Meta cuts its AI prices by 75 percent
Jared Blank hosted alongside Andy Hoar this week, with Brian Beck on vacation. The breaking news picked up a thread the show has followed for weeks: the cost of AI tokens. Meta has priced its developer platform at roughly one-fourth the rate of Claude and OpenAI’s comparable offerings, a 75 percent cut. After spending heavily to recruit AI talent, Meta has spent the past year trailing the frontier labs on adoption, and Andy’s read was that this is the standard move for a company that is behind.
What are two things that are going to happen here? Either they capture this budget-sensitive developer group that wants to save money, or they embarrass themselves.
Andy Hoar, Master B2B
Jared’s read was that the economics make sense regardless of outcome. Meta is cash-rich, funds a massive advertising business, and can treat a multi-month price experiment as low-risk, similar to the early streaming wars where services undercut each other for share before eventually raising prices once customers were locked in.
A market for cheap tokens is emerging
The hosts pointed to a real market response taking shape. Anthropic has begun offering lower rates for off-peak Claude Code usage, echoing old mobile-carrier night-and-weekend minutes. And a new category of token arbitrage has appeared: OpenRouter, which recently raised a large funding round, routes a request in real time to whichever model offers the cheapest available token for that job. Jared connected this to his own workflow, since he already picks different platforms for different tasks, image generation on one, everything else on another, based on which does the job better at the right price.
Andy’s closing point on the topic was the one worth holding onto: token price is not the only variable that matters.
Ultimately, this is all about ROI. Even the expensive tokens are still cheaper than paying a human, and the ROI on the cheap token will be better still, assuming you can do the same thing.
Andy Hoar, Master B2B
Get big, get small, or get out
The main topic returned to a theme the hosts have tracked for months in B2B distribution: the middle is disappearing. Two deals from the past week illustrated the “get big” side. Ferguson, the plumbing and HVAC distributor, agreed to acquire industrial flow-control distributor FloWorks for about 1.6 billion dollars, a move Ferguson says expands its total addressable market to roughly 400 billion dollars. Separately, buying group Affiliated Distributors, whose eCommerce Solutions lead Caroline Ernst is a past guest of the show, agreed to acquire supplyFORCE, its second such acquisition in the past year following last year’s Commonwealth Group merger. AD aggregates independent MRO distributors so they can compete with larger players like Grainger on buying power.
Jared’s view is that scale keeps winning because the old advantage held by regional players, faster local delivery, has largely evaporated now that next-day and overnight shipping works from adjacent states. Andy drew the parallel to what he watched happen across 30 years in consumer commerce.
You can be small, you can be big, you cannot be in the middle. All of the middle disappeared over the last thirty years. That is what happened to Circuit City, to Macy’s, to Bed Bath and Beyond.
Andy Hoar, Master B2B
The case for going small: a doorknob distributor
The counter-argument is specialization. Andy told the story of a small Akron, Ohio hardware distributor that was being outcompeted by Home Depot and Lowe’s, and instead of trying to match their scale, narrowed itself down to a single category: doorknobs. That focus turned it into one of the most successful doorknob distributors in the country, competing on its own terms rather than head to head with the big-box chains. Jared called this a brave move for a CEO to make.
It is a rare CEO who is willing to make a bold move like that, to say, I see where the future is going in our industry, and we have to niche down.
Jared Blank, Master B2B
Smart scale versus dumb scale
Both hosts drew a line between adding capability that reinforces a business and adding it just to get bigger. Andy called the first smart scale and the second dumb scale, describing manufacturers who keep bolting on adjacent categories until, twenty years later, one of their own business units is competing with another. Jared connected this to the private-equity cycle: firms with abundant capital buy conglomerated companies precisely because they can find the inefficiencies, break the parts apart, and often generate more value broken up than whole. The lesson for a diversified distributor is that scale alone no longer carries an argument. Being part of a large parent company means little if one division is not competitive on its own merits.
IBM’s historic stock drop: canary in the coal mine?
With two minutes left, the hosts flagged IBM’s stock falling roughly 25 percent on July 14, its worst single trading day in 115 years, worse than Black Monday in 1987. IBM said clients are redirecting spending toward AI infrastructure, servers, storage, and memory, ahead of expected price increases and supply constraints, pulling budget away from IBM’s software and consulting business. Andy’s read leaned skeptical of IBM specifically.
I think IBM is fundamentally flawed. They never miss an opportunity to miss an opportunity. This is yet another example.
Andy Hoar, Master B2B
Jared’s open question was whether the drop signals something broader: that AI infrastructure spending, which some assumed was leveling into a maintenance phase, may still be in its early, capital-intensive innings in a way the market had not fully priced in. Andy’s answer leaned toward IBM being the exception rather than the signal, though both agreed the debate is far from settled.
What this means for B2B leaders
Token prices are entering a genuine market, with arbitrage tools and off-peak pricing both emerging in the same month, which argues for evaluating AI spend on ROI rather than sticker price alone. And in distribution, the middle continues to shrink. Whether the right response is Ferguson’s and AD’s path toward scale, or the Akron Hardware path toward deep specialization, the businesses least likely to survive are the ones that do neither.