Anthony Pompliano

The Biggest Pivot In AI History Is Happening Right Now

🇬🇧 EN🇪🇸 ES
48:36 min youtube 2026 Semana 25 🇪🇸 ES

Resumen

YouTube: https://www.youtube.com/watch?v=qezOp08hwPQ  |  Duración: 48 min

âš¡ TL;DR

  • Jordi Visser cree que la IA ya entró en otra fase: hubo más de 20 lanzamientos de modelos en dos semanas, apareció intervención regulatoria tras el episodio Mythos/Fable, y modelos abiertos como GLM 5.2 ya están lo bastante cerca como para presionar a los modelos cerrados.
  • La tesis alcista de CAPEX en IA sigue viva, pero el punto débil es evidente: si la eficiencia del software mejora más rápido de lo previsto, hyperscalers y semis pueden comerse un air pocket. Su señal de alerta es que el crecimiento del dólar en semiconductores pasaría de cerca de 100% este año a alrededor de 30% en 2027.
  • En macro va por otro lado: Kevin Warsh le parece más enfocado en productividad que Powell, cree que el próximo CPI podría salir cerca de 0% mes contra mes, y sigue diciendo que Bitcoin está en mercado bajista hasta recuperar la media móvil de 200 días.

◆ El pivote de la IA ya empezó

La tesis central del episodio es que el mercado todavía no ha asimilado la velocidad del cambio. Para Visser, las dos últimas semanas importan más de lo que parece: en vez de esperar meses por un gran release, el sector vio una ráfaga de modelos nuevos, iteración mucho más rápida y señales claras de que la industria se mueve hacia agentes y, más adelante, hacia la mejora recursiva.

El disparador clave fue la secuencia de Anthropic: Mythos se consideró demasiado potente para salir tal cual, se liberó una versión limitada como Fable, y poco después esas barreras se rompieron. Su lectura no es solo que hubo un jailbreak, sino que la intervención del gobierno llegó antes de lo esperado y que la idea de mantener los frontier models perfectamente contenidos se está rompiendo.

▶ El open source está atacando la soberanía del modelo

Visser plantea el open source como algo estratégico, no ideológico. Si una empresa construye encima de un modelo cerrado y ese modelo cambia, se retira o se encarece, todo el negocio queda expuesto. Por eso cree que los modelos abiertos ganan peso justo ahora que sistemas chinos como GLM 5.2 se están acercando al rendimiento de los lanzamientos estrella de Estados Unidos.

La implicación de mercado es bastante seria: en cuanto las empresas crean que pueden correr algo “suficientemente bueno” en su propio hardware, se debilita la escasez económica de los modelos cerrados. No significa que los líderes caigan mañana, pero sí que la capa de modelo puede perder poder de pricing antes de lo que descuenta el mercado.

★ La demanda de tokens pasa de consumo bruto a output

Una de las mejores observaciones del episodio es que muchas empresas ya están cerrando el grifo del gasto desordenado en IA. Los rankings internos de consumo y el “usa todo lo que quieras” desaparecen; ahora importa más eficiencia, output y retorno que cantidad de tokens gastados. Eso permite que la demanda por usuario baje aunque la adopción total siga creciendo.

Aquí entra la paradoja de Jevons: una herramienta más eficiente puede acabar generando más demanda total porque sirve para más tareas. A corto plazo, sin embargo, esa transición puede presionar los precios del token y la narrativa de que más compute siempre equivale a más monetización.

► Dónde se puede romper la historia del CAPEX

Riesgo a vigilar: si un hyperscaler grande empieza a recortar o incluso a frenar el gasto en infraestructura de IA, el mercado puede repricing rápido a semis y memoria alrededor de una sobreoferta temporal o “air pocket”.

Visser no dice que el buildout de IA sea humo. Dice que el mercado puede estar extrapolando en línea recta el gasto actual sin respetar lo rápido que mejora la eficiencia algorítmica. Su marcador concreto es el crecimiento del dólar en semiconductores: cerca de 100% este año y luego alrededor de 30% en 2027. Sigue siendo crecimiento, pero la pendiente cambia muchísimo.

También interpreta movimientos como Microsoft tomando distancia de OpenAI como señal de que la capa de modelo empieza a parecer más intercambiable. Si el model layer se comoditiza antes de lo previsto, el mercado tendrá más motivos para cuestionar la intensidad de capital en la infraestructura.

◆ Por qué no está totalmente bajista en memoria e infraestructura

Su contrapeso de largo plazo es geopolítico. Cuando el mercado ve que un modelo frontier puede restringirse o apagarse, todos los países que van por detrás reciben el mismo mensaje: construye tu propio stack. Por eso cree que la soberanía tecnológica y la lógica de seguridad nacional pueden mantener viva la demanda de memoria, energía e infraestructura aunque el mercado público pase por un susto intermedio.

▶ La agencia importa más de lo que parece

Fuera de la parte de mercado, Visser insiste mucho en la agencia. Usa ejemplos como Revolut entrenando un foundation model y Midjourney entrando en hardware de escaneo médico para argumentar que la IA no solo recorta costes: también amplía brutalmente el techo de ambición para empresas y personas.

Eso conecta con el punto social del final: mucha gente se siente atrapada por el coste de vida aunque la macro oficial no se vea desastrosa. Para él, la IA es una de las pocas herramientas reales que pueden aumentar el apalancamiento personal lo bastante rápido como para importar.

★ Amplitud de mercado, Fed e inflación

Coincide con la idea de que gran parte de la fortaleza del mercado está concentrada en IA y energía. Si quitas eso, la amplitud es bastante peor. En política monetaria, le pareció que Kevin Warsh sonó diferente a Powell: menos obsesionado con microgestionar cada dato, más abierto a la productividad y más dispuesto a mirar medidas como el trimmed mean en vez del ruido headline.

Además, Visser cree que el telón de fondo de inflación está más blando de lo que descuenta el consenso. El petróleo no logró sostener la subida ni con tensión geopolítica, los inflation swaps a un año se han desinflado y dice que el próximo CPI podría salir cerca de 0% mes contra mes. Eso no arregla la presión real sobre el coste de vida, pero sí cambia la lectura sobre la Fed y los activos de riesgo.

â–º Bitcoin sigue siendo un trade de mercado bajista

En cripto es bastante directo: Bitcoin sigue en mercado bajista hasta volver por encima de la media móvil de 200 días. El problema de fondo es la atención. Ahora mismo el capital y el relato están en la IA porque ahí está la aceleración de beneficios. En su marco, Bitcoin y SpaceX son activos muy ligados a la creencia en el futuro, pero ese tipo de trade sufre mientras el mercado premie earnings inmediatos de IA.

â—† Buscar el alpha

El alpha real de la conversación no es “la IA es buena” o “la IA es mala”. Es que el mercado puede estar pagando demasiado por una narrativa lineal de CAPEX justo cuando la tecnología se vuelve más eficiente, más abierta y más fragmentada geopolíticamente.

  • Vigilar semis y memoria por el lenguaje de CAPEX, no solo por earnings. Si los hyperscalers suavizan el gasto, la rotación puede ser dura.
  • El open source es la válvula de presión. Cuanto mejores sean los GLM y compañía, más difícil será para los proveedores cerrados defender economics premium.
  • La soberanía de IA es el soporte de largo plazo. Aunque tiemble el gasto privado, gobiernos y países rezagados seguirán construyendo.
  • Bitcoin necesita un cambio de régimen en atención. Hasta que la euforia de IA afloje y BTC recupere la 200-day, no ve momentum real.
La vuelta de tuerca: Básicamente está diciendo que los próximos ganadores de la IA pueden salir de la misma fuerza que amenaza a los ganadores actuales: la eficiencia. Si los mejores modelos reducen la necesidad de gasto bruto, la narrativa favorita del mercado puede convertirse en su próximo problema.

► Resumen por capítulos

0:57 — AI pivot & hyperscaler weakness

Visser dice que el ritmo de progreso en IA ya rompió el ciclo viejo de releases. El episodio Mythos/Fable, junto con el jailbreak y la respuesta regulatoria, le confirma que estamos más cerca de agentes y mejora recursiva de lo que parecía.

5:47 — Open source models & the US/China race

Los modelos abiertos se están convirtiendo en la cobertura práctica contra la dependencia de proveedores cerrados. GLM 5.2 importa porque muestra que un competidor con menos recursos puede acercarse mediante eficiencia algorítmica.

7:19 — Token demand, Jevons y adopción

Las empresas pasan del “consume todo” al “enséñame el output”. Eso puede bajar el gasto por consulta aunque el uso total de IA siga creciendo.

14:22 — Cuándo el CAPEX se convierte en problema

La señal de alerta es clara: si los hyperscalers empiezan a dudar de cuánto gasto sigue siendo necesario, semis puede corregir rápido. Visser destaca sobre todo la desaceleración del crecimiento de ~100% a ~30% hacia 2027.

