Networks that Think.
Palo Alto, CA
Joined October 2025
- Tweets149
- Following181
- Followers290
- Likes121
A great event in Berlin connecting with our partners!
Thank you to our partners for the meaningful conversations, valuable insights, and continued collaboration. Looking forward to building on the momentum together!
@mansourkaram
@djspry
Thank you for covering @AriaNetworks!
A fast-growing AI startup needed more than office space in Sunnyvale. It needed serious power. bizjournals.com/sanjose/news…
Something is wrong in your network. You just can't see it yet. 👀
👻 Welcome to ScAria. Between now and Halloween, we will share a few true horror stories from AI networks.
#ScAria
Heading to @stacresearch Summit New York this week? So are we. Aria Networks is proud to be a Silver Sponsor!
In the AI Factories, every microsecond counts, and the network is often where they’re won or lost. That’s where Aria’s Deep Networking platform comes in, with microsecond-level telemetry and fabric intelligence for AI and high-performance infrastructure.
📍 New York Marriott Marquis, 5th Floor
🗓 Wednesday, October 21
🎤 Catch our very own Tim Murphy on stage at 12:10 p.m. ET
Our team is on site from the 8:30 a.m. exhibit opening through the evening reception. If you’re attending, come find us. We’d love to talk about what your network is really doing.
stacresearch.com/events/fall…
"We have an aggressive growth plan." 🚀
That's our CEO @mansourkaram in today's @svbizjournal, talking about Aria Networks, Inc. move into our first office.
🏢 35,000 sq ft at 400 W. California Ave., up from under 10,000 in Palo Alto. We officially outgrew our home at Sutter Hill Ventures.
🛠️ We launched in early 2025 to build networking for AI factories, and most of our hiring this past year has been hardware and software engineers.
📈 We're going from 70+ people today to 200+ worldwide by this time next year, with the Bay Area team growing from ~50 to 120–150.
🚆 We're less than a mile from Caltrain and near downtown Sunnyvale, so commutes are easier for the team across the Bay Area.
Our hardware team moves in this month, and everyone else follows in November.
As Mansour put it, "Well be working on the plane as it flies." ✈️
From day one, @shv didn’t just back Aria, they gave us a home. 🏠
We spent our first chapter building out of their Palo Alto office, and a lot of what Aria is today started in that space. Thank you to the whole Sutter Hill team for making room for us while we grew, and for putting up with the hardware, the late nights and the ever-growing headcount.
Want to help build the networking company for the AI era? We're hiring 👉 arianetworks.com
@mansourkaram @stefandyckerhof
A fast-growing AI startup needed more than office space in Sunnyvale. It needed serious power. bizjournals.com/sanjose/news…
Apparently, the slow network walkers are becoming famous. And increasingly in demand. 🤭
OCP 2026 has been quite a week!
Jokes aside, there were some great conversations with customers, partners and people building the next generation of AI infrastructure.
One thing is clear: the race to build AI factories is accelerating. And as these systems scale, the network becomes ever more critical.
But remember: Demand will fill your AI factory. Performance will define its value.
Thanks to everyone who connected with us in San Jose this week.
What caught your attention at OCP this year? Let us know in the comments.
#AriaNetworks #OCP2026
🏆A win and a runner-up finish for Aria Networks at the 2026 @SONiC_Fund 2026 Hackathon!
Our team took on multiple challenges and came away 🥇 Winners in AI + Debugging & Root-Cause Analysis with our SONiC Root Cause Analyzer, and 🥈 Runner-up in AI Agents for Networking.
32 teams from 13 organizations took part. We’re building a team with deep technical talent, people who know networking and AI and aren’t afraid to tackle the hard problems. This recognition is proof of that. And we're only getting started.
Huge congratulations to Krishnamurthy Mayya, Sudharshan, Shankar Vaideeswaran, Saravanakumar Venkat, Shrey Shah and everyone who made this happen. Great work!
We're planning a technical deep dive on what the team built, so stay tuned for that on our blog.
