📊 Portafolio público de inversión
📈 Empresas seleccionadas por sus fundamentos
🔗 Consulta mi portafolio y las posiciones que mantengo en el siguiente enlace 👇
https://t.co/QDlgi5NvHB
For me, $NOW. I prefer its cleaner balance sheet and more asset-light model. $ORCL is growing fast, but its much higher leverage and capex intensity make me more cautious.
📊 Semana del 21 al 25 de septiembre — Agenda de mercado
🔘 Lunes 21/9: inicio tranquilo en EE.UU., con solo el índice de actividad nacional de la Fed de Chicago (agosto). El mercado sigue digiriendo la primera subida de tasas de la Fed desde 2023. Fuera de EE.UU., el Banco Popular de China anuncia su decisión de tasas.
🔘 Martes 22/9: Fed de Richmond (manufactura de septiembre) y subasta de US$78.000 millones en notas a 2 años. Antes de la apertura reporta $AZO (AutoZone) y tras el cierre $KBH (KB Home), útil para medir cómo afectan las hipotecas caras al sector vivienda.
🔘 Miércoles 23/9: PMI preliminares de septiembre en EE.UU., zona euro, Alemania y Reino Unido, la primera lectura de actividad tras la subida de la Fed. Subasta de US$70.000 millones a 5 años. Antes de la apertura reportan $PAYX (Paychex), $CTAS (Cintas) y $GIS (General Mills).
🔘 Jueves 24/9: solicitudes de desempleo, ventas de viviendas nuevas de agosto y Fed de Kansas City. Subasta de US$44.000 millones a 7 años. Antes de la apertura reporta $DRI (Darden Restaurants) y tras el cierre $COST (Costco), un buen termómetro del consumo.
🔘 Viernes 25/9: órdenes de bienes duraderos de agosto y lectura final de confianza del consumidor de la Universidad de Michigan, con foco en las expectativas de inflación a 1 y 5 años.
En resumen: semana con poca data pero con más de diez intervenciones de miembros de la Fed. Tras subir a 3,75%–4,00%, el dot plot apunta a una subida más antes de fin de año, y el mercado está dividido entre octubre y diciembre. Con el Treasury a 10 años cerca del 5% y tres subastas de deuda en agenda, las tasas largas seguirán marcando el tono para las tecnológicas. La semana siguiente llegan el PCE de agosto (30/9) y los resultados de $MU (Micron).
📊 My week in the markets — September 14–18
Main portfolio moves:
⚡ $BE — Took partial profits after the strong run while keeping a meaningful position.
☁️ $NBIS — Rebalanced the position and took partial profits, reducing its overall portfolio weight.
🔎 $GOOG — Continued building the position with additional staggered entries.
⚡ $VST — Kept adding gradually during the pullback.
🟢 New positions: $NU, $NOW
🟠 Reduced: $MRVL
🔴 Closed: $IREN
💵 Cash: ~19% of the portfolio
📌 Current portfolio: $BE $NBIS $META $VST $AAPL $RKLB $STRL $GOOG $MRVL $NU $NOW
This week I focused on reducing concentration, taking some profits, and increasing liquidity so I have capital available for new opportunities.
Not financial advice.
Why is the market rising today if the Fed RAISED rates yesterday?
Because markets are forward-looking.
Yesterday’s 25 bps rate hike was already largely priced in. Once the decision was confirmed, part of the uncertainty disappeared.
Two other factors are also helping stocks today:
🔹 Long-term Treasury yields are falling, which is especially supportive for technology and growth-stock valuations.
🔹 Oil prices are pulling back, slightly easing inflation concerns.
At the same time, the U.S. economy continues to show resilience, with solid consumer spending, productivity and business investment.
⚠️ This does not mean the Fed has suddenly turned dovish. Further rate hikes are still possible.
$TSLA $NVDA $AMZN $GOOGL $MSFT $AAPL $CSCO $AMD
Last month I wrote about how we can build a positive and safe future for everyone: https://t.co/eoLGVY8yad
Every lab has the responsibility and incentive to move at the pace required to train its models safely, and the ability to take its own actions to ensure that happens.
The reality is:
- People won't want to use agents that are misaligned with them and that don't do what they ask, so labs have a strong natural incentive to make their models more aligned.
There is a lot of debate about slowing progress on capabilities until alignment catches up. My view is that trust and alignment are quickly becoming the most important capabilities that will differentiate agents and models. Any lab that doesn't focus on alignment will fall behind.