19:25 — Agencia para individuos

La IA no es solo una herramienta de recorte. Cree que eleva muchísimo el techo de ambición para empresas y personas, y por eso insiste tanto en la agencia como unlock real.

28:32 — Warsh, inflación y lectura de mercado

Warsh le parece más orientado a productividad y menos rígido que Powell. Junto con señales de inflación más suaves y un posible CPI casi plano, eso le sugiere una macro que puede aflojar antes de lo que cree el mercado.

44:51 — Mercado bajista en Bitcoin

Su línea es limpia: Bitcoin sigue siendo un activo de bear market hasta romper la media de 200 días, y probablemente necesita menos euforia en IA para que la atención vuelva a rotar.

Generado con algoritmo v2.1-anchor-first · modelo openai-codex/gpt-5.4 · 2026-06-20T14:14:51Z

Transcripción

[0:00] When SpaceX happened this week, it
[0:02] became obvious. SpaceX and Bitcoin are
[0:04] basically to me the same thing. I love
[0:06] when people send me things and they're
[0:07] going, "You're wrong. SpaceX is
[0:09] overvalued." I'm like, "When something
[0:10] doesn't have a valuation, it can't be
[0:12] overvalued or undervalued. It has no
[0:14] valuation. It's a complete guess. We're
[0:16] talking about flying to Mars. We're
[0:18] talking about flying to the moon and
[0:19] building space stations." Like, this is
[0:21] a this is a dream. So, Bitcoin has no
[0:24] energy because it is a vehicle meant for
[0:27] two things. One is
[0:29] >> what's going on guys? Today we got a
[0:30] great conversation with Jordi Visser. In
[0:32] this conversation we talk about what's
[0:33] going on with the hyperscalers, why
[0:34] their stocks have been a little weak,
[0:36] what's going on with Fable 5 and many of
[0:38] the open source models, how the
[0:39] competition is heating up, what that
[0:40] means for you and for investors. We talk
[0:42] about Bitcoin trueflation, Kevin Worsh's
[0:44] first meeting, and what the Fed is
[0:46] likely to do going forward. Jordi is
[0:48] thinking a lot differently than he was
[0:50] maybe a couple of weeks ago, and he
[0:51] explains what's changed, what's stayed
[0:52] the same, and how his portfolio is
[0:54] adopting. Hope you enjoy my latest
[0:56] conversation with Jordi Visser. All
[0:58] right, Jordy, there's a big pivot going
[0:59] on in the AI world. It feels like the
[1:01] hyperscalers, there's a lot of weakness
[1:03] there, but there's a lot of technical
[1:05] innovation that's going on. How do you
[1:06] evaluate the current state of AI and why
[1:09] so much of a pivot is occurring?
[1:12] >> You know, a lot's happened in the last
[1:16] two weeks. That's probably more
[1:18] important than what um everyone
[1:21] realizes. And I think this is just a
[1:23] world where we sit here, we talk about
[1:24] AI every week. I think the reason we
[1:26] always have a new topic to talk about is
[1:28] because uh the changes happen so fast. I
[1:30] wrote a paper this week and
[1:33] so people know this because the the the
[1:36] number one story from let's say the end
[1:38] of February until now has been the
[1:40] straight of Hormuse
[1:42] >> Iran the US
[1:44] >> during that period where oil prices
[1:45] peaked in the first week
[1:48] literally the first week and then
[1:50] progressively have gone lower despite
[1:52] all of the doom and gloom forecast by um
[1:56] people who get paid to sell
[1:58] subscriptions on oil. Uh there's been
[2:02] over 20 model releases in AI, including
[2:06] Opus 4.7, 4.8, and Fable. So, uh the
[2:11] progress in AI last year, we had to wait
[2:14] a long time for Chat GPT5 to come out.
[2:17] Uh it was the most anticipated model
[2:21] release and the most nothing release of
[2:24] any model yet. And now we're releasing
[2:26] them so fast that they don't get any
[2:28] attention. So, in the last two weeks, we
[2:30] hit a point, and I think for everyone,
[2:33] um, they should go back to the the piece
[2:34] by Leopold, who's gotten all this
[2:37] attention now because of his hedge fund
[2:38] going up astronomically,
[2:40] >> situational awareness,
[2:41] >> situational awareness, and they should
[2:42] go back and read it. Um he highlighted
[2:44] in there that an important moment would
[2:47] occur in 2027 2028 when we would reach
[2:50] recursive self-improvement but also when
[2:53] the government would be forced to
[2:56] basically regulate AI and it would no
[2:58] longer be this everyone has this model
[3:01] and that happened in the last two weeks.
[3:03] >> Now recursive self-improvement is
[3:05] basically the idea that the model
[3:07] doesn't need feedback from humans. It is
[3:08] learning autonomously essentially and
[3:10] making improvements to itself.
[3:12] >> Yeah. Yeah. So we have to hit AI agents
[3:14] first and then the question is how long
[3:15] between AI agents growing in numbers. So
[3:18] think of it as digital employees just
[3:20] sprouting everywhere and the models
[3:22] getting so good that they can
[3:25] create new models on their own and all
[3:27] the algorithmic efficiencies and things
[3:29] that'll come with that. So he wrote that
[3:33] and we've got the government now getting
[3:35] involved and it's specifically involved
[3:38] because a third party researcher in this
[3:41] case it appears to be Amazon figured a
[3:44] way to get around the guardrails. So all
[3:46] that time we heard where mythos was
[3:48] released then it was pulled back during
[3:51] that time period they put up guard rails
[3:53] to make sure that they were preventing
[3:55] bad actors from being able to do bad
[3:56] things. Just so for people who aren't
[3:58] deep in the weeds, uh, Mythos was a new
[4:00] model from Anthropic. It was deemed
[4:02] incredibly powerful. People were using
[4:04] it internally in testing and they were
[4:06] finding a lot of cyber security uh, kind
[4:08] of vulnerabilities. And so the
[4:10] government and a couple of cyber
[4:11] security companies came together, said,
[4:12] "Hey, this may be too powerful to
[4:14] release in its current form." And so,
[4:16] Anthropic essentially took a watered
[4:18] down version. They were able to create
[4:20] these guardrails uh, around it. and they
[4:22] released it under the name Fable, but
[4:25] within I don't know what was it, three
[4:26] days. Yeah.
[4:27] >> Somebody at it appears Amazon was able
[4:29] to basically jailbreak it. So same way
[4:30] you could jailbreak a phone, you they
[4:31] jailbroke the model and then of course
[4:33] everyone freaked out cuz they're like,
[4:34] "Oh my god, this mythos thing is
[4:36] actually out in the wild in a jailbroken
[4:38] form."
[4:38] >> Yeah. So for those of you looking for an
[4:40] analogy who've played golf, if you take
[4:42] the governor off your golf cart, it goes
[4:43] a lot faster. So that's what they
[4:45] basically were able to do.
[4:46] >> No one would ever do that.
[4:47] >> No one would ever do that. Not not as a
[4:49] young kid playing golf. Um but what did
[4:52] end up what ends up happening is Leopold
[4:54] talked about the government getting
[4:56] involved which has happened and
[4:58] recursive self-improvement. Now he
[5:00] forecasted that that would happen in
[5:02] 2027 2028. So again I I want to make
[5:04] sure people understand that we've
[5:06] reached a point where the next step is
[5:08] AGI which is something we've all talked
[5:10] about. So you've got the government
[5:12] getting involved shutting it down in a
[5:14] way that makes it very very difficult.
[5:16] I'm sure it'll be re-released again
[5:18] after new guardrails are put in and they
[5:20] find some way to go through it. But you
[5:22] just opened up a can of worms and what
[5:24] he wrote about is once you get to this
[5:26] stage, everything changes. Now, at the
[5:27] same time,
[5:29] Z.AI, another one of these Chinese open
[5:32] source models, released GLM 5.2,
[5:36] which ended up basically getting close
[5:40] to Fable 5 Mythos. And so you've got
[5:44] you've lost sovereignty where you don't
[5:46] know if a model can be shut down. So if
[5:47] you built your entire business on Fable
[5:49] 5 now it only been quick. But let's
[5:51] assume you used it and you've already
[5:53] made upgrades and then all of a sudden
[5:54] they say it's shut down. Okay. But
[5:58] you've got an open source model which
[5:59] you can just download onto your
[6:01] hardware, run it, and it's almost as
[6:03] capable as that. Is this the point now
[6:06] where we start to see the drive to more
[6:07] open source? And that's become a story
[6:09] that's bigger and bigger. Now, at the
[6:11] same time with recursive
[6:13] self-improvement, you had OpenAI
[6:15] basically say maybe we're not going to
[6:17] do an IPO and the speculation on this
[6:21] point when they brought up codeex and
[6:22] all these points is are we reaching a
[6:25] point where capital is not going to
[6:26] become as important because these models
[6:28] have improved at this pace without the
[6:31] data centers being completed. So from my
