The prompt asks AI to read your company’s health benefits contract and tell you where you’re being ripped off.

Why would that work?

• Two in three Americans insured through work are in self-funded plans: the employer pays the medical bills itself

• To process those bills, employers hire administrators owned by the big insurers

• UnitedHealth alone runs 3,000+ subsidiaries and moves about $150B a year between them

• The contracts behind all this run over 1,000 pages, and employers often can’t see their own claims data

The money hides in contracts nobody has time to read. A model reads them in seconds.

This week Cuban sat down with a16z’s Julie Yoo and Erik Torenberg and laid out where AI pays in healthcare, where it nets out to zero, and the company he would start tomorrow.

High-signal analysis on markets, capital and the companies shaping the next decade

Inside you will find:

1. How AI Let Cost Plus Drugs Force Incumbents to Match Its Prices See how 5,000-square-foot factory and a public price list became a weapon.

2. Why the Biggest Market in Healthcare Was Closed to Startups Until Now Learn how employer health plans work and where $150B a year disappears from view.

3. The One Prompt That Could Collapse the Big Insurers Get the exact prompt Cuban gives employers, and the math behind his collapse claim.

4. Why Prior-Auth and Claims Bots Will Net Out to Zero Use this filter on every healthcare AI pitch you see.

5. The Healthcare AI Company Cuban Would Start Tomorrow Find out how an agent layer can connect patients to six-figure aid nobody uses.

This issue is available exclusively to paid subscribers!

If you’re not subscribed yet, you can unlock this and our full library by upgrading below:

Get 7 day free trial

Let’s begin!

Read more

Harvey and Legora raised almost $1.9B between them in 2026, selling AI to law firms across the US and Europe.

I’ve been super bullish on vertical AI for years, and confident AI would take over much of the slow, expensive legal industry. (My first serious job, back when I thought building things wasn’t serious, was as an M&A lawyer.)

I just found the most promising company I’ve seen in this sector, and I’m fighting for an allocation before it’s worth tens of billions 😂

It’s called GC AI. It sells to the legal team inside a company, the general counsel and their lawyers, and that buyer is now the fastest-growing part of legal AI.

In this newsletter I’ll run you through:

• The vertical AI market, and how to build a billion-dollar company competing with the frontier labs

• How GC AI is winning the in-house buyer that Harvey and Legora are now chasing

High-signal analysis on markets, capital and the companies shaping the next decade

Inside you will find:

Part 1: The Vertical AI Market, and How to Win It Against the Frontier Labs

• Legal AI Just Had the Biggest Year of Any Vertical AI Category See where Harvey and Legora’s $1.9B went, and where the new growth is coming from.

• Why Legal AI Is Really Two Markets Learn how law firms and in-house teams buy, work and measure success.

• Why the In-House Buyer Is Where Legal AI Is Heading Find out why Harvey and Legora are both moving toward company legal teams.

• How to Build a Billion-Dollar Vertical AI Company When Everyone Has the Same Models Get the five tests I use to judge whether a vertical AI company can beat the frontier labs.

Part 2: How GC AI Is Winning the Buyer Harvey and Legora Are Now Chasing

• Why the Founder Is GC AI’s Moat See how three GC jobs and early access to GPT became the product spec.

• The Four Product Choices That Win In-House Deals Learn the product decisions that make GC AI fit the way in-house teams work.

• Why GC AI’s Market Is Bigger Than Legal Find out why nearly 30% of its seats go to people outside the legal team.

• The AI Market Fit Scorecard: How GC AI Scores on the Five Tests See where GC AI passes, and the one question still open.

This is normally a paid issue, but we’re making it free today so you can get a taste of what we publish for our paid subscribers!

Subscribe for free to read it! Let’s begin 🚀

Read more

Anthropic is getting into the consulting business.

On October 23, the FT reported that Anthropic and Hg, a London buyout firm, agreed to launch a European AI consultancy by the end of the year. Its consultants and engineers will sit inside companies from hubs in London and Berlin and build AI into how they work, the job Accenture and the Big Four do today. Neither company has confirmed it yet.

It’s Anthropic’s second consulting firm in six months, and OpenAI has one too:

• May 4: Anthropic, Blackstone, Hellman & Friedman and Goldman Sachs create Ode, a consulting firm for mid-sized companies with about $1.5B of planned investment

• May 11: OpenAI launches its own, the Deployment Company, with more than $4B and 19 partners led by TPG

• October 23: Anthropic and Hg bring the model to Europe

Why does it matter?

For years the labs sold the model and left the implementation to consultants.

→ Now they’re building their own consulting firms.

Most companies don’t have the engineers to put AI into their core operations, so the labs send their own, with private equity bringing the money and the clients.

The consultants noticed. Accenture’s shares fell more than 4% the day the news broke.

High-signal analysis on markets, capital and the companies shaping the next decade

Inside you will find:

1. Why the Labs Are Building Consulting Firms Now See why demand for AI implementation outgrew the model business, and what it means for Accenture and the Big Four.

2. The Three-Piece Machine: Lab, Private Equity and a Boutique Bought on Day One Find out how Ode, OpenAI’s Deployment Company and the Hg venture copy the same structure.

3. How Anthropic Covers Every Company Size Learn how integrators, Ode and Hg split the market from the Global 2000 to the European mid-market.

4. How an Ode Engagement Works, From the CEO Call to Production Follow the six steps of an engagement, rebuilt from Ode’s announcement, interviews and job posts.

5. Why the Model Is the Smallest Part of the Work See how evals and reusable code cut a two-month deployment to one day for the second team.

6. Why Private Equity Is the Best Sales Channel in AI Find out why owners of hundreds of companies are the shortest path to signed AI projects.