- Labs face significant liability if their models cause harm, so they have a strong incentive to prevent this as well.
Meta delayed shipping Muse for several months to focus on safety and security. We didn't call for everyone else to do this before we would. We just did it as part of our day-to-day work because it was clearly the right thing for people and for us. I'm proud of the security foundations we've built.
- Engaging independent evaluators and advisors is industry best practice. MSL already does this today in several areas because it helps produce better work. Other labs can just do this too. In general, it would be helpful for there to be a larger and more diverse ecosystem of evaluators.
- Committing the significant majority of compute towards serving people rather than racing towards recursive self-improvement is one of the best ways to ensure we develop this technology safely. Meta has made this commitment and other labs can do this as well.
I believe the key to building a positive future for everyone is maintaining the right balance of power. This is within our power to do.
🚨 The Fed HIKED rates for the first time since 2023: +25 bps to 3.75%-4.00%, unanimous.
And it may not be the last: 16 of 18 members see at least one more hike in 2026 (4 of them see two).
The cutting cycle is over… for now.
This reinforces my conviction in $VST. I’ve been building a position around the current levels, and acquiring 5.5 GW of operating capacity at roughly $730/kW looks highly accretive—especially as power demand continues to rise.
The key now is execution and how efficiently management integrates the assets, but the risk/reward remains compelling.
$VST Vistra is buying power plants for less than half their implied value:
-Vistra is acquiring 10 operating natural gas plants through Cogentrix for approximately $4 billion
-Those plants provide 5.5 GW of power capacity, meaning Vistra is paying roughly $730 per kilowatt
-Building the same amount of new natural gas generation would cost roughly $1,500-$2,000+ per kilowatt before accounting for the land, permits, fuel access, and grid interconnections
-Elon Musk said consensus estimates suggest roughly 15 GW of AI compute scheduled for 2027 may not have enough electricity to turn on because power generation and grid connections cannot be built fast enough
Owning secured power is owning the single most important bottleneck in the AI infrastructure buildout.
Long $VST Vistra.
The Fed decides tomorrow, and the market seems to have already reached its verdict: roughly an 88% probability of a 25 bps hike on Polymarket and 87% on CME FedWatch.
That would lift the target range from 3.50–3.75% to 3.75–4.00%—the first rate hike since July 2023.
But with the move largely priced in, the decision itself may not be the main catalyst. The real focus will be on the dot plot, the statement, the vote split, and Warsh’s tone.
If the dot plot signals more than one additional hike this year, yields could reprice higher, putting pressure on tech and semiconductor valuations: $MU $MRVL $GOOGL.
Palo Alto ya estaba muy posicionada en seguridad de agentes con Prisma AIRS.
Ahora CrowdStrike entra de lleno con Falcon Guardian, llevando su fortaleza en endpoint al runtime de los agentes.
La batalla por asegurar la IA agéntica se pone mucho más interesante. 👀
$CRWD $PANW
📊 Semana del 7 al 11 de septiembre
S&P 500 −0,8% | Nasdaq −0,7% | Russell 2000 −2,4%
El mapa deja una señal clara: los semiconductores no se movieron en bloque.
$NVDA cayó 4,45%, mientras:
• $AMD +13,15%
• $MRVL +13,06%
• $INTC +12,29%
• $MU +1,78%
$DELL también destacó con +9,86%.
La semana anterior escribí sobre el regreso del dinero a la infraestructura de IA. Esta vez, el matiz importa: estar en el mismo sector no garantiza el mismo resultado.
Entre las grandes tecnológicas también hubo diferencias:
$META +6,12% · $AAPL +1,24%
$MSFT −2,84% · $GOOGL −1,16%
Y el software siguió bajo presión:
$PLTR −8,38% · $NOW −8,97% · $CRM −6,32%.
Mi lectura: hablar de “la tecnología” como una sola apuesta explica cada vez menos. La selección de empresas está marcando la diferencia.
Dario Amodei, CEO de Anthropic, propone moderar el avance de los modelos de IA más poderosos para que la alineación, las auditorías externas y la ciberseguridad puedan seguirles el ritmo.
Anthropic dará a evaluadores independientes acceso permanente a sus sistemas para verificar sus medidas de seguridad y evaluar la alineación de los modelos durante el entrenamiento.
Esto abre una posible tesis de inversión: la seguridad de la IA podría convertirse en una ventaja competitiva y una fuente creciente de gasto tecnológico.