[6:33] side the pivot point is we have a stock
[6:36] market which has just raced higher based
[6:38] on the capex buildup.
[6:41] every single part of its capex buildout.
[6:43] I think again we're at a point where in
[6:46] when the market shows the next sign of
[6:48] weakness in some of these semiconductor
[6:50] names and stuff, I think people should
[6:52] pay attention. What I just laid out is a
[6:54] narrative that will be a bigger
[6:55] narrative sometime. It might take 3
[6:57] months, it might take 6 months, it might
[6:58] take 12 months. But when you combine
[7:00] what you said, which is the hyperscalers
[7:03] are weak.
[7:05] The hyperscalers are weak because
[7:07] they're spending tons of money. And the
[7:09] question is, are they coming under
[7:10] pressure now to maybe cut their capex,
[7:15] especially a Microsoft and a Meta whose
[7:16] stocks are extremely weak.
[7:18] >> Now, as we watch this play out, there's
[7:20] a couple things that I think are
[7:21] happening. And I've talked in the past
[7:22] about, you know, the mandate from heaven
[7:24] 12 months ago was everyone go use AI.
[7:26] And every company went and ran crazy
[7:29] with it. We have now since seen uh a
[7:31] number of different companies come out
[7:33] and either say we blew through our
[7:34] budgets, we're spending too much. Hey,
[7:36] we're taking down the token leaderboards
[7:38] and we're saying it's not about
[7:39] consumption, it's about output and
[7:40] efficiency and effectiveness. Um we
[7:43] inside of Sylvia have seen this where we
[7:46] said hey look this user generated that
[7:48] means everyone can just go spend as much
[7:49] as they want of our money for
[7:50] subsidizing. Let's stop doing that.
[7:55] I see the per customer per query demand
[8:00] actually shrinking because everyone's
[8:02] becoming smarter about how efficiently
[8:03] do you use the model and the best way
[8:05] that I use this is like you know uh if
[8:07] you're using your regular computer and
[8:09] you type something in and you measured
[8:11] how much computational power is used for
[8:13] that query it is getting more and more
[8:15] efficient over time that's basically
[8:16] what's happening with the models
[8:18] >> but their revenue is skyrocketing still
[8:20] so each individual customer is becoming
[8:23] moreffic efficient which means it's
[8:24] actually less revenue for the company on
[8:26] a per query basis but the demand is not
[8:28] slowing in fact actually it may be
[8:30] accelerating because people are
[8:31] realizing wait I can do this more
[8:32] efficiently I get more productivity out
[8:34] of this and so I actually want to
[8:35] consume more overall it's just that I'm
[8:37] getting more productivity is that kind
[8:38] of how you see this playing out
[8:40] >> yeah so the whole point with Jevans
[8:43] paradox and this whole belief that if
[8:44] token prices start going lower because
[8:46] people get more efficient with going on
[8:47] the models get more efficient all this
[8:49] stuff happens you end up with place that
[8:51] you're just going to get more demand um
[8:53] which I agree with over the long term. I
[8:55] I I think in the Q1 of this year we had
[8:58] an explosion of people diving into
[9:02] anthropic partly to because they felt
[9:04] like they were falling behind. So one of
[9:06] the things that I think with Jeff
[9:07] German's paradox that people assume uh
[9:10] especially the people who say it on the
[9:12] AI side is that it's going to be a
[9:13] linear growth adoption and I don't think
[9:15] that's the case. I think now we're in a
[9:17] place where people are analyzing just
[9:19] like you just said with Sylvia, how much
[9:21] money am I spending and what ways can we
[9:23] get around that? Now it happens at a
[9:24] time when open source models are way
[9:26] ahead. And when I mentioned Leopold's
[9:29] situational awareness, if people go back
[9:30] and read it, he didn't expect open
[9:33] source models to be where they are. He
[9:35] literally in this piece he wrote about
[9:37] how important it is for the US
[9:39] government to take over these models if
[9:40] they get so good to pres to prevent
[9:43] China from the ability to catch up
[9:45] because it's a military dominance thing.
[9:47] Well, the Chinese have been able to keep
[9:49] up no more than 6 months behind
[9:54] without having our chips to the degree
[9:56] that is without having all of the things
[9:58] of the model. So, I think we've kind of
[10:01] reached a point in in AI where it's
[10:04] amazing how fast this moves. A year ago
[10:08] is when the GPT5 thing happened. It's
[10:10] it's we're we're moving so fast that I
[10:13] think people are getting caught up
[10:14] particularly people investing and this
[10:15] is a warning to everyone out there. Um
[10:18] this is why I do my weekly video. It's
[10:20] like anything can change during the
[10:22] week. And this narrative that I'm
[10:23] talking about and what you're saying we
[10:26] could see a period where token prices
[10:27] drop because the adoption side pulls
[10:30] back for a variety of reasons even
[10:31] though in the long term German paradox
[10:33] is going to work. So I think the cost of
[10:35] the data centers is becoming the bigger
[10:37] issue.
[10:37] >> There's two other aspects to this. um if
[10:39] the Chinese open source models are able
[10:41] to keep pace with the US models, I think
[10:44] the generalized uh view is that they
[10:47] don't have access to the chips, the
[10:48] power, you know, kind of all the the
[10:50] training ingredients, if you will, but
[10:52] they're keeping pace to a degree. Does
[10:54] that actually is that like indicative
[10:56] that the United States is not innovating
[10:58] fast enough? like we should the gap be
[11:00] bigger because of the advantages that we
[11:02] have in terms of chip access power uh
[11:04] and the computational um uh kind of
[11:07] aggregate size the training data all
[11:09] these things.
[11:10] >> So again this argument has been brought
[11:12] up now for since deepseek in January of
[11:15] last year which was okay when you take
[11:18] away something the the raw brute force
[11:21] of data centers and colossus and all of
[11:24] this stuff and you just say you're going
[11:25] to have to get smarter without it. The
[11:27] interesting thing about this week, there
[11:28] was another thing that Open Router came
[11:30] out with which was called Fusion.
[11:32] Now Fusion
[11:35] is somewhat similar to Z.AI
[11:40] GLM 5.2. GLM 5.2 is a mixture of
[11:45] experts. So with inside the model, it's
[11:49] kind of like having a judge and then a
[11:51] board that's giving things and then one
[11:54] person is, you know, one one part of it
[11:56] making the decision. On Fusion, it's
[11:59] taking all the best models and kind of
[12:02] taking all their opinions and going up.
[12:04] Now, when I originally showed people and
[12:06] gave prompts out on on on my payw wall,
[12:09] the thing that got the most attention
[12:11] was, "How did you come up with this
[12:12] concept of deep research?" So when I do
[12:14] deep research on Vera Rubin, which I
[12:16] just did one this week, I actually use
[12:19] all five models to do the deep research.
[12:21] So 30 pages each from five different
[12:23] models and then I consolidate them into
[12:25] one that is somewhat similar to what
[12:28] Fusion does, but it also has a similar
[12:29] construct to the way GL GLM 5.2. The
[12:32] reason I bring that up with the Chinese
[12:33] open source models, they're forced to
[12:35] take these other techniques using
[12:37] reasoning, using reinforcement,
[12:38] learning, human feedback, mixture of
[12:40] experts, all these different things and
[12:41] come up. And I just think that that has
[12:43] led to more innovation on them side on
[12:45] the algorithmic and the efficiency side
[12:47] while we've just focused our attention
[12:49] on all the spending on the data center
[12:51] side to getting bigger models and going
[12:53] through it. That's where I think it
[12:55] hasn't been that long. It takes a long
[12:57] time to get these data centers built
[12:58] except for Elon Musk. So I think maybe
[13:00] we're at that pivot point where people
[13:02] should start to pay attention that there
[13:03] will be another narrative at some point
[13:04] this year. There's no doubt in my mind.
[13:06] >> Today's episode is brought to you by
[13:07] Simple Mining. Bitcoin mining has a
[13:09] reputation for being complicated, risky,
[13:11] and hard to evaluate as a real
[13:12] investment. If you're considering mining
[13:14] in 2026, what actually matters isn't
[13:17] headline profitability. It's uptime,
[13:19] repairs, and whether the operation is
[13:21] run like a real business. That's why
[13:23] I've been using Simple Mining. They're
[13:25] based in Cedar Falls, Iowa, and they run
[13:27] a white glove hosting operation where
[13:29] you own your miners. You choose your own
