Stack L5
CoreWeave
Pre-IPO AI neocloud; the public-comp benchmark for SpaceX/xAI compute-as-a-service
Recent Business Signals
Oct 1, 2026
product_launch
Introducing CoreWeave Notebooks for Interactive AI Development CoreWeave Notebooks are now available in CoreWeave Forge, with reactive Python notebooks connected to your data and saved in your project so your team can build on them.
source: coreweave-blog
Oct 1, 2026
product_launch
ART-Optimized Megatron (AOM): 12x higher training throughput ART-Optimized Megatron (AOM) improves RL training throughput by up to 12x using shared-prefix training, optimized attention, sequence packing, and tuned parallelism for ART worklo…
source: coreweave-blog
Oct 1, 2026
product_launch
What Is ARIA, and How Did We Build It? Putting an agent next to your experiments was the easy part. The real work was making its research outlast the chat without taking authority from the researcher.
source: coreweave-blog
Oct 1, 2026
product_launch
Why I Built Fugue: Controlled Experiments for Agent Systems Generating code got cheaper. Knowing which change actually helped did not. Fugue tests one agent-system change at a time, with human approval and evidence that has to reconcile.
source: coreweave-blog
Oct 1, 2026
product_launch
CoreWeave ARIA: The Coding Agent That Drives Continuous Improvement Across the AI Loop CoreWeave ARIA is generally available in Forge. Whether you're building an agent or training a model, ARIA analyzes your experiments, proposes what to tr…
source: coreweave-blog
Oct 1, 2026
product_launch
How We Benchmark ARIA—and Why It Runs on Weave A successful retry proves little. Here's how we benchmark ARIA on Weave, testing proposed changes against the production version so real improvements ship and regressions don't.
source: coreweave-blog
Sep 30, 2026
product_launch
CoreWeave ARENA Delivers Results, Welcomes New Teams See what customers have achieved in CoreWeave ARENA and what your team could learn. CoreWeave ARENA Pass opens expert-guided validation to qualified enterprise teams new to CoreWeave.
source: coreweave-blog
Sep 30, 2026
product_launch
Cognition Becomes First Customer for NVIDIA Vera Rubin NVL72 on CoreWeave Cloud NVIDIA Vera Rubin NVL72 is now in limited availability on CoreWeave Cloud. Cognition becomes the first customer running production agentic AI workloads and benc…
source: coreweave-blog
Sep 30, 2026
product_launch
CoreWeave Forge: Turn AI Iteration Into Compounding Improvement CoreWeave Forge keeps your AI loop connected and free to build on any model, framework or cloud, so every version compounds on the last.
source: coreweave-blog
Sep 30, 2026
product_launch
Automate Agent Observability and Improvement with CoreWeave Agent Lens Stop reading traces one by one. See how CoreWeave Agent Lens automatically clusters production failures and user intent, turning failures into test cases to prove each f…
source: coreweave-blog
Sep 30, 2026
product_launch
NVIDIA Vera CPU Is Coming to CoreWeave: Pack In More Agents The first CPU built for AI agents is coming to CoreWeave. See how NVIDIA Vera packs 11,000 active environments into a single rack, absorbs bursty demand, and joins a multi-vendor C…
source: coreweave-blog
Sep 30, 2026
product_launch
What We Already Know, Waiting to Be Connected AI can link findings across disciplines that no one person could search, turning existing knowledge into new discovery.
source: coreweave-blog
Sep 30, 2026
product_launch
CoreWeave Partner Network Jumpstarts Your AI Stack Announcing the CoreWeave Partner Network: a curated set of partner integrations, each one proven on CoreWeave Cloud under production load before it reaches you.
source: coreweave-blog
Sep 29, 2026
product_launch
Agentic workflows in production: Mechanisms, trade-offs, costs and design patterns Explore how production agentic workflows are designed, orchestrated, evaluated, and governed, with practical guidance on patterns, trade-offs, costs, and fai…
source: coreweave-blog
Sep 29, 2026
product_launch
Kimi K3: A Claude clone or something else? Take a closer look at Kimi K3, its architecture, benchmark performance, and reported similarities to Claude, and examine what the evidence says about how distinct the model really is.
source: coreweave-blog
Sep 29, 2026
product_launch
Introducing CoreWeave ARIA: AI Research and Iteration Agent Meet CoreWeave ARIA, the AI research agent inside Weights & Biases. It reads your runs, backs its analysis with live dashboards, and helps turn every experiment into the next impro…
source: coreweave-blog
Sep 29, 2026
product_launch
Getting started with the Agent Reinforcement Trainer (ART) Get started with Agent Reinforcement Trainer (ART), an open-source framework for training LLM agents with reinforcement learning, including rollouts, rewards, GRPO, and W&B integrat…
source: coreweave-blog
Sep 29, 2026
product_launch
From Simulation to Reality: Why Your Robot Works Perfectly in the Lab But Fails in Real Life Explore the simulation-to-reality problem in robotics and learn how domain randomization, system identification, reinforcement learning, and evalua…
source: coreweave-blog
Sep 29, 2026
product_launch
Tutorial: Benchmarking GPT-6 Astra vs Claude Fable 5.1 vs GPT-5.6 Sol using W&B Weave Learn how to benchmark GPT-6 Astra, Claude Fable 5.1, and GPT-5.6 Sol with Weave, comparing model quality, latency, cost, and performance across practical…
source: coreweave-blog
Sep 28, 2026
product_launch
CoreWeave Becomes the Only Provider to Earn Three Consecutive Platinum ClusterMAX™ Ratings ClusterMAX 3.0 set a tougher standard for AI clouds. See why only CoreWeave has earned Platinum three times and what it means for performance and rel…
source: coreweave-blog
Funding & valuation history
Apr 1, 2026 · funding-research
CoreWeave upsized and priced a $3.5 billion convertible senior notes offering within 48 hours of the Meta deal announcement.
