Torch Capital

Investing in tech-enabled companies that transform how consumers and businesses operate.

[Torch Capital @torch]

Three years ago the assumption was that smarter models would need less data. The opposite happened. Frontier labs now spend an estimated $10-15B each on it.

@umachalik broke down what that money is buying, and how long it lasts.

[Uma Chalik @umachalik]

https://t.co/wp6YQUn2C0

[Uma Chalik @umachalik]

https://t.co/wp6YQUn2C0

[Torch Capital @torch]

Anam just launched cara-4, its latest real-time avatar model. Avatars can now adapt expressively throughout live conversation. Independent testing ranked cara-4 no.1 vs Tavus, HeyGen, and LemonSlice. Incredible milestone from the team and a big step forward for the category.

[Torch Capital @torch]

@lmstudio @yagilb

[GREG ISENBERG @gregisenberg]

The takeaway from Fable 5 being BANNED by the government: GET GOOD AT LOCAL MODELS SO YOU HAVE 100% CONTROL.

My entire weekend was going to be building my craziest ideas with Fable 5. That's now cancelled.

So instead of building with Fable this weekend, I've decided I'll go deep on local models:

1. Start with the runtime. Download Ollama or LM Studio first. This is the thing that actually runs models on your machine.

2. Match the model to your hardware. A model's size is measured in billions of parameters (7B, 32B, 70B). Bigger is smarter but needs more memory. Rule of thumb: a 7B model runs on almost any laptop, a 32B needs a good Mac with 32GB+ RAM, a 70B needs serious hardware like a DGX Spark or a maxed-out Mac Studio.

3. Know which model for which job. Qwen 3 is the best all-around choice for most tasks. DeepSeek for reasoning and coding. Gemma 4 when you need something tiny that runs on a phone. Llama when you want the biggest community and the most fine-tunes.

4. Quantization. You can shrink a model to run on weaker hardware with barely any quality loss. Look for versions labeled Q4 or Q5. This is how a model that "needs" a server runs on your laptop. Learning this one concept changes everything.

5. Connect it to your agent. Point Hermes or your agent stack at a local model.

6. Context window is your real constraint locally. Cloud models give you huge context for free. Local models make you pay for it in memory. A bigger context window eats RAM fast. Keep your sessions tight and your prompts lean or your machine chokes.

7. Learn to give local models tools. A smaller local model with web search, file access, and code execution beats a giant model with none. The capability gap closes fast when you wire up the right tools. The model is the engine but the tools are the wheels.

8. Fine-tuning is more accessible than you think. You don't need this on day one, but know it exists. You can take an open model and train it on your own data so it gets good at your specific domain.

I'll probably do a breakdown at some point on this @startupideaspod if people are into it.

The lesson from this ban is basically don't build your entire workflow on something that can disappear with a single letter. Own part of your stack. Local models are insurance.

It reminds me when people realized they don't own social media accounts. And then you saw people build email lists etc.

I remember running a startup and my biggest traffic source was organic FB. All of a sudden, algo changed, and I lost 99% of my traffic.

Same sorta moment (but bigger) for AI.

This is a wake up call.

[Torch Capital @torch]

F2 raised $24M in funding. Purpose-built for private credit and commercial banking, 100+ funds and banks are live in 10 months, including Jefferies, Bain Capital, and Live Oak Bank. MAUs up 650%. Deals underwritten are up 10x YTD. Proud to be pre-seed and seed investors. @F2aiHQ

[Torch Capital @torch]

Great to see @ZReitano on @FoxBusiness talking about the Wegovy oral pill and the future of telehealth. Big moment for @ro .
https://t.co/0cbGp9715T

[Torch Capital @torch]

Incredible milestone for @solacehealth_ - $130M Series C at a $1B valuation led by @IVP 💥 Proud to co-led the Seed and back a team building key infrastructure for patient advocacy. Huge validation of the mission, the scale of the problem, and what’s ahead
https://t.co/jFDEY16QPb

[Z Reitano @ZReitano]

https://t.co/bnxvBTWk45

[LM Studio @lmstudio]

Introducing LM Studio 0.4.0 🎉✨

It's the next generation of LM Studio.

