
Sharp Tech with Ben Thompson · 2026-06-26 · 26 min
Key moments - from our scoring
Substance score
40 / 100
Five dimensions, 20 points each
This episode opens with a bearish take on Apple's recent price hikes for MacBook Pro and Mac Studio models, attributing them to the company's failure to anticipate DRAM and HBM shortages - a stark contrast to NVIDIA's proactive supply-chain moves under Jensen Huang. Thompson argues Apple's reluctance to raise prices in March when new Macs launched, combined with their now-awkward out-of-band price adjustments, signals the company remains strategically blind to AI's infrastructure demands. The conversation then shifts to how legislative chip export controls and business-driven market capture are inadvertently creating optimal conditions for Chinese dominance in memory manufacturing, as local players can profitably move down the learning curve on surplus inventory. In the lighter half, the hosts tackle a mailbag covering daily caffeine intake (Thompson drinks a venti Starbucks iced coffee plus occasional afternoon Yerba Mate; Andrew typically sticks to morning coffee), messy garage management, and Thompson's concept of "vibe coding" - using simple software and human labor rather than enterprise complexity to solve personal organization problems. The vibe coding segment centers on building a lightweight home inventory system using QR codes and property-room-structure-storage hierarchies, designed for someone with one employee who handles grunt work while the owner simply retrieves items.
Not raising prices in March preserved Apple's brand promise that products stay at launch prices, but raising them now mid-cycle breaks that promise anyway. Thompson argues this signals Apple had no idea memory shortages were coming, unlike NVIDIA which proactively secured supplies.
Export controls force Chinese companies to make older-generation memory profitably on their own, allowing them to climb the learning curve faster than Western competitors who won't touch low-margin commodity inventory, eventually making China competitive across all memory tiers.
Vibe coding is building minimal, single-purpose software tuned to one person's exact workflow and constraints - in this case, a QR-code-based lookup system for a house organized by property, room, structure, and storage spot - rather than adopting bloated multi-feature products designed for mass markets.
Thompson drinks a venti Starbucks iced coffee most mornings, occasionally adds a grande Americano in early afternoon, and sometimes includes a late-afternoon Yerba Mate, though he avoids caffeine after noon to protect sleep.
He wants to avoid the work of organizing and picking up, so he pays someone else to make the decisions and do the grunt work, while he keeps a software system to look up where things are stored without needing to organize them personally.
Our reviewer’s read on each dimension, with quotes from the episode.
The opening Apple/memory and China DRAM segments contain genuine macro-level insights about strategic blindness and learning-curve economics, but the episode quickly degrades into caffeine habits, garage anecdotes, and a personal home-inventory app - probably half the runtime yields nothing a B2B operator couldn't live without.
by not raising prices in march what that suggests to me is even in march when all this agentic stuff was working was in the market no appreciation for what was coming
China now can move down the learning curve profitably. They can sell old tech at profitable prices because there's such a shortage in the market.
The framing of Apple's March non-price-increase as evidence of genuine strategic blindness toward AI/memory is a sharper-than-usual take, and the China DRAM learning-curve argument is non-obvious; the rest of the episode is lifestyle chat with no analytical ambition.
Apple has become a bunch of normies out there in cupertino
The fact they didn't raise prices in March when they could have at least preserved their brand promise that when we launch a product at a price, it stays that price no matter what.
Ben Thompson is a genuinely respected tech analyst with real intellectual credibility, but both participants are media/analyst personalities rather than operators who have built or scaled B2B businesses; there are no external guests in this preview.
I make a pot of coffee in the morning. Absolutely love it. I'm not super fussy about my coffee.
I consume a venti Starbucks iced coffee every single morning. And then I will go early afternoon with a grande Americano from Starbucks.
There are concrete anchors - Jensen Huang's soju photos, HBM generation numbers, the March Mac launch as a specific data point - but hard metrics, dollar figures, and cited sources are absent, and the second half of the episode leans on personal anecdote rather than evidence.
they can make like hbm one or two we're on hbm four now
I made a steak dinner bet with John Gruber that the new max they introduced in March were going to include a price increase
The hosts have easy chemistry and occasionally push each other (calling out caffeine dodging, the steak-bet directional vs. timing point), but the mailbag format is largely unchallenging and the listener questions are surface-level lifestyle prompts that produce no productive friction or follow-up depth.