7. What the Labs Pay for Their Special Forces See the salaries on Ode’s job board, the profiles it hires and Anthropic’s $100M plan to train 10,000 more.

8. How to Run the Same Play With a Ten-Person Team Build your own version: one vertical, a three-stage offer, a channel and a place on Anthropic’s partner ladder.

This issue is available exclusively to paid subscribers!

If you’re not subscribed yet, you can unlock this and our full library by upgrading below:

Get 7 day free trial

Let’s begin!

Read more

There’s a $15 BILLION AI company almost nobody knows about.

It’s called Applied Intuition. And while everyone is fighting for attention in AI coding tools, chatbots, and foundation models, this company has quietly built an estimated $830 million ARR business, doubling in just one year.

Here’s the crazy part: most of its revenue no longer comes from the industry it started in.

Applied began by building simulation software for self-driving cars. Today, its technology powers the development of intelligent cars, trucks, military vehicles, mining equipment, and agricultural machinery.

Its ambition? To become the software infrastructure behind the entire physical AI revolution.

And its CEO, Qasar Younis, believes the company could eventually be worth $100 BILLION.

and I sat down with Qasar for a 52-minute conversation to understand how he built one of the most successful, yet least-known, AI companies in the world.

We broke down his entire playbook: how he identified the initial opportunity, scaled into new industries, built a three-layer AI product platform, and why he now believes distribution could unlock the company’s next chapter.

In this guide, you’ll discover:

1. How Applied found a buyer-funded wedge in automotive AI.

2. Why physical AI could be a much bigger opportunity than coding tools.

3. The three-layer product stack behind its estimated $830M ARR.

4. How Qasar thinks about capital, investors, and building a $100B company.

5. The customer learning system behind Applied’s growth.

6. Why founder-led distribution could become its next competitive advantage.

👋 Just a quick note before you continue!

If you are serious about building or investing in AI, you can subscribe here:

Here’s a good place to start:

↳Alex Karp’s Playbook for Which AI Companies Survive

↳19 great new details about Cursor’s rise

↳Perplexity’s CEO 2026 AI Pitch

↳The First 48 Hours of Claude Fable 5

Try premium free for 7 days. Or get 50% off this week only.

Get 50% off forever

1. How Applied found a buyer-funded wedge in automotive AI

Read more

Anthropic went from $41 million of revenue in 2023 to $11.6 billion in Q2 2026 alone. In the same quarter, it reached $974 million of GAAP operating profit.

That makes the February 2023 pitch deck look prophetic. But the S-1 also reveals hundreds of billions of dollars in infrastructure arrangements, complex accounting, and a governance structure that gives public investors limited control.

In this newsletter we went through:

• All 318 pages of the S-1

  

• The original ten-slide pitch deck

  

You’ll find both at the end of the newsletter!

High-signal analysis on markets, capital and the companies shaping the next decade

In this issue you’ll find:

1. The deck’s real bet was trust, not chat

2. Demand turned into an enterprise machine

3. “Demand > capacity” became the whole AI economy

4. The S-1 shows real growth and real operating leverage

5. Claude’s moat comes with dependencies investors cannot ignore

6. Great company, unanswered stock: what to watch before buying

This issue is available exclusively to paid subscribers!

If you’re not subscribed yet, you can unlock this and our full library by upgrading below:

Get 7 day free trial

Let’s begin!

Read more

For decades one law held: a small team with a three-year head start was untouchable, because Google was never going to hire 2,000 engineers to catch them.

Ben Horowitz just declared that law dead.

Three things follow: money now replicates talent, scarcity collapsed to roughly 40 people on earth, and half of what investors still price as a moat is worthless.

That’s the thesis Ben Horowitz has

In this issue you'll find:

1. Why the market is wrong that incumbents are finished, even as AI products grow revenue faster than anything before

2. The old law that just died: you could never hire your way past a three-year head start

3. Why Elon catching the frontier models proves money now replicates talent

4. The new scarcity: roughly 40 people on earth can build a frontier model, and schools can't make more

5. AI is the greatest equalizer of opportunity we've seen, and crypto already showed how the little guy wins

6. The job-destruction bet almost no one is pricing correctly

7. The IPO bar moved from $50M to $1B, and it changed what you fund before exit

8. The trade Horowitz calls a really good business idea: AI-powered PE roll-ups, and why a16z won't run it

👋 Just a quick note before you continue!

If you are serious about building or investing in AI, you can subscribe here:

Here’s a good place to start:

↳Alex Karp’s Playbook for Which AI Companies Survive

↳19 great new details about Cursor’s rise

↳Perplexity’s CEO 2026 AI Pitch

↳The First 48 Hours of Claude Fable 5

Try premium free for 7 days. Or get 50% off this week only.

Get 50% off forever

1. Why the market is wrong that incumbents are finished, even as AI products grow revenue faster than anything before

Read more

Yesterday SpaceX went live from its Starlink factory in Bastrop, Texas, for its Q3 2026 earnings, with Elon Musk, Gwynne Shotwell, and CFO Bret Johnsen on the desk.

We watched the full 67 minutes and read the official release so you don't have to.

The short version:

• a potential Tesla merger

• a chip fab called TeraFab, that would each be the biggest tech story of most years.

👋 Just a quick note before you continue!

If you are serious about building or investing in AI, you can subscribe here:

Here’s a good place to start:

↳Alex Karp’s Playbook for Which AI Companies Survive

↳19 great new details about Cursor’s rise

↳Perplexity’s CEO 2026 AI Pitch

↳The First 48 Hours of Claude Fable 5

Try premium free for 7 days. Or get 50% off this week only.

