The AI chip war just went real dirty

Today, we’re looking at Big Tech’s quiet rebellion against Nvidia on building their own custom AI chips. From Google’s TPUs to Amazon’s Trainium, this shift could redefine who controls the future of AI and who profits from it. Also looking at Ubisoft’s first real attempt to bring generative AI into gameplay not as a gimmick, but as a core mechanic.

In today’s post:

• Nvidia’s AI dominance is real but it’s getting crowded

• Ubisoft’s AI teammates don’t just follow

• This AI learns how you use CAD

What’s Trending Today

AI CHIPS

Big Tech is building its own AI chips

Nvidia’s AI chips are selling faster than ever. But a quiet shift is underway: Google, Amazon, Microsoft, and others are designing custom AI chips to reduce their reliance on Nvidia and maybe even outmatch it.

Here’s everything you need to know:

• Nvidia’s GPU sales are “off the charts,” with systems like the 72-GPU Blackwell rack going for $3 million apiece.

• These GPUs power everything from LLM training to real-time AI inference but they’re expensive, power-hungry, and in short supply.

• Enter ASICs: custom-built chips from Google (TPUs), Amazon (Trainium), and Microsoft (Maia), tailored for specific AI tasks.

• ASICs trade flexibility for efficiency, they’re faster and cheaper for narrow jobs, but can’t be reprogrammed.

• Google’s 7th-gen TPU, Ironwood, just dropped, and Anthropic plans to train Claude on 1 million of them.

• Amazon’s Trainium chips now power its largest AI data centers, serving clients like OpenAI and Anthropic.

• Other players are joining in: Meta, OpenAI (with Broadcom), Tesla, Intel, and a wave of startups like Groq and Cerebras.

• On-device AI is rising too with Apple, Qualcomm, and Samsung baking NPUs into phones, laptops, and cars.

Nvidia isn’t going anywhere, but the future of AI chips won’t belong to one company. As training plateaus and inference takes center stage, hyperscalers want chips that are faster, cheaper, and purpose-built. The era of “one chip fits all” is fading and we’re about to find out what happens when everyone builds their own.

CTV ads made easy: Black Friday edition

As with any digital ad campaign, the important thing is to reach streaming audiences who will convert. Roku’s self-service Ads Manager stands ready with powerful segmentation and targeting — plus creative upscaling tools that transform existing assets into CTV-ready video ads. Bonus: we’re gifting you $5K in ad credits when you spend your first $5K on Roku Ads Manager. Just sign up and use code GET5K. Terms apply.

Use code GET5K now

UPGRADE

This experiment might change how you play

Ubisoft just dropped “Teammates,” a playable AI research project with voice-activated NPCs. It’s a glimpse at how generative AI could reshape the way we interact with games not just how they look or play.

Here’s everything you need to know:

• Teammates isn’t a demo or a tech showcase, it’s a playable experiment built inside a first-person shooter.

• You play as a resistance fighter, guiding two AI squadmates and an AI assistant named Jaspar through a mission using only your voice.

• Say the word, and Jaspar can highlight threats, tweak settings, explain lore, or pause the game mid-fight.

• Your NPC allies, Sofia and Pablo, don’t just follow orders, they interpret them in real time based on your tone and timing.

• Ubisoft’s team is clear: this isn’t about replacing writers or devs, it’s about letting human-made characters respond with more depth and flexibility.

• “We create the fences,” says narrative director Virginie Mosser, “but the AI improvises within them always grounded in the story.”

• Early playtests suggest players quickly formed a bond with Jaspar not just as a tool, but as a character.

Voice-led AI NPCs aren’t just a gimmick they’re a test of how natural interaction could unlock deeper immersion. If done right, this tech could reshape how players relate to characters, not just control them. It’s early days, but this feels like the start of something new not the end of something old.

RESEARCH

MIT’s new AI agent could become every engineer’s design co-pilot

MIT researchers just trained an AI model to use CAD software like a human clicking, dragging, and building full 3D models from simple sketches. It’s a quiet breakthrough that could change how engineers learn and work.

Here’s everything you need to know:

• CAD software is powerful and notoriously hard to learn, with thousands of tools and commands.

• MIT’s team created an AI that learns to use CAD by watching how real users build designs, click by click.

• The secret: a dataset called VideoCAD, with 41,000+ annotated videos showing every move a human makes in CAD.

• Instead of coding geometry directly, the AI mimics human actions selecting tools, clicking menus, and dragging shapes.

• It can take a 2D sketch and turn it into a 3D model autonomously, even in complex scenarios.

• The goal isn’t to replace designers, but to co-pilot with them speeding up workflows and removing tedious steps.

• Researchers believe this could make CAD more accessible to students, creators, and engineers alike.

AI that learns how we work not just what we want might be the next big shift. By lowering the barrier to entry and simplifying design steps, tools like this could unlock new levels of creativity for engineers and makers everywhere. We’re not just teaching AI to build, we’re teaching it to build with us.

🚀 Founders & AI Builders, Listen up!

If you’ve built an AI tool, here’s an opportunity to gain serious visibility.

Nextool AI is a leading tools aggregator that offers:

• 500k+ page views and a rapidly growing audience.

• Exposure to developers, entrepreneurs, and tech enthusiasts actively searching for innovative tools.

• A spot in a curated list of cutting-edge AI tools, trusted by the community.

• Increased traffic, users, and brand recognition for your tool.

Take the next step to grow your tool’s reach and impact.

Reach 140,000+ READERS:

Expand your reach and boost your brand’s visibility!

Partner with Nextool AI to showcase your product or service to 140,000+ engaged subscribers, including entrepreneurs, tech enthusiasts, developers, and industry leaders.

Ready to make an impact? Visit our sponsorship website to explore sponsorship opportunities and learn more!