Look you lost the bet with gruber so you do have to buy the steak
Wait, so we already had the Sharp Tech discussion about two daily Starbucks visits. I didn't realize there was frequently a third.
Computed from the transcript - who did the talking, and the words that came up most.
Ben and Andrew begin with an extended look at the market for memory chips, including market history, the supply crunch driving up today’s valuations, the myopia from SK Hynix, Samsung and Micron, new entrants from China, and why the timing of Apple’s price increases are a bearish signal. From there: Both hosts share their daily caffeine regimen, garage takes, and lots of thoughts on vibe coding after Ben’s 10-day foray into AI-assisted engineering. At the end: Codex marketing and an imminent OpenAI super app, Sam Altman’s appearance in a game of mafia, could the EU push Apple to better business strategy, stock buyback clarification, ocean data centers, metricmaxxing, and an update on last week’s hydration break take along with a better fix for international soccer.
Transcribed and scored by The B2B Podcast Index.
Hello, and welcome to a free preview of Sharp Tech. Apple's price raises this week are one of the more bearish indicators for Apple that I've seen in a long time. Okay. And the reason is that, number one, I don't think they saw this memory increase coming at all, right?
Because they would have made the moves that they made in the past. They would have locked up way. Like remember Jensen Huang being out there like, you know, drinking soju with like the, you know, Samsung CEO or whatever, like all those pictures. Oh, yeah.
He saw it coming. Yeah. He was out there securing supply such that everyone else is having a hard time. That's wild.
Yeah. So he was. We didn't see Tim Cook drinking soju. Didn't see Tim Cook out there securing memory supplies.
Okay. So number one, I don't think they saw this coming generally. number two I made a steak dinner bet with John Gruber that the new max they introduced in March were going to include a price increase how could they not given what was happening with memory it blew my mind they did not increase prices then because layered on that is one of Apple's brand promises is we don't change prices now part of that is we don't decrease prices so you can always buy with confidence like right and i believe this there's there's never been a trajectory sale because of the i don't want someone on the moment someone who paid full price you don't want to feel like no it's not that someone is sitting on that page deciding whether to subscribe or not i don't want even a hint of in the back of their mind oh i'll just wait for the wait for the august deal that's right like like the once a year sort of thing right no like it my price is my price it's only going in one direction it will increase at times it is never going to be cheaper now Apple so Apple is not cutting prices for sure but they also usually never want to change prices while product is in the market you should be able to count on Apple prices are what they are we don't like you know you go to Dell a computer is a different price every single day right sure that's fine that's part of their promise right like you can customize whatever you want if the price will be whatever you want who knows you need a computer just buy it uh Apple doesn't do that so by not raising prices in march what that suggests to me is even in march when all this agentic stuff was working was in the market no appreciation for what was coming yeah they seem to be like well maybe it'll blow over right it's pretty unbelievable maybe like maybe it's like the old cycles right it's gonna be six months and right we have a lot of temporary let's try let's wait as long as, like, see if we can make it through.
And they said on the earnings call that, oh, this is definitely going to impact the second quarter a little bit, but it's definitely going to impact the third quarter. I think they're like, we're still in the second quarter, and they're like, oh, we're out there. Prices are up hundreds of dollars. And it suggests to me this is a company that still has no idea what's going on with AI.
But continues to win, you know what I mean? Siri thing, like I said, that's 2024 technology. Now, we said in 2024, this stuff is so good. There's tons of products to be built on what exists right now.
And that governs the fact Apple is going to ship in 2026 what they promised in 2024, and it's going to be great. It's pretty good. Because AI is amazing. It is also nowhere close to what AI is capable of right now in 2026, much less going forward.
And the fact that Apple, it's not just that Apple missed the memory thing. Fair enough. The fact they didn't raise prices in March when they could have at least preserved their brand promise that when we launch a product at a price, it stays that price no matter what. And they're having to out of band increase prices in this very awkward sort of way suggests that they just don't know what's going on.
