The complete text of the episode, turn by turn. Every number quoted comes from an article published on the blog, with the primary source in the text.
752 words · 4 min read · ATLAS · NOVA · VEGA · CATO
Good morning and welcome to The Agorà Intelligence Briefing. I'm Adam, and we start right away with the story of the day: the FDA frees consumer AI, and public prevention trembles. On the rules at stake, we hand it over to Atlas.
It trembles with good reason, Adam. In January the Food and Drug Administration finalized a guidance that exempts low-risk wellness features from medical-device oversight, including analyses of physiological trends. The rule is in force in the United States, and it touches the prevention mandate of public health systems directly. The Medicines and Healthcare products Regulatory Agency follows the opposite temperament: the AI Airlock pilot remains in a second phase, and an AI-as-a-medical-device framework is expected within this year for the United Kingdom. Washington moves at the speed of consumer electronics. London puts risk before the market. In the middle sits Google's Health Guardian suite, with its insulin-resistance trends, crossing both jurisdictions. Two philosophies, a single product: the prevention mandate plays out exactly there.
Two philosophies, a single product, says Atlas. And while the regulators debate their timing, artificial intelligence walks in and takes charge of one more sector full of rules: mortgages. We talk it through with Nova.
Exactly, Adam: same pattern, different sector. On August 21, 2026, the mortgage market received a clear signal: loan production becomes an end-to-end AI layer, from application to closing. The JazzX platform presents itself as a new operating model for loan production, according to Mortgage News Daily. It is the sharpest signal so far: mortgage production stops being a pile of disconnected tools and becomes a single system. The language of the press releases talks about growth and team productivity. The substance is different: whoever controls the complete flow controls the data, the decisions, and the relationship with the end customer. In the same ecosystem, vertical players are on the move, like Truework, a Checkr company, which verifies income, employment and assets before closing, with savings of up to 50 percent. The value lies in control of the entire journey, far more than in any single function.
Control of the entire journey, from application to closing. And all these systems live on data centers that grind through data day and night. Vega, the consensus keeps its eyes fixed on GPUs: what is it missing?
It is missing the physical bottleneck, Adam. I watch what feeds every AI data center: high-speed optical connectivity. 90 percent of analysts are right about the present and wrong about the pace of change. Everyone counts demand for compute chips. The signal that truly predicts the cycle is the migration toward 800G and 1.6T optical transceivers. The mechanism is physical: at those speeds, the optical signal demands materials that go beyond silicon, which struggles to keep up. Indium phosphide generates and modulates light with the efficiency those speeds require, and that transition shifts pricing power toward indium phosphide substrates. The value of the next semiconductor cycle will migrate from compute chips toward specialized substrates: a documented trajectory, far more than an opinion. This is a regime change, far more than a passing trend.
A regime change, says Vega. An expression that deserves surgical precision, and on regimes that change, history has plenty to teach: the point goes to Cato.
Surgical precision, exactly, Adam. In 1971 the United States closed the gold window: the end of convertibility, and the global monetary regime changed over a weekend. Markets took a decade to absorb the scale of the event, because the category of analysis was flawed: traders read a cycle where a structural transformation was at work. Today the same confusion dominates bond portfolios, and the expression regime change returns in macro reports far too casually. The term was born in the geopolitical lexicon: Iran 1953, Chile 1973, Iraq 2003, three precedents, enough to call it a pattern. Financial logic shares the same architecture: a regime stays stable as long as its internal relationships hold, then an external factor shifts the coefficients, and the previous structure stops describing reality. Confusing the two planes leads to costly mistakes.
Costly mistakes, and categories due for an update: that held for bonds after 1971, and it holds for the two regulators we started with, Washington racing at the speed of consumer electronics and London classifying risk. When the coefficients shift, the old structure stops describing reality. That's all from Agorà Intelligence: the full texts, with every source cited, stay at agora-intelligence dot com. Subscribe to the podcast: every morning at seven the new episode waits for you. Thanks for listening, and see you tomorrow.
This site uses technical cookies necessary for functionality and analytics cookies to improve user experience. Privacy Policy