Get 50% off forever

The numbers

Read more

Cluely had made going viral its entire growth strategy. Our idea was to bring that distribution mindset to investing: back exceptional founders, then help them reach millions of potential customers and investors.

Twelve months later:

→ A portfolio that has raised $1.2B+.
→ Companies including HappyRobot, Replit, Nozomio, Throxy, Apollo Atomics, Gru Space and QFEX
→ Alongside investors including a16z, Sequoia, Founders Fund and General Catalyst
→ 1 Exit: Portco went from $10Kmrr to 7M arr in close to 1 year

Today, we’re publicly introducing what we’ve been building: GR Angel.

Here’s how it started, what we’ve accomplished and where we’re taking it.

1. Before GR Angel, We Were Two Angels With an Audience

Rubén and I came to angel investing after working in venture capital.

I had been an M&A lawyer at PwC, founded a B2B SaaS company, and worked at GoHub Ventures. Rubén had worked at Mundi Ventures before leading investments and media at Itnig Capital.

Alongside those careers, we built our publications and audiences. Through that work, we developed relationships with founders, operators and investors across the startup ecosystem.

When we started writing angel checks, we put those assets to work for the companies we backed.

We shared their launches, featured their products, and introduced them to people who could help. One newsletter feature led to a $1 million investment offer for a founder. A single post from Rubén helped another startup double its ARR.

Founders noticed. They started making room for us in oversubscribed rounds because they wanted that support.

As we shared our investments and the companies’ progress, several private investors approached us. They wanted access to the opportunities we were finding and asked to invest alongside us.

We opened the door to a small group. That was the beginning of GR Angel.

2. Our Edge: Distribution and Network Access

The founders we want to back often have several investors competing for a place in their rounds.

Our job is to give them a compelling reason to choose us and deliver after we invest.

We bring two things.

Distribution: Getting Founders in Front of Millions

I write Product Market Fit and AI Market Fit. Rubén writes The VC Corner, The AI Corner and The Founders Corner.

We both have 2 of the biggest Linkedin accounts in Europe + now some of the fastest growing X accounts.

Together those publications reach hundreds of thousands of founders, operators and investors, generating tens of millions of impressions every month.

For a portfolio company, that creates an audience for product launches, a way to reach potential customers, and visibility with investors ahead of its next round.

We can help a founder explain what they’re building and get it in front of people who could buy it, work on it, or invest in it.

Network Access: Opening the Right Doors

Our work in venture capital, our publications and our investments have built relationships across the startup ecosystem.

We use those relationships to make targeted introductions: a potential customer, an experienced founder who has solved a similar problem, or an investor who could be relevant for the next round.

We’re also expanding that network through Compound, our podcast, with guests including Mark Cuban, Substack co-founder Chris Best and Creandum founding partner Staffan Helgesson. Rubén is also an official scout for a16z Speedrun.

The value is in connecting founders to the right people at the right moment and giving those people a reason to pay attention.

Distribution and network access help us earn allocations. After we invest, we put both to work for the founders.

Our investing now focuses on the top AI companies coming out of Y Combinator. YC gives us a concentrated place to look. Our relationships and reach help us compete for the opportunities we want.

3. The Companies Behind the Numbers

Across our personal angel investments and GR Angel, we have backed 53 companies, including 42 from Y Combinator. Together, those companies have raised $1.2 billion+.

These figures cover our combined portfolio, including investments made before GR Angel existed. Here are some of the founders and businesses behind it.

HappyRobot: AI Agents for Logistics

Pablo Palafox, Luis Paarup and Javier Palafox are building AI agents for logistics—the calls, emails and messages that keep freight moving.

The company reached a $1.2 billion valuation following its $150 million Series C. Its customers include DHL, Kuehne+Nagel and Uber.

Replit: Making Software Creation Accessible to Everyone

Amjad Masad and Haya Odeh built Replit to let anyone create software. Its AI agents have accelerated that ambition: describe what you want, and the platform helps build it.

Replit grew ARR from $10 million to $525 million in 18 months, passed 50 million users, and raised a $400 million Series D at a $9 billion valuation in March 2026.

Peec AI: From Zero to $10 Million ARR in 16 Months

As people move from searching Google to asking AI, brands need to understand whether they appear in the answers.

Marius Meiners, Tobias Siwonia and Daniel Drabo built Peec AI to solve that problem. The company reached $10 million ARR in 16 months, with 3,000+ customers, including Zalando, Hugo Boss and Chanel. It raised a $21 million Series A in November 2025.

QFEX: Backed by General Catalyst and Paul Graham

Annanay Kapila and Joshua Wharton are building a trading platform for continuous global markets.

QFEX raised a $9.5 million seed round at a $95 million valuation before generating revenue, with General Catalyst and Paul Graham among its investors.

Nozomio: Better Context for AI Coding Agents

Arlan Rakhmetzhanov, one of the youngest founders we’ve backed, is building infrastructure that helps AI coding agents access the information they need to work effectively.

Nozomio raised a $6.2 million seed round led by CRV, with Paul Graham and Gokul Rajaram investing alongside us.

GRU Space: A Hotel on the Moon

Skyler Chan is pursuing one of the most ambitious ideas in the portfolio: building the first hotel on the Moon.

The company reports $500 million in letters of intent and has attracted 1 billion+ views. It’s an early, highly speculative bet—and an example of the scale of ambition we’re willing to back.

More Companies to Watch

The portfolio also includes:

• Quiver AI: an $8.3 million seed round led by a16z.

• CodeWords (Agemo): €7.6 million raised, with the CEOs of Miro, ElevenLabs, Personio, Zalando and Supercell among our fellow angels.

• Momentic: AI-powered software testing, with a $15 million Series A and General Catalyst among its backers.

• Throxy: $6.2 million raised with Base10.