[13:31] pool, and you have Bitcoin sent directly
[13:33] to your wallet. They were featured on
[13:35] the Inc. 5,000 list as the fastest
[13:36] growing company in Iowa with over 40,000
[13:39] machines under management. What stands
[13:41] out to me is execution. They have the
[13:43] number one rated ASIC repair center. And
[13:45] for the first 12 months, repairs are
[13:47] included. If mining margins get tight,
[13:49] you can pause with no penalties. And if
[13:51] you want to resize or upgrade your
[13:52] fleet, there's a marketplace to resell
[13:54] equipment instead of being stuck. To
[13:56] help people think it through whether
[13:57] mining actually makes sense right now,
[13:59] they put together a short resource
[14:01] called the 2026 Bitcoin mining
[14:03] blueprint. It walks through the five
[14:04] mistakes investors make when allocating
[14:06] to mining and they also explain how to
[14:08] avoid them before deploying capital. If
[14:10] it sounds interesting to you, you can
[14:12] get it for free at simplemining.io/p.
[14:15] That's simplemining.io/p.
[14:19] Go check it out today and see if you
[14:20] should get into the mining game. All
[14:22] right, now let's talk about the capex
[14:24] spending. Um, there has been a lot of
[14:27] debate as to are they spending too much?
[14:29] if they're not using free cash flow,
[14:31] they start to take a lot of debt, then
[14:32] there's the kind of leverage question.
[14:34] Um, but I think what you're really
[14:35] highlighting is maybe the amount of
[14:38] power and uh compute that we thought we
[14:41] needed, we may not actually end up
[14:43] needing if there's this recursive
[14:44] self-improvement, the software side is
[14:46] becoming so efficient. How do you
[14:48] analyze what the tipping point is when
[14:52] you would say, okay, all of this capex
[14:54] spending, maybe we only need 50% of what
[14:56] we thought we were going to need.
[14:58] So the tipping point from memory stocks
[15:03] is honestly going to be if someone says
[15:05] they're cutting capex. Uh the second
[15:08] derivative is a very powerful thing in
[15:10] investing. Um I say it all the time rate
[15:13] of change. And if you look at um
[15:16] semiconductor dollar growth for this
[15:19] year according to Gartner it's going to
[15:21] be close to 100% relative to last year.
[15:23] The forecast for 2027 is about 30%. So
[15:27] we went from 100% growth to 30. That's a
[15:30] problem. Um now again it's a good
[15:32] problem to be growing 30%. But if you
[15:34] question whether the capex is necessary,
[15:37] that is the point where I think people
[15:39] start to worry and I think it starts to
[15:40] become a bigger story. Now I I will say
[15:42] this to everyone out there. We have such
[15:44] a shortage of memory that it would be an
[15:47] air pocket. Uh and that's what I expect
[15:49] to happen at some point. I expect capex
[15:51] air pocket where everyone freaks out.
[15:53] >> Yeah. What does that mean air pocket?
[15:54] That means that if one person pulls back
[15:56] on their capex
[15:59] and again we have no signs of that right
[16:01] now. If anyone would be the likelihood
[16:03] it would be Microsoft is is my guess.
[16:05] Sacha Nadell has been very outspoken
[16:07] about things and even said in the last
[16:08] week
[16:10] they're moving to DeepSeek possibly as
[16:13] some for
[16:13] >> like a Microsoft hosted deepseek version
[16:15] or something right.
[16:16] >> Uh and again what that is basically
[16:18] saying is he made the decision a while
[16:21] ago to kind of break ties with open AI.
[16:23] He has said that the models will
[16:25] eventually be commoditized and getting
[16:26] involved in this. He said this not that
[16:28] long ago either. So when people go that
[16:30] was a long time ago. It was a long time
[16:31] ago in AI time. It was like a decade ago
[16:33] in AI time but it was about a year ago
[16:35] in in human time. Um
[16:37] >> mere mortal time.
[16:38] >> Exactly. If they and again they have a
[16:41] problem with pulling back on this not
[16:43] because they're a model company but
[16:45] because they have Azure and they need
[16:47] more and more stuff for to house this
[16:50] because they've got a lot of revenue to
[16:51] come in. But if capback starts to slow
[16:54] down, then all of a sudden you've got
[16:56] all these memory that's sitting there,
[16:58] it will not stop. Because what shutting
[17:00] down Fable 5 said to every country in
[17:03] the world who's way behind the US, you
[17:05] better go build your own thing. You
[17:07] better have open source. You better have
[17:08] whatever because you cannot depend on
[17:10] the United States to just give you
[17:11] models cuz we shut down the rest of the
[17:13] world. Non US people cannot use it. Um,
[17:16] the ironic thing is philanthropic and I
[17:18] think for all of the model companies
[17:20] like 70% of the people that work there
[17:22] are non US
[17:24] people from the moonshots uh episode
[17:27] this week. So I just think that we're
[17:29] going to hit a point where the risk
[17:31] finally and I didn't think this was
[17:33] going to happen is that there'd be an
[17:34] air pocket in capex because one of the
[17:36] companies would say we don't need it.
[17:38] You'd have all this memory. The stories
[17:39] would start going they have all this
[17:41] memory they're going to need to dump.
[17:43] But Apple said they're raising prices
[17:44] because memoryy's got it more expensive.
[17:46] At some point here, the price of all
[17:49] this stuff that's been hoarded could end
[17:51] up in an air pocket, but eventually
[17:53] everyone will build their own AI
[17:54] factories and every country will and
[17:55] there'll be demand for memory for the
[17:57] next 5 years. But I do expect an air
[17:58] pocket at some point.
[17:59] >> The open- source uh or open systems seem
[18:03] to win out over the long run. Um you
[18:05] know, AOL versus open uh systems. Um
[18:07] you've seen a lot of open source
[18:09] technology, Linux, etc. that that
[18:11] continue to dominate. is your belief
[18:13] that these closed AI models or the open
[18:16] source models end up winning over the
[18:18] long run?
[18:19] >> So I I do, but I think for scientific
[18:22] discoveries and stuff, they'll just
[18:24] continue to be there. I think anything
[18:26] that
[18:27] >> the closed sourced ones.
[18:28] >> Yeah. I I So I think um you have to
[18:31] break this down and say we need more
[18:34] Einsteins to solve living forever, to
[18:37] solve energy, to solve space, to solve
[18:40] having data centers on the moon. like
[18:42] Elon Musk is great, but it would be
[18:43] great to be able to do much better
[18:45] simulations and actually have scientists
[18:47] that are uh at 400 500 IQ. So, we're
[18:51] going to need those for scientific
[18:53] breakthroughs. But I think we're coming
[18:54] to the point of what do you actually
[18:56] need for all this other stuff. And I I
[18:58] think housing your own models, training
[19:00] your own models, which is different than
[19:02] what's happening. Um I do think we're
[19:04] getting to that point. I've always
[19:06] envisioned that there'd be more
[19:07] specialized models for particular things
[19:10] that each company that was big enough
[19:13] and was involved in something important
[19:15] would have their own factory just like
[19:17] Eli Liy does with Lily Pod. Uh I think
[19:19] we're going to get to that faster than
[19:21] we were. I never thought you could
[19:22] depend on the cloud for a variety of
[19:24] reasons. But now what's happened is it's
[19:26] not just security, it's not just
[19:28] latency, you're also dealing with the
[19:30] issue that you're at the mercy of a
[19:31] company that just says you can't use it
[19:33] anymore. Mhm. The thing that I kind of
[19:36] come back to is uh Revolute is a
[19:38] business that usually isn't talked about
[19:40] here in the United States. Uh Revolute
[19:41] is um maybe for the the most basic
[19:43] description for people to understand is
[19:45] the Robin Hood of Europe, right? So it's
[19:46] retail brokerage. They have a lot of
[19:48] banking services etc. Um they actually
[19:50] trained a foundation model and it seems
[19:53] like okay they've got their like retail
[19:55] offering and now they've got this model.
[19:57] We'll see what they do with it but that
[19:59] seemed like a very big breakthrough
[20:00] moment. Midjourney this week came out
[20:03] with uh hey we've been doing this image
[20:05] uh you know creation. It was this cute
[20:07] creative tool. Oh now we're going into
[20:09] hardware medical services and they seem
[20:12] to have a pretty big breakthrough in
[20:15] terms of the ability to scan your body.
[20:17] The way my I understand this works is
[20:19] like you stand on almost like a graded
[20:21] platform and then they submerge you in
[20:23] water and they're shooting light into
[20:26] your body. um and they're able to do
