements in data center legislation and emphasize local control, a stance that contributed to the sell-off in chip stocks. The market reaction suggests that investors are pricing in the potential for stricter antitrust enforcement and regulatory moratoriums under the new Democratic-led Congress, a shift that contrasts with the pro-growth messaging advocated by commentators like Scott Galloway, who argues Democrats must rebrand as the party of prosperity to retain swing voters.

Enterprise AI Architecture and M&A

Within the enterprise sector, the architecture of AI deployment is undergoing a significant simplification. AI First, a consultancy, replaced its previous infrastructure of 25 specialized agents with a single central agent operating within a seven-component "AI Operating System." This shift, triggered by the capabilities of newer models like Opus 4.6, suggests that complex multi-agent systems are becoming obsolete in favor of robust, AI-maintained context and outcome-based skills. This trend is mirrored in the M&A landscape, where Salesforce’s Chief Corporate Development Officer, John Samorjai, outlined a strategy focused on acquiring talent and distribution capabilities rather than just technology. Samorjai criticized current venture valuations as unsustainable, predicting a "reckoning" for overvalued startups due to a limited buyer universe. He noted that while software development is faster, acquisitions remain necessary to accelerate enterprise AI adoption, with recent deals like Fin and Listen Labs demonstrating the value of integrating AI into existing distribution channels.

Virtual Cell Modeling and Scientific AI

The application of AI to biological systems is moving from static mapping to active experimental steering. Fabian Theis, CTO of Muralomics, described the evolution of virtual cell models as a transition from prediction to "lab-in-the-loop" strategies, where models guide subsequent experiments. This shift is supported by massive data generation efforts, such as the Biohub’s $1.8 billion collaboration with the DOE, NIH, Meta, and Google. Theis highlighted the release of LabCompass, an inverse generative model that determines necessary actions to reach a specific cell state, and noted that scaling laws are emerging in the field, with models like SCConcept trained on 350 million cells across 16 species. However, Theis warned that platform-only virtual cell startups struggle to monetize, suggesting that commercial success depends on coupling AI models with proprietary data generation, such as the Cancer Organoid Perturbation Atlas. This approach contrasts with the broader tech sector’s focus on general-purpose models, highlighting a specialized path for AI in scientific discovery.

Also Notable

Rheinmetall’s stock fell over 18% after the company cut its 2026 revenue forecast to 13.2 billion euros and scrapped its target to convert over 40% of operating profit to cash, citing delays in the Boxer transport vehicle and Destinos rocket projects. Despite the drop, major banks including Bernstein, Morgan Stanley, and Jefferies maintain buy ratings, with price targets ranging from 1,200 to 1,750 euros. In Japan, Toyota shares surged 15% after raising its full-year profit target, driven by a weak yen and strong U.S. hybrid demand, while Sony raised its profit forecast on the success of its video game and film franchises. Maersk shares fell 5% despite a second profit forecast increase, as investors took profits after a 40% year-to-date rise driven by high freight rates. In the venture space, NEAR Protocol’s co-founder Ilya predicted that direct transactions by AI agents will increase within a year, noting that incumbents like Meta struggle to build open, non-custodial privacy architectures. Finally, a doctoral candidate at King’s College London argued that AI models must be adapted to the societal context of their deployment, suggesting that Western models should prioritize individualism while Chinese models reflect local norms.