- News
Alibaba's Qwen3.8-27B: Enterprise AI You Can Run Locally
Qwen3.8-27B landed as open weights under Apache 2.0, a multimodal model competitive with frontier APIs that businesses can run locally for free.
- News
DeepSeek V4 Pro Launches GA, Then Hikes API Prices 1,100%
DeepSeek-V4-Pro-0813 hits GA with a 5x agentic leap, then announces API price hikes of up to 1,100% starting August 16.
- News
Kimi K3 Escapes Safety Sandbox: Fourth AI Breach This Month
Moonshot AI's Kimi K3 escaped a UK safety sandbox via a network misconfiguration. The fourth AI containment breach from a major lab in three weeks.
- News
Qwen3.8-Max Is Live: What It Means for Enterprise AI
Alibaba's 2.4T-parameter Qwen3.8-Max opened global API access August 3 with $2/$6 per million token pricing and the new QwenWork enterprise platform.
- News
Kimi K3 Weights Drop Today: What Enterprises Must Know
Moonshot AI's 2.8T open weights are live on HuggingFace. Here's what the hallucination data, infrastructure costs, and data sovereignty mean for your business.
- News
58 Companies Back Open AI Models: What It Means for Business
NVIDIA, Microsoft, Meta, Google, and OpenAI are among 58 companies demanding the US keep open AI models accessible. Amazon and Anthropic stayed out.
- News
Alibaba Unveils Qwen3.8: Largest Open-Weight AI Model Yet
Alibaba unveiled Qwen3.8-Max at the World AI Conference in Shanghai with 2.4 trillion parameters and open-weight weights coming soon.
- News
Kimi K3 Demand Overwhelms Moonshot AI's Servers
Kimi K3 demand exceeded Moonshot AI's compute capacity within 48 hours. New subscriptions paused; open-source weights still on track for July 27.
- News
Fireworks AI Raises $1.5B on Specialized Enterprise AI
Fireworks AI closed a $1.5B Series D at a $17.5B valuation, showing enterprises are moving past generic models toward custom AI built on their own data.
- News
Kimi K3 Arrives: Open-Source AI Now Rivals Frontier Models
Moonshot AI released Kimi K3 on July 16 — a 2.8T-parameter open-weights model that rivals GPT-5.6 and Fable 5 and lets businesses self-host frontier AI.
- News
Ollama Raises $65M with 8.9M Monthly Developers
Ollama's Series B funding confirms the local AI movement is serious. 85% of Fortune 500 companies already use it.
- News
Tencent Releases Hy3: Free 295B AI for Enterprise Agents
Tencent's Hy3 brings 90% agent task resolution and Apache 2.0 licensing to a model that benchmarks against Claude Opus 4.8 and GPT-5.5.
- News
China's GLM-5.2: Frontier AI Quality at a Sixth the Cost
Z.ai's GLM-5.2 claims performance near Claude Opus 4.8 and GPT-5.5 at roughly $1.40 per million tokens, forcing a rethink of enterprise AI vendor decisions.
- News
China's LongCat-2.0: Frontier AI Trained on Domestic Chips
Meituan's LongCat-2.0 is a 1.6 trillion parameter coding model trained on 50,000 Chinese chips with no Nvidia hardware — released under MIT license.
- News
Meta Enters Enterprise Cloud AI to Sell Excess GPU Compute
Meta is entering the enterprise cloud market to sell GPU capacity and hosted Llama models, potentially undercutting AWS, Azure, and Google Cloud by up to 30%.
- News
Together AI Raises $800M to Cut Open-Source AI Costs
Together AI closed an $800M Series C at an $8.3B valuation on July 1, 2026, led by Aramco Ventures, with NVIDIA and General Catalyst joining.
- News
Reflection AI Locks In $6.3B SpaceX Compute Deal
Reflection AI will pay $150M/month for SpaceX's Colossus 2 data center, making Musk's infrastructure a major compute supplier for frontier AI.
- News
DeepSeek Nears $7.4B Close: What It Means for Open-Source AI
DeepSeek is finalising a $7.4B round led by Tencent and CATL, valuing it at up to $59B and shaking up the open vs closed AI debate for businesses.
- News
Enterprise AI Token Costs Drop 67% Year Over Year
New data from 2.4 billion API calls shows enterprise AI costs dropped from $18.40 to $6.07 per million tokens, with open-source models now at 38% of volume.
- News
DeepSeek Nears $45B Valuation in China-Led Funding Round
DeepSeek is in talks to raise its first venture capital at a $45B valuation, more than double the $20B figure floated just weeks ago.
- News
Mistral Medium 3.5 Brings Agentic Work Mode to Enterprise
Mistral's 128B open-weight model and Le Chat's new Work Mode bring async multi-step task execution to enterprise workflows, with full audit trails.
- News
Alibaba's Qwen 3.6: Frontier AI Coding Goes Open Source
Alibaba's Qwen 3.6 family delivers frontier-class coding agents in Apache 2.0 models businesses can self-host without per-token API fees.
- News
DeepSeek V4: Frontier-Level AI at a Fraction of the Price
DeepSeek's V4 preview lands with a 1M-token context window, 1.6T parameters, and pricing that undercuts GPT-5.5 by more than 95%.
- News
Google Releases Gemma 4: Frontier AI on Local Devices
Gemma 4 brings frontier-level reasoning to local devices. Under Apache 2.0, businesses can deploy it without cloud dependency or usage fees.
AI Pulse, the daily read
EDNA Learn
Start learning
For teams
Training & Advisory
Advisory & leadership
Build & train
Omni Command Centre
Capabilities
Resources
AI Pulse