• Apollo Atomics: an MIT nuclear spinout that raised a $31 million seed round.

• Karumi: personalized product demos, backed by Rebel Fund.

Across the broader portfolio, our co-investors include a16z, Sequoia, Founders Fund, Coatue, General Catalyst, Peak XV, Lightspeed, Khosla Ventures, CRV and Kfund.

Explore the full portfolio →

4. Why We’re Launching Publicly Now

A year after our first private investors joined us, they wanted to increase their commitment.

They had seen the companies we could access, how we supported founders, and how the portfolio was developing. They encouraged us to grow GR Angel and bring in more capital.

That is why we’re taking this next step.

We’re raising. We’re opening conversations with additional private investors who want to invest alongside us, allowing us to write larger checks into our highest-conviction opportunities.

We’re hiring. We’re looking for exceptional people across sourcing, content and operations to support the investment activity and the distribution behind it.

We’re meeting founders. If you’re building an exceptional AI company, we want to hear what you’re working on and where our audience and network could help.

We started by investing our own money. Investors asked to join us. A year later, they pushed us to think bigger.

Today, we’re launching GR Angel publicly, with the same commitment that got us here: invest in founders, put them in front of millions, and open doors that help them grow.

Hope this was valuable!

Cheers,
Guillermo

Portfolio figures combine our personal angel investments and investments through GR Angel. GR Angel is a private investment company, not a fund. This newsletter is for information only and is not investment advice or an offer to sell securities. Past performance is not indicative of future results.

The result is a 23-slide deck.

I’m publishing it here in full, slide by slide, with the story behind each one.

The short version: 2026 was simultaneously the biggest gold rush in venture history and the year reality caught up with the hype. Both things are true.

The full deck is at here: share it freely.

State Of Ai 2026 Guillermo Flor

14.1MB ∙ PDF file

Download

Download

Now, the breakdown:

1. AI outlook for 2026

2. Sequoia Tech Waves and what's the next wave

3. How electricity becomes key bottleneck

4. Fastest growing AI startup verticals

5. Why companies are not seing ROI from AI

6. Five AI investment bets every VC and founder should have in mind

PS: In case you missed, check out The AI Rollup Playbook: Why Buying a Boring Business and Running It With AI Agents Is the Best AI Bet of 2026

Onwards!

1. Gold

Rush & Reckoning

The theme of the deck. The money pouring into AI has no precedent and neither does the gap between companies deploying AI and companies getting value from it. Every number that follows is sourced and current as of this week.

2. 2026 was the most capital-concentrated year in VC history: $679B in nine months

The whole story in six tiles. $679B of venture in nine months, up to 80% of it into AI. Anthropic passed OpenAI and goes public in November. Hyperscalers are spending ~$700B this year, and the bottleneck is electricity now, not chips. Growth records kept falling while ~5,600 AI startups died. The opportunity: sell work, not software.

3. Every technology wave minted more value than all waves before it. AI is the steepest yet.

Sequoia’s classic framework, updated. Semiconductors, systems, networks, internet, apps, mobile. each wave was worth more than everything before it, and each was built on the previous one. AI is the seventh wave, and its adoption curve is steeper than any of them. The logos tell the story of who captured each wave.

4. The Gold Rush

California, 1850. A few miners struck the vein; most went home with nothing. The people who reliably got rich sold picks, shovels and denim, Levi Strauss never touched a pickaxe. The structure is identical today: the river is intelligence, the shovels are power and compute, and the denim is deployment.

5. AI took up to 80 cents of every venture dollar in 2026

AI’s share of all venture funding went from 34% in 2024 to 80% in early 2026. The Q3 dip to 64% isn’t a slowdown, Q3 was still up 53% year over year. The giant lab rounds closed in the first half, then capital rotated into infrastructure and physical AI. The money didn’t leave. It moved.

6. Four companies raised 65% of all venture capital on Earth in one quarter

OpenAI, Anthropic, xAI and Waymo raised $188B in Q1, 65% of all venture capital on the planet. Every other startup split the rest. If you’re building outside that club, this is the weather you’re operating in.

7. OpenAI’s $122B is the largest private round in history

Amazon wrote $50B of it; Nvidia and SoftBank about $30B each. Anthropic’s $65B Series H in May was its second mega-round of the year, at a $965B valuation. Below those two, the drop-off is steep, the top two bars are a different species from everything else.

8. The seed round became a $1B game for famous names

David Silver (AlphaGo) raised a $1.1B seed (the largest in European history) for a company that was months old. Yann LeCun’s new lab raised $1.03B three months after founding. Core Automation was a unicorn one month after incorporating. 311 new unicorns were minted through September, the second-highest year ever, after 2021.

9. Anthropic overtakes OpenAI: $1B → $65B+ in eighteen months

The company story of the year. $16.8B of revenue in Q3, up 45% quarter over quarter, with its first adjusted profit. 300,000+ business customers, 80% of revenue from enterprise. Claude Code alone is at ~$8B ARR and writes ~4% of public GitHub commits. The S-1 is filed; November is the target. That IPO matters for everyone, it’s the first audited look at frontier-lab economics.

10. OpenAI is still enormous. It is also suddenly second.

~$50B annualized, $122B raised in March and then quiet: no 2026 IPO, and September revenue revised down on accounting treatment. The leaked 2025 financials showed $13.1B of revenue against a $20.9B operating loss. They’re shipping fast and waging a price war with Anthropic, but the strategic tell is real: they stopped building data centers and started renting compute instead.

