· Update · 4 min read


Hello there!
Q2 2026 gave us plenty to talk about at DataPacket. More places to deploy. More capacity in our global network. More hardware for AI and high-throughput workloads. More insightful blog entries.
Let’s take a closer look at what’s new for you.
Welcome Barcelona in northeast Spain, Calgary in Western Canada, Helsinki in southern Finland, and Perth in Western Australia among our data center locations; now 67 in total across the world.
All the new servers operate from Equinix data centers, known for their exceptional uptime, uncompromising security, and full alignment with leading industry certifications. Above all, you can expect us to have your back 24/7/365.
Not only did we expand our location coverage; our global network interconnections got denser as well.
Across Q2, we added 13.2 Tbps of regional provider, IX, PNI, and transit capacity. North America led the quarter with new Microsoft capacity in Ashburn, Los Angeles, and New York; new T-Mobile capacity in New York; and multiple Tata and Cogent upgrades across major U.S. locations.
Europe added new or upgraded capacity in Amsterdam, Budapest, Copenhagen, Dublin, Frankfurt, Helsinki, London, Marseille, Milan, Paris, Prague, Sofia, Vienna, and Zürich.
Asia-Pacific added 1.0 Tbps, including a 600 Gbps Airtel expansion in Singapore, bringing that PNI to 2.5 Tbps total. Latin America added 600 Gbps through upgrades in Bogotá, Lima, and Querétaro. Middle East & Africa added 900 Gbps, all through new Istanbul PNIs.
We also worked on our private global backbone: 6.7 Tbps improvement across new links and upgrades.
That brings Q2's total capacity work to 19.9 Tbps.

In May, we added two high-end GPU server configurations built for demanding AI and HPC applications: 8× NVIDIA HGX B200 and 8× NVIDIA HGX B300. These Blackwell additions are built for training, tuning, and serving models at scale.
Throughout July, we're adding 500 CPU servers to production, with on-demand GPU servers for heavy compute coming next. Those GPU servers are planned primarily around 8× NVIDIA HGX B300 and 1–2× RTX PRO 6000 Blackwell configurations.
In May, we officially sponsored the Prague Crawl event in Prague, revolving around the web scraping community. We took this opportunity to publish a new case study with our client Massive, a company running real-time web access for AI agents across 195+ countries. By moving to dedicated servers and direct cloud peering, Massive cut $2.9M a year in cloud egress costs while still growing 4–5× year over year. Read it here.
If you’re a DataPacket client and have a success worth spotlighting, let’s make it heard. Get in touch, and we’ll collaborate on a story that showcases your achievements and the network powering them.
Our ClickHouse engineering blog series continues with a second part focused on performance tuning at scale. This part aims at getting the most out of your current setup. Because, sometimes, extra performance comes from knowing which settings are worth touching.
Give your ClickHouse setup some extra care. It's all written purely by human engineers, for engineers.
We’ll keep building a global infrastructure that’s ready to support your growth, anytime, anywhere. Whether you’re scaling up globally, launching a new product, or looking for a top-quality provider, our team is ready to deliver the finest service for you and your team.
See you soon, and thank you for your trust!
Founded in 2014, DataPacket is a dedicated server provider operating a global low-latency network. With a footprint of 67 locations across 6 continents, DataPacket helps businesses–including gaming and video streaming companies–to deliver great online experiences.