AI Newsletter

Mistral released Large 4, a 1-billion-parameter model with native multimodal support – the strongest open-weights model from Europe or the US according to aggregated benchmarks. Particularly noteworthy: the highest cybersecurity score of all models (82%) and superiority over Chinese competitors in critical workloads. The release tweet achieved 40,000+ likes and signals Europe's return to the frontier AI front.

TSMC reported a net profit of NT$763 billion for Q3 2026 – an increase of 69% year-over-year, significantly exceeding analyst expectations. The driver is unbroken strong demand for AI chips, particularly for training and inference. This confirms that the AI infrastructure investment wave continues to accelerate despite capacity concerns.

With $2 million trading volume and 72% probability for Google, Polymarket sees the search giant as clear market leader in frontier models by end of October 2026 – Anthropic follows with 25%. Simultaneously, Google, OpenAI, Anthropic, and Mistral announced new frontier models within days, including Gemini Nano Banana 2.1 and Claude Haiku 5.5. Competition is measurably escalating in an unprecedented release pace.

A new private equity fund that explicitly uses AI to increase revenue in portfolio companies is testing its strategy on the first acquisition – according to WSJ, a first in the PE market. In parallel, investors are discussing how $5 trillion in US companies could change hands by 2035, with AI serving as a central value creation instrument. This trend signals that AI adoption has moved from tech niches into classic financial capital.

IMF Managing Director Kristalina Georgieva stated on October 12 that strong AI infrastructure investments are a key factor sustaining global economic growth at around 3%. The Washington Post confirms that AI is already supporting key sectors like manufacturing, advertising, and energy – not just tech balance sheets. This marks the transition from AI as an industry phenomenon to a systemic economic factor.