[20:28] some imaging and it's supposed to be,
[20:29] you know, faster, more accurate, etc.
[20:32] Forget whether it actually works or not.
[20:35] A retail brokerage creating a foundation
[20:37] model that is a cranking of the ambition
[20:40] dial from like 20 to 100.
[20:42] >> Mhm.
[20:43] >> Midjourney going from creative image
[20:45] generation to medical device cranking of
[20:48] the ambition dial to 100.
[20:51] I h I just don't see anything else other
[20:52] than like AI is now giving them the
[20:55] ability to do this stuff in a way that
[20:57] they couldn't have dreamed of trying to
[20:58] accomplish these way bigger more
[21:00] ambitious projects.
[21:03] Is that what we should expect now as the
[21:04] new normal is like people will just say
[21:06] what can I do that's bigger better you
[21:08] know go faster? Yes. Yes. Yes. I mean
[21:12] last week just as like a a jump off
[21:16] point here. So when you brought up Jeff
[21:19] Bezos and Bezos and Prometheus
[21:22] >> Mhm.
[21:23] >> crazy.
[21:24] >> Yeah. Crazy. And
[21:26] again
[21:28] intelligence is involved in all it
[21:30] crosses sectors. Creativity in my
[21:34] opinion crosses like you can have a
[21:36] person who's an engineer who is the
[21:39] greatest artist painting and cook and
[21:41] writer and all this stuff and you just
[21:43] would never know and you view them as
[21:45] being this math person that's good this
[21:48] what AI allows you to do is take
[21:49] intelligence and combine it with
[21:51] creativity and with that it's very
[21:52] powerful as someone who spends his time
[21:54] trying to think and I had to find a word
[21:57] for this because I've said it a lot on
[21:59] on on this show because I know we we
[22:01] reach a an audience based on the
[22:04] response that I'm getting that might be
[22:05] different than a lot of my my viewers,
[22:07] which is people that have children that
[22:10] care about their children that are not
[22:11] just in the financial world.
[22:13] >> And the issue with that is for all kids,
[22:16] they need agency. Um, and that word to
[22:20] me is very powerful and I think it fits
[22:21] into uh a degree what you're saying. If
[22:24] you're a person at home, agency to me
[22:26] fits directly with empowerment. How do
[22:29] you use the tool yourself to create a
[22:32] business, to create a life? And
[22:34] curiosity will lead you down rabbit
[22:36] holes that you never expected, but
[22:37] you've created the idea combined with
[22:40] the interaction with AI. It's the reason
[22:43] why the most powerful people in
[22:45] artificial intelligence that I listen
[22:46] to, and I've run into them at
[22:49] conferences, and they're at all ages,
[22:51] and you probably fit in with this to
[22:53] some degree, although your your your
[22:54] kids are younger.
[22:56] I don't have free time anymore.
[22:58] >> And the reason is because I can have a
[23:00] conversation with the smartest person
[23:02] I've ever met and any question that
[23:05] enters my mind, I'm in the habit of what
[23:08] do you think about this? And by the time
[23:10] we're done, if it's a great idea, it
[23:12] might be in my video over the weekend.
[23:14] I'm spending more time trying to help
[23:16] people with using artificial
[23:17] intelligence to get to your point
[23:19] because they can create ideas that will
[23:20] be better than the job they have. And
[23:23] the world needs people doing that
[23:25] because it speeds up innovation. It
[23:27] speeds up curing health. When I first
[23:29] talked about Eli Liy and I mentioned uh
[23:32] something on Nolan podcast that Gavin
[23:34] Baker talked about about a hedge fund
[23:35] manager that basically spent all of his
[23:37] time to figure out how to help his child
[23:41] that had a disease
[23:43] and he found a medicine that existed
[23:46] that came back and I don't remember I
[23:48] don't think he got into specifics but
[23:49] let's assume it was something that was
[23:51] being used for something else but AI
[23:54] said this will probably work for that
[23:55] and it ended up helping his kids and I
[23:57] mentioned other people that have had
[23:59] mold issues and all kinds of different
[24:00] things that have ended up there. You
[24:02] have to understand that if you're not
[24:04] using artificial intelligence, number
[24:06] one, you're falling behind on
[24:07] everything. And that's the reason why
[24:08] now I've broken my my weekly to there's
[24:11] a signal. How do I avoid the noise? This
[24:13] is a bubble. This is bad. Okay, let's
[24:14] get through it and show you the facts on
[24:16] this. The alpha, okay, if you're going
[24:18] to invest, what should you invest in?
[24:19] Why are the hyperscalers not working?
[24:21] And why are they an issue on the
[24:22] spending side? And why are all these
[24:24] memory? What's the risk that the memory
[24:25] names fall? And then you've got the
[24:27] agency side. Once you start with AI and
[24:29] you start using it, you get more
[24:31] powerful. And I think you can end up in
[24:32] the same situation as the story you
[24:33] described.
[24:34] >> Now, uh Torsten over at Apollo, uh he
[24:37] has uh sometimes charts heard around the
[24:39] world. You know, it's like a canon. He
[24:41] puts the he puts it in, he packs in all
[24:43] the uh uh the powder and then he says 3,
[24:46] two, one, fire. uh today he fired one
[24:50] which is uh if you take out AI and
[24:53] energy stocks from the S&P the S&P 500
[24:55] is down. Now I jokingly said earlier
[24:57] well if you take all the water out of
[24:58] the ocean then there's no ocean right
[25:00] obviously. Um but I do think that there
[25:04] was even a day this week Ryan Dietrich
[25:06] pointed this out that like 428 of the
[25:09] 500 stocks sold off in the same day. One
[25:12] of the biggest sell-offs where you know
[25:14] majority sold off. And so we've been
[25:17] talking about this for a long time now,
[25:18] how there's weakness in a lot of other
[25:20] industries. Do you get worried that it's
[25:23] only AI and energy or are you like, no,
[25:26] this is what happens. There's sectors
[25:27] that get hot, everything else kind of
[25:29] cools as capital rotates and you know,
[25:31] this is more normal than people would
[25:32] think.
[25:34] >> Um, I mean, we watched this the prior 15
[25:38] years.
[25:39] The S&P 500 was driven by the Mag 7. So
[25:42] this is not something new. And this is
[25:44] what happens when innovation reaches the
[25:45] point where it's very concentrated. Um
[25:49] the the the funny thing is the Russell
[25:50] >> indexes work.
[25:52] >> So in mentioning the S&P a lot of times
[25:55] what I'll I'll look at is a combination
[25:57] of equal weight or the Russell 2000. I
[26:01] mean these are small cap stocks. The
[26:04] Russell 2000 made an all-time high this
[26:05] week. The S&P 500 did not. So the
[26:08] problem is for people with statistics,
[26:11] you can always find statistics that
[26:13] support your case. You can always find
[26:15] things that show up. Here's the reality.
[26:17] AI has exploded higher in a pace that no
[26:21] industry has ever seen. I believe it is
[26:23] incredibly disruptive. I believe it has
[26:27] been a problem for a lot of things, but
[26:28] I also believe that a lot of the S&P 500
[26:32] are companies that are based on the
[26:33] industrial revolution. And now we've got
[26:35] companies that are based on AI. The
[26:38] industrial revolution companies which
[26:40] make up the majority of the names. I I
[26:42] don't see how they're going to compete.
[26:44] It'll really get bad when crypto is
[26:46] working too because then crypto will
[26:48] have the same impact. So I have said
[26:49] before, I will continue to say it at the
[26:52] pace that we are going and the fact that
[26:53] we are closer to we are at RSI now. AGI
[26:57] is behind it. Super intelligence behind
[26:58] that. We will have missed the forecast.
[27:00] Everyone's moving their forecasts up on
[27:02] AGI. everyone including Deis Sabis who
[27:04] was the most skeptical
[27:06] I'd say realistic person someone who I
[27:09] actually believed the most he is not
[27:10] hyperbolic in any way so when he goes
[27:12] from I think it's in the early 2030s to
[27:14] I'm now thinking 2029 that's a big deal
[27:17] to move up the timeline that far for it
[27:20] to to shatter Liupold I have said 20 by
[27:24] 2030
[27:26] all companies that exist all large
[27:28] companies will have an issue with AI
[27:31] they will be disrupted in some way. If
[27:34] we don't need as much capex,
[27:37] well, then there's nothing left in the
[27:38] stock market. I I mean, you're
[27:40] commoditizing AI, which is the model
[27:42] companies are having trouble with that.
[27:45] If we don't need the capex, but AI is
[27:47] accelerating without the capex, what's
[27:49] going to happen to all the capex stocks?
[27:50] So, my belief has been for all investors