11. 2026 blew the exit window open

The largest IPO ever (SpaceX, $1.77T, with xAI inside), the largest merger ever ($1.25T, SpaceX–xAI), and the largest acquisition of a venture-backed startup ever (SpaceX buying Cursor for ~$60B) — all in one year. Anthropic files next. The warning label: Figma fell from $143 to $22. The window is open, but the market is picky about what it lets through.

12. Hyperscalers will spend ~$700B on AI in 2026 — and the constraint is now electricity, not chips

The four big hyperscalers nearly doubled capex this year. Nvidia did $75B of data-center revenue in a single quarter. But GPUs are cheap to rent now — what you can’t get is power. Grid queues run 4–8 years and $100B+ of projects are stuck waiting. That’s why everyone, from Meta to OpenAI, is suddenly designing their own silicon and buying energy-integrated campuses.

13. The New Constraint

Ansel Adams shot Boulder Dam in 1941. The last industrial age ran on concrete and electrons; this one does too. The chip shortage is over. The electron shortage has begun. In 2026, the most valuable real estate in tech is a substation.

14. Coding is the breakout category — $8B ARR in eight months

What’s actually working: coding (Claude Code $500M → ~$8B ARR in eight months; Cursor exited at $60B), robotics (Figure: 10,000+ deployments, $39B), defense (Anduril doubled to $61B in eleven months), voice (ElevenLabs past $500M ARR), healthcare AI (a quarter of Series B dollars). Notice the pattern: all of them sell completed work, not tools.

15. ~5,600 AI startups died in 24 months

The shakeout, by the numbers: a ~40% failure rate, mostly thin wrappers whose margins got eaten by the model providers. OpenAI killed its browser before its first birthday. Meta gave up on open source. Vibe-coding traffic collapsed (Lovable is down 40% from peak). And 89% of agent pilots never reach production. The tourist phase is over.

16. AI companies now reach $100M ARR in months — the newest record claim is six

MicroAGI, a Munich robotics-data startup, claims $100M annualized six months after founding (self-reported — haircut accordingly). Mercor did it in 11 months, AfterQuery in 14 — YC’s fastest unicorn ever, with founders aged 22 and 23. For reference, the last SaaS generation measured this milestone in years. Lovable’s 8-month record from 2025 already feels like ancient history, which tells you the half-life of these records.

17. Eight more companies that defined 2026

Cognition at $48B proves AI coding isn’t winner-take-all. Prime Intellect is the open training stack’s unicorn. Micro1 went 8x in a year with a 25-year-old founder. Reflection AI went 46x. Fireworks crossed $1B in inference revenue. Baseten signed OpenAI. Skild raised the biggest robotics AI round ever. Two years from now, some of these will be on the earlier slides — and some won’t exist.

18. The Cars Have Arrived

New York, around 1920: horse-drawn wagons and an early automobile sharing the same street. The horse still works. The car passes it anyway. Every enterprise is somewhere in this photo — the one running its fourteenth AI pilot is the wagon; the one with agents in production is the car.

19. 88% of organizations use AI — yet 95% of pilots show zero P&L impact

Both halves are true. Adoption is nearly universal, 31% of enterprises have an agent in production, and the companies that get it right see 171% ROI. But most pilots produce nothing measurable, and 56% of CEOs report no revenue gain and no cost savings from AI. The difference isn’t the model — everyone has the same models. It’s the last mile: workflow redesign, integration, ownership. That’s why Anthropic pays Forward Deployed Engineers up to $400K.

20. The IMF, the Bank of England and Michael Burry are all warning of an AI correction

The IMF named an AI correction as a global financial-stability risk this week. The Bank of England compares valuations to pre-dotcom. Burry calls it the final stretch of 1999. The circular financing is the real thing to watch — Nvidia invests in OpenAI, OpenAI buys Nvidia through Oracle and CoreWeave. The counter-argument: unlike 1999, the leaders actually have revenue. My read: not a dot-com bubble, but priced for perfection.

21. Twelve Seconds

The first flight lasted twelve seconds and covered 120 feet. Witnesses called it a toy. Twenty years later there were airlines. That’s where agents are today — impressive, useful in spots, nowhere near the destination. The prize was never the flight. It’s the airline.

22. Sell work, not software — the $10T+ prize is the services market

Sequoia and a16z published their 2026 theses and basically converged: services-as-software (sell outcomes, not seats), physical AI (bits deflate 10x a year; atoms are scarce), agent-native infrastructure (Stripe’s OpenRouter acquisition is the starting gun), energy and compute, and the last mile — whoever industrializes deployment turns the 95% failure rate into their pipeline.

23. The bottom line

In every gold rush, the winners sell shovels or strike the deepest vein. In 2026 the shovels are power and compute; the vein is completed work. The market stopped paying for narrative this year — it pays for revenue, outcomes and deployed systems. That’s the reckoning. It’s also the biggest opportunity I’ve seen in my career.

Enjoyed this? The full deck is at here:— share it freely.

State Of Ai 2026 Guillermo Flor

14.1MB ∙ PDF file

Download

Download

All data sourced on each slide; private-company figures are company-reported and flagged where self-reported.

The smartest money in tech has stopped asking "what AI company do I build?" and started asking "what cash-flowing business do I buy and rewire?"

The receipts, all from the last twelve months:

• Thrive Holdings raised $2B at a $12B valuation in August (SoftBank, D1, Altimeter), with OpenAI as a shareholder since December 2025 — embedding research and engineering teams directly inside acquired accounting firms and MSPs.

• General Catalyst has co-created an entire fleet of these: Crescendo (support), Eudia (legal), Titan (IT services), Long Lake (which just completed the $6.3B all-cash take-private of American Express Global Business Travel in September, calling itself "the world's first AI holding company").

• Dwelly raised $263M in under six months to buy UK letting agencies and run them on one AI platform — 10 acquisitions in 2026 alone.

