A proud moment in Times Square last week - our CEO @mansourkaram, was there to sit down with @Nasdaq's The Signal for an interview on building networking purpose-built for AI clusters.
Seven weeks post-GA, we're heads-down doing exactly that: a Deep Networking platform made for the demands of modern inference and training.
The full interview drops in a few weeks - stay tuned.
Thank you to our team, partners, and customers putting Deep Networking to work.
2: Gacha - Inseon Hwang, Steven Yang, Ransith, and David Song (@HwangInseo22523, @syang0624, @_mrdavidsong)
Gacha built a kind of Facebook Marketplace-style storefront for used bikes — but with agentic workflows baked in.
The concept is quite cool: scan Marketplace, surface the top 3 deals, and then negotiate in parallel, how handy is that?
This is such a strong example of what makes this category exciting: taking a messy, time-consuming consumer workflow and turning it into something far more efficient, intelligent, and usable. Wonderful work here from team Gacha!
@jukan05 honestly, this makes a lot of sense. people can laugh at xAI all they want, but they are incredible at building data centers faster than anyone else
@jasminewsun extremely well thought and written piece, which does a good job summing up SF culture. reading through the comments on NYT is something else though
@Google just confirmed our strategy.
Their new Virgo Network fabric is built on a telemetry-first approach to handle AI's sub-millisecond traffic bursts, synchronized collectives, and massive bandwidth demands purpose-built for TPUs.
At @AriaNetworks, we've been saying this all along: telemetry-first is the only way to run accelerated infrastructure at scale.
What Google does for TPUs in their backend network, we do for ALL accelerators and for every cloud and enterprise customer running them.
https://t.co/P1Gin5WbA8 @googlecloud@mansourkaram@stefandyckerhof
Breaking News: The Supreme Court struck down President Trump’s sweeping tariffs on imports from nearly every U.S. trading partner. https://t.co/Pn7AmTmzsd