AGI to AGI: How Artificial General Intelligence Became Advertising Generated Income

Part 1 of 3: The collision between AI, the U.S. consumer economy, and the battle for attention. Part 2: the attention hyperinflation crisis. Part 3: what operators should actually solve & build.

Three waves. Same pattern.

Web. Mobile. AI.

Each one promised to change everything. Each one did but not the way the headlines said.

The pattern is always the same: new technology creates new surfaces for attention, attention gets monetized through advertising, and advertising funds the next cycle of innovation. The American economic engine doesn’t run on technology. It runs on consumer attention. Technology just builds better capture mechanisms.

Today, we’re watching that pattern complete its fastest cycle yet. The industry promised Artificial General Intelligence. What it’s actually delivering?

Advertising Generated Income. :)

AGI to AGI.

The $19 Trillion Truth Nobody Wants to Talk About

Let’s start with a number that should embarrass every “AI is the new economy” take you’ve read this year.

U.S. GDP is nearly 70% personal consumption expenditure hovering between 67% and 69% depending on the quarter. That ratio hasn’t meaningfully changed in over 30 years. Not during the internet boom. Not during mobile. Not during cloud. Not during AI.

In January 2026, MRB Partners published research showing that consumption not AI investment was the #1 driver of U.S. GDP growth in 2025. Adjusted for imports, AI-related capital expenditure contributed just 40 to 50 basis points to real GDP — roughly 20–25% of total growth. Consumer spending drove the rest.

MRB’s strategist Prajakta Bhide told CNBC: “AI is an important part of the growth story, but it’s not the only part of the growth story. That’s a narrative that’s out there, that if we didn’t have the AI capex, GDP would have slumped last year. And that’s simply not true.”

Meanwhile, Alphabet, Amazon, Meta, and Microsoft are planning roughly $650 billion in combined AI capital expenditure for 2026 a 60% jump from last year’s $410 billion and a 165% increase from 2024’s $245 billion.

But where does the return on $650 billion come from? Consumer engagement. Ad efficiency. Attention monetization. The same place it always comes from.

Every Tech Wave Builds a New Attention Surface

Every platform wave creates what I think of as a new attention surface a front door where consumers discover, browse, compare, and buy. A surface is anywhere intent gets expressed and harvested, whether or not it looks like a traditional ad.

The web created the first digital attention surfaces at scale. Search turned a blinking cursor into the most powerful intent surface ever built Google sold that intent for $265 billion a year. Social turned the feed into a surface. Facebook’s 2012 mobile pivot saved the company and built Meta’s $160B+ annual ad business. The web era proved the formula: build a surface where consumers express intent, monetize it through advertising.

Mobile multiplied those surfaces and made them personal. The home screen, the notification tray, the app store all became attention auctions. Streaming turned the recommendation algorithm into a surface: Netflix, Spotify, TikTok monetized through subscription or ads. The “For You” page is an attention surface. So is autoplay. Mobile didn’t replace the web. It made attention portable, persistent, and vastly more valuable.

AI is creating the newest surfaces. Chat interfaces, copilots, and agents. “Find me a hotel.” “Compare these products.” “Draft an outreach plan.” These are intent-rich, conversational surfaces where consumers express needs in natural language. They don’t look like traditional ad units. They won’t stay ad-free.

The pattern never breaks: new tech → new attention surface → intent harvesting → advertising monetization → revenue funds the next tech cycle.

Every row is the same story: a new place where consumers express intent, and a new way to monetize it.

The Singularity Was a Fundraising Deck

Now let’s talk about the narrative that was supposed to make this cycle different.

Artificial General Intelligence. The singularity. Existential risk. “We’re building God.”

Andrej Karpathy has argued we’re at least a decade from real general intelligence and that current agents are often clumsy “slop” that don’t match the marketing. He’s right. But it doesn’t matter for the economic argument. Even narrow, commoditized AI is already powerful enough to reshape monetization and create new attention surfaces.

The hype, though the hype was extraordinary.

OpenAI marketed GPT-5 as near-PhD-level intelligence. It launched in August 2025. The world shrugged. MIT Technology Review called 2025 “The Great AI Hype Correction.” OpenAI co-founder Ilya Sutskever now running his own lab says LLMs “generalize dramatically worse than people.”

