650Group

Marvell has been recognized for two key optical connectivity innovations in the 2026 @lightwaveonline Innovation Reviews. The 1.6T silicon photonics light engine and the active copper cable linear equalizers each received a 4.0 outstanding honoree rating for delivering clear technical advantages in performance, power efficiency, and system integration. The 1.6T light engine provides eight lanes of 200G PAM4 optical connectivity in a single package and is purpose‑built for AI scale‑up, supporting future CPO architectures. The ACC linear equalizers extend copper reach while maintaining ultra‑low latency, offering a practical on‑ramp to 1.6T deployment in existing data center designs. These recognitions demonstrate how Marvell continues to advance the optical technologies that enable efficient, scalable AI infrastructure. Read the full article for more details: mrvl.co/474d6Vp

GPUs/XPUs are getting faster at an exponential rate, but the network infrastructure feeding them data is falling behind. Without a breakthrough, AI data centers face: ⚠️ Underutilized GPUs ⚠️ Performance and throughput bottlenecks for AI training and inference ⚠️ Ballooning TCO Read more from @650Group: 650group.com/blog/the-widen… #AI #GPU #Networking

Jensen @nvidia says every company needs an @OpenClaw strategy and thanks Peter Steinberger (@steipete) for joining at #gtc26. And it introduced nemoclaw!! @cdepuy @650group @alanweckel

Today, @CatoNetworks announced its new SASE and #ZTNA security offering supporting AI and GPUs with @nvidia technology. Exciting stuff. Read @cdepuy thoughts here! 650group.com/blog/cato-netw…

Eridu exits stealth with $200M to rebuild AI networking spr.ly/6014B6GZTG

@Meta's latest blog gives a solid look at its MTIA plans and how it is thinking about both scale-up and scale-out. Big takeaway: the future bandwidth demands are huge. It’s a reminder that networking will be one of the key factors shaping how fast AI infrastructure can scale — and how quickly the Ethernet market will need to grow to keep up. Also, credit to @credosemi that purple cable definitely stands out in the photo Meta used. #AIInfrastructure #AINetworking #Ethernet #Meta #MTIA #DataCenter

AI factories are hitting a networking wall. The industry is pouring hundreds of billions into GPUs and XPUs, but the fabric connecting them is not scaling at the same pace. @650Group data shows compute has surged 5x+ since 2022, while effective scale-out bandwidth per GPU has grown only 4–5x. That mismatch is becoming a tax on utilization, efficiency and AI economics. In the next wave of scale, the bottleneck is not the GPU. It is the network. Read the analysis: 650group.com/blog/the-widen… #AIInfrastructure #AINetworking #Ethernet #DataCenter #Hyperscale

New blog; XPU/GPU having an impact: A growing disparity. Compute #FLOPs are scaling exponentially, especially as #GPUs go multi-die. Compute is hungry for more scale-out and scale-up bandwidth. Read about it! 650group.com/blog/the-widen…

At Mobile World Congress today, Nokia and Amazon Web Services presented today about how both companies are working together to deliver Telecom Core functions both on cloud and on premises #MWC26 #Nokia #NokiaCore #NokialAR @AWS @nokia @650Group

🚨 24 hours to go! 🚨 Join Gilad Shainer (NVIDIA) and Alan Weckel (@650Group) tomorrow for a live conversation on how enterprises are scaling AI infrastructure and driving real ROI with NVIDIA Spectrum-X Ethernet. They’ll share practical insights on building high-performance, efficient AI factories that scale from on-prem deployments to gigascale systems. 👉 Register now to save your spot: nvda.ws/3Ot5aq7

@Cisco launched Silicon One G300 this week Focus on Network #ROI and #Performance @AlanWeckel weighs in concerning Back End Scale for AI #datacenters and the power of 1.6Terabit systems Cisco emphasizes: ✅ Intelligent Collective Networking ✅ Future-Proofed Infrastructure

OIF’s Q1’26 Technical & MA&E Committees Meeting is underway in Palm Springs. Big thanks to @AlanWeckel (@650Group) for helping kick things off yesterday with a market view to frame the week ahead. #OIF #Interoperability #AINetworking

Ethernet is in the driver's seat in AI Data Centers Production data from @drivenets shows that Ethernet isn't just catching up—it’s winning: To achieve ROI in 2026, AI data centers must mix future silicon #GPU/ #XPU without networking bottlenecks - bit.ly/4ksUGCY

@NetworkWorld roundup for 2026 trends (by @NealWeinberg ) includes my take: SONiC is entering its next adoption phase. @650Group data: • SONiC-based DC switching rev >$5B in 2026 (hyperscalers + neocloud + enterprise) • Enterprise SONiC adoption +25%+ YoY thru 2027 • Next: SONiC expands toward the edge as inference scales SONiC is shifting from “hyperscaler default” to a commercial battleground: open NOS flexibility + validated interoperability + silicon optimization. networkworld.com/article/412658… #NetworkWorld #SONiC #DataCenterNetworking #Ethernet #AINetworking #Neocloud #WhiteBox #Telemetry #650Group

@Extreme Networks delivered strong Q2 results today, posting $317.9 million in revenue (up 14% year over year) and 25% growth in SaaS ARR. In the earnings call, CEO @Ed Meyercord highlighted @650Group data as the benchmark for tracking their continued market share gains in the enterprise. Full news here: investor.extremenetworks.com/news/news-deta…

@gigamon says it hit 50% market share in deep observability (1H’25) and it’s a strong signal of a bigger data center shift. @650Group data was cited, but the real driver is the surge in lateral east-west traffic that legacy tools often miss. As AI scales, those flows get more encrypted and complex, making network-derived telemetry the baseline for detecting threats without adding performance drag. gigamon.com/company/news-a… #DeepObservability #NetOps #SecOps #AIInfrastructure #Gigamon

Upscale AI has raised a $200M Series A funding round, bringing our total funding to $300M! The investment reflects strong confidence in our differentiated approach to AI networking. Our full-stack, turnkey networking solutions are slated to ship this year. upscaleai.com/from-100m-seed…

In our latest blog we discuss the exploding data center market, namely, scale-up and scale-out networking providing a line of sight to a $200B market. Meanwhile, #Upscale AI AI today announced it has raised a $200 million Series A funding 650group.com/blog/navigatin…

To hit $1T in AI CAPEX, we’ll need ~12 advanced fabs dedicated to AI. That’s why @TSMC’s accelerated Arizona roadmap matters — we’re entering a “Fab+” era where manufacturing capacity has to scale as fast as @Amazon, @Google, @Meta, @Microsoft demand. Why the potential $265B+ investment is pivotal: • Scale: enables multi-trillion parameter models • Velocity: strengthens the supply chain for @NVIDIA and @Broadcom with localized 3nm The physical foundation for turning AI forecasts into reality. Full analysis: 650group.com/blog/taiwan-tr… #Semiconductors #TSMC #AIInfrastructure #Hyperscalers #NVIDIA #Broadcom #650Group

650 Group founder @AlanWeckel recently had an insightful conversation with Sandeep Bharathi, President of the Data Center Group at @MarvellTech - #AI #XPU #Chiplets - Give it a view!