The Watson Weekly is sponsored by Avalara — the agentic AI platform automating global tax and compliance for leading eCommerce brands with simple plans, transparent pricing, and complete coverage across Shopify, Stripe, WooCommerce, BigCommerce, and more.

Public software companies shed something in the neighborhood of $2T in valuation between January and February of this year. The explanation that stuck was that AI writes code now, so nobody needs to buy software.

That is the wrong thing to be afraid of.

The evidence people pointed at was real. Salesforce and Atlassian both reported per-seat licenses declining, which by their own telling had not happened before in their history. Seats going backward at companies built on selling seats is a genuine signal. But it is a signal about how software gets billed, not about whether it gets bought. Inside of a year most of this looks like a utility bill. Usage, updates shipped, outcomes delivered. The revenue finds a path back. The multiples follow it.

What does not come back on its own is the ability to sell anything to anybody.

Roughly half of consumers are now using generative AI. Around 18 percent say they trust the companies providing it. I have not seen a technology run that far ahead of its own credibility before. Social media got a decade of goodwill first. Facebook was unambiguously pleasant for ten years while grandparents found their grandchildren and everyone posted baby photos, and only then did the reckoning arrive. AI got about five seconds. Sam Altman said he was not going into the advertising business roughly six months ago. He was headlining at Cannes not long after.

So the question for a software vendor is no longer whether the product can be built. It obviously can, by your customer, over a weekend, for about twenty dollars in tokens. The question is whether anyone believes you when you tell them what it does.

Some things still resist that pressure. Nobody is vibe coding a payment processor. NetSuite is not going anywhere. Systems of record hold because the cost of being wrong is unbearable and trust in them was earned across decades of boring reliability. Design holds too, though not design as a nice-looking screen. Design as Steve Jobs meant it, form following function, which requires knowing your user well enough to say no to them. That knowledge sits in a feedback loop nobody can generate synthetically.

Brands are running into their own version of this. The AI committee has twenty projects and cannot get any of them approved because everything routes through legal and compliance. The instinct is to call that a technology adoption problem. It is not. Building was never the constraint. What the committee is actually doing is deciding how much it trusts a system it cannot fully audit, and it keeps deciding not very much.

The pattern underneath all three groups is the same. We compressed an enormous amount of adoption into a very short window and skipped the part where anyone worked out whether to believe in it. Trust is the input everything else runs on. Software cannot be sold without it. Consumer goods cannot be sold without it.

Which leaves an uncomfortable position. The companies best placed to repair the credibility of this technology are the ones being paid the most to keep moving without it.

Citi's read of StockX data puts Nike at about 70% of sales volume in June, down from roughly 77% in late 2023. On most dashboards that reads as decline. Set it against what Nike said it was trying to do and it looks like inventory discipline finally catching up to intent. Prices on Air Jordan 1s and Air Force 1s have recovered on the platform. That happens when supply stops outrunning demand.

The performance side is where the real money question sits. Nike says it has now strung together five straight quarters of double-digit running growth, roughly a billion dollars in added sales, and picked up five points of running share across Western Europe and North America in the fiscal year ended May 31. Those are company numbers. Weigh them accordingly.

Asics is running the same trade backwards. Performance and sports slid to 55.4% of revenue from 64.5% in 2022, and the lifestyle margin is doing the heavy lifting on earnings. Onitsuka Tiger spins off in early 2027.

Wall Street still expects two more quarters of constant-currency declines at Nike. At 23 times forward, the stock is either early or wrong. Two prints from now we find out which.