And yet the money kept flowing. Safe Superintelligence: $2 billion at a $32 billion valuation. No product. Thinking Machines Lab — Mira Murati’s new venture — $2 billion raised at a $10 billion valuation. No product, no revenue. The largest seed round in Silicon Valley history, for a six-month-old company.

Sequoia’s David Cahn ran the math: the AI industry needs $600 billion in annual revenue to justify current investment. OpenAI’s actual revenue? $20B+ ARR as of late 2025 still a fraction of the gap. The math doesn’t math.

Unless you add advertising.

Sam Lessin said it on More or Less: “People want reasons to have a zero-sum race because they know what the fun part of zero-sum races are... the narrative is the product.” Eric Schmidt coined “the San Francisco consensus” the belief in a winner-take-all AGI race. It’s not a technology forecast. It’s a marketing position.

We over-rotated from “AGI soon” to “AI is overhyped.” Both miss the more mundane, tectonic shift: who captures the incremental attention surfaces AI is unlocking?

The Ad Pivot Was Survival, Not Strategy

Here’s what makes the AGI-to-AGI reframe feel inevitable rather than just clever.

Classical SaaS economics are under siege. Model costs are falling. Competition is driving per-unit AI prices toward zero. Usage-based and outcome-based pricing are eroding the “seat-based” margin stacking that funded the last decade of software. One founder recently described rebuilding a $200/user/month SaaS product in a single day with AI — better than the original, customized to his business. When software can be built at near-zero marginal cost, not just scaled at near-zero marginal cost, the moats that funded the last decade of software are gone. Durable profit pools migrate to where you can tax transactions, attention, or outcomes.

That means advertising. Or something that looks a lot like it.

OpenAI made it official in January 2026. Their blog post stated: “Our pursuit of advertising is always in support of [our AGI] mission.” Ads at the bottom of ChatGPT responses. A new ad-supported tier ChatGPT Go at roughly $8/month.

But the real tell wasn’t the announcement. It was the hire.

OpenAI brought in Fidji Simo the architect of Facebook’s mobile advertising machine in the 2010s as CEO of Applications. Twenty percent of OpenAI’s workforce now has Meta or Facebook on their LinkedIn profiles. Axios called it “OpenAI’s Meta Makeover.”

You don’t hire the executive who built Facebook’s ad juggernaut because you’re focused on the singularity.

Google’s entire AI roadmap Performance Max, automated targeting, AI Overviews with sponsored results is ad efficiency. Meta’s massive Hyperion data center part of its planned $115–$135 billion in 2026 capex is funded by and built for ad revenue optimization.

The IAB’s 2026 Outlook projects U.S. ad spend growing 9.5% this year. Digital: social +14.6%, CTV +13.8%, commerce media +12.1%. Five of six top advertiser priorities are now AI-driven. Two-thirds of advertisers are focused on agentic AI for ad buying and campaign execution.

The new attention surfaces are already being wired for monetization. This wasn’t a strategic choice. It was structurally inevitable.

And if you think this is just an AI company story, look at the grocery aisle.

Retail media networks are the second-order proof. Walmart’s ad business now accounts for nearly a third of its $6.7 billion operating income. Walmart Connect grew 27% year-over-year. Amazon Ads is expected to hit $56 billion in 2025 — commanding roughly 75% of the U.S. retail media market. Globally, retail media hit $140 billion in 2024 and is projected to reach $165 billion by 2026, growing at 5× the rate of the overall ad market. McKinsey projects it will surpass television ad spend by 2028.

The margins tell the real story: retail media networks generate 60–70% profit margins compared to traditional retail’s 5–10%. Retailers realized their product pages, checkout screens, and loyalty data are attention surfaces worth more as ad inventory than as commerce infrastructure. There are now 277 retail media networks globally. Chase Bank, United Airlines, Saks, and Wawa have all launched ad networks. When a convenience store chain launches something called “Goose Media Network,” the thesis isn’t theoretical anymore.

OpenAI adding ads to ChatGPT isn’t the anomaly. Walmart making more margin on advertising than on selling you groceries that’s the pattern. Every company with an attention surface eventually becomes an advertising company. AI just accelerates how fast they figure it out.

---

The Hype Machine Is the Product

One final irony.

Every major AI model drop o1, o3, GPT-5, Sora, Gemini is staged for maximum media coverage. Benchmarks are designed to go viral: “96.7% on AIME” isn’t information for users. It’s content for headlines.

