AI News – August 30, 2026: The Closed Supply Chain, Anthropic's Trillions, and the Vertical Awakening

Is artificial intelligence becoming a closed-door ecosystem? In the week of August 24–30, 2026, the concentration of power reaches new heights. We analyze the m

The end of August 2026 captures an artificial ecosystem rapidly contracting upward. Data, capital, chips, and regulations are intertwining in the hands of an increasingly exclusive club. From Nvidia's financial expansion to Anthropic's sci-fi estimates, and on to the strategic responses from Alibaba and Thomson Reuters, the news from August 24-30 raises a crucial question: does it still make sense to talk about an open market when the foundational infrastructure dictates terms for everything else?

1. Nvidia and Perplexity: A $30 Billion Axis

Nvidia is in advanced talks to participate in a new funding round for Perplexity. The deal would project the AI search startup's valuation beyond $30 billion, marking a 50% jump from roughly $20 billion the previous year. The impact: This is not a simple venture capital investment. Nvidia is verticalizing its influence: it is not just selling the chips needed for training, but is directly entering the capital of those who consume that computing power. An integration that consolidates a strategic role across the entire supply chain and risks dictating development priorities and access to compute for years to come.

2. Anthropic and the Mirage of $30 Trillion

According to leaks from the Wall Street Journal, Anthropic is preparing to communicate to its investors a stratospheric Total Addressable Market (TAM) estimate: over $30 trillion. To put that in perspective, it is a figure higher than the $28.5 trillion estimate outlined by an aerospace giant like SpaceX. The impact: These figures do not reflect current revenues, but serve as a magnet for capital, justifying off-scale company valuations. If the industry begins pricing models on TAMs of this size, the risk that the financial bubble inflates much faster than the real economic utility produced becomes a systemic problem.

3. Alibaba Relaunches on Video with Wan3.0

After completing a monumental share placement of around $10 billion to finance its rapidly growing infrastructure, Alibaba has officially released Wan3.0, a deeply enhanced video generation model. The impact: China proves it is not merely a follower. Alibaba is building a perfect self-sufficient ecosystem: it raises immense capital, develops proprietary models, integrates them into its cloud infrastructure, and produces vertical applications. Wan3.0 also certifies that the true frontier of commercial competition has definitively shifted from text and code toward generative audiovisual content.

4. EU AI Act: Transparency Becomes Operational

The era of future regulation is over. The transparency rules of the European regulation have been fully in force since August 2, and the Commission has published the final operational guidelines. The obligations impose stringent practices: immediate disclosure for interactive systems, marking of synthetic content, notices on emotion recognition, and explicit watermarking for deepfakes of public interest. The impact: Compliance is no longer a statement of principle for sustainability reports, but an immediate technical requirement. Those operating in Europe (or offering services aimed at European citizens) must adapt their workflows today.

5. Thomson Reuters and the Value of Vertical Data

While the giants fight over general intelligence, Thomson Reuters has unveiled Thomson 1.0. Developed internally from open-source foundations, it is a language model trained exclusively on its own invaluable proprietary legal, tax, and professional databases (Westlaw, Practical Law, Checkpoint). The declared investment is around $40 million. The impact: It is the revenge of vertical models. Not everyone needs algorithms capable of writing poetry or passing general medical tests. Thomson 1.0 proves that the true competitive moat lies not in mathematical parameters, but in the quality, exclusivity, and context of the data used to train systems focused on professional tasks.

Our Take: The Perfect Oligopoly

This week's news outlines an unequivocal trajectory: the ruthless concentration of power across the entire Artificial Intelligence supply chain.

When those who make the chips finance those who develop the algorithms, and those who develop the algorithms project $30 trillion markets to absorb global venture capital, the space for independent innovation disintegrates. In parallel, the sacrosanct European transparency rules impose compliance costs that often only tech giants can absorb without slowing down.

If access to cloud infrastructure, billion-dollar capital, and proprietary data pools remain the only keys to participating in the sector, the idea of a competitive, open, and democratic AI ecosystem risks being confined to marketing manuals.

Bibliographic References and Sources

  1. Investments and Valuations:
    • Reuters – Nvidia discusses Perplexity investment.
    • Wall Street Journal / Reuters – Anthropic expected to tell investors its TAM exceeds $30 trillion.
  2. Models and International Development:
    • Reuters – Alibaba launches Wan3.0 AI video model.
    • Reuters / AIdapted – Thomson Reuters launches proprietary LLM.
  3. European Regulation:
    • TLT / Simmons & Simmons – AI Brief & View August 2026 (EU AI Act transparency guidelines).

Article by the Editorial Team of La Bussola dell’IA – AI News Section.