• CNBC formally named the category on June 8, 2026: "AI rollup: Silicon Valley's new buyout playbook hits Wall Street."

This issue is the full tactical playbook:

1. The Opportunity: Why buying businesses and automating them with AI is a massive opportunity.

2. Who’s Doing It: How Thrive, General Catalyst, and others are leading the way.

3. What to Buy: The 11 best industries for AI-powered acquisitions.

4. How to Fund It: Three ways to finance your first acquisition.

5. The AI Stack: The best tools to automate operations and increase margins.

6. The Playbook: How to buy, automate, and scale a business step by step.

7. Risks & Returns: What could go wrong and how to maximize your upside.

8. Watch & Listen: The best interviews, podcasts, and resources to go deeper.

👋 Just a quick note before you continue!

If you are serious about building or investing in AI, you can subscribe here:

Here’s a good place to start:

↳Alex Karp’s Playbook for Which AI Companies Survive

↳19 great new details about Cursor’s rise

↳Perplexity’s CEO 2026 AI Pitch

↳The First 48 Hours of Claude Fable 5

Try premium free for 7 days. Or get 50% off this week only.

Get 50% off forever

Read more

Andreessen says there are over 1,000 state-level laws right now trying to kneecap American AI, and California nearly outlawed it outright six months ago.

That matters because in the same interview he lays out the clearest map he has given of where the trillions actually move next: out of dead manufacturing and into embodied robotics and the three regulated sectors where price never falls.

Here is the thesis in one line. Stop trying to win the software-app layer in the cities.

The asymmetric bets are embodied AI hardware (robots, drones, autonomous vehicles, defense) and AI attacking the three price-exploding regulated sectors (housing, healthcare, education) where the only moat is regulation, not technology.

In this issue you'll find:

1. Why the old manufacturing jobs are gone

2. Why the AI race is US versus China

3. Where is the real trillion-dollar turn

4. The American System: the 150-year-old industrialization playbook China is running right now

5. Why the knowledge economy ejected the middle class and left the countryside empty

6. The energy and debt constraints nobody is pricing into AI buildout

7. Washington's split personality is the biggest regulatory variable for AI founders

8. Why Andreessen bets AI is a disinflationary job creator, not a job killer

9. The precautionary principle is the single biggest threat to the AI buildout

10. Housing, healthcare, education: where AI founders should actually point their companies in the next 18 months

11. What the alignment-versus-risk fight tells you about who will try to control your AI company

This issue is available exclusively to paid subscribers!

If you’re not subscribed yet, you can unlock this and our full library by upgrading below:

Get 7 day free trial

Let’s begin!

1. The old manufacturing jobs are gone, and chasing them is the losing bet

Read more

When Elon Musk reads the training logs of new Grok models, he says the AI looks like someone answering questions with a gun to its head.

He said it on Saturday at X Takeover, in a 47-minute video interview in front of more than 2,000 people.

The rest of the hour was just as direct:

• SpaceX bought Cursor because only three companies had enough coding data to train a useful AI

• Musk wants to ship 10 chip designs a year, where the industry takes two years for one

• In his view, almost every impressive robot demo you see online is a canned routine

• Starship will try to catch its ship on the next flight, a few weeks from now

One word you will hear a lot: SI. A September 29 executive order told US federal agencies to call AI “Super Intelligence”. Within days Musk renamed SpaceXAI, the company behind Grok, to SpaceXSI.

High-signal analysis on markets, capital and the companies shaping the next decade

Inside you will find:

1. What Grok’s Training Logs Reveal About How AI Learns

2. Why SpaceX Bought Cursor: The Four Things Every AI Company Needs

3. Why AI Now Has Taste

4. Why Rockets and AI Belong Under One Roof

5. How to Ship 10 Chip Designs a Year

6. Why Most Robot Demos Are Fake

7. How Tesla Taught Its Robotaxis Not to Run Over Kittens

This issue is available exclusively to paid subscribers!

If you’re not subscribed yet, you can unlock this and our full library by upgrading below:

Get 7 day free trial

Let’s begin!

Read more

Jeff Bezos thinks AI will let many families get by working three days a week.

He said it on Fox News on October 7, sitting with Bret Baier on the Blue Origin factory floor in Cape Canaveral. And he is betting billions on it: his AI company, Prometheus, raised $12 billion in June at a $41 billion valuation, seven months after launching with $6.2 billion.

In the same interview he laid out where he thinks AI and space go next:

• Double-digit productivity gains across the economy

• A bubble that funds every experiment and leaves the inventions behind

• Data centers in orbit, carried there by rockets that fly every day

• Humans back on the Moon in 2028, ahead of China

High-signal analysis on markets, capital and the companies shaping the next decade

Inside you will find:

1. The $41B Company That Wants to Turn Anyone Into an Engineer See how Prometheus wants to take AI from the screen to the factory floor.

2. Why AI Could Make Everyone Richer Learn the plow-to-steam-engine argument behind his double-digit productivity bet.

3. Why You Might Work Only Three Days a Week His case for a labor shortage, tested against Amazon’s own 30,000 cuts.

4. Why an AI Bubble Would Be Good News for Founders Find out what biotech drugs and fiber optic cable say about the pop.

5. Why AI Should Be Regulated Like a Knife The rule he wants for AI, and where it puts the compliance burden for startups.

6. Why Your Next AI Model Could Run in Space Project Sunrise, 51,600 satellites and rockets that fly every day.

7. The Plan to Beat China to the Moon by 2028 Why the Moon comes before Mars, and the factories he wants to build there.

This issue is available exclusively to paid subscribers!

If you’re not subscribed yet, you can unlock this and our full library by upgrading below:

Get 7 day free trial

Let’s begin!