Empresas cotizadas para seguir:
🔹 $PANW — Prisma AIRS protege modelos, datos, aplicaciones y agentes de IA.
🔹 $MSFT — Azure AI Content Safety ofrece controles para detectar contenido dañino y ataques mediante instrucciones maliciosas.
🔹 $AMZN — Amazon Bedrock Guardrails permite incorporar filtros y controles de seguridad a aplicaciones de IA.
🔹 $GOOG — DeepMind aplica un marco de evaluación y prevención de riesgos para modelos avanzados.
La conexión comercial es distinta en cada caso: desarrollar IA con controles de seguridad no implica necesariamente generar ingresos adicionales por vender protección.
¿Hasta qué punto esta necesidad se convertirá en un negocio rentable?
Today I'm publishing a new essay, Policy on the AI Exponential. AI is progressing extremely fast—much faster than the policy process was built to handle. The essay lays out where I think the technology is now, and the action needed to close the gap: https://t.co/Lh6PWae178
The world’s 10 richest people right now 💰
Elon Musk sits at #1 with a staggering $915.8 billion, more than 3× the fortune of #2 Larry Page.
Which name on this list surprises you most? 👀
I cap my own portfolio at 15 names, ideally 10. Not because concentration is magically better — it cuts both ways — but because I can only genuinely follow around 10 businesses.
With 23, your best idea is already diluted. Before adding anything, I'd sell the 7 or 8 I trust the least.
🚨 MY FULL PORTFOLIO 👇🏻
I have $5,000 cash ready to deploy into my 23-holding portfolio.
If you were managing this:
- Which holdings are you adding $5k to right now?
- Which position are you trimming or selling?
🔥 Drop your picks below 👇🏻
$ZETA – Zeta Global Holdings Corp
$META – Meta Platforms Inc.
$MA – Mastercard Inc
$MSFT – Microsoft Corporation
$AMZN – https://t.co/Pk9Y7K8q1J Inc.
$CRM – https://t.co/KINHYZTPTv Inc.
$SOFI – SoFi Technologies Inc.
$CELH – Celsius Holdings Inc
$ELF – e.l.f. Beauty Inc
$ADP – Automatic Data Processing Inc.
$NOW – ServiceNow Inc
$INTU – Intuit Inc
$SPGI – S&P Global Inc
$ACN – Accenture plc
$UBER – Uber Technologies Inc
$NFLX – Netflix Inc
$VICI – VICI Properties Inc
$TSM – Taiwan Semiconductor Manufacturing Co.
$NVO – Novo Nordisk A/S
$UNH – UnitedHealth Group Inc.
$ROL – Rollins Inc
$RBRK – Rubrik, Inc.
$GOOGL – Alphabet Inc. Class A
The most interesting part of that chart is what nobody is reading: if $314B of the $319B comes from the owned fleet, JPMorgan is saying $TSLA has to become Waymo.
Buying, operating, insuring and maintaining cars, city by city.
$GOOGL already does exactly that: ~500k paid rides a week across 10 cities.
JPMorgan sees $TSLA robotaxi revenue reaching ~$320B by 2035 but nearly all of it comes from Tesla owning + operating fleet rather than the owner network.
That puts the whole opportunity on how fast Tesla can scale Cybercab & clear regulators but prize is enormous if it works.
📊 My week in the markets — September 8–11
What moved my portfolio:
🔋 Bloom Energy — S&P Dow Jones confirmed its addition to the S&P 500 on September 21, replacing Molson Coors. UBS raised its price target from $300 to $325 and Clear Street from $290 to $330. The stock went from ~$206 on 8/31 to $252 on 9/04. I used the move to take partial profits.
☁️ Nebius — announced a strategic partnership with Palantir on September 8 for sovereign AI infrastructure, and the stock jumped more than 6% that day. I also took partial profits.
🧠 Marvell — Piper Sandler initiated coverage with a $270 target and the stock closed Friday +5% at $237, backed by its growing role in custom AI silicon. Investor Day is October 6: that's where the thesis gets tested.
🚀 Rocket Lab — a week full of news: Berenberg initiated coverage with a Buy and an $83 target, Raymond James with Outperform, and the company launched its 16th Electron of the year.
My moves this week:
🟢 Bought: GOOGL (3 entries), VST (2 entries), RKLB
🟡 Partial profits: BE, NBIS
🔴 Closed: JNJ
📌 My portfolio: $MRVL $BE $NBIS $RKLB $META $AAPL $STRL $VST $IREN $GOOGL