⚡ Nvidia makes its biggest AI investment yet

Nvidia is snapping up key assets from Groq in a massive $20B deal focused on AI inference. The agreement brings Groq’s core technology and its founder and senior leaders into Nvidia’s orbit, signaling how serious Nvidia is about owning the full AI stack. Let’s dive in!

In today’s insights:

LATEST DEVELOPMENTS

Evolving AI: Nvidia struck a $20B asset deal with Groq to deepen its AI inference stack.

Nvidia agreed to acquire Groq’s core assets in a $20 billion cash transaction centered on inference technology. The agreement includes a non-exclusive license and the transition of Groq’s founder and senior leaders to Nvidia. Groq will continue operating independently under new leadership, with GroqCloud unaffected. Nvidia plans to integrate Groq’s low-latency processors into its AI factory platform, expanding real-time and inference workloads. The move follows a pattern of talent-and-IP deals as Nvidia builds out its AI stack.

Groq’s official announcement →

AI spend is shifting fast from training to inference, where speed, cost, and predictability decide who wins real workloads. Nvidia pulling Groq’s low-latency designs into its stack shows where demand is coming from right now: chat systems, agents, search, and real-time apps that need instant responses at scale. For developers and enterprises, this points to faster inference options baked directly into Nvidia’s platform, tighter hardware-software coupling, and fewer reasons to experiment outside its ecosystem. The question is how much room remains for independent inference challengers once Nvidia folds these designs into its stack.

Some companies make lofty promises to investors and never deliver. Others use those dollars to unlock new levels of scale.

That’s Westbound & Down’s story. Already Colorado’s most-awarded craft brewery, they opened their doors to investors for the first time to help open a flagship Denver-metro-area location.

With 2,800% distribution growth since 2019 and a retail partnership with Whole Foods, it’s no shock investors maxed out that campaign in less than 60 days.

But it’s what comes next that’s even more exciting. Fresh off Brewery of the Year honors at the 2025 Great American Beer Festival, W&D is scaling toward 4X distribution growth by 2028.

Become an early-stage W&D investor before their Denver-metro location opens and momentum accelerates further.

Invest in W&D Now

_{This is a paid advertisement for Westbound & Down’s Regulation CF Offering. Please read the offering circular at https://invest.westboundanddown.com/}

Evolving AI: OpenAI is exploring ads inside ChatGPT responses as pressure mounts to monetize free users.

OpenAI is considering ads that appear as natural recommendations during ChatGPT conversations. Internally called intent-based monetization, the approach would surface products or services only when they fit the user’s question. The company is weighing generative ads, affiliate-style revenue, and sponsored GPTs, all aimed at funding rising infrastructure costs tied to model training and inference.

OpenAI already rolled out “shopping research” in ChatGPT, so ads wouldn’t arrive as random banners, they’d show up right where people ask for help picking products and services. That shifts ad power from keywords to conversation: the model can spot intent, write the pitch, then steer the next click or purchase, which is gold for retailers and a direct threat to classic search ads. Reuters reported OpenAI has been testing shopping-related ad formats, even as leadership talks about delaying ad work during internal “code red” focus periods. For users, the daily change is simple: recommendations start needing labels and receipts, or you’ll never know if you’re getting clean advice or paid placement.

A festive Emirates clip just went viral, not because of where the plane was flying, but because of how unreal it looked. The holiday video showed an A380 decked out with Santa’s sleigh visuals, and a huge chunk of the internet immediately assumed it had to be AI-generated. The effects looked too clean, too perfect, too 2025. But according to Emirates, there was no generative AI involved at all.

The fact that so many people got it wrong says a lot about where visual trust is right now. If you saw it without context, would you have clocked it as human-made, or would you have sworn it was AI too?

Evolving AI: Salesforce is dialing back generative AI after trust and reliability issues.

Salesforce says large language models proved unpredictable in real business use. Executives point to missed instructions, task drift, and failures once prompts grow complex. After rolling out AI agents and cutting around 4,000 support roles, the company is now prioritizing deterministic triggers and data foundations. Agentforce remains a revenue bet, though the tone has shifted from AI-first hype to controlled automation.

Salesforce is basically saying out loud what a lot of enterprise teams already learned the hard way: LLMs are great at talking, shaky at doing. When an agent can ignore one instruction after prompt #9, or “drift” the moment a customer goes off-script, you end up adding rules, triggers, and data grounding just to keep workflows stable. That shift toward guided determinism shows where real spend is heading next year: less “AI everywhere,” more controlled automation that can pass audits, hit SLAs, and ship the same outcome every time.

“When I first opened Attio, I instantly got the feeling this was the next generation of CRM.”
— Margaret Shen, Head of GTM at Modal

Attio is the AI-native CRM for modern teams. With automatic enrichment, call intelligence, AI agents, flexible workflows and more, Attio works for any business and only takes minutes to set up.

Join industry leaders like Granola, Taskrabbit, Flatfile and more.

👉Start for free today

QUICK HITS

🤖 Google Introduces An Open Sourc Protocol for Agent Driven Interfaces.

🧠 Deep-learning electronic structure calculations.

🇮🇹 Italy Doesn’t Want Meta’s WhatsApp AI Rules.

*partner link