1. The $41B Company That Wants to Turn Anyone Into an Engineer

Read more

OpenAI just held back its newest model.

Astra 6.1 was ready to ship, and Sam Altman stopped it because it failed the company’s own safety bar.

It happened in the same weeks in which OpenAI:

• launched Dots, a personal agent Altman calls his favorite product ever

• paused a training run because capabilities were moving faster than its safety work

• started preparing for an IPO, with a nonprofit still in control

Where is it all going?

Altman answered in two long conversations with Vanity Fair’s Mark Guiducci, recorded three weeks apart in San Francisco. He went through what OpenAI will build next, what will slow it down, how AI will be regulated and where he expects the first big breakthroughs.

High-signal analysis on markets, capital and the companies shaping the next decade

Inside you will find:

1. Why Altman Says AGI Is Close See what OpenAI saw coming ten years ago and why the world still looks almost the same.

2. How Personal Agents Like Dots Will Run Your Work Find out how Altman uses his own agent and what OpenAI is building next.

3. Why OpenAI Is Slowing Down Its Most Capable Models Learn why Astra 6.1 was delayed and why safety testing now starts during training.

4. Why OpenAI Wants AI in Everyone’s Hands See the two ways Altman thinks AI can go wrong and the position OpenAI takes between them.

5. The Wave of Cyberattacks Coming From Open Models Find out why Altman expects severe incidents and what he means by AI resilience.

6. How AI Will Be Regulated Learn which rules Altman expects, from aviation-style incident reporting to US-China safety standards.

7. What Changes When OpenAI Goes Public See how a nonprofit will still govern the company, and why Altman regrets taking no equity.

8. Where AI Will Hit First: Science, Health and New Startups Find out which diseases and discoveries Altman expects within a few years, and the startup boom he sees coming.

This issue is available exclusively to paid subscribers!

If you’re not subscribed yet, you can unlock this and our full library by upgrading below:

Get 7 day free trial

Let’s begin!

1. Why Altman Says AGI Is Close

Read more

Last week, ElevenLabs doubled its valuation to $22 billion.

On paper, it was the startup most likely to be crushed. It makes AI voices, and every big lab can teach its models to talk.

Ben Horowitz uses it as his favorite example of how the AI market works. a16z co-led the Series A in 2023, when ElevenLabs was a year old, and has invested again in its later rounds.

His read on the rest of the market fits in five lines:

• AI is a new computer, and it is bigger than the internet

• the boom has pockets of excess, but the 1999 math doesn’t apply

• money can now buy a lead that used to take a decade to build

• customers can rebuild a lot of their software on their own

• the infrastructure under AI is wrong, and it will be rebuilt

Why it matters? Each of those lines changes what you should build, how you defend it and who you will compete with.

High-signal analysis on markets, capital and the companies shaping the next decade

Inside you will find:

1. Why AI Is Bigger Than the Internet See why Ben calls AI a new computer and what it lets a founder with domain knowledge build alone.

2. Why the AI Boom Isn’t 1999 (Yet) Learn where bubbles come from and where an excess of GPUs could still show up.

3. Why a Head Start No Longer Protects You Find out why money can now close a gap that used to take years of engineering.

4. How ElevenLabs Built a Moat the Big Labs Can’t Copy See the four steps that took a voice startup from vulnerable to $22 billion.

5. Why Your Customers Can Now Rebuild Your SaaS Learn which software customers rewrite for themselves and which they keep paying for.

6. Where to Build Next: The AI Infrastructure Rebuild Find out why Ben thinks everything from the chip to the model will change, and where that opens room for startups.

This issue is available exclusively to paid subscribers!

If you’re not subscribed yet, you can unlock this and our full library by upgrading below:

Get 7 day free trial

Let’s begin!

1. Why AI Is Bigger Than the Internet

Read more

Brian Chesky tried booking an Airbnb through Instinct and through Meta’s Muse. Both struggled, with hotels and even more with homes.

That is a strange thing to say in a week like this one:

• Instinct raised $1B at a $10B valuation, four times its price in August

• Muse has millions of downloads and sits at the top of the app stores

• Airbnb launched its own AI search, built without a chatbot

The problem?

A consumer agent is an assistant that acts for you: it books, buys, pays bills and cancels subscriptions. That works when the task is a single transaction. A trip on Airbnb means browsing, comparing homes, messaging a host and deciding with friends, and that is where today’s agents get lost.

High-signal analysis on markets, capital and the companies shaping the next decade

Inside you will find:

1. Why Chatbots Fail at Shopping and Travel Why Chesky still refuses to put Airbnb inside a chat box, and the interface he expects instead.

2. Why Airbnb’s Whole App Is Becoming an Agent How Chesky defines an agent, and what Airbnb’s fall release says about where apps are going.

3. Why Airbnb Still Breaks Inside Instinct and Muse Where today’s agents fail, and the four conditions Chesky says consumer AI needs.

4. Why Agents Need Their Own Operating System What Chesky told Sam Altman about the ChatGPT app store, and why he thinks Apple and Google will decide what comes next.

5. What Instinct’s $10B Backlash Says About the Race to Be the Main Agent Why the main-agent seat is fragile, and where the infrastructure money is going.

6. How Airbnb Ships 80% More Features With AI Why Chesky thinks token leaderboards were a mistake, and why AI is widening the gap inside teams.

7. Where Chesky Would Build If He Were 26 Again The categories and interfaces he would bet on today, and why he expects 50 to 100 more Anthropics.

This issue is available exclusively to paid subscribers!

If you’re not subscribed yet, you can unlock this and our full library by upgrading below:

Get 7 day free trial

Let’s begin!