[27:53] that at some point here, it becomes an
[27:55] issue. We're not there yet. Earnings are
[27:58] growing, the stock market's up. I
[28:00] learned a long time ago as my first rule
[28:02] before I ever put a dollar into the
[28:04] market and I ever traded. I studied the
[28:07] Elliot wave theory for a reason. As long
[28:10] as the market is trending a certain
[28:11] direction, everything's fine.
[28:13] >> If the market has issues like breath and
[28:16] all this stuff, the S&P 500 will start
[28:18] to go down. But the S&P 500 is like
[28:20] Bitcoin is completely amorphous. It gets
[28:24] rid of the weak and it keeps the money
[28:26] with the strong. right now. AI are the
[28:29] strong at some point. I'm not sure all
[28:30] the AI companies are going to win.
[28:32] >> All right. Kevin Wars had his first big
[28:34] day in front of the bright lights. Uh
[28:37] started off hot. Didn't say good
[28:38] afternoon. Said good day. Everyone was
[28:40] freaking out about that. Um I don't
[28:43] know. I thought he did a pretty good
[28:44] job. Uh he didn't say a lot, which maybe
[28:46] was why he did a good job because he
[28:48] didn't have a big attack surface. What
[28:50] was your take on his press conference,
[28:51] some of the decisions they made, uh
[28:53] their focus on, you know, this task
[28:55] force? I think what we learned is he's
[28:57] going to be very different than Jerome
[28:59] Pal and he has said that um he has very
[29:02] different views. Uh I think the reaction
[29:06] by the media was just like the oil
[29:09] doomers and just like every other issue
[29:12] that has popped up. Tariffs are going to
[29:13] take down the country. Oh, he was
[29:15] hawkish. He wasn't hawkish. Um it's
[29:18] ridiculous. Uh tenure rates have been
[29:20] stuck in a range now since 2022. And in
[29:24] this weekend, I'm going to highlight to
[29:25] people the only reason rates went higher
[29:28] over the course of the last since 2021
[29:31] is because the Fed raised rates.
[29:34] Yes. We
[29:35] >> Funny how that works.
[29:36] >> Yeah. But that's it. Like everyone is
[29:38] worried about the debt. Everyone is
[29:40] worried they're going to go higher.
[29:41] Every time it ticks higher for a week,
[29:42] then everyone comes out and says the
[29:43] world's going to end. Then it ticks
[29:44] right back down. Two things that
[29:47] happened. Uh number one, he's a believer
[29:50] in productivity. He announced task
[29:52] forces. Now my kind of thought on the
[29:56] task force is he believes which I do as
[29:58] well and I think a lot of I think a lot
[30:01] of younger people non-academic and I
[30:04] consider myself non-academic because of
[30:05] much how much I hated school but I've I
[30:07] I I understand macroeconomics extremely
[30:10] well.
[30:11] I think the Fed has been in in the habit
[30:14] since 2009 of micromanaging everything
[30:17] >> every word every speech. It's
[30:19] ridiculous. At the end of the day, they
[30:22] do very little unless they have to come
[30:24] in and save something. That's it. Like,
[30:26] they're not moving rates much at all.
[30:29] They're going to cut 25. Nope. Now
[30:30] they're going to raise 25. Who cares?
[30:32] It's not going to change anything. It's
[30:33] a function where the market gets overly
[30:35] dramatic with it. So, the task force to
[30:37] me is, are we looking at this the right
[30:39] way? And I have said repeatedly GDP by
[30:42] itself is a horrible measure of an AI
[30:44] world. But let's use inflation. He has
[30:47] been very outspoken that he thinks the
[30:48] way we think of inflation and goes
[30:50] through it is wrong. He did not change
[30:53] and he won't change the target inflation
[30:56] rate. And the reason is cuz we're kind
[30:58] of trapped into that. If he came out and
[30:59] said, "I don't actually think the 2%
[31:01] means anything. We're going to go back
[31:02] to something else." But what he did say
[31:04] with inflation and what he said publicly
[31:06] is, I think the trimmed mean is a much
[31:10] better measure. And I completely agree.
[31:14] I believe in getting rid of these things
[31:16] that are way on the bottom, way on the
[31:18] top, and just kind of sticking with
[31:19] where the bulk of things are, using the
[31:22] median, and every single part of the
[31:24] median and trimmed mean, it's been it's
[31:27] been getting in a tighter range, meaning
[31:28] most things are kind of in this point of
[31:31] like 3%.
[31:33] And the extremes with oil and with all
[31:35] this stuff, it blew out. I think one of
[31:36] the most important things that's
[31:38] happened in the last
[31:40] month is as the news got worse on the
[31:43] straight of Hormuz, oil prices continued
[31:46] to move lower. We've seen no ships go
[31:49] through. Yet the ura price in New
[31:51] Orleans has collapsed back to lower than
[31:53] it was, meaning fertilizer.
[31:56] Somehow or another, the world had the
[31:59] most important thing shut down and
[32:01] nothing changed. How is oil ever going
[32:04] to sustain an up move at this point if
[32:06] it didn't just do it now? Now, maybe
[32:07] that'll change over the summer as we hit
[32:09] rock bottom, but I learned with prices,
[32:11] that's a warning sign. And the reason
[32:13] that matters is one-year inflation
[32:16] swaps, which are the only thing that has
[32:18] led the rise up in 2022 from the Fed. It
[32:22] led the entire time said, "Okay,
[32:24] inflation's going higher." It collapsed
[32:27] over the course of the last two weeks.
[32:29] So all the crap I gave to true
[32:31] inflation, even though we did get above
[32:32] 4% and they broke away from their CPI
[32:36] kind of guide, it stayed near what the
[32:40] median did and what the mean did. And
[32:41] now we have a Fed chair that's trying to
[32:43] create a task force to agree that okay,
[32:45] we shouldn't be watching this thing. I
[32:48] think that means actually at the end of
[32:49] the day when they focus on inflation
[32:50] that they're going to realize that it's
[32:52] actually lower than what we thought it
[32:54] would be. And if it couldn't go higher
[32:56] with that, how is it actually going to
[32:58] go higher when you strip out memory
[32:59] prices and you strip out all these
[33:01] one-off things?
[33:03] >> Are you saying that you think true is
[33:04] more accurate now
[33:06] >> or no?
[33:07] >> Well, again, when you say more accurate,
[33:10] >> let me rephrase. Are you using
[33:12] trueflation as a data point in anything
[33:14] you do or not really?
[33:16] >> This weekend, I'm going to show it
[33:17] overlaid with
[33:18] >> Oh, that's that's big.
[33:20] >> Yes. I don't think I've ever shown it in
[33:22] there. Um, but I'm also going to show
[33:23] how it deviate deviated from CPI. So, if
[33:27] if you were using it to forecast what
[33:29] would happen with inflation, it didn't
[33:31] work.
[33:32] >> If you were using it to say what the
[33:34] true level of inflation is that you
[33:37] should be using for where things will
[33:39] settle once you get rid of a disruption,
[33:41] it was far better than the headline CPI.
[33:44] >> But core CPI did the same thing. And I
[33:46] didn't think core CPI would go higher.
[33:47] The question is on this whole thing is
[33:49] CPI went up to 4.2. too. The next
[33:52] reading, which will come out in early
[33:55] July,
[33:55] >> you think up or down?
[33:57] >> It's going to be around zero.
[33:58] >> Yeah. Like zero 0% month- over-month
[34:00] >> growth. And that'll make the
[34:02] year-over-year down if that happens down
[34:04] to like four from 4.2. So, you'll have a
[34:07] peak in inflation, which will come back
[34:09] down. And and that was not not only not
[34:12] that, there were there were shows saying
[34:13] we're going to see double digit
[34:14] inflation. Double digit. I remember
[34:16] sitting on
[34:17] >> when listen on the the next CPI report.
[34:20] If that happens, everyone should just
[34:21] stay off the internet for the day. I'm
[34:24] going be uh you know what what's the
[34:26] saying? Uh uh trigger fingers turn to
[34:28] Twitter fingers.
[34:29] >> I I would be surprised
[34:32] if it doesn't come out close to zero
[34:34] just because gas at the pump has gone
[34:36] from 455 to below four now.
[34:38] >> Yeah.
[34:39] >> And that's one of the major drivers.
[34:41] >> All right, guys. Let's talk about
[34:42] something that's actually moving the
[34:43] needle in crypto right now. And that's
[34:45] my guys over at Arch Public. These guys
[34:47] are killing it. They continue to lead
[34:49] the way in everything that's got to do
[34:50] with Agentic Trading. And everyone knows
[34:52] this. Nothing else is being talked about
[34:54] in the market other than let your AI
[34:55] agents trade whatever asset you want.
[34:58] Now, with the addition of Arch AI, the
[34:59] market wave algorithm, and their tax
[35:01] harvest tool, they have turned this into
[35:03] a true one-stop shop for automated
