- Generative AI Art
- Posts
- DeepSeek V4 Is Coming on Huawei Chips
DeepSeek V4 Is Coming on Huawei Chips
PLUS: Google Gemma 4 Beats 400B Rivals at 31B, and OpenAI Buys a Media Company
Gen AI Art
April 11, 2026 • Estimated Reading Time: 9 minutes
In partnership with
China’s most controversial AI lab is about to ship its biggest model yet — on Huawei chips, bypassing NVIDIA entirely. Google just dropped an open-weight model that punches way above /its weight class. And OpenAI made its first-ever media acquisition. Here’s what you need to know.
Today in AI:
DeepSeek V4 targets late April launch — 1 trillion parameters on Huawei’s Ascend chips
Google’s Gemma 4 ranks #3 open model worldwide at just 31B parameters
OpenAI acquires TBPN, its first media company purchase
The IT strategy every team needs for 2026
2026 will redefine IT as a strategic driver of global growth. Automation, AI-driven support, unified platforms, and zero-trust security are becoming standard, especially for distributed teams. This toolkit helps IT and HR leaders assess readiness, define goals, and build a scalable, audit-ready IT strategy for the year ahead. Learn what’s changing and how to prepare.
What’s new? DeepSeek is preparing to launch V4, a 1-trillion-parameter Mixture-of-Experts model built entirely on Huawei’s Ascend 950PR chips — the first frontier model to bypass NVIDIA’s CUDA ecosystem. The model features a 1M-token context window, native multimodal generation, and only 37B parameters active per token to keep inference costs low.
What matters?
DeepSeek spent months working with Huawei and Cambricon Technologies to rewrite core pieces of V4’s code for Huawei’s CANN architecture, proving frontier AI can be built outside the NVIDIA ecosystem.
V4-Lite has been live-tested on API nodes since early April, with developers reporting 30% inference speed increases and 94% context recall at 128K tokens — up from 45%.
At roughly $0.28 per million input tokens, V4 is 20–50x cheaper than GPT-5.4 or Claude Opus 4.6 — a pricing earthquake for the industry.
Why it matters?
This is the model that proves China’s AI independence is real, not theoretical. If DeepSeek can train and serve a frontier model on domestic chips at 20x lower cost, the entire Western assumption that export controls would slow Chinese AI development is wrong. The geopolitical implications are enormous.
GUIDE
What’s new? Google released Gemma 4, a family of open-weight models from 2B to 31B parameters under Apache 2.0 with zero commercial restrictions. The 31B Dense model ranks #3 on the Arena AI text leaderboard — outperforming Meta’s Llama 4 on math, coding, and reasoning despite being a fraction of the size.
What matters?
This is the first Gemma family that handles text, image, audio, and video natively across all sizes — true multimodal at the edge.
Available in four sizes (2B, 4B, 26B MoE, 31B Dense) with 256K context windows and 140+ language support, making it viable for everything from phones to enterprise servers.
Apache 2.0 license means no usage restrictions, no commercial gates, no phone-home requirements — the most permissive licensing of any frontier-class open model.
Why it matters?
Google just commoditized a capability that cost millions to develop a year ago. A 31B model matching 400B rivals means the performance gap between open and closed models is collapsing. For startups and enterprises, Gemma 4 eliminates the biggest argument for paying frontier API prices — and it runs on your own hardware.
PRESENTED BY CYTONICS
The Key to This $240B Market Is in Your Bloodstream
Every year, $240B is spent on treating the symptoms of osteoarthritis, yet not a single therapy has stopped it. Cytonics not only discovered why, they found an answer hiding inside the human body all along. Their first-generation therapy has already treated 10,000+ patients. Now they're pushing a 200% more potent version toward FDA approval. Claim a piece as an early-stage investor today.
What’s new? OpenAI acquired TBPN (Technology Business Programming Network), a daily live 3-hour tech talk show founded by John Coogan and Jordi Hays. The deal, reportedly valued in the low hundreds of millions, marks OpenAI’s first acquisition of a media company. TBPN will sit within OpenAI’s strategy org under chief global affairs officer Chris Lehane.
What matters?
TBPN is one of the fastest-growing media companies in tech, with a daily live format that convenes influential voices across tech, business, and culture.
OpenAI says TBPN will maintain full editorial independence — choosing their own guests and making their own editorial decisions, with that protection written into the agreement.
The acquisition gives OpenAI a direct media channel to shape narratives around AI development at a time when public opinion is increasingly polarized on the technology.
Why it matters?
This is OpenAI playing the long game beyond model development. Owning a media platform gives them something no API can buy — cultural influence. As AI regulation debates heat up globally, having a trusted editorial voice in the room matters more than another benchmark score.
Everything else in AI
Perplexity hit $450M ARR after a 50% revenue surge in a single month, driven by its Computer agent platform that orchestrates 19 frontier models — the fastest monthly revenue jump in the company’s history.
Anthropic’s Model Context Protocol crossed 97 million installs in March, with every major AI provider now shipping MCP-compatible tooling as the default mechanism for agents to connect to external tools.
Microsoft announced a $10 billion investment in Japan’s AI infrastructure, following its $17.5 billion India commitment — signaling a massive bet on Asian AI compute capacity.
The USGS unveiled AI technology capable of predicting drought conditions with unprecedented accuracy up to 90 days in advance — one of the most impactful real-world AI applications this year.
Essential AI Guides - Reading List:
Let us know!
What did you think of today's email?Before you go, please give your feedback to help us improve the content for you! |
Work with us
Reach 100k+ engaged Tech Professionals, Engineers, Managers and decision makers. Join brands like MorningBrew, HubSpot, Prezi, Nike, Ahref, Roku, 1440, Superhuman, and others in showcasing your product to our audience. Get in touch now →