April was slightly less hectic for the team, with only one major show post KubeCon EU London — RSAC 2025 Conference (saying RSAC Conference isn’t redundant as it’s no longer “RSA Conference” — RSA Security sold it in 2022 to a consortium led by Crosspoint Capital Partners, a PE firm). Jim Carroll of Converge Digest, our media partner, together with us, debuted our “En Route” series with “En Route to RSAC,” featuring interviews with thought leaders in self-driving Waymos on the streets of San Francisco! Check out Jim’s beta ride with Adtran last month at OFC. When the rest of the series posts, we’ll let you know. BTW, the OFC 2025 Interviews are on NextGenInfra.io — those of you interested in the latest in optical connectivity (like for AI data centers) should check it out. And in case you missed it, we’ve passed 50K subscribers on the NextGenInfra.io YouTube channel! 🥳

Our industry kicks off May 2025 with continued uncertainty, though the world seems eager to glom onto any glimmer of positive news to justify continued investment in the US stock market. Nevertheless, portfolios appear to be shifting out of US assets (stocks, bonds, and likely real estate) — hey, if BlackRock says so, it must be true. Plus, we’re seeing signs of wobbliness (a technical term) in AI data center investments, though there’s still a lot of money flowing that way.

We have a new report out this month courtesy of SoftBank, our sponsor. It’s … drumroll … on AI RAN. (What? Were you expecting a non-AI report?) More details are below, but if you’re eager to get your hands on it, download from SoftBank here. It’s an optimistic but measured take on the space. We thank the SoftBank team (particularly Dr. Alex Choi and Dr. Ryuji Wakikawa) for making this report possible.

Speaking of data centers, another call to download our 2025 Data Center Networking Report, which covers the impact of AI and cloud workloads on data centers — it’s now UNLOCKED, meaning no need to provide any personal information to download. For the shy ones out there or the super privacy-conscious, it’s the moment you’ve been waiting for!

Here’s what the rest of our newsletter has in store for you:

• Roy’s Travelogue — RSAC 2025
• Latest Reports — AI in the Mobile RAN
• Recent Articles— Insights from MEF’s GNE 2024 plus 2025 look ahead
• NextGenInfra — OFC 2025, MPLS WC 2025
• April News Roundup — SKT breach, tariffs, Zero-Trust, AI optimizations for telcos
• Other Content — AvidThink in the News, Webinars, Meet Us Live!

Roy’s Travelogue

After returning from KubeCon EU in London, I had a good couple of weeks to catch up with work before this year’s RSAC Conference 2025. The first RSA conference was in 1991 and organized by the eponymous RSA Security firm, famous for licensing public-key (and shared key) encryption solutions (remember RSA BSAFE SDK?) and the ubiquitous tokens we learned to love/hate. I’ve attended almost every RSA since the late 1990s as an attendee (scrounging free expo passcodes from sponsors), an exhibitor, and, in the last decade, a media/analyst. It’s still the pre-eminent security show with a good mix of content and exhibits.

(L to R: Moscone Center during RSAC 2025, with Eric Hanselman of 451/S&P, a driverless Waymo, Jim Carroll with Larry Lunetta of HPE Aruba Networking – one of our En Route guests)

Under the theme “Many voices. One community,” this year’s conference focused on forging relationships among governments, vendors, and academia and recruiting new voices into the fold. With 44,000+ attendees (3,000+ more than in 2024), RSAC is a proxy for the continued growth in all things cybersecurity.

Here are a couple of major themes for me:

• AI is still the star of the show — This year, though, there’s a focus on implementation and “agentic AI” (defined in as many different ways by as many different vendors). Enterprises prioritize specific, achievable outcomes with measurable benefits, including documentation assistance, incident triage, and threat analysis.

• Security for AI and AI for security — This fits the general “X for AI and AI for X” catchphrase that you could slot any major IT component into (storage, networking, etc). Jeetu Patel, EVP and CPO of Cisco, said this aloud in his keynote, which was repeated numerous times throughout the conference. The use of AI/ML in security is far from new, but GenAI and Agentic are fresher. Cisco also released on HuggingFace their Llama-3.1-FoundationAI-SecurityLLM-base-8B (Foundation-sec-8b), an open-weights model built on Llama-3.1–8B fine-tuned for security use cases.

• Outcome-focused approach for AI applications—More discussions on specific uses of AI with measurable benefits. Like customer service, marketing, and sales for general enterprise use cases or coding, the security industry has found GenAI helpful for incident triage, threat analysis, and documentation assistance.

• Trust and reliability in AI—There are more conversations and research around understanding the limits of AI technology and approaches to ensuring reliable output from AI-assisted or AI-powered systems.

• Resilience-based strategies—There’s a broader approach to security that doesn’t purely involve prevention of every single incident but also effective recovery and adaptation after security incidents/breaches. Part of this is an ROI and practicality mindset.

• Identity (+ zero-trust) as the foundation—An increasing number of vendors and speakers have discussed the importance of identities (users, devices, applications), both static and ephemeral, as a key security element. Identity management is back in vogue again tied closely to zero-trust, the other ongoing hot topic. It’s only 21 years after I launched my ill-fated startup, Identity Engines, or as my former security architect called it, “zero trust before zero trust”. What’s old is new again.

• Critical infrastructure and operational technology (OT): There’s a greater spotlight on vulnerabilities in essential systems that power industries and our everyday lives. Finally, we’re getting increased government (and vendor) attention on this crucial space. I just hope it’s not too late to prevent catastrophic failures due to cybersecurity attacks soon.

• Platform consolidation vs. strategic tool selection—There are vestiges of the “platformization” trend that started in 2024. This trend involves vendor and feature consolidation and a desire by the larger players to own the platform, including the data, workflow, and AI engines. There are still plenty of new security categories, and with the hype around AI, I anticipate the ongoing redefinition of the platform will make it a moving target.

Up next for me is FutureNet World in London later this week. I’m looking forward to chatting with the CxOs and thought leaders across major European carriers and at the vendors. Check out the conference agenda, and if you’re in town, join us!

After London, I’m at the AI Networking Summit (ONUG) in Dallas at the end of May (28th, 29th), where we’ll discuss AI Networking, AI Infrastructure, AI Security, and much more. Hope to meet you there!

Also, I’m happy to chat more about what I’ve learned on the road. Drop us a line at [email protected].

– Roy for the AvidThink Team