- Uncovering AI
- Posts
- 📸 74% of AI profit is seized by 20% of companies
📸 74% of AI profit is seized by 20% of companies
What side are you on?
In partnership with
My fellow AI explorers
It’s Monday morning, and the AI world just dropped three grenades on my news feed — all in the last 48 hours.
Stanford just released its annual AI report card. Google is now using AI to make full-length songs. And a major new study confirmed what you probably suspected: Most companies are playing AI checkers while a tiny few are playing chess.
In today’s edition:
📊 The Stanford AI Index 2026: China has caught up, and your benchmarks are broken
🎵 Google Lyria 3 Pro: AI just became a music producer
💸 PwC's new study: 20% of companies are eating 74% of AI's value — are you in that group?
The ones showing up in LLMs convert 3× better than Google
They optimized for LLMs, not just Google.
FAQs. Comparison pages. Transparent pricing. LinkedIn presence. These aren't vanity plays. They're what gets you cited in ChatGPT, Gemini, and Claude when your buyers are researching, your investors are looking, and your future hires are deciding where to work.
Download the free AEO Playbook for Startups from HubSpot and get the exact checklist. Five minutes to read.
Stanford HAI
📊 The 2026 AI Index Just Dropped — Read This Before Anyone Else Does
Stanford's annual AI Index is out today, and it's the clearest snapshot we have of where this technology actually stands. The headline number? Generative AI reached 53% population adoption within three years — faster than the personal computer or the internet. The estimated value of generative AI tools to U.S. consumers already reached $172 billion annually by early 2026, with the median value per user tripling between 2025 and 2026. Stanford HAI
But here's what's really going on underneath that growth:
On SWE-bench Verified, a key coding benchmark, model performance jumped from 60% to near 100% in a single year. Organizational AI adoption reached 88%, and 4 in 5 university students now use generative AI. MIT Technology Review
The most powerful models are also the least transparent. The Foundation Model Transparency Index dropped from 58 to 40 points this year — the most capable models disclose the least about how they work. Stanford HAI
The talent pipeline stat should be a flashing red alarm for US policymakers: the number of AI researchers and developers moving to the US has dropped 89% since 2017, with an 80% decline in the last year alone. Stanford
The talent pipeline stat is the one that should be flashing red for US policymakers: the number of AI researchers and developers moving to the US has dropped 89% since 2017, with an 80% decline in the last year alone. Stanford
🔮 Prediction: The AI race is no longer just about who builds the best model. It's about who controls the infrastructure, the narrative, and the pipeline. America is winning on dollars. China is winning on volume and patents. And everyone is running on benchmarks that the Stanford report itself admits are poorly designed — one popular math benchmark has a 42% error rate. We're measuring the wrong things, too slowly.
Google DeepMind
🎵 Google Just Shipped a Full-Length AI Music Studio — And It's Going Enterprise
This one flew under the radar for many, but it shouldn't. Google released Lyria 3 Pro, a music generation model that lets users create tracks up to three minutes long — a six-fold jump from the 30-second cap of its predecessor, Lyria 3, which launched just one month earlier. TechCrunch
What makes this release different:
Lyria 3 Pro understands full musical composition. Users can now prompt for specific structural elements like intros, verses, choruses, and bridges, not just a vibe and a genre. Google
It's available across six platforms: the Gemini app, Google Vids, ProducerAI, Vertex AI, the Gemini API, and AI Studio. This means it’s the first major cloud provider to put a music generation model directly into its commercial API stack. Google Cloud
Google trained Lyria 3 Pro on licensed and permissible data from YouTube and Google partners — a deliberate legal differentiation from Suno and Udio, which are both facing copyright lawsuits from major labels. TechCrunch
This is the moment music AI crosses from demo to infrastructure. The same trajectory we saw for text generation in 2021 and image generation in 2022. Audio just hit that same inflection point.
🎧 Use case for right now: If you create video content (YouTube, ads, social), you no longer need stock music. Open Google AI Studio, select Lyria 3, describe your vibe, and generate a custom track in minutes. No royalty issues, no licensing headaches.
PwC
💸 74% of AI's Value Is Going to 20% of Companies — Which Side Are You On?
PwC dropped a bombshell study today. After interviewing 1,217 senior executives across 25 sectors, they found that nearly three-quarters of AI's economic value is being captured by just one-fifth of organizations — revealing a stark divide between a small group of AI leaders and the majority still stuck in pilot mode.
The gap between those two groups isn't about spending more on AI. It's about how they're using it:
Top-performing companies are twice as likely to redesign workflows around AI, rather than simply layering AI tools onto existing processes. Apacoutlookmag
Leaders are 2.6 times more likely to say AI helps them reinvent their business model, and 2 to 3 times more likely to use it to identify growth opportunities from industry convergence. PwC
The best AI performers are 2.8 times more likely to have increased the number of decisions made without human intervention — they're not using AI as a suggestion engine, they're letting it run workflows autonomously. Apacoutlookmag
The brutal takeaway: PwC warned that the gap will probably grow. Leaders are learning faster, scaling successful use cases, and automating decisions at a pace the laggards can't match. Storyboard18
🔮 Prediction: We're entering the consolidation phase of AI adoption. The companies that jumped in early and built real systems — not just Copilot licenses — are about to pull so far ahead that the gap becomes structural. The 80% stuck in pilot mode aren't just behind on tools. They're behind on organizational learning, and that gap is far harder to close.
30-Second AI Play
🧠 Mine the Stanford AI Index for Stats That Win Rooms
The 2026 AI Index is 400+ pages. You don't have time. Here's how to get the signal in under 5 minutes:
Jump straight to the Executive Summary — the biggest findings are condensed in the first 10 pages
Skim the Workforce chapter — this is where the job disruption data lives. Perfect for any client conversation
Pull 2-3 stats relevant to your industry and paste them into a running notes doc
Use them in your next meeting when someone asks "but how big is AI really?"
💡 Pro tip: The single most jaw-dropping stat in the report is that the median consumer value per user has tripled in one year. Lead with that. It stops every skeptic cold.
Other Relevant AI News!
🌏 China has effectively erased the US lead in AI, according to the Stanford Index — while the US still leads in capital and chips, China now dominates in patents, publications, and physical AI/robotics. Full story here.
🌿 AI's environmental cost is exploding — xAI's Grok 4 training alone generated an estimated 72,000 tons of CO2, and running GPT-4o inference may use enough water annually to sustain 12 million people.
📱 Apple is designing its first AI smart glasses with multiple frame styles and a unique vertical camera system — with the foldable iPhone also on track for September.
🔬 AI in science publishing has exploded — the number of natural sciences papers mentioning AI grew nearly 30-fold between 2010 and 2025, now appearing in 6–9% of all publications across biology, physics, and earth sciences.
🤖 Anthropic's Cowork — its AI agent platform for automating professional tasks like contract review — reportedly triggered a sharp sell-off in legal tech and SaaS stocks, in what analysts are calling the SaaSpocalypse.
Golden Nuggets
📊 AI adoption is outpacing the internet — but the benchmarks we use to measure AI capability are broken
🎵 Music AI just went enterprise-grade. If you're still paying for stock music, you're overpaying
💸 74% of AI's economic value is going to 20% of companies. The window to join that group is narrowing fast
Would love to hear your thoughts! Send me your thoughts by replying to this email (yes, I read them all :)
Until our next AI rendezvous,
Anthony | Founder of Uncover AI