@digitalocean Feedback: on Cloudways, my acocunt is temporarily blocked and almost got suspendend because of an unpaid bill. Why don't you send any email warnings? I was lucky that I logged in today.
The AI infrastructure conversation is shifting.
It is no longer just “who has GPUs?”. It is becoming:
Own your weights.
Own your data.
Own your intelligence.
AI-native companies want openness, portability, performance, and control without giving up developer velocity or cost efficiency. That is exactly why we built @digitalocean as the AI Native Cloud: infrastructure, cloud services, inference, data, and agents in one open, integrated stack for production AI workloads.
With Q2 revenue growth expected to accelerate to ~29% and RPO up more than 10x YoY, the signal is clear: builders are choosing platforms that go beyond bare-metal GPU rental and help them scale real AI businesses.
The AI Native cloud era is here and it will be open, flexible, and built for builders.
Agents are starting to change the shape of cloud infrastructure. The current wave of AI is about generating tokens. The next wave is about executing work. Agents will think, call tools, use memory, query databases, coordinate workflows, and run inside secure CPU-based environments.
So the infrastructure stack for AI can’t stop at generating tokens. It needs CPUs, memory, storage, networking, databases, sandboxes, inference, and orchestration working together. As token demand explodes, agentic workloads could become orders of magnitude larger than human-driven workflows.
See my full interview here https://t.co/yuexZaZWfk
3 new DigitalOcean datacenters are being connected.
Three new PeeringDB facilities, all added June 25:
•Digital Realty Seattle SEA10 — Seattle
•PTOR1 — Portland
•Equinix DA3 — Dallas
Looks like everything planned will be going online on time end Q3 or Q4.
$DOCN
One startup shifted their entire work schedule to 1am to 10am local time.
The rationale: 80% of the world’s population is asleep during those hours and they believe ChatGPT and Codex make fewer mistakes in that window of time as all of the compute is under-utilized.
The strongest developer pull we’re seeing is simple:
“Give me agentic coding where my code actually runs.” Today, @grok Build is available as a 1-Click App on the @DigitalOcean Marketplace.
SSH into a Droplet, run grok, and you have xAI’s terminal-native coding agent pre-configured with DigitalOcean Gradient serverless inference.
It supports plan/review/approve workflows, clean diffs, headless automation with grok -p, OpenAI-compatible inference, one Gradient model access key, and optional Intelligent Inference Router support across leading models.
No heavy setup. No web UI required. No moving work out of the terminal. Developers wanted AI agents closer to their repos, shells, and infrastructure.
We’re making that one click away.
https://t.co/UiTiIz8cFX