boldstart ventures

This account is verified because it's an official organization on X. Learn more

@Boldstartvc

the inception fund. helping the bold start. first check for technical founders. Cybersecurity, AI infrastructure, physical AI, agents, the weird. since 2010

Miami, Boston, SF, NY Inception.vc Joined September 2010

665 Following 9.1K Followers 4.7K Posts

Latest posts 5

boldstart ventures @Boldstartvc 13h

Huge congrats @spectrocloudinc!

"Why Snowflake's former revenue chief is joining a little-known AI infrastructure startup"

businessinsider.com/spectro-cloud-…

www.businessinsider.com Why Snowflake's former revenue chief is joining a little-known AI infrastructure startup Snowflake's former revenue chief Mike Gannon has joined Spectro Cloud, a $1 billion startup helping companies deploy AI.

boldstart ventures @Boldstartvc 2d

Day-0 model availability is becoming table stakes. Kimi K3.1 Pro is live on Systalyze with an OpenAI-compatible endpoint, and that's how inference platforms win routing share. @systalyze
x.com/systalyze/stat…

Manya Ghobadi @systalyze 3d

🚀 Kimi K3.1 Pro is live on Systalyze, Day 0.

OpenAI-compatible:
Base URL: inference.systalyze.com/v1
Model: moonshotai/Kimi-K3.1-Pro

Sign up: console.systalyze.com
@Kimi_Moonshot
View quoted post on X

boldstart ventures @Boldstartvc 2d

Physical AI just crossed into millimeter-scale dexterity. Thinner O-rings aren't a demo flex. They show the perception and control stack catching up to what factories actually need. @GeneralistAI
x.com/GeneralistAI/s…

Generalist Verified account @GeneralistAI 3d

Even thinner O-rings.

These tasks test dexterity and the ability of the models to perceive fine-grained millimeter-scale details.

Read more about GEN-1.5 and sign up for the waitlist in the comments below.

View quoted post on X

boldstart ventures @Boldstartvc 3d

Agents carry long histories, yet each decode step touches a sliver of it. Memory, not compute, becomes the next AI infra bottleneck. @netpreme is building the warm tier for KV cache.
x.com/netpreme/statu…

Netpreme Verified account @netpreme 3d

Our XM-Sparse article explores how Netpreme X-Mem™ integrates with SGLang HiSparse to provide high-bandwidth access to offloaded KV caches.

Read it here:
netpreme.com/blog/xm-sparse…

Sparse attention is changing what AI systems need from memory.

DeepSeek and Z.ai’s combined share of OpenRouter’s publicly reported tokens rose from 23% in July to 41% in October.*

Meanwhile, research is advancing on several fronts:

🔹 Smarter selection of relevant context
🔹 Hierarchical search for efficient token retrieval
🔹 Reuse of KV data and token selections across layers
🔹 Hybrid architectures combining sparse retrieval with local or linear attention

Agents accumulate long histories. Video creates enormous token sequences. Yet each decode step often needs only a small subset of that context.

The hardware opportunity follows: retain KV history in a large warm memory tier, keep the active working set in HBM, and fetch selected data over a high-bandwidth fabric.

At Netpreme, we’re building toward memory capacity that scales independently of compute, with scale-up bandwidth connecting GPUs to a shared warm memory tier.

*The chart uses DeepSeek + Z.ai lab totals as a proxy for sparse-attention adoption, not a model-by-model count. Source: OpenRouter, via CrbonFree: crbonfree.com/case-studies/o…

View quoted post on X

boldstart ventures @Boldstartvc 3d

Our partner @ellenchisa spent ~120 hours building her own AI operating system and wrote up the architecture. Living inside agents is the fastest way to learn where they go next. Worth a read for builders.
x.com/ellenchisa/sta…

Ellen Chisa @ellenchisa 3d

Since January I have spent ~120 hours (three full work weeks, a lot of it while "on vacation") building a robust personal operating system.

It focuses more on interlinkages vs. one-offs. I'd love to trade notes on architecture with like-minded folks.

ellenchisa.substack.com/p/trading-note…
View quoted post on X