🎬 Hollywood tests AI | Axios AI+ Summary

1. Key Themes

🎬 Hollywood Has Crossed the Rubicon on AI Adoption

The entertainment industry has moved decisively from resistance to active integration, framing AI as production infrastructure rather than a creative threat.

"Like most industries, those in entertainment have moved from resistance to experimentation, and now selective adoption."

🏭 "Creator in the Loop" Is Becoming the Standard Operating Model for AI in Creative Industries

Amazon MGM Studios has formalized a governance framework that embeds AI across workflows while preserving human creative authority — a model likely to be replicated across studios.

"AI [use] must be human-centered. That is a North Star. We had to think through how AI can be applied to each of our workflows, and also make sure that people are driving the creative process." — Albert Cheng, Amazon MGM Studios

💰 AI Is Dramatically Compressing Entertainment Production Economics

Cost and time savings are not being pocketed — they are being reinvested to expand content volume, which signals a broader content supply shock coming to streaming.

"We can actually fit five movies into what we would typically spend on one. If anything, I think [AI] can actually increase, improve and expand the possibilities." — Albert Cheng

📡 Nvidia's China Re-entry Is a Major Geopolitical-Commercial Flashpoint

Nvidia is actively restarting H200 chip manufacturing for Chinese customers while also preparing a Groq inference chip variant for the market — representing a high-stakes bet on navigating dual-approval export complexity.

"We've been licensed for many customers in China for H200. We have received purchase orders from many customers, and we're in the process of restarting our manufacturing." — Jensen Huang

☁️ Cloud Hyperscalers Are Making Massive AI Revenue Bets

Amazon CEO Andy Jassy told employees AI should allow AWS to reach $600 billion in annual revenue by 2036 — up from ~$130 billion today — doubling a prior estimate. Nvidia separately forecasts $1 trillion+ in chip revenue by 2027 from Blackwell and Rubin systems alone.

Huang told reporters "his prediction of $1 trillion in revenue by 2027 only reflects sales of Blackwell and Rubin systems and suggests there is plenty of upside from other products, including CPU-only systems and storage."

2. Contrarian Perspectives

🚨 Jensen Huang Thinks AI Safety Alarmism Is Itself a Problem — Not a Virtue

While the consensus narrative praises companies like Anthropic for drawing "red lines" around AI use in warfare and surveillance, Huang pushes back hard, arguing that fear-mongering crosses a line from responsibility into arrogance. This is a minority view among high-profile tech leaders and warrants watching as a potential regulatory flashpoint.

"Scaring everybody about a science fiction version of AI is a little bit too arrogant... Warning people is one thing. Scaring people is a completely different thing." — Jensen Huang

His comments came directly after Anthropic drew red lines around U.S. government use of AI for "fully autonomous warfare and for mass surveillance of Americans."

🎥 AI in Hollywood Is a Content Supply Expansion Story, Not Just an Efficiency Story

The conventional framing is that AI will cut costs in Hollywood. The more important and underappreciated implication is that it could flood the market with content, fundamentally altering content scarcity economics for streaming platforms and advertisers.

"The time and cost savings decrease the amount of time between seasons and are being reinvested in the creation of more content and storytelling." — Albert Cheng

🧑‍💻 AI Studios Could Democratize Content Creation at a Structural Level

The mainstream focus is on how AI affects existing studios. Cheng's view suggests a more disruptive outcome: individual creators becoming their own content studios, which would challenge the incumbent distribution and production model from below.

"Cheng believes AI studios could be the next wave of entrepreneurship as the tech allows more individuals to 'be their own content studio.'"

3. Companies Identified

Netflix Streaming platform Acquiring AI startup to support post-production; positioned as an early mover in production-side AI integration.

"Netflix recently announced a deal to acquire Ben Affleck's startup that uses AI to support the post-production process."

Peacock (NBCUniversal) Streaming platform Deploying AI-generated avatars of talent to drive in-app content discovery.

"Peacock rolled out a new AI avatar of TV personality Andy Cohen to help viewers discover content within its app."

Amazon MGM Studios / Prime Video Studio and streaming platform Deepest case study in the newsletter; using AI for pre-production scene mapping, post-production effects, personalized recommendations, AI-generated synopses, and accessibility features.

"You can shoot something and see near-final visuals the same day." — Albert Cheng

Nvidia AI chip manufacturer Restarting H200 production for China; forecasting $1T+ revenue by 2027; featured as the central infrastructure player powering the AI economy.

"Nvidia will generate more than $1 trillion in revenue by 2027 from its mainstay AI chips." — Jensen Huang

Anthropic AI safety company Named as the direct foil to Huang's anti-alarmism argument; drew public red lines on AI use in warfare and domestic surveillance.

"His comment comes after Anthropic drew red lines around U.S. government use of AI, specifically around fully autonomous warfare and for mass surveillance of Americans."

Huawei Technologies Chinese chip manufacturer Identified as the domestic competitor Beijing is pushing Chinese companies toward, in direct competition with Nvidia's re-entry.

"Beijing has to approve any sales into the country, and has been pressing domestic companies to use Chinese-made chips — part of an effort to boost the domestic industry led by Huawei Technologies."