[35:05] trading. They have now over 25,000
[35:07] users. That includes several corporate
[35:09] treasuries, asset managers, and
[35:10] businesses. And so the testing phase for
[35:12] Arch Public that is over. They have
[35:15] perfected how to automate advanced
[35:16] trading strategies without ever taking
[35:18] custody or asking you for your private
[35:20] keys. It's a big deal obviously. So Arch
[35:23] Public is an action speak louder than
[35:25] words type company. That's what I like
[35:27] to see. They give every user access to
[35:29] their platform absolutely free. You get
[35:31] to test it. You run the strategies. You
[35:33] watch the performance in real time. And
[35:34] then only after all of that do you put
[35:36] real money in to go put it to work when
[35:38] you're 100% comfortable. That's it. No
[35:41] pressure, no hidden fees, no gotchas,
[35:43] none of that nonsense. Go to
[35:44] archpubic.com right now and you can get
[35:47] started for free. Once you see how you
[35:49] can accumulate and manage crypto this
[35:50] way, you'll never trade the same again.
[35:52] archpubic.com. Go check them out today.
[35:57] Today's episode is brought to you by
[35:59] Blowfin. If you're actively trading
[36:00] crypto, then you already know the
[36:02] platform matters. Execution speed
[36:03] matters, liquidity matters, and
[36:05] reliability during volatile markets.
[36:07] That definitely matters. That's why more
[36:08] traders are starting to use Blowfin.
[36:10] They built a trading platform focused on
[36:12] fast execution, deep liquidity, advanced
[36:14] futures products, and a really, really
[36:16] clean user experience. They don't over
[36:18] complicate things. You can trade
[36:19] hundreds of spot and perpetual pairs.
[36:21] They integrate directly with Trading
[36:22] View, and they've built a platform for
[36:24] both active traders and people who are
[36:25] just getting started. Security is a
[36:27] major focus. Obviously, assets are
[36:29] custody with Fireblocks. They provide
[36:30] proof of reserves, and they've built a
[36:32] strong reputation for platform stability
[36:34] during those high volatility
[36:35] environments. And to celebrate the
[36:37] partnership with us, Blofin is now
[36:39] giving away $100,000. That's right.
[36:41] Blowofin is going to give away $100,000
[36:43] in deposit and trading rewards. So, if
[36:45] you're looking for a better place to
[36:46] trade crypto, go check them out today
[36:48] using our link in the description.
[36:50] Again, that's Blofin, the trading
[36:52] platform built for serious crypto
[36:53] traders, go sign up. Pay attention cuz
[36:56] they're giving away $100,000 in deposit
[36:58] and trading rewards. Now, here's one
[37:00] thing that I will say, and I I forget if
[37:01] we talked about this last week or not.
[37:03] Uh I think two things are true.
[37:04] Everything you're saying about
[37:05] inflation, I think, is right. There's
[37:06] kind of been this peak. I think the end
[37:08] of the war, I early on was like, hey, I
[37:11] don't think the war is going to be that
[37:13] long. I think that we probably hit a
[37:16] point where I could say I was wrong
[37:17] about that, right? Like it was longer
[37:18] than I thought it was going to be, but
[37:19] it's still not like it's like a
[37:20] three-year thing, but I was like, hey,
[37:22] if we're going to do this, let's do it
[37:23] in like 90 days and be out. So, it's I
[37:26] don't know how long it's been now. Maybe
[37:27] it's 150 days or or whatever, but it's
[37:28] longer than I I thought it was going to
[37:30] be. And so there's been a little bit
[37:31] more persistent uh pressure, but the
[37:33] fact that we're now kind of near the end
[37:35] and it's now rolling back over feels
[37:38] pretty good. And it wasn't that crazy
[37:39] inflation. Very similar to the tariff
[37:41] thing, right? Some areas you saw spikes,
[37:42] whatever.
[37:44] >> At the exact same time, what is also
[37:46] true, if you go on the streets of New
[37:48] York and you talk to people about
[37:49] grocery prices, gas prices, all this
[37:52] stuff,
[37:52] >> they are in pain. If you go to any city
[37:55] in America, everything is way too
[37:56] expensive. They still can't afford a
[37:58] home. they still can't, you know, uh,
[37:59] buy all the things they want, all the
[38:01] stuff. And so,
[38:03] I've actually maybe changed my mind a
[38:05] little bit on when I talk to people in
[38:08] the finance world, it's all about the 2,
[38:10] you know, up or down or this or
[38:12] whatever. The numbers don't matter.
[38:14] >> Mhm.
[38:15] >> People psychologically think everything
[38:16] is too expensive. And it goes back to
[38:19] the Bessant, you know, do you believe
[38:20] the economic data or not? What he said
[38:22] is, no, I believe the people over the
[38:23] data. What I think he really was saying,
[38:26] you know, in hindsight was the data
[38:28] doesn't matter because ultimately people
[38:30] are going to act the way they feel. And
[38:33] so the uh the data point now that I've
[38:37] been using for the last week in
[38:38] conversation, I've got young kids. My
[38:41] wife went on Amazon recently. She bought
[38:43] two boxes of diapers. For those of you
[38:47] young men and women who don't have kids,
[38:49] you could do a lot of diapers with kids.
[38:52] >> I don't know. when I should ask her,
[38:53] were they like superized boxes or some
[38:55] whatever, right? But she bought two
[38:56] boxes of diapers. $150.
[38:59] >> Yep.
[39:01] >> I don't care how many diapers are in the
[39:02] two boxes. If you're going through 8 to
[39:04] 10 diapers a day, right? That's
[39:07] unsustainable.
[39:08] >> And so you look and you say, like, by
[39:10] the way, like I don't think he's buying,
[39:11] you know, super organic, non-GMO,
[39:13] whatever crazy diapers. That is one tiny
[39:16] anecdote across an entire economy where
[39:18] people are saying to themselves, "This
[39:20] is nuts." And World Cup ticket prices,
[39:23] Knicks, you know, NBA Finals tickets,
[39:25] all of that stuff, I think, is the
[39:27] extreme end
[39:28] >> where you just have this massive
[39:30] divergence of the K-shaped economy.
[39:34] But you know, the crazy data point in
[39:37] 2006, 20 years ago, guess how many
[39:40] Americans were considered under the
[39:41] poverty line? 36.5 million.
[39:45] 20 years later, guess how many Americans
[39:48] are considered under the poverty line?
[39:50] 35.9.
[39:51] It's basically the exact same number.
[39:53] >> Mhm.
[39:54] >> 20 years later, now people celebrate
[39:56] because the poverty percentage has
[39:58] dropped because the overall population
[40:00] has grown, but it's only dropped by 2%.
[40:03] >> And so, you look at that and you say,
[40:04] there's 35 million Americans who live
[40:07] below the poverty line. the poverty line
[40:09] for those that don't know $15,000 a year
[40:11] in income. A family is $33,000.
[40:16] >> 35 million Americans. And so when Kevin
[40:18] Worse gets up and starts talking about
[40:19] inflation and all this stuff, it's like,
[40:21] dude, we Yes. As an investor, you got to
[40:23] pay attention, the average American,
[40:25] Kevin Worsh, Kevin, right? They're just
[40:27] like, "Look, man, groceries are
[40:28] expensive, diapers are expensive." And I
[40:30] think that's ultimately where you get so
[40:32] much uh debate in uh in society now.
[40:36] So, I I have a uh
[40:38] >> I hope you disagree.
[40:39] >> No, no, no. I I think I have a different
[40:41] um
[40:43] a different take. And
[40:44] >> well, that's called a disagreement,
[40:46] Jordy.
[40:46] >> No, no, no, no. Um so, I I think
[40:50] inflation is um is an excuse for for
[40:54] people at this point. Uh it's not that
[40:57] it's not real, but it's always real. So,
[40:59] let's go through it this way. Wages are
[41:00] growing right now at 3.6% and inflation
[41:03] is four.
[41:04] >> So, not good. That's not good. That
[41:06] means you can't keep up with what's
[41:07] going on. Uh I think the bigger issue
[41:10] and the biggest issue and I hear this
[41:12] repeatedly from young people. So when I
[41:14] say young, because I live in
[41:16] Williamsburg, so I knew who was going to
[41:17] win the mayor the the election in in New
[41:20] York.
[41:20] >> I know you
[41:21] >> as opposed to being in Manhattan. Um uh
[41:26] that group of people.
[41:27] >> I'm going to be walking the streets of
[41:28] the Bronx this weekend, Jordy.
[41:31] Um, I've never seen young people more
[41:34] dissatisfied with the work environment
[41:35] than now.
[41:36] >> And here's what I think it's feeling.
[41:38] And this is a word that everyone hate
[41:40] hates. Um, Bruce Springsteen sang about
[41:44] this, trapped. There's no worse feeling
[41:46] for a human being than feeling trapped.
[41:49] >> If they're trapped at a job they can't
[41:50] get out of, you used to be able to go
[41:52] interview. And sometimes what would