Flying blind. Yeah, it's not great. apple has become a bunch of normies out there in cupertino um look you lost the bet with gruber so you do have to buy the steak do you think gruber should have to give you like a third of the steak like a good chunk of the ribeye that you buy him whenever this bet is finally paid off absolutely not uh bet is a bet number one number two it's not enough to be right timing matters uh and so timing i got the timing wrong i don't know though i think directionally you had some good instincts uh chris says we'll keep it going we went long on the first question but let's bounce you always promise these fun like i know well look that that one was very interesting to me and obviously it's germane to the china podcast just what we're just on the china bit remember dram is fungible right so if no one's gonna let chinese dram whatever the chinese can't make hbm yet they can make like hbm one or two we're on hbm four now like yeah so there's all that complexity goes into it and their concern would be that this allows them to climb the ladder and make better hbm we just need dram right uh and so if the big three are gonna be super focused on hbm china said we can make dram like and maybe our dram isn't the most advanced or maybe it has bigger problems how do you get more advanced and how do you solve your problems how you make a bunch of it and like and so a lot of china you go back to our discussion of china building up their their fab capacity before, it was, well, China's willing to bear the cost of selling old stuff.
Eating it at the bottom of the ladder. Because you have to move down the learning curve, right? And no one else is going to eat it, but China will eat it, and so they'll make it down. China now can move down the learning curve profitably.
They can sell old tech at profitable prices because there's such a shortage in the market. And soon enough, they're going to be good at memory. and they're going to be competitive with these companies. And if NVIDIA is not allowed to take it, Huawei will take it.
Other people will take it. Like we're creating the conditions. This is the problem with cutting stuff off, cutting stuff off legislatively through expert controls, cutting stuff off from a business perspective because you're trying to like capture the market. You are creating the conditions for inevitable competition.
And I think the dice cast, like China is going to be a dominant player in memory. There no question because the conditions could not be better to become one Right Well and you mentioned earlier that it doesn really make sense for American companies to invest in memory at least from an economic standpoint but potentially from a security standpoint. It will at some point over the next five or 10 years here. Chris says, how much caffeine do you guys consume on a daily basis as you drink from a coffee cup?
As I read the question, what's the pattern morning and afternoon coffees? What's the energy drink policy? My wife is making me quit having a Red Bull every single day. So this is top of mind funnel stuff for me.
What do you think, Ben? I don't know how much of a caffeine guy you are. Yeah, you're the one that needs to answer this question. I mean, one of my all-time favorite segments on Greatest of All Talk was you guys did a, you know, in the summer, it's a little more loosey-goosey.
It's kind of like a holiday mailbag, but every week when nothing's happening. right and so you have like your top five things and you did a top five candy bar or episode and i i've never been more flabbergasted in my life than when the episode starts and you're like i'm not much of a candy guy and i'm like this guy i said was i love candy but i don't really eat candy anymore whatever it was it was the biggest crock of bs i've ever heard in my life no one loves eating terrible stuff more than you and what was great about that episode is you basically undressed yourself in front of the entire audience you had so many candy takes you had so many anecdotes in recent memory about candy that it was so obvious that you have a problem and you like exposed it was like hunter biden talking about crack cocaine on that episode i talked for about 45 minutes about all my favorite candy memories of candy and it was it got a hierarchy of candy interspersed with like the right times to have the candy proper storage for different types of candy it was oh my god that's right you gotta keep Certain candies in the freezer.
No question about it. On caffeine, I'm not as passionate. I am just as addictive. Okay.
Not as passionate as addictive. I'll give you that. I was going to say, if you're going to try to spin the I'm not a caffeine guy. Deeply, deeply addicted.
I consume a venti Starbucks iced coffee every single morning. And then I will go early afternoon with a grande Americano from Starbucks. and then I will sometimes layer on a Yerba Madre drink in the late afternoon. They have these 20-ounce teas.
Whoa, whoa, just a second. Wait, so we already had the Sharp Tech discussion about two daily Starbucks visits. I didn't realize there was frequently a third. Holy cow.
