An economics control layer that Finance can use to govern AI cost and revenue Al is the first spend category where execution is implicitly authorized and billed after the fact. What follows makes Al behave like every other budgeted expense - approved before money is committed. * Pre-execution provisioning saves money:
AI is not ready for business. That isn’t an opinion. It’s what the data, the vendors and real-world trials now show: * MIT researchers report that roughly 95 percent of generative AI pilots inside companies fail to move beyond experimentation. * A recent McKinsey survey found that nearly 80 percent
Despite endless comparisons to Amazon, Google, and Facebook, AI is not like monetizing eCommerce, Search, or Social – because unit economics improve with scale for them. But not for AI. Rubber duckies are cheap to make one at a time, and cheaper to make a million at a time. Amazon loses
Last week, The Information ran a long, unusually candid piece about Anthropic’s new “agentic” tool, Cowork. It’s one of the rare stories that doesn’t just repeat the promise — it documents the experience. And the experience matters. The reporter tried to use Cowork for two very normal tasks:
Unlike ecommerce, search, and social, AI doesn’t get cheaper with scale. It gets more expensive. More users means more compute, and more compute means more cost. So margins don’t expand unless compute per request falls. This earnings season will expose a simple divide: who can reduce compute cost
1. AI companies talk endlessly about revenue growth — is AI cost rising faster than AI revenue? 2. If an AI system rehydrates context on every turn and unit cost rises with usage, how is that scaling rather than a cost curve pointing the wrong way? 3. If token pricing is
* Microsoft revealed it has 15 million paying users of Microsoft 365 Copilot, priced at $30 per user per month. On its face, that’s several billion dollars a year in high-margin software revenue. * But Microsoft also has more than 400 million Office 365 users worldwide. Do the math: 15 million
By Alan Jacobson, Systems Architect & Analyst Can LLMs “scale” their way to profitability by charging a flat rate of $20/month for unlimited compute? * If every man, woman and child in America pays $20/month = $81.6B/year * Capex depreciation: $50B/year * Staff: $11.25B/year * Cost of electricity
By Alan Jacobson, Systems Architect & Analyst The first wave of AI adoption is over. Not because the technology failed, but because users hit its limits faster than the industry expected. Models forgot. Systems hallucinated. Outputs couldn’t be trusted. And when reliability drops, adoption at scale is imp. That
Many assume AI will follow the path of ecommerce, search, and social — lose money for years, achieve massive scale, then discover profitability. This will not work for AI. Advertising only works when content is free * Ecommerce doesn’t pay for its content. Sellers provide it for free. * Search doesn’t
Two major 2025 studies reach opposite conclusions about enterprise AI. * Wharton says AI is succeeding. * MIT says it isn’t. They’re both right — but only one of them answers the question that matters. Ironically, the role of each institution flipped: * Wharton, a business school, framed AI success in terms
By Alan Jacobson, Systems Architect & Analyst OpenAI’s prototype ads look harmless enough. They sit politely near answers. They look like ads – sort of. They promise relevance without surveillance. Nothing screams “feed takeover.” That’s the problem. They don’t look like they’ll make any money—because they
By Alan Jacobson, Systems Architect & Analyst The AI industry is optimizing the wrong scoreboard, and the consequences are starting to show up in public. The industry keeps celebrating the wrong wins For the past two years, AI “progress” has been defined almost entirely by research metrics: * benchmark scores * capability
By Alan Jacobson, Systems Architect & Analyst OpenAI’s decision to introduce advertising inside ChatGPT isn’t a bold new growth chapter. It’s a pressure move. The structure of the rollout — CPM pricing, no self-service console, vague targeting, small trial commitments — looks less like the launch of a serious
By Alan Jacobson, Systems Architect & Analyst Many people assume AI will follow the same path as Amazon, Facebook, and Google — lose money for years, achieve massive scale, and then discover profitability on the other side. That assumption is wrong. More than 95% of AI users pay nothing. And why
By Alan Jacobson, Systems Architect & Analyst * Subscriptions cannot deliver ROI: More than 95% of users do not pay, do not convert to paid and are unlikely to pay for something they get now for free. * Advertising cannot deliver ROI: Unlike social platforms, AI cost does not decline per user
By Alan Jacobson, Systems Architect & Analyst Definition. Governed, pre-execution provisioning (G-PEP) optimizes LLMs to reduce inference cost and capacity requirements up to 90 percent. It is an architectural control layer that authorizes, constrains, routes, defers, or denies AI execution before inference based on policy, entitlement, and cost constraints. Unlike
By Alan Jacobson, Systems Architect & Analyst * LLMs do not disclose active user growth or retention. (Analysis of each company suggests why, below) * Independent reporting by five data sources on web traffic and app downloads indicates adoption flattened months ago. * Chip sales are reported, but end-user adoption is not. * Media
By Alan Jacobson, Systems Architect & Analyst Before AI can be adopted, it must work as a business. Today it doesn’t. The economics are broken on both sides of the ledger: revenue is mispriced and costs are uncontrolled. While chipmakers may prop up the NASDAQ by selling infra to
By Alan Jacobson, Systems Architect & Analyst The railroad crash of 1873 corrected capital excess without changing the product. The AI crash will correct capital excess by forcing the product itself to change. Why the 1873 railroad analogy is attractive — and incomplete The surface analogy is strong: * massive capital inflows
By Alan Jacobson, Systems Architect & Analyst There are three markets for AI. They are routinely conflated, incorrectly marketed as one, and that mistake explains why AI adoption, trust and monetization are breaking down at the same time. 1) Consumers: the mass market Consumers don’t want to think better.