OpenAI AI model developer Released mini and nano versions of GPT 5.4, expanding low-cost access to its latest engine. Signals a continued push down-market to capture broader developer and consumer segments.

"OpenAI released mini and nano versions of GPT 5.4, offering faster and lower-cost access to its latest AI engine."

Google AI and cloud company Expanding its "personal intelligence" Gemini/Gmail integration from paid to free U.S. users — a distribution move that increases lock-in at scale.

"Google is expanding access to its 'personal intelligence' features to all U.S. users after previously limiting the Gmail integration feature to paid users."

Microsoft AI and cloud company Restructured its AI division: former Snap executive Jacob Andreou now leads Copilot consumer and enterprise products; Mustafa Suleyman shifts focus to internal model development.

"Microsoft reorganized its AI efforts, putting former Snap executive Jacob Andreou in charge of consumer and business versions of Copilot, with AI chief Mustafa Suleyman now focused on the company's internal model development effort."

4. People Identified

Jensen Huang CEO, Nvidia Central figure driving Nvidia's China re-entry narrative and publicly reframing AI safety discourse as potentially counterproductive.

"We have to just continuously learn and be a little bit more humble about what we know and don't know."

Albert Cheng Head of AI Studios, Amazon MGM Studios Most substantive operator voice in the newsletter; architect of the "creator in the loop" AI production model and key source on economics of AI-driven content creation.

"AI [use] must be human-centered. That is a North Star."

Raf Soltanovich Head of Technology, Prime Video and Amazon MGM Studios Works in tandem with Cheng to build AI tools; frames the goal as a virtuous cycle between technology capability and creative ambition.

"Powerful, forward-thinking creators are really leaning in to deeply understand what is possible, and how this can actually help elevate their storytelling."

Steven Spielberg Filmmaker Represents the softening of high-profile creative resistance to AI — still cautious, but no longer categorically opposed.

"I am not for AI if it replaces a creative individual," but added he sees value "in many disciplines."

Andy Jassy CEO, Amazon Made a bullish internal projection linking AI to a 5x growth target for AWS over the next decade.

"Amazon CEO Andy Jassy told employees that AI should allow the company's AWS unit to reach $600 billion in annual revenue by 2036, double an earlier estimate."

Mustafa Suleyman Chief AI Officer, Microsoft Shifting from consumer-facing Copilot leadership to internal model development — a notable strategic repositioning within Microsoft's AI org.

"AI chief Mustafa Suleyman now focused on the company's internal model development effort."

Jacob Andreou Former Snap executive; now leading Copilot, Microsoft Elevated to lead Microsoft's highest-visibility AI consumer and enterprise product.

"Putting former Snap executive Jacob Andreou in charge of consumer and business versions of Copilot."

Ben Affleck Actor/filmmaker and AI startup founder His unnamed AI post-production startup was acquired by Netflix, reportedly for ~$600M — a signal that Hollywood talent is building, not just reacting, to AI.

"Netflix recently announced a deal to acquire Ben Affleck's startup that uses AI to support the post-production process."

5. Operating Insights

✅ Map AI to Individual Workflows Before Deploying It Broadly

Cheng's approach at Amazon MGM is instructive: rather than applying AI as a blanket tool, the team audited each production workflow independently to identify where AI adds value without displacing human creative ownership. This reduces labor friction and copyright risk simultaneously.

"We had to think through how AI can be applied to each of our workflows, and also make sure that people are driving the creative process."

✅ Use AI to Compress the Feedback Loop Between Production and Visualization

One of the highest-leverage applications at Amazon Studios is using AI to generate near-final visuals on the same day of shooting — dramatically reducing iteration cycles and late-stage production surprises.

"You can shoot something and see near-final visuals the same day." — Albert Cheng

✅ Copyright Protection Requires Human Inputs at Every Stage of the Creative Process

For any company building AI-assisted creative workflows, this is a structural legal and commercial guardrail: human involvement at each stage is not just good practice — it is the mechanism by which copyright ownership is preserved.

"In order for us to protect copyright, we need to have human inputs in all parts of the creative process." — Albert Cheng

6. Overlooked Insights

🔍 Nvidia's $1T Revenue Forecast Excludes a Significant Portion of Its Business

Huang's headline $1 trillion figure is widely reported — but less noted is that it is deliberately conservative, covering only Blackwell and Rubin chip systems. The total addressable opportunity is materially larger when CPU-only systems and storage are included.

"His prediction of $1 trillion in revenue by 2027 only reflects sales of Blackwell and Rubin systems and suggests there is plenty of upside from other products, including CPU-only systems and storage."

🔍 SAG-AFTRA Contract Negotiations Are the Coming Regulatory Chokepoint for Hollywood AI

The article briefly flags ongoing negotiations between SAG-AFTRA and the Alliance of Motion Picture and Television Producers as an active battleground for AI governance in entertainment — the outcome will likely set enforceable industry-wide standards for how and where AI can be deployed in production.

"AI has been a critical piece of the ongoing negotiations between SAG-AFTRA and the Alliance of Motion Picture and Television Producers, which began last month."