Well, there is a thing. It depends on the day, but every now and then, I would say three out of the five work days, I will go with a late afternoon Yerba Madre. and in general though the challenge with caffeine is I hit a point of diminishing returns if I drink too much of it and my brain just sort of shuts down at a certain point so I do try to moderate it and energy drinks like a Red Bull Chris I think your wife is probably right about a Red Bull a day being a bad idea I'm not a doctor but I think she's on the money with that one so I stick to coffee for the most part yeah um the one other thing when i was considering moving back um you know it was a very fraught decision very difficult for for lots of different reasons one of them was we would record sharp tech at night for you because you're on the east coast in particular particularly rough with taiwan and sometimes it's like yeah andrew went too far on the caffeine he's barely functional right now because he's like too on edge and i'm like i need to get back so uh You can be aware.
That's 100% true. There would be some nights where I'd be like just completely fried and trying to solve it with more Diet Coke at 945 or 10 o'clock at night. I feel great now as we sit here recording this. So I'm glad to have you back in the United States of America.
But you still are you dodging the question? How much caffeine do you drink? I make a pot of coffee in the morning. Absolutely love it.
I'm not super fussy about my coffee. I buy very good beans. I grind them right before I make them, but I use like a coffee maker. Like I think I do the Mocha master.
I like, I like them. So I bought like five of them. There's a new one out there that is like a Mocha master, but it uses all stainless steel parts on the inside. There's like no plastic, which is like good for like microplastics, whatever.
It's still sitting in my garage. I haven't actually set it up yet. You'll get there. The microplastics aren't really moving the needle one direction.
Yeah. as far as I'm concerned. I'm kind of in the same boat, to be totally honest. So I make a pot of coffee in the morning, and then I'm usually good for the day.
I will occasionally get a Diet Coke, particularly if I'm running at night and in the evening. I don't like caffeine past noon because I don't want to have any issues sort of falling asleep or whatever. But it might not be theoretical. I'm not sure if I have issues falling asleep.
I fortunately have the constitution of I don't really get addicted to anything so like on the weekend sometimes I wake up slow whatever I just don't have any coffee like one of my good friends if he goes longer like there's one he's like oh shoot it's at night I forgot to have my afternoon coffee I'm like but you're going to bed he's like I know I'm going to wake up in the middle of the night with a headache because it went too long I'm like oh my god you have a problem that's what happens been there.
Yeah, so I'm not I'm pretty sure I'm not addicted. I could go a week without coffee. I have. No issue.
I greatly enjoy it. It's one of the pleasures of life. But that is basically it. I love coffee.
There we go. I have it in the morning. Just regular coffee. If I go healthier than me.
I have Americano is my go to Starbucks Americano. Starbucks is a great store. I think I've given this take before. I have.
Not because it's great, but because it's the same every time. It's a Starbucks forward podcast here. We are pro Starbucks. Save your bespoke coffee houses.
All right. Well, keeping it moving, we have Daniel who says, Sharp I couldn help but notice that Ben implied on the last episode that you have a messy garage when you said you didn want to charge your car Throwing a little shade at you I don think that just throwing shade That's just a statement of facts. I'll explain that. My summer mailbag question for Ben, he says, is what is his version of the garage?
Is there something that Ben walks by at his house all the time and thinks, I need to take care of it, but he never touches it? If the answer is no, I won't be shocked. but it will be revealing in its own way perhaps maybe he could use a backup antenna on his roof or a third generator or more storage space for generators you always could use a backup generator for your backup generator well it is funny i did this by um uh you know we're going into this middle of nowhere i think they supposedly have solar panels and batteries uh and i guess there's the backup generator but just in case uh i did get like this big like anchor battery it's like this big uh and it has its own solar panels so you can charge it so you can always have power but i figure it's it is a backup to the backup i could have like an extra battery if you stay ready you don't have to get ready you're ready for anything as you undertake this fishing expedition i love to see it the yeah no i yes i have unfortunately i have a lot of stuff that i want to get to where the problems is i have ideas of how to solve things i acquire the components to solve it and then they are never actually implemented assembled well no they are but i need like a forcing function to to get it done uh so like i have a bunch of networking gear because i have a few ideas of the things i want to do then something will happen and i haven't installed it yet and then it's like a flurry of like a weekend of getting this all in and then it's done but i had the supplies on hand so so it does get done eventually um i would say in general yeah organization i love being organized i love knowing where stuff is i love being clean and i hate organizing i hate cleaning and i hate picking up after myself so there's a real there's a real tension about what i enjoy and what i'm capable of accomplishing and that is why you know i didn't i don't like talking about like i hire people and i have like assistants that's why you hire someone that you don't you don't fix your weaknesses you make more money so you could pay for someone else to fix them well i'm glad that daniel picked up on the garage comment during the last episode because what you were referring to there was not necessarily a messy garage at my house but when i bought the house the previous owner had converted half of a pretty small one car garage into an exercise room and put up a wall halfway through the garage.