If you want to help shape the future of AI networking alongside a team like this, come talk to us. 👇🏻
arianetworks.com
#SONiC #SONiCHackathon #OCPGlobalSummit #OpenNetworking
Still waiting for the network to catch up?
Meanwhile, Aria Networks, Inc. is running circles around it.
It’s been a great start to OCP!
#OCP #AriaNetworks #NetworksThatThink
Demand will fill your data center. Performance will decide what it’s worth.
When supply catches up, the winners will be the ones getting the most useful work out of every gigawatt.
In his new blog, @mansourkaram explains why performance matters, and why the network is key to making it measurable.
Read the blog: arianetworks.com/articles/de…
At OCP this week? Come meet the Aria team and see our demos.
[email protected]
Aria Networks, Inc. retweeted
SemiAnalysis's Dylan Patel warns networking incumbents could be disrupted because networks are only 15% to 20% of cluster cost but can determine far more performance
“People are not just doing fat-tree networks anymore. They’re doing a lot of other kinds of network designs. That means a lot of software for the workload to be aware of that network design, for that network to be managed and for that network to be controlled.
“Networks are ultimately the lifeblood, so how you control and optimize those is very, very important. There are numerous companies popping up to focus on that, whether it is data center to data center, within the data center, WANs or security for networks—all these different things.
“There are a lot of problems to tackle here because, while networks are only 15% to 20% of the cost of a cluster, they can have impacts on performance way larger than that. Then there is security.
“All these things, I think, are very important. I think there are a lot of advancements to come, and the incumbents better move faster, or else they will be disrupted.”
SemiAnalysis's Dylan Patel: enterprises have a "broken philosophy" because they'll spend six months to a year building custom AI systems that are already worse than foundation models by the time they finish
“I think a lot of people’s philosophy, especially a lot of enterprises’, is just broken. They think, ‘Okay, where AI is today is where it is going to pause, and then we’re going to be able to build our own internal capabilities that are way better.’ That often ends up not being the case.
“They decide to fine-tune a model or whatever, and then they end up creating a model that's worse than just the foundation model by the time they’re done.
“Across the stack, you have to measure your performance, your TCO, your cost and everything across every aspect of the supply chain, whether it be your infrastructure, your serving or your inference, but also evaluating your models.
“Evals for models are the Wild West. People don’t know how to measure models appropriately. They don’t know how to think about model progress appropriately. That's leading them to make bad engineering decisions which then end up being wasted six months or a year down the road, when the foundation model is so much better.”
AI clusters are only as fast as the network connecting them. 🧠⚡
We'll be at OCP Summit next week in San Jose, Oct 12–15, talking Deep Networking: microsecond telemetry and fabric intelligence for the AI factory.
See you all in San Jose. 🌉
[email protected]
Aria Networks is a Silver Sponsor at #STACSummit New York, Oct 21 at the Marriott Marquis 🗽
Catch our very own, Tim Murphy on stage at 12:15 p.m. ET.
@stacresearch
stacresearch.com/events/fall…
Networks that think.
The Aria team is heading to the 2026 OCP Global Summit in San Jose, October 12–15.
Meetings are by appointment. Book a time with our team: [email protected]
📅 October 12–15, Monday–Thursday
📍 San Jose Marriott, 301 S Market St
@OpenComputePrj
Your network isn't broken. It just can't see clearly. 👀
AI clusters slow down in microseconds, and most tools never notice. Dashboards stay green while GPUs wait and costs climb.
With Aria, you get microsecond-level visibility, so it can spot issues and act on them before performance takes the hit.
See it for yourself.
Watch our latest product demos, expert conversations, event highlights and more 👉 youtube.com/@Aria_Networks
@mansourkaram
Inference now runs across the network: KV cache, FFN layers, and MoE experts all on the wire.
Microsecond stalls → multi-second agent hangs.
Part 3 of @djspry inference series on why distributed inference needs self-tuning networks:
arianetworks.com/articles/ho…