🪄 Deploy on servers, in CI, or anywhere
🚄 Parallel requests for high throughput use cases
🔨 New stateful REST API: use local MCPs
🎨 Complete UI revamp

See what's new in this release👇🧵

[Torch Capital @torch]

Proud to have backed @SalvoHealth early and excited to see the team close a strong $8.5M Series A. Real outcomes, real scale, and a clear path forward for GI care. Onward 🚀 https://t.co/H30NtSMuu7

[Torch Capital @torch]

The ongoing shift from passive Systems of Record to active Systems of Agency will be profound. This analysis outlines the companies building the patient-first stack, and the traits of the inevitable winners within Consumer Health. https://t.co/JLOnLatX6b

[Torch Capital @torch]

Excited to share that Jake Moross was named to Alpha Partners’ 2025 Rising Star Pre-Partner VC list.

Well deserved recognition. We’re lucky to have him on the team and excited to watch what’s next.

https://t.co/DTmUGV60uz

[Amy Wu Martin @amywumartin]

I’ve been chatting with young people trying to break into VC so sharing what I generally advise:

- To make the best impression, don’t ask VCs for coffee, send them *good* deal flow instead
- Bad deal flow is worse than sending nothing. If you don’t know the difference, absorb more content from great founders
- Build your founder network authentically. If you’re in college it’ll prob be with other student founders
- Few firms hire straight from college. Don’t feel bad if you don’t get the job. Go get experience at a fast-growth startup in any role, FAANG, or found a company first. Learn what it feels like to work with the best
- Get fluent with the latest AI tools. Build, ship projects and write about your learnings
- Unsure if VC is for you? I always say the closest feeing role at a startup is the sales rep
- If you’re young and risk-on, it’s a great market environment to be a founder (vs investor). Take the plunge!

[Torch Capital @torch]

🚀 Big milestone for @takearecess closing a $30M Series B led by CAVU Venture Partners.

Recess has pioneered a new category of relaxation + wellness drinks, offering functional beverages and nutraceuticals that help people feel cool, calm, and collected.

https://t.co/hKqxKDvdao

[Torch Capital @torch]

Excited to share that Liftoff, the hiring & networking platform that helps you find incredible people through trusted connections, launched out of stealth. Even in an AI world, the best opportunities still come through people you trust.

👉 Check it out: https://t.co/3CsG5KOOfc

[Torch Capital @torch]

Excited to share that @LittleOtterCare and @hazelhealthinc have merged! Together, they’re breaking down silos and building an integrated ecosystem that connects schools, families, and payers to deliver better mental health outcomes for kids.

https://t.co/cfg6uv7jRm

[Torch Capital @torch]

Big congrats to @qihealthapp and @edwardchiu on an incredible launch, and exciting start to what's ahead. Qi is building the world's first Social Health Network. Check it out here 👉 https://t.co/j4wfecjoxb

[Edward Chiu @edwardchiu]

Woke up and we're officially the #2 health app in the nation...😭🙏

Qi: The world's first Social Health Network

[Edward Chiu @edwardchiu]

Everyone has a health story. The late-night searches. The trial and error. The fear you might never get better.

The hardest part isn’t the symptoms.

It’s feeling like you’re in it alone.

Introducing Qi Health👇

[Brett Adcock @adcock_brett]

Another massive week of AI and robotics news.

I summarized everything from OpenAI, xAI, Google, Cognition, Figure, AWS, Meta, UBTECH, Pollen Robotics, Agility Robotics, and more.

Here's everything you need to know and how to make sense out of it:

[Z Reitano @ZReitano]

1/ @serenawilliams is our newest ambassador at @Ro. She’s sharing her weight loss story to inspire others. She has more willpower than most people on the planet and is the perfect example that weight is not just a matter of “diet and exercise”.