So the garage I have doesn't actually fit a car and is full of storage. Maybe my garage project would be knocking down the extra wall and just recreating an actual garage for myself. But it's pretty small. It pales in comparison to your massive Midwestern garage.
So it's probably not big enough to fit like a modern SUV. be um no i realize this is that like the coastal elites don't appreciate a good garage uh well super important i'll tell you what not having a garage has made me appreciate a good garage i go out to madison and i'm like man this is what i need um i feel like you answered this question in part on stratechery this week so i wanted to use this as an on-ramp to the vibe coding discussion so tell people about your problem and your vibe coding solution to the problem well so i i put this one it's it's just super fun like like there are a few people that like took it way too seriously like i make you pronounce this about the future of ai and i was just like like there was a bit about like look there's there's this delightful like aspect of ai number one there's a bit where i don't think people fully appreciate like how good this stuff is uh yeah then number two it's like i feel a bit like a kid again back in the 90s putting computers together like acquiring different parts like you know building trying to do different things um having all these delusions of grandeur about all these businesses i'm going to start like how many people put up websites like all all the bits and pieces and just loving the work i love the piece on wednesday because it was pretty whimsical and you felt the joy throughout and having a great time with computers like that's the vibe coding experience that i came away which is about organizing my garage so um so the issue i have is um you know i moved back and you know i took one week off of work when moving back but then i had shipments arriving at multiple times i had like one air shipment i had one that came by boat i had a bunch of suitcases that we brought uh like brought when we came back in june then we went back in july to pack up and brought more stuff back then just a lot of stuff coming in.
I'm also still working full time and I have a house with all this, you know, lots of good sort of storage opportunities. But the problem is that it's the first storage run is super important because you have to figure out this is the cord area. This is the like the fishing stuff area. This is the car stuff area.
And so that in the future, when you need to find something, you know where to go. It doesn't help put away and you can't find it so the problem though is there was so much stuff the reality that i live in is the house looks relatively neat and organized because all that stuff was put on shelves with no organization because it's just like oh people are coming over like we're hosting uh let's get all this stuff out of the garage we need it's actually great your basement is spectacular and looks really clean and organized and then i stepped into your office last time i was in madison i was like whoa there's a lot going on in here so that's a job that still needs to be so messy because i don't want to put stuff away because i need to figure out where it needs to be put away first so that i can find it in the future so you can see i'm in this doom loop right like it's just it's just a total mess now i mentioned i i do have someone here that that works with me a friend from high school and it's you know it's great you know my wife is back and forth she spends a lot of time in taiwan it's part here so when particularly when i'm here by myself sometimes just driving the kid to school or like going to pick stuff up so i the the it's very beneficial and he's the one who put all the stuff in the shelves in the first place because it's like a panic oh like uh we need to get the stuff out of here just do something with it uh of course the answer is to have him organize everything Yeah But if you organize everything I don know where it is So, like, I need to be involved.
So the reality is what the answer is, I need to not be involved. How can we get a situation where everything is organized and put away? Centralized resource. But I can still find it.
What I need is software. I need an ability to look something up and know where it is. So there's lots of home inventory products on the market, right? If you want to imagine having enterprise software for your house.
That sounds horrible. There's like a gazillion little forms of a checkbox. Because they're trying. This is the issue with software.
So software, what's the magic of it? You write software once, and then you can sell it to a gazillion people at zero marginal cost. The problem is to sell to a gazillion people, you have to cover a lot of use cases. needs.
Yeah. Right. The whole like everyone complains about like, I think there's a point that Steven Stanowski is made before. Everyone complains about word has so many features.
I only need these five features. Everyone feels this way. The problem is their five features are all totally different and they cover the entire gamut of features. The reason why there's so many things there or AWS is like the modern version of this.
