Andy Martínez

AI & Cloud GTM | Enterprise AI | Building partnerships across North America

[Andy Martínez @h1andyc]

@AmericaAeterna @TokenGremlin ever heard of windsurf?

[Andy Martínez @h1andyc]

@ns123abc total bs. block now

[Pedro Sánchez @sanchezcastejon]

La solidaridad y la empatía son opcionales.

El respeto a los tratados europeos y a los datos, no.
👇🏼

Número de entradas irregulares entre 2021 y 2026 según Frontex:

Por Italia: 478.600
Por los Balcanes occidentales: 340.600
Por Grecia: 259.800
Por España: 234.760
Por la frontera oriental: 48.000

No es tiempo de dividir. Es tiempo de seguir construyendo una UE fuerte y unida.

[Jensen Huang @JensenHuang]

For my first post, I’m sharing a letter @NVIDIA signed on why open models matter.

AI will transform every industry, power every company, and be built by every country.

Open models strengthen safety and cybersecurity, accelerate innovation and diffusion, and enable sovereignty.

The world needs both frontier closed models and frontier open models.

https://t.co/AUKzoQ5Ikb

[Mayor Zohran Kwame Mamdani @NYCMayor]

Benjamin Netanyahu is a war criminal.

[Sam Altman @sama]

there are a lot of benchmarks that suggest 5.6 sol is the best model in the world right now, but the most reliable way to tell is that elon is obsessed with me again

[Sam Altman @sama]

@iliketeslas i am not afraid of apple, but i have tremendous respect for them. s-tier company.

[Sam Altman @sama]

homeboy you're the one sellling public market investors on short-term space datacenters

[Andy Martínez @h1andyc]

@gelunding 怎么什么弱智都会出现在我的 fyp

[Andy Martínez @h1andyc]

talked to 4 AI teams this week. every single one complained about the same thing: inference costs they didn't budget for. the model was the easy part. keeping it running at scale is where the money disappears.

[Andy Martínez @h1andyc]

Hot take: most AI startups will fail because of infrastructure, not because of bad models.

The model is 20% of the work. Serving it reliably at scale is the other 80%.

[Andy Martínez @h1andyc]

Every AI team I talk to has the same gap:

→ Model works great on a single A100
→ Falls apart at 100 concurrent users
→ Nobody budgeted for the infra to bridge that

I've seen this pattern kill more launches than bad models.

[Andy Martínez @h1andyc]

3 things killing AI teams right now:

→ GPU quotas that take weeks to approve
→ Spot instances disappearing mid-training
→ Cloud bills that double every quarter

Compute access is the new bottleneck. And nobody talks about it during fundraising.

[Andy Martínez @h1andyc]

I talk to AI teams every week who spent 6 months building a model and 2 days thinking about deployment.

Then they wonder why inference costs eat their margin alive.

[Andy Martínez @h1andyc]

"We'll figure out deployment later."

This sentence has mass-produced more failed AI products than bad training data ever will.

[Andy Martínez @h1andyc]

@_The_Prophet__ This doesn't show up in earnings slides. Capex doubles, but AI revenue hasn't caught up. The companies spending most on AI infra right now have the widest gap between investment and monetization. That math has to close.

[Andy Martínez @h1andyc]

@EmmanuelInvest See this every week on calls. Teams budget for GPUs and completely forget about HBM and interconnect costs. Then the real quote comes in 3x higher than planned. The "AI infra" line item is never just compute anymore.