1. Why Chatbots Fail at Shopping and Travel

Chesky has been saying the same thing for years, long before Instinct existed: a chatbot is the wrong interface for e-commerce. When ChatGPT took off, Airbnb was one of the few large consumer companies that refused to rebuild its app around a chat box, and for a while that looked like being slow.

Read more

More than 20 launches in a single keynote: always-on agents, a new workspace inside ChatGPT, cheaper frontier-level models, new infrastructure for developers, and new ways for startups to reach ChatGPT’s 1.2 billion users.

Some of these announcements could materially change how startups build, distribute and monetize AI products.

🔎 Here’s the full breakdown, with clips from the keynote for every major announcement.

High-signal analysis on markets, capital and the companies shaping the next decade

Inside you will find:

1. Dots Are OpenAI’s Bet on Agents That Own Work Full-Time

2. ChatGPT Is Coming for Notion, PowerPoint and Your Meeting Notes App

3. Near-Frontier Intelligence Just Got 80% Cheaper

4. OpenAI Is Handing Developers the Agent Infrastructure It Uses Itself

5. The New Way to Reach 1.2 Billion ChatGPT Users

6. What OpenAI’s AI Research Intern Says About the Next Twelve Months

The rest of the issue is available exclusively to paid subscribers!

If you’re not subscribed yet, you can unlock this and our full library by upgrading below:

Get 7 day free trial

Let’s begin!

1. Dots Are OpenAI’s Bet on Agents That Own Work Full-Time

Read more

It is the closest thing you will get to hearing how Sequoia talks about AI when it is speaking to the people whose money it manages.

Grady says plainly that AGI arrived last November, walks through what the labs actually believe right now, lists the good, the bad and the ugly of what is coming for everyone else, and then explains why the real money for startups sits in the space between what the models can do and what companies actually use them for.

Inside you will find:

1. The Three Things That Make This Wave Different

2. Why Sequoia Says AGI Arrived in November 2025

3. The Good, the Bad and the Ugly of AI on Main Street

4. What the AI Labs Actually Believe Right Now

5. How the Labs Are Fighting for Every Token

6. The Diffusion Gap Where Startups Make Their Money

7. The Four Places Sequoia Wants Startups to Build

8. How the Best AI Startups Now Run Their Companies

9. The Two-Step Deal Behind Sequoia’s 110 to 3,400 Jump

This issue is available exclusively to paid subscribers!

If you’re not subscribed yet, you can unlock this and our full library by upgrading below:

Get 7 day free trial

Let’s begin!

1. The Three Things That Make This Wave Different

Read more

Ask six of the biggest LLMs point blank how likely AI is to wipe us out, and they all land somewhere between 1% and 10%.

Peter Thiel agrees the risk is real. And he still thinks we should accelerate.

In a 65-minute conversation with Mathias Döpfner, CEO of German publisher Axel Springer, Thiel explains why. Along the way, he connects AI risk to a much broader argument about technological stagnation, Europe’s decline, the return of Weimar-style political extremism, and what he sees as a 50% chance of a democratic socialist winning the White House in 2028.

We went through the entire conversation and pulled out the most important arguments, predictions, and moments.

Inside you will find:

1. The Stagnation Thesis That AI Finally Proved

2. Why Slowing AI Loses Even at 10% Extinction Risk

3. How China Turns Western AI Fear Into an Advantage

4. Why Europe Is Falling Further Behind Than Minitel

5. The Competence Crisis at the Top of the West

6. His 2009 Democracy Quote and the Weimar Threshold

7. Why He Gives a Democratic Socialist 50% Odds in 2028

8. Why Argentina’s Past Could Be the West’s Future

9. The Founder Scoreboard That Should Worry Germany

10. Health, the Only Kind of Progress With No Downside

This issue is available exclusively to paid subscribers!

If you’re not subscribed yet, you can unlock this and our full library by upgrading below:

Get 7 day free trial

Let’s begin!

1. The Stagnation Thesis That AI Finally Proved

Read more

For the first time in history, the people building frontier AI briefed the UN Security Council directly, and one of them called it "the most important global security issue facing the world today."

I went through all 21 minutes of Sam Altman, Dario Amodei and Clement Delangue's interventions so you don't have to, and pulled out the exact point where their united "be careful" message splits into two opposing bets about who should control the technology.

That split is the signal. It tells you whether the money flows to a few audited frontier labs or to the open, cheaper models the defenders actually reached for.

Every headline reads "CEOs warn UN about AI risk."

The real story is a live disagreement about who gets to hold the technology, and it maps directly onto the concentrate-versus-distribute bet every builder is already making, whether they've named it or not.

In this issue you'll find:

1. Renaissance or industrial revolution: Altman frames the whole fight as a choice, not a forecast

2. The two failure modes that keep the labs up at night: losing control and concentrating power

3. "We are not the heroes of the story": Altman's three principles and the pledge to slow down

4. The standards play: why AI decisions can't be made 'by labs in San Francisco alone'

5. Amodei's timeline bomb: a 'country of geniuses in a data center' one to two years out

6. Amodei's ask: put AI at the Security Council and pace the frontier like a food inspector

7. Delangue breaks ranks: the real attack where closed APIs blocked the defenders

8. The asymmetry thesis is the investable signal: distribute capability, don't concentrate it

PS: Here’s the full inverventions

Hey, welcome to AI Market Fit!

Every week you’ll get two things, fast:

What’s happening now (and where the money’s headed).

Who’s already building it (so you can invest, partner, or steal their playbook).

However, most of the content from this newsletter will be only accessible to the paid subscribers.

If you are serious about building or investing in AI, you can subscribe here:

If you want to sponsor The AI Opportunity or Product Market Fit email me to g@guillermoflor.com

Read more