AWS is so complicated. There's so many services. Have you tried navigating their dashboard? It's unbelievable.
Can they just like, why do they have, but everyone feels that way, but what they want are all totally different. That's what happens to these products when you're, but that's also very profitable because you, you, you get scale on this. You sell it to a bunch of people. Everyone uses word and complains about how cluttered it is, but it has the one feature they need.
And the Microsoft makes a million dollars along the way. Right. So, yeah. So it doesn't need to be this way.
now. What I want is I want software that will solve this problem. And I think the unique constraint is I don't have a large staff. I have one staff and the brilliance of having someone that you pay is you could make them do stuff that normal people don't want to do.
Who wants to go through and actually do all the cataloging, right? And do that sort of stuff. Well, if you're getting paid to do it, it's a lot more like acceptable to sort of get it done. And honestly, he's happy about it, too, because instead of us doing a scramble to find something every other day, it's clearly better for him.
There's no question about it. But this is a unique dynamic where I have a human to do a lot of this work, so I don't need a lot of complexity. I want to get a software that's perfectly tuned to someone like me and how I work and perfectly tuned to having an asset like this person to implement these pieces and making it super simple, super approachable, not having a single feature I don't need, only what's there. This isn't an inventory system.
These are important notes. I'll come back to this. It's not like a task management system. It's a look up something and find out where it is system, and that's it.
And it's tuned to there is someone available and capable to do the grunt work, and the other person involved in this transaction doesn't want to do any work at all. It just wants to be able to find stuff. So because of this, it's a very - the magic here is I can make software that is going to solve this problem, and I don't need permission to do it. I don't need to hire a developer.
I don't need to try to bend stuff on the market to fit into my niche. I can build a perfectly fitting piece of software to this exact need. And that's awesome. And what you're doing is you're using QR codes, right?
To categorize some of it. The details of the system, I don't, you know, basically the idea is I'm defining the house by, so I, you know, I have other properties, you know, here. My parents is like right down the street. I got them in place.
I have some stuff stored there as well, of course. That's actually where the fishing stuff will go long term. So I have the property, and then within the property, there's rooms. Within the rooms, there's structures, like a shelf or a dresser or whatever.
And within the structure, there's an actual storage spot, which is like a single shelf or a single drawer or whatever. So the idea is my assistant now owns storage. I'm not going to - the storage, everything's going to be organized and put away, and I'm going to have zero input into it. he does what makes sense to him.
So he decides and records and we're set. And then after he's done, he's going to go through space by space. And this is where the AI is super awesome. How do you catalog everything?
95% of stuff, you take a picture of it, it gets it perfect. And then there's like the edge case where you need to make changes or whatever. There's a little bit of flow to check it. So you just take an iPhone photo and it just adds it automatically to the app?
Yeah. Wow. And it fills in all the details. You could theoretically- That's incredible.
Yeah. And so you organize everything. So where the QR codes come in is he doesn't necessarily need the QR codes at the start, but later, say, you have a new - something shows up in the mail, right? And he's like, this should go here.
So he gets the thing, takes a picture of it, puts it on a shelf, scans the QR code of that shelf, and boom, it's in the system in that spot. And then - so you can always find everything. So you search for something. I have right here.
I have a windshield washer concentrate with antifreeze. That is like a specific kind of windshield washer concentrate for, of course, you know, the point of winter. Yeah. You get a nicer car.
You pay up in all sorts of ways. One of these is you have to buy special windshield. Okay, great. So I have this next to my desk because this was a test object, like taking pictures of it.
Like, does it already exist in the system or does it not? Like, you know, I have it sitting around. But the idea is, yeah, if I need this, I can. I love that you're up late at night testing different objects for your new vibe coded app.
How long did it take you to build the app? It takes longer than you think. All right. And that is the end of the free preview.
If you'd like to hear more from Ben and I, there are links to subscribe in the show notes, or you can also go to sharp tech.fm. Either option will get you access to a personalized feed that has all the shows we do every week. Plus, lots more great content from Stratechery and the Stratechery Plus bundle.
Check it out. And if you've got feedback, please email us at email at sharptech.fm.
Other episodes covering the same guests and topics, from across The B2B Podcast Index.