AI News September 1, 2026: Google Courts Hollywood for AI Licensing, Salesforce Puts Its CRM Inside Claude, Perplexity + Nvidia Ship a Local AI Computer

Today’s AI news cycle moves from Los Angeles soundstages to Wall Street. Google is making its most aggressive push yet to bring Hollywood’s intellectual property into the AI era, Salesforce has decided its future lives inside Claude rather than its own app, Perplexity and Nvidia want to move your AI workloads off the cloud entirely, and the two most valuable AI startups in the world are racing toward public markets. Meanwhile, the cost of the chips underneath all of it is going up.

1. Google Courts Hollywood: License Us Your Films (LA Times)

Google is pitching Disney, Universal, and other major studios on licensing characters and films for its AI tools, framing generative AI as a way to streamline production and visual effects, according to a detailed report from the Los Angeles Times published August 31.

The push faces stiff resistance. The same studios Google is courting have actively sued over AI, and union workers and creators are raising concerns about copyright, job displacement, and artistic control. The report lands at a moment of maximum tension between Hollywood and Silicon Valley: studios want the cost savings of AI-assisted VFX, but licensing their crown-jewel IP to a frontier AI lab could set precedents they can’t walk back.

Why it matters: If even one major studio licenses its catalog for AI training or generation, the copyright landscape that has defined the generative AI legal wars shifts overnight — and Google would gain a licensed-content moat that its rivals can’t easily replicate.

2. Salesforce Embeds Its Entire CRM Inside Claude

Salesforce has launched “Salesforce in Claude,” a plugin for Anthropic’s Claude CoWork that ships with 37 pre-built sales skills — covering meeting preparation, deal health reviews, and pipeline analysis. Sellers can query, update, and act on live CRM data without ever opening the Salesforce app itself, VentureBeat reports.

The product is available to select pilot customers now, with an open beta planned for September and additional business-function skills rolling out through Q3.

Why it matters: This is a remarkable strategic concession by Salesforce: the CRM giant is acknowledging that the primary interface for enterprise work is shifting from dedicated SaaS apps to AI agents. If your data lives in Salesforce but your workday happens in Claude, the app stops being the product — the agent does.

3. Perplexity + Nvidia Launch a Zero-Token-Cost Local AI Computer

Perplexity, in close partnership with Nvidia, has launched “Portable Computer” — a fully local AI agent device where the model, the user’s files, and the work itself all stay on the machine. Work completed locally consumes no billing credits, and every task starts on-device by default, with the system asking permission before sending any individual step to a more powerful frontier model in the cloud.

It’s one of the most aggressive attempts yet to move serious AI agent workloads off the cloud and onto local hardware.

Why it matters: Between this device, the open-weight GLM-5.3-Flash at 320B parameters, and Qwen3.8-Flash-Next’s open preview of the Qwen4 architecture, the “local-first” AI stack is assembling fast. Privacy-sensitive enterprises and developers watching API bills compound monthly now have a credible off-ramp from cloud-only AI.

4. The Trillion-Dollar IPO Race: OpenAI and Anthropic Both Filed

Both OpenAI and Anthropic now have confidential S-1 filings on file at the SEC. OpenAI submitted its confidential draft S-1 on June 8, saying it “has not decided on timing” but targeting a valuation of up to $1 trillion in a debut that Reuters reported could come as early as September. Anthropic filed its own confidential paperwork on June 1, shortly after closing a $65 billion Series H at a $965 billion valuation that put it above OpenAI’s $852 billion mark.

The NYT reported in late June that OpenAI was leaning toward waiting until 2027, citing tech-share volatility after SpaceX’s record June 12 listing — but the option to go public sooner remains on the table, and September is the earliest window it has left itself.

Why it matters: A trio of trillion-dollar-valuation debuts (SpaceX, then potentially Anthropic and OpenAI) would be the most consequential test of investor appetite for high-growth tech in a decade. For the AI industry, public markets mean public financials — and scrutiny of the massive compute losses every lab is currently running.

5. Nvidia Raises Flagship AI Chip Prices 17%

Server makers tell The Information that Nvidia is raising prices on its flagship AI chips by 17% — more bad news for data-center developers, arriving just as AI labs lock in gigawatt-scale compute commitments. Nvidia is simultaneously on a dealmaking spree, discussing an investment in Perplexity at a $30 billion-plus valuation and investing in data-center power firms to secure the electricity its chips will need.

Why it matters: Rising chip prices flow directly into AI API prices and the economics of every AI startup. Combined with the OpenAI/Anthropic race to public markets — where profitability questions will be unavoidable — the compute-cost story is becoming the central tension of the AI economy in late 2026.

Coverage draws on reporting from the Los Angeles Times, VentureBeat, Reuters, CNBC, and The Information. This digest is published daily at AI Tools Hub.