WEKA

🏙️ Chicago, here we come! We're headed to @Supercomputing 2026 with faster AI, GPU-powered innovation, and plenty of fun. See WEKA NeuralMesh in action and score some exclusive WEKA swag! 🗓️ #SC26

👏 Our DACH team brought the energy to Munich for the #HPCRoundtable26! Great conversations, new connections, and a reminder of just how fast this space is moving.

Cheaper AI tokens mean more AI, not less demand. In this article, WEKA’s Val Bercovici explains why memory is becoming the next AI bottleneck. Take a look ⬇️

🗓️ Join WEKA & @Microsoft at the inaugural WEKA Executive Summit at #SC26 at Wrigley Field. Explore the agenda and reserve your spot 👇

WEKA and Oasis Technologies are bringing WEKA NeuralMesh to Israel’s AI ecosystem, helping organizations keep GPUs fed and AI moving at full speed. Get the full story ⬇️

RT @mmarshall: Not everyone sees a glut. @weka's @valb000 says demand for tokens now outstrips supply at any price, as security agents run around the clock. We’ll debate it Wednesday at our Agentic Infrastructure Stack Summit in SF. Apply by Monday:

👏 Thanks to our team, customers, partners, and @NVIDIAGTC for an incredible experience in Berlin! 💜 Great conversations, fresh ideas, and big momentum for AI. #NVIDIAGTC

WEKA Exchange Dubai brought the AI infra community together last week to discuss what holds AI back at scale. Thanks to our speakers, guests, and co-host @vadtechnologies! 🗓️ Next up: Sovereignty by Design, exploring AI infra for regulated markets:

🤝 WEKA and @smartbird_AI are partnering to bring WEKA NeuralMesh to dedicated AI clusters, helping enterprises run production AI with greater GPU utilization, efficiency, and data sovereignty. Get the full story 👇

WEKA’s Val Bercovici joins @nvidia’s Jason Hardy on The Five Nine podcast to discuss GPU utilization and memory tiering. Catch up on the conversation ⬇️

ICYMI: In #DeepGeeks Ep. 7, WEKA’s Val Bercovici and @Cognizant’s Vernon Turner explore why traditional metrics fall short—and what measuring value per token reveals about AI’s real costs. ➡️ Check it out:

☝️ More GPUs won’t fix a data problem. If your training runs are slow because GPUs are waiting on data, the answer isn’t a bigger cluster. It’s a pipeline that keeps the GPUs you already have fed. 😄 See how WEKA NeuralMesh does it:

🧠 The next AI breakthrough may be about memory, not bigger models. WEKA’s Val Bercovici recently joined John Kim on the MindMakers podcast to explore the AI memory wall, context engineering, and the future of agentic AI. Check it out 👇

👻 Our Campbell crew got a little spooky this week! From creative costumes and tasty treats to games that brought out everyone’s competitive side, we had all the ingredients for a fang-tastic WEKAween celebration. 🎃 #LifeAtWEKA

📌 ICYMI: WEKA Federal is now an Awardable vendor in Tradewinds, helping government agencies access #AIInfrastructure that makes every GPU, megawatt, and rack count. 🔗

🇩🇪 Just one week until @NVIDIAGTC Berlin! We’re excited to connect with the innovators shaping what’s next in #AIInfrastructure. Attending? Come say hello to the team! 👋 🗓️ #NVIDIAGTC

⚡ @andromeda_ai + WEKA: 6x faster environment load times, metadata ops cut from hours to minutes, and lower storage costs. Check it out ⬇️

🤝 WEKA and @vadtechnologies are teaming up to accelerate AI adoption across the Middle East, bringing NeuralMesh to partners in the UAE, Saudi Arabia, Qatar and beyond. Full story 👇

📊 AI broke the yardstick. So how do we measure its true value? Dr. Serena Huang, Vernon Turner (@Cognizant) and Val Bercovici explore AI ROI, inference costs and outcomes per token. 🎧 #DeepGeeks