Boris Renski

In the last month, $NBIS: - $2B Investment from $NVDA - $27B AI Cloud Deal with $META - Affirmed $7-9B 2026 ARR Guidance - Increased 3GW 2026 Power Guidance - Confirmed 16 Data Centers in Development - Acquired Tavily Pure Execution from Arkady and the Team!

6

You can't fix what you can't see—and partial visibility simply isn't enough anymore. 👀 In today's intricate environments, conventional network monitoring tools are insufficient. Our newest article delves into why integrating flow data with performance metrics and service state is not merely a "nice to have," but essential for contemporary NetOps. 🔗 Read the full post:

3

We shipped 100,000 ISP-tagged IPs, 100 Gbps of internet transit, and 30 physical servers for one enterprise data team. Not a slide. Not a promise. A live system moving traffic today. Read that again. Most vendors sell access. We deliver ownership. Here is why that matters. ISP-tagged IPs are not rented scraps. They belong to a real network. They announce cleanly on the public internet. They behave like native infrastructure because they are. I have seen teams burn months chasing “large pools” that collapse under load. This one didn’t. It held. Now the transit. 100 Gbps is not a talking point. It is raw headroom. It means the pipes stay open when demand spikes. It means no silent throttling when things get busy. That capacity came online day one. No warm-up period. No excuses. Then the servers. Thirty dedicated machines. No neighbors. No shared risk. Each box exists to do one job well and keep doing it. The result is simple. The data team runs at full speed. Routing stays stable. Addresses stay clean. Operations stop breaking at scale. I will be blunt. Most infrastructure companies sell hope. HypeProxies ships outcomes. If you need proof, this deployment is it. If your data team is hitting scale walls and needs real infrastructure, reach out for a free data assessment and we’ll tell you if this setup even makes sense for you. #DataInfrastructure #dataops #IPTransit #InfrastructureEngineering #IPTransit #BareMetal #colocation #Datacenter #ISP

18

2 Comments

Real Agent Learning - courtesy of Varick Agents and Vasuman M. "I've deployed agents for over a year now. I started my software career at Meta but left 6 months ago to build a company that does nothing but deploy production agents for enterprise. We're at $3M ARR and growing, not because we're smarter than anyone else, but because we've built and failed enough times to know what the formula is now. This is everything I've learned about building agents that work. It should apply at any level, whether you’re a beginner, an expert, or somewhere in between. My goal with this article is to share my biggest learnings from a few years of being in the AI space. My hope is that you walk away with useful information that you can use to build better agents. Let's begin..." Direct link to article: https://lnkd.in/gwH7kVNx

10

11 Comments

Green KPIs. Red Reality. I just published the full NovaWireless synthetic call-center research stack — a reproducible environment built to study KPI drift, proxy gaming, and trust-signal decay under AI optimization. This isn’t just a dataset. It’s a complete synthetic system: • Telecom population model   • Representative (agent) generator   • Call interaction engine   • Proxy vs. true 30-day audit outcomes   • Integrity validation instrumentation   • KPI Drift Observatory + governance papers  The lab contains 82,000+ call records across 12 months and 250 representatives, with structured scenarios, repeat-contact windows, and synthetic transcripts for NLP testing. Each interaction includes both a proxy “resolved” flag and a true audited outcome — enabling direct measurement of when performance metrics diverge from durable outcomes. Accompanying papers document the modeling architecture, robustness testing, and a System Integrity Index (SII) framework for detecting when optimization conditions become unsafe. All data are fully synthetic. All methods are reproducible. 🔎 Explore the full research stack: GitHub: https://lnkd.in/gSciwTsS Website: https://lnkd.in/gc6eqPvc Performance dashboards optimize. Integrity systems constrain. Both matter.

1

1 Comment

Kora Logix’s journey their recent $115M Series E caught my eye. Their AI platform learns from traffic logs and device metrics, then takes action rerouting traffic, scaling capacity, or patching configs automatically. In a carrier pilot, incidents dropped 35% and MTTR fell to 20 mins. This approach reminds me of how Neura AI uses Router and RDA agents to handle complex workflows seamlessly. There’s real power in letting AI handle routine tasks so humans can focus on strategy. #meetneuraai #neuraai #telecom #AI #5G #networkmanagement #techtrends Read more 👉 https://lnkd.in/d3YqF4J4

2

Smart Communications, Inc. and Amdocs are basically saying something far simpler and far sharper than what their press release admits; there’s no room left for blind launches. “Launching without understanding your user is just spending; launching without measuring is improvisation.” And that’s exactly where their three pieces lock together, full-plan personalization through KiQ, the connectX “telco-in-a-box” setup that speeds up MVNO creation, and eSIM quietly running underneath to keep everything seamless. That mix real personalization, fast rollout, and a digital backbone with no loose stitching turns Smart into an operator that doesn’t guess; it watches. KiQ isn’t some pretty app; it’s a sensor. Every tweak a user makes, every blend of data, calls, duration becomes live information that dismantles the old model where operators imagined what people wanted and just crossed their fingers. Meanwhile, Filipino Gen Z with that instinctive demand for control and consistency ends up being the clearest reminder of which operators will actually make it, not the loudest ones, but the ones that listen. Sure, a platform can run on AWS and have generative AI baked in… but if it doesn’t measure, it doesn’t learn. And if it doesn’t learn, it falls right back into the original sin launching on impulse and waiting for miracles. And as Smart shapes the future from the Philippines, a question slips in almost on its own, what about Mexican Gen Z, Brazilian Gen Z, the broader Latin American crowd? When do they get something that treats them like protagonists instead of points on a spreadsheet? Anthony Goonetilleke, Group President of Technology and Head of Strategy at Amdocs. FYI. Shuky Sheffer, President & Chief Executive Officer at Amdocs. PLDT.

5

AI touches on everything. Actually, most of Iron Nation Fund portfolio companies are leveraging AI tools to be able to move faster, and provide superior value to their customers.  Gil Friedlander Jason Wolf Chen Linchevski Ariella Raviv Chemi Peres Izhar Shay Charlie Federman

9