Piris Labs

@ycombinator @terranoxai Great team! Great execution speed.

Terranox (@terranoxai) is building the first AI-powered uranium discovery company. Nuclear energy is experiencing a global resurgence, and the world needs to 4x uranium production by 2050 to fuel it. Traditional exploration has a hit rate below 1%, driving extremely high costs and decade-long timelines. Terranox uses AI trained on decades of exploration data to find the deposits needed to power the next century. Congrats on the launch, @jadecheclair and @leeavlipton! ycombinator.com/launches/PX9-t…

If you could go back and had the opportunity to see Groq at day one, what would you do? Join them? Invest in them? Partner with them? Now assume not just faster compute, but fundamentally better unit economics at the system level. Faster chips alone don’t matter anymore. What matters is whether you can reduce the cost of inference without compromising model accuracy, size, or speed. That’s what generational companies are built on. That’s exactly what we’re working toward at Piris Labs (YC W26). Stay tuned for our inference demo in the coming week. #Inference

.@PirisLabs delivers low-latency, low-cost AI inference by eliminating the memory wall using full-stack optical interconnects and a purpose-built software stack. Congrats on the launch Ali & Keyvan! ycombinator.com/launches/PQd-p…

Seeing Systems (@S2_INC) builds inexpensive strike drones designed to adapt as warfare evolves, with modular hardware, simple operation, and an agentic control system built in collaboration with Royal Marine Commandos and other NATO forces. Congrats on the launch @Alex_10200 and Matthew! ycombinator.com/launches/PLA-s…

I find this picture very interesting—not because it’s a selfie with an extremely inspiring person, but because of the backstory. @LisaSu and I both attended the same department at @MIT, in the same building. Years later, I worked in Lisa Su’s building at MIT (the MIT.nano cleanroom). She worked on copper interconnects during her PhD; I work on photonic interconnects. We’ve now crossed paths again in San Francisco and discussed interconnects. Life really does come full circle in amazing ways—so many interconnects, all in a good way, coming together. Let’s solve inference problems the same way: with better interconnects. Fun fact: she likes boxing! 🥊 💪 #AI #AIHardware #AIInference #AMD #DeepTech #HardTech #Semiconductors #Compute #AIInfrastructure #Engineering #Startup #StartupLife

@LisaSu @AMD — we’re interested in exploring a technical collaboration to test our optical interconnect with AMD hardware for our optimized inference stack. Would appreciate connecting with the right team

@paulg Then why we haven’t met yet.

Someone asked about tricks for meeting eminent people. I said the best plan is just to do really good work. Then you'll tend to meet them organically. In the worst case you'll see one whenever you look in the mirror.

@sama We all know

We love working with NVIDIA and they make the best AI chips in the world. We hope to be a gigantic customer for a very long time. I don't get where all this insanity is coming from.

@Sakshi50038 I ask AI to remove what it added

What you will do in this situation! 🥹🥲

@albysjourney I can help. Followed you

Who unfollowed me. I had 4,200 followers and now I have 4,199. I will find you.

Want to know a Y Combinator SECRET?? Over the last two weeks, I spoke with 200+ AI founders, engineers, and application builders. When I ask about their biggest hurdles, it almost always comes down to the same three things: 1️⃣ Cost/Token 2️⃣ Speed (Time to first token) 3️⃣ Throughput (Tokens Per Second) The ranking seems to shift depending on the company stage. I'm curious—what is your order? 1-2-3? 2-3-1? What’s your stage? Drop the sequence below 👇 #AI #GenAI #LLM #ArtificialIntelligence #MachineLearning #DeepLearning #Founders #StartupLife #Startups #Entrepreneurship #VentureCapital #Engineering #SoftwareEngineering #Developers #CTO #TechLeaders #MLOps #AIInfrastructure #CloudComputing #DevOps #PlatformEngineering #DevTools #SaaS #B2B #ProductEngineering #TechStrategy #DigitalTransformation #FutureOfAI #Innovation #Inference #AICompute #Scalability #Performance #CostOptimization

@KevinSzabo14 Be my guest

Give me an account with: 0 followers 0 Credibility 0 Influence 0 voice And I guarantee I can have them monetized in a week. I'd even pay for their blue tick if needed.

@saylordocs It’s a psychology issue not a money issue.

HOT TAKE: 9-5 with a high salary is better than owning your own business.

Nvidia’s $20B acquisition of Groq signals a massive shift in the landscape: 1. Inference is the ROI engine. 2. General purpose compute is no longer enough. Optimized stack is a MUST. At Piris Labs (YC W26), we build Inference at the speed of light using a fully optimized stack—silicon, software, and optical networking fabric engineered for ultra low latency, high bandwidth, and ultra low power consumption / bit. The result: Ultra low time to first token at a fraction of cost / token. Pictured: Mohsen Moazami (ex Groq President) with Igor Arsovski (VP @nvidia). #AI #Inference #Semiconductors #NVIDIA #Groq #DeepTech #AIInfrastructure #YCombinator