Dhenenjay Yadav (@dhenenjay) header
Dhenenjay Yadav (@dhenenjay) profile picture

Building @axionorbital (YC W26) | Computer Vision Research

San Francisco, Cal
Joined August 2017
Photo by Dhenenjay Yadav (@dhenenjay)Photo by Dhenenjay Yadav (@dhenenjay)Photo by Dhenenjay Yadav (@dhenenjay)Photo by Dhenenjay Yadav (@dhenenjay)Photo by Dhenenjay Yadav (@dhenenjay)Photo by Dhenenjay Yadav (@dhenenjay)Photo by Dhenenjay Yadav (@dhenenjay)Photo by Dhenenjay Yadav (@dhenenjay)Photo by Dhenenjay Yadav (@dhenenjay)Photo by Dhenenjay Yadav (@dhenenjay)
Update: @atharva_peshkar and I got into @ycombinator W26 (after getting rejected and then reversing it) to build @axionorbital Huge thanks to our YC partners @bosmeny , @ChristinaG325 , @gustaf , and @dazzeloid for believing in the vision. At @axionorbital , we're building foundation models to generate high-resolution optical satellite imagery from Radar (SAR). (1/3)
16
3
46
3,270
See you at the @ycombinator Demo Day tomorrow! We’ve got a banger live demo ready for you. 🚀 @dhenenjay @atharva_peshkar @rubbrbandHQ
2
3
1
14
1,164
Dhenenjay Yadav retweeted
YC is the YC for space tech
🎉 Ask me why I am so excited for this year's @ycombinator Winter '26 batch: - GRU Space (YC W26) building out a lunar tourism economy - Beyond Reach Labs (YC W26) building stiff-yet-compact deployable structures for space - seriously under-served market IMO - AxionOrbital Space (YC W26) building foundational models for the Earth
15
9
9
133
31,141
Dhenenjay Yadav retweeted
Introducing NeoEarth by @axionorbital: the first world model for Earth. They forecast surface changes days to months out to track floods, crops, and forests, and predicted Hurricane Helene with 25% more accuracy than legacy forecasters. Congrats on the launch, @dhenenjay & @atharva_peshkar!
27
30
1
291
30,809
Accepted to CVPR 2026. Dropping a model that can predict the future of our planet at unprecedented accuracy tomorrow. LFG🚀 @axionorbital @atharva_peshkar
4
2
1
16
999
Dhenenjay Yadav retweeted
🛰️ Legacy optical satellites are blind nearly 70% of the time due to clouds and darkness. Meanwhile, radar satellites can see through weather, but the data is difficult to interpret and doesn’t work with standard vision systems. @axionorbital is solving this by translating raw radar signals into analysis-ready optical imagery in real time, enabling persistent 24/7 visibility for applications like defense, commodities, and disaster response, even when the sky is blocked! Imagine the capabilities. Launch video: youtube.com/watch?v=qWYjqk26… @pioneer_fund is pumped to support @atharva_peshkar and @dhenenjay on their journey!
1
2
238
3000+ GPU hours in feb. 12 failed experiments (multiple weeks). 1 worked. Hardtech for a reason. @axionorbital @atharva_peshkar @ycombinator
1
8
317
RIP Autoregressive Models!
Mercury 2 is live 🚀🚀 The world’s first reasoning diffusion LLM, delivering 5x faster performance than leading speed-optimized LLMs. Watching the team turn years of research into a real product never gets old, and I’m incredibly proud of what we’ve built. We’re just getting started on what diffusion can do for language.
4
239
Discussed AGI with @sama at the @OpenAI Office! We’re building something different. Not an LLM. Not diffusion. A model that reasons across state & time. Stay tuned. @axionorbital @atharva_peshkar
4
2
18
874
Dhenenjay Yadav retweeted
We were voted the #1 most promising startup when our @ycombinator W26 batch kicked off. We were voted #3 during the YC W26 Product Showcase by our batchmates. @axionorbital @dhenenjay
4
2
14
1,024
Dhenenjay Yadav retweeted
⚡ LIVE on Vesting: @ycombinator backed startup founders @atharva_peshkar and @dhenenjay recount getting rejected and accepted by YC in the same batch. "We applied to YC and were actually rejected in the current batch [and still got in]..."