Mar 31, 2026 · funding-research
CoreWeave closed an $8.5 billion delayed draw term loan facility (DDTL 4.0) with an investment-grade rating (A3 by Moody's, A (low) by DBRS), co-structured by MUFG, Morgan Stanley, Goldman Sachs, and JPMorgan.
Feb 1, 2026 · funding-research
CoreWeave secured an $8.5B delayed draw term loan facility using Meta AI contracts as collateral, its first investment-grade rated HPC financing.
Jan 1, 2026 · funding-research
NVIDIA invested $2 billion in CoreWeave at $87.20 per share to expand its data center build-out and deepen the partnership.
Jan 1, 2026 · funding-research
NVIDIA invested $2B in CoreWeave at $87.20/share to expand their data center partnership.
Mar 28, 2025 · funding-research
CoreWeave raised $1.5 billion in its IPO on Nasdaq under ticker CRWV, the largest AI-related listing by amount raised per Dealogic.
Mar 28, 2025 · funding-research
CoreWeave completed its IPO on Nasdaq (CRWV), raising $1.5B.
Mar 27, 2025 · funding-research
CoreWeave reduced its IPO size from $2.7B to $1.5B.
Nov 1, 2024 · funding-research
CoreWeave closed a $650M secondary share sale with investors including Jane Street, Magnetar, Fidelity, and Macquarie Capital.
Nov 1, 2024 · funding-research
CoreWeave closed a $650 million secondary share sale with investors including Jane Street, Magnetar, Fidelity, and Macquarie.
Strategic Position
What they do best
Packaging GPU capacity as a financeable neocloud whose valuation now anchors how peers like SpaceX are priced ahead of an expected June 2026 NASDAQ listing (per peer_candidates/spacex).
Their bet
That a public-market neocloud comp can out-raise and out-build vertically integrated rivals before SpaceX monetizes the Colossus 100K+ GPU site and its first external customer Anthropic (May 2026, per peer_candidates/spacex).
Top risk
If SpaceX's June 2026 $SPCX listing and Anthropic Colossus deal land, CoreWeave loses its 'only liquid AI-compute comp' premium within 6 months — directly falsifiable by that listing date.
Forward view
Derived from public capacity & product announcements — not a company-authored roadmap.
product launch · Oct 1, 2026
Introducing CoreWeave Notebooks for Interactive AI Development CoreWeave Notebooks are now available in CoreWeave Forge, with reactive Python notebooks connected to your data and saved in your project so your team can build on them.
product launch · Oct 1, 2026
ART-Optimized Megatron (AOM): 12x higher training throughput ART-Optimized Megatron (AOM) improves RL training throughput by up to 12x using shared-prefix training, optimized attention, sequence packing, and tuned parallelism for ART workloads.
product launch · Oct 1, 2026
What Is ARIA, and How Did We Build It? Putting an agent next to your experiments was the easy part. The real work was making its research outlast the chat without taking authority from the researcher.
product launch · Oct 1, 2026
Why I Built Fugue: Controlled Experiments for Agent Systems Generating code got cheaper. Knowing which change actually helped did not. Fugue tests one agent-system change at a time, with human approval and evidence that has to reconcile.
product launch · Oct 1, 2026
CoreWeave ARIA: The Coding Agent That Drives Continuous Improvement Across the AI Loop CoreWeave ARIA is generally available in Forge. Whether you're building an agent or training a model, ARIA analyzes your experiments, proposes what to try next, and launches it with your approval.
product launch · Oct 1, 2026
How We Benchmark ARIA—and Why It Runs on Weave A successful retry proves little. Here's how we benchmark ARIA on Weave, testing proposed changes against the production version so real improvements ship and regressions don't.
Compared to peers
Direct competitor
SpaceX
SpaceX owns the xAI Colossus 100K+ GPU site and ground systems; CoreWeave rents/operates third-party capacity without a captive frontier-lab anchor tenant.
Substitute
SpaceX
An anchor-customer-funded operator (xAI/Anthropic Colossus) can substitute for a merchant neocloud by absorbing demand internally rather than buying CoreWeave capacity.
Why someone would join
- 1.Joining ahead of the IPO when CoreWeave is the public-comp reference cited for SpaceX's expected June 2026 NASDAQ listing ($SPCX) — equity timing tied to a dated liquidity event, not vague growth.
- 2.Direct competitive pressure is concrete and scheduled: SpaceX's Colossus 100K+ GPU cluster takes Anthropic as first external customer May 2026, so roadmap work here is measured against a named rival and date.
Recent Hiring (60 days)
No specialty signals in the last 60 days.