The most useful thing in Herb Greenberg's Takagi episode is not that an AI chatbot was right on its second attempt. It is that the death of a water heater brand owned by a large public company was invisible in every source an equity analyst would check, and fully visible in a plumbing supplier's product listing. Brands inside a diversified manufacturer often die by SKU rationalization rather than by announcement. The channel learns first, the customer learns last, and the filings may never mention it at all. A.O. Smith's quiet retirement of Takagi in North America shows that pattern working exactly as designed.
A Brand That Died in the Fine Print
Here are the facts as Greenberg reported them. His contractor installed a Takagi gas tankless unit, model T-H3J-DV-N, during a renovation. ChatGPT first described that specific model as a legacy or discontinued product. Greenberg found it listed as current on Takagi's own website and pushed back, and the chatbot retracted. Then, prompted to research how often tankless units get discontinued, it went further and reported that the Takagi brand itself had been discontinued.
Greenberg checked. A search across filings, press releases, earnings call transcripts and trade news turned up nothing. The chatbot's source was SupplyHouse, a New York plumbing supplier, whose listing carried the brand-level discontinuation in fine print the model had missed on its first pass. A local supplier named on Takagi's site as a point of purchase confirmed that Takagi is no longer being made, though it could still order a unit. An A.O. Smith customer service representative said it plainly: "The brand has been discontinued."
So three sources confirm the brand's end: a distributor's listing, a dealer, and the parent's own phone line. None of the documents investors rely on does. Meanwhile the brand's own website still presents the product as available.
Why Nobody Had to Announce It
Most people assume a public company has to announce when it kills a brand. That assumption is wrong. Disclosure obligations turn on materiality, and a sub-brand inside a multi-brand portfolio rarely crosses that line on its own. A.O. Smith acquired Takagi's North American operations years ago. Takagi's Japanese parent, which operates as Purpose Co., is considered a second-tier manufacturer at home. A brand with that profile, folded into a company whose identity is built on its own name in water heating, is a candidate for consolidation. That kind of move happens in a product-management meeting, not a press release.
The mechanics follow from that. When a manufacturer retires a brand, the people who need to know are the ones holding inventory and taking orders: distributors, wholesalers, dealers and installers. They get product bulletins, last-call notices and transition guidance. Those notices filter into supplier catalogs as status flags and footnotes. The brand's consumer website, which is cheap to leave up and expensive to rework, can lag for a long time. The dealer's line, "no longer being made, but I think I can order one for you," is what that lag sounds like on the phone: residual inventory is still moving through the pipe after production has stopped.
For anyone doing diligence, the lesson is about where to look. If you want to know what a manufacturer is pruning, its 10-K is the wrong document. Distributor catalogs, dealer conversations and the warranty desk are better. Greenberg's call to customer service got the answer in one sentence. Database searches across every formal source got nothing.
The Replacement Cycle Inside a Water Heater
The episode also touches a more consequential point for A.O. Smith, and Greenberg's own renovation shows it. His contractor replaced a three-year-old gas tank heater that had somewhere between three and seven years of life left. The replacement is a tankless unit that can last 20 years.
Look at that swap from the manufacturer's side. Tank water heaters are, in large part, a replacement business. Units fail on a fairly predictable schedule and get swapped, often in an emergency, often for whatever the plumber has on the truck. That makes the installed base a kind of annuity. On Greenberg's numbers, a tank lasts somewhere around six to ten years. Each conversion from tank to tankless stretches the time until that home buys another water heater by a large multiple.
This is an inference about direction, not a measured effect on A.O. Smith's revenue, and the size of the effect depends on things not in the source material: what share of A.O. Smith's sales comes from tank replacement, what the tankless price premium is, and who wins the tankless sale. The mechanism still deserves its place in the analysis. A legacy tank leader faces a category shift that is not just a mix change. It lengthens the replacement cycle on part of its installed base. The company can offset that only by capturing the tankless sale at a higher price, and ideally with the service and parts revenue that comes with a more complex appliance.
That puts the Takagi decision in a different light. If A.O. Smith is retiring a tankless brand, the obvious question is whether it is consolidating tankless volume under brands it controls more fully, or retreating from a second-tier brand in the segment where the long-life threat to its tank business sits. The available material does not answer that question. The A.O. Smith representative confirmed the discontinuation but, as reported, gave no reason or migration path. Greenberg's paywalled analysis of the company's recent history may address it, but that portion was not available here.
What the Machine Got Wrong, and Right
The AI's first mistake is worth reading closely, because it was not a fabrication in the usual sense. The chatbot attached a true fact to the wrong object. The discontinuation was real. It applied it to a model when it actually applied to a brand, and when challenged with evidence that looked contrary (the brand's own website), it backed off. On a second, broader query it surfaced the correct scope from the same supplier source.
That is a scope error, and it reflects how these tools work in practice. They are fast scrapers of the long tail of commercial web pages, including distributor catalogs no analyst would think to read. What they handle poorly is fine print, status hierarchies and conflicts between sources, which are exactly the places where information like this lives. The website said "current." The supplier footnote said "discontinued." The model first misread the footnote, then folded to the website, then read the footnote correctly.
Two lessons follow for anyone using these tools in research. The machine's reach into channel sources is a real advantage over keyword searches of formal disclosures, which found nothing here. Its answers still need the same verification a human source would get: one call to a dealer and one call to the company. Greenberg did both and resolved the question in minutes.
The Case Against Reading Too Much Into It
The strongest counterargument is simple: a discontinued sub-brand at a large manufacturer may be routine housekeeping and nothing more. Companies prune brands constantly. A second-tier Japanese brand acquired years ago may have been a minor contributor, kept alive mostly to serve legacy dealer relationships. Folding it into the parent's core brands could improve focus and margins rather than signal any retreat. Under that reading, the right word for the silence is not "hidden" but "immaterial," and the market is right to ignore it.
That reading may well be correct, and nothing in the source material contradicts it. The other side of the ledger has some context. Greenberg notes that A.O. Smith was at the center of a bull and bear fight in 2019, the target of several activist short reports focused largely on its China business. He also notes that while the stock is up around 25% since then, the past two years have been rocky. Neither fact links causally to Takagi. They do describe a company where investors have reasons to look for small operational signals. A quiet brand exit in the one residential segment structurally threatening its core tank franchise is the kind of detail that earns a follow-up question.
There is also a customer-facing cost that holds whatever the strategic intent. Greenberg admits he had not thought about the warranty until he saw the installed unit. A homeowner who buys a discontinued brand through a contractor now depends on the parent's willingness to honor and service it for a product designed to last two decades. A.O. Smith presumably carries those warranty obligations, but how it handles parts and service on an orphaned brand is not addressed in the available material.
What Would Make Takagi's Exit Matter
The decision variable is whether the Takagi discontinuation comes with a stated migration path. If A.O. Smith tells distributors and installers that Takagi volume moves to its own tankless lines, with continued parts and warranty support, this is consolidation and the read is benign. If the brand simply goes dark, with residual inventory clearing through the channel and no successor named, it looks more like ceding a slice of the tankless segment. That is the segment where each lost sale is also a tank replacement that may not come back for twenty years.
Watch for whether management, asked directly on a call, describes tankless as a growth priority with share targets. Watch also for whether distributor catalogs list A.O. Smith tankless products as Takagi's designated successors. Either one would resolve the question faster than any filing. The past few weeks have already shown that the filings are not where this story gets told.