1
2
1
8
1,329
Dhenenjay Yadav retweeted
🚀 @axionorbital launched! SOTA Foundation Models for 24/7 Earth Observation “Introducing Orion : The most advanced SAR-to-Optical Model ever built ~ FID=30.24 || SSIM=0.6” 🌐 fondo.ai/49ozEjW Congrats @dhenenjay @atharva_peshkar!!
3
1
17
579
Dhenenjay Yadav retweeted
Replying to @atharva_peshkar
5
5
2
55
11,433
Dhenenjay Yadav retweeted
"see through clouds" is a pretty epic 1-liner
Optical satellites don't work 70% of the time. Axion (@axionorbital) built advanced SAR-to-optical models to solve this. "Hubble" delivers 10m resolution, while "Orion" reaches 0.5m, with an FID of 30.24. Now, you can track any asset—day or night, in any weather. Congrats on the launch @dhenenjay & @atharva_peshkar! axionorbital.space/
1
1
11
769
Dhenenjay Yadav retweeted
They can literally see through the clouds ...
Optical satellites don't work 70% of the time. Axion (@axionorbital) built advanced SAR-to-optical models to solve this. "Hubble" delivers 10m resolution, while "Orion" reaches 0.5m, with an FID of 30.24. Now, you can track any asset—day or night, in any weather. Congrats on the launch @dhenenjay & @atharva_peshkar! axionorbital.space/
2
3
13
2,110
Dhenenjay Yadav retweeted
What is satellites could see perfectly through clouds, smoke, and darkness? @axionorbital has built a SOTA foundational model (SAR-to-Optical) to do just that. Congrats on the launch @dhenenjay & @atharva_peshkar! 🚀
Optical satellites don't work 70% of the time. Axion (@axionorbital) built advanced SAR-to-optical models to solve this. "Hubble" delivers 10m resolution, while "Orion" reaches 0.5m, with an FID of 30.24. Now, you can track any asset—day or night, in any weather. Congrats on the launch @dhenenjay & @atharva_peshkar! axionorbital.space/
6
5
48
13,750
Dhenenjay Yadav retweeted
Optical satellites don't work 70% of the time. Axion (@axionorbital) built advanced SAR-to-optical models to solve this. "Hubble" delivers 10m resolution, while "Orion" reaches 0.5m, with an FID of 30.24. Now, you can track any asset—day or night, in any weather. Congrats on the launch @dhenenjay & @atharva_peshkar! axionorbital.space/
32
27
11
340
52,814
Dhenenjay Yadav retweeted
🚀SonicMoE🚀: a blazingly-fast MoE implementation optimized for NVIDIA Hopper GPUs. SonicMoE reduces activation memory by 45% and is 1.86x faster on H100 than previous SOTA😃 Paper: arxiv.org/abs/2512.14080 Work with @MayankMish98, @XinleC295, @istoica05, @tri_dao
24
111
21
637
249,589
For decades, the space industry has had a dirty secret: 70% of the time, satellites are blind. Clouds, smoke, and darkness turn billion-dollar sensors into expensive, useless metal. We fixed the game. We use deterministic generative AI to translate Synthetic Aperture Radar (SAR), which sees through everything, into crystal-clear optical images. Input: unintelligible radar noise. Output: photorealistic vision in 0.06 seconds. (2/3)
1
3
264
We are replacing "weather permitting" intelligence with ground truth. This means tracking military convoys through storms and mapping floods at night at 1/100th the cost. We are actively onboarding partners in Defense, Finance, and Disaster Response. Watch us turn noise into signal here : axionorbital.space/ #YCombinator #DeepTech #Space #AI #AxionOrbital #Startups (3/3)
2
230