Product launches follow entertainment industry logic. Teasers. Countdowns. Embargoed reviews. OpenAI’s “12 Days of OpenAI” was literally an advent calendar marketing format.

AI companies have become media companies. Their primary output isn’t intelligence. It’s attention.

And the AGI narrative itself? That’s an attention surface too. Every “we’re close to AGI” announcement generates clicks, coverage, investor confidence, and valuation bumps which is advertising generated income in its purest form.

The singularity is a content strategy.

So What Does This Actually Mean for You?

The macro reframe is fun. But the pattern above isn’t abstract it’s already reshaping the P&L of every company that sells to, markets to, or competes for consumers. If you’re running a company, investing in one, or leading a marketing org, here’s why this matters on Monday morning:

If you’re a CEO or founder stop benchmarking against SaaS multiples that assume 80% gross margins. AI is compressing those margins toward commodity levels. Your real competitive question isn’t “how do we ship features faster?” it’s “which attention surfaces do we own or access, and how do we monetize them?” Distribution, data, and demand capture are the only durable moats left. If your business doesn’t have a clear answer to “where does our attention come from and how does it compound?” you have a strategy problem, not an AI problem. Ask yourself: if Google changes its AI Overview format tomorrow, or if Meta shifts its algorithm again, does your revenue model survive? The companies that own their attention surfaces their own data, their own customer relationships, their own distribution don’t have to ask.

If you’re an investor the $600B revenue gap between AI investment and AI returns will close through advertising economics, not SaaS subscriptions. The next wave of value creation won’t look like the last one. Watch for companies that control attention surfaces (not just build models), that have proprietary demand data (not just training data), and that can monetize outcomes (not just seats). The companies still pitching “we’re building AGI” without a monetization path are the ones that will need another round or an acquirer. The tell? Look at unit economics. Retail media networks generate 60–70% margins. Most AI startups are still burning cash to subsidize usage. The money will follow the margins, not the narratives.

If you’re a CMO or marketing leader this is the most directly urgent. Customer acquisition costs have risen roughly 40% between 2023 and 2026. Across industries, CAC increased 60% over the last five years. SaaS companies are spending $2.82 to acquire $1 of new ARR. And it’s about to get worse because every company whose software margins are compressing is now entering your attention auction. More bidders, same pool of eyeballs, every surface getting noisier. The brands that win won’t be the ones that spend the most on attention. They’ll be the ones that learn the most from every dollar of attention they buy and compound that learning into lower CAC and higher LTV every quarter. Here’s a concrete test: can your team answer “We spent $500K on this audience last quarter what did we learn, and what changed this quarter as a result?” If that requires stitching dashboards and chasing the analyst who left, you have a decision problem, not a data problem.

If you’re a data or ops leader you’re sitting on the most strategically important function in the company and most people don’t know it yet. The shift from SaaS economics to attention economics means that decision infrastructure the ability to capture what was decided, why, what happened, and what to change becomes the core operational advantage. Not dashboards. Not more agents. The ability to learn systematically from attention spend and compound that into better outcomes. That’s your mandate. Translation: every campaign that runs without a closed feedback loop what was decided, what happened, what to change is decision debt. And AI agents are about to accelerate that debt at machine speed.

Three Waves. Same Destination.

Whoever captures consumer attention, wins. The rest is theater.

AGI - Artificial General Intelligence might arrive someday. But AGI Advertising Generated Income is already here. It’s funding the data centers, the model training, the $650 billion in capex, and every breathless announcement about the future of intelligence.

The smartest builders are quietly rediscovering that they’re in the media and advertising business whether they admit it or not. Because in a world where AI is compressing every software margin, the only durable moat is distribution, data, and demand capture. The new attention surfaces are being built right now. The question is who monetizes them.

Coming Up in Part 2: “The Noisiest Economy in History”

If AI is building new attention surfaces faster than ever and SaaS margin compression is pushing more companies into the attention auction what happens when everyone fights for the same fixed pool of human eyeballs, but AI agents are the ones doing the bidding autonomously, at machine speed, with your budget?

Attention inflation. A trust collapse. And a structural shift in who wins and who wastes money.

Part 2 drops next week.

Thanks for reading Data to Decision Activation. If this reframe landed, share it with someone still chasing the singularity. And subscribe if you want the rest of the series it gets more actionable from here.

Author's note: An LLM helped with research, citations, and light proofreading without changing content. All ideas, arguments, voice, and em dashes are mine. :)