[41:54] satisfy people enough is knowing they
[41:56] could get a job with a raise.
[41:57] >> Yeah. But if nobody is going to hire you
[41:59] for more money or the probability of
[42:01] finding it, you're trapped. If you're
[42:03] trapped in a bad relationship and you're
[42:05] trapped, if you're trapped in a city
[42:07] that you can't get out, if you're a
[42:10] Democrat and you're trapped in a
[42:12] Republican world, if you're a Republican
[42:13] and you're trapped in a Democrat world,
[42:15] I just think these have been amplified
[42:17] because of technology
[42:19] >> and it has coincided with the post
[42:22] period of the great financial crisis. So
[42:23] the great financial crisis took the
[42:25] unemployment rate up to 10. The iPhone
[42:28] comes out. So there's a grace period. We
[42:30] went from 10% all the way down to three
[42:32] and change. It took a decade. I don't
[42:34] think people realize that. If people are
[42:36] listening, we were at 10% unemployment.
[42:38] It took us literally a decade to get
[42:41] back down there. Now during that time,
[42:43] you could become a Door Dash driver. You
[42:45] could become an Uber driver. Like none
[42:46] of that stuff existed before the great
[42:47] financial crisis because we didn't have
[42:49] the iPhone. So I think there was a grace
[42:51] period for people. Then COVID happens
[42:53] and the reality sits in. We dump a ton
[42:56] of money on people. They yolo it for a
[42:58] little while. But then 2022 happens. We
[43:00] move rates higher. Insurance goes
[43:02] higher. And all of the money that we
[43:04] gave to people that they spent on cruise
[43:06] lines and traveling through Europe,
[43:08] well, they didn't work. That was meant
[43:10] to offset the inflation that comes.
[43:12] There's a direct relationship between
[43:14] this is money because we're shutting
[43:16] down the economy. We know it's going to
[43:17] create inflation. So don't spend it all
[43:20] now because you're going to need it. and
[43:21] then we reset the bar and then the job
[43:23] situation was worse. That is what I
[43:25] think happened and AI happened. When did
[43:26] it happen? 2022,
[43:29] right? When rates went higher. So, I
[43:31] just think there's a chain of events
[43:32] here that have left people um angry and
[43:35] they're blaming inflation. They're
[43:36] blaming politics. They're blaming a
[43:38] bunch of stuff. The reality is uh I
[43:40] believe in empowerment and that gets
[43:42] back to that agency side. You have a
[43:43] chance to make more money. you have a
[43:45] chance to not have to work for someone
[43:47] or make money on your own, just like you
[43:48] did as an Uber driver, but you have to
[43:50] use AI and you have to embrace it. And
[43:52] if you keep saying AI is a bubble and I
[43:54] don't want to use it,
[43:55] >> good luck.
[43:55] >> You're just being a victim then.
[43:56] >> Yeah. All right. So, we got two things
[43:58] real quick before uh uh we got to go.
[44:00] But uh first is Nick's parade happened
[44:01] this uh uh this week. Um my number one
[44:05] takeaway was New York forever. I mean
[44:08] that this is the most New York thing of
[44:10] all time. My second takeaway, do these
[44:12] people not have jobs? There's two
[44:14] million people in lower Manhattan right
[44:17] now. We gave our entire office the day
[44:19] off. And I said, "You want to go to the
[44:20] parade? Knock yourself out. You don't
[44:21] want to go to the parade, that's fine.
[44:22] You get the day off." Very unlike me to
[44:24] normally do this. But I said, "Okay, I
[44:26] at least know our people have the
[44:28] opportunity to go because we're giving
[44:29] them the day off."
[44:31] >> 50% of the people there, they playing
[44:32] hookie. Like I there's two million
[44:34] people, right? I was like, "What the
[44:36] hell is going on here?" Um, I saw one uh
[44:38] one image was a big picture of a huge
[44:40] crowd and it said uh for those of you
[44:42] watching this from your office on
[44:44] Instagram right now, just know your tax
[44:45] dollars are supporting these people. I
[44:48] don't think that's actually true, but
[44:49] that was the sentiment. That's first.
[44:50] Second, Bitcoin uh has actually been
[44:53] somewhat weak in the last two or three
[44:55] uh weeks. What's your take as to what's
[44:57] going on with Bitcoin?
[44:58] >> Um, so
[45:00] first of all, it's in a bare market. I'm
[45:02] going to say it every single day until
[45:03] it's not in a bare market. When we break
[45:04] above the 200 day moving average, it
[45:06] will change. Until that happens, it is a
[45:08] bare market. Every single time it runs
[45:10] into any moving average, the most recent
[45:12] one was the 20 moving average, it falls
[45:14] back down. Uh you mentioned the money
[45:18] going into AI
[45:20] when SpaceX happened this week and it
[45:24] became obvious.
[45:26] SpaceX and Bitcoin are basically to me
[45:29] the same thing. Like SpaceX is a future.
[45:32] >> They're both going to the moon. Well,
[45:34] they're both based on a belief that
[45:36] people have in the future. They're
[45:38] they're not about anything fundamental
[45:40] right now. Like, it's very hard to make
[45:42] an argument as to the fundamentals
[45:44] behind this. I love when people send me
[45:46] things and they're going, "You're wrong.
[45:47] Space S is overvalued." I'm like, "No."
[45:49] When something doesn't have a valuation,
[45:51] it can't be overvalued or undervalued.
[45:53] It has no valuation. It's a complete
[45:55] guess. We're talking about flying to the
[45:57] Mars. We're talking about flying to the
[45:59] moon and building space stations. Like,
[46:00] this is a this is a dream. So, Bitcoin
[46:03] has no energy because it is a vehicle
[46:07] meant for two things. One is for people
[46:09] to hide their money from the government,
[46:11] but that is always going to be a small
[46:12] portion of it because the wealthiest
[46:14] people on the planet, which is where it
[46:15] draws its money from, aren't hiding
[46:17] their money from anyone at this point
[46:19] because they're not taking it. On the
[46:20] other side is the energy that comes from
[46:22] retail and the energy that comes from
[46:24] momentum. It has no momentum last right
[46:27] now. It every week there's more people
[46:29] now. You've got uh strategy has brought
[46:32] people's attention. They're looking at
[46:33] the way STRC is trading and it's down to
[46:37] every everybody has a viewpoint on it.
[46:40] I'm going to keep saying the same thing.
[46:43] It is very difficult for Bitcoin to be
[46:45] traveling higher if all the money is
[46:48] going into stuff that is based on
[46:49] earnings.
[46:50] >> Mhm.
[46:50] >> We need and we're running into the
[46:52] second quarter. I believe the second
[46:54] quarter will be more of a disappointment
[46:56] for earnings than the first quarter. Not
[46:58] that it won't grow, but now the
[47:00] expectations are so high.
[47:02] >> 22% earnings.
[47:03] >> Exactly. So if if the stock market is
[47:05] unchanged from now until September,
[47:08] well, that's a better environment for
[47:09] Bitcoin than one where AI continues to
[47:11] go at 50% a month or 50% a quarter
[47:14] because if it keeps doing that, I do
[47:17] believe like you said six weeks ago,
[47:19] you're talking to people in Korea. It's
[47:21] like they used to be a big player in
[47:22] Bitcoin. Not there. Retail will migrate
[47:25] to what's working. And right now,
[47:26] Bitcoin is in a bare market. We need to
[47:28] see it change. And I'm gonna emphasize
[47:29] this point again and again.
[47:32] Never believe you're smarter than the
[47:33] market. Ever. The market knows. I had a
[47:36] day on Friday in my own or Thursday on
[47:39] my own portfolio. I finished up on the
[47:42] day.
[47:43] I have 20 stocks that I own including
[47:46] Bitcoin. 18 of the 20 were down. Two
[47:49] were up. Those two, Marll and
[47:52] Entrogress, they made up for everything
[47:55] of the other ones. But I silver was
[47:57] down, Bitcoin was down, Eli Liy was
[48:00] down. All of these things that we've
[48:01] talked about over the time, they were
[48:02] all down. When I'm hoping Bitcoin and
[48:06] these other things are working, it's
[48:08] when there's a pause in all of the AI
[48:10] stuff, but we're not there yet.
[48:12] >> I don't think that uh anything you're
[48:13] saying is crazy. I actually agree. I
[48:16] like Bitcoin here. We'll see what
[48:17] happens. um your video this weekend. If
[48:21] you get any value whatsoever out of
[48:22] Jordy, go to Jordy Visser YouTube on
[48:25] your little Google machine, AI machine,
[48:27] whatever. Go hit the subscribe button.
[48:28] It's a digital thank you, a little
[48:30] handshake. Dap them up, tell them I
[48:32] appreciate all the information. And uh
[48:33] we'll do this again next week. See you
[48:35] next week.

← Volver al listado de vídeos

Scroll al inicio