Welcome back to LLM Decode 👋
The AI race is expanding across both capital and capability. Nvidia is reportedly considering a major investment in Perplexity at a valuation above $30 billion, while DeepSeek is adding visual understanding to V4 Flash, making lower-cost AI increasingly competitive with premium multimodal models.
Here’s what matters today.
Google pushes deeper into the AI agent race
Nvidia is reportedly discussing an investment in Perplexity as part of a funding round that could value the AI search and agent startup at more than $30 billion.
The potential valuation would represent a significant jump from Perplexity's previous financing. The company's annualized revenue has also grown rapidly, helped by rising adoption of its AI search products and cloud-based agent tools for professional workflows.
For Nvidia, the potential investment fits a broader strategy of supporting AI companies that create more demand for computing infrastructure and advanced GPUs.
Why It Matters
Perplexity is evolving from an AI search engine into a broader agent and productivity platform.
Nvidia is strengthening its influence across both AI hardware and the applications running on it.
Rapid revenue growth shows businesses are increasingly willing to pay for AI-powered research and automation.
Strategic investments are becoming another way chipmakers can shape the AI ecosystem.
Sponsored by
Smarter CRM. Less Busywork.
Disconnected data and tools make it harder to understand your customers. HubSpot's Agentic Customer Platform brings your data, teams, and tech stack together with AI built in to help your business work faster and create more personalized customer experiences.
Why HubSpot and what's new
Use AI powered tools to take action faster
Unify your data, teams, and tech stack in one place
Create one shared view of customer data
Connect teams around the same customer context
Bring your business tools into one place
Connect more of your business in one place and give every team a smarter way to work. Get set up quickly and start checking off your hardest tasks.
DeepSeek Gives V4 Flash the Ability to See
DeepSeek has introduced an experimental multimodal version of V4 Flash that adds visual understanding while retaining the model family's text, reasoning, and agent capabilities.
The new model can process images alongside text, allowing it to interpret screenshots, charts, documents, interfaces, and other visual information. This makes it more useful for browser agents and workflows that need to understand what is happening on a screen.
DeepSeek says the experimental model maintains V4 Flash's core text capabilities while significantly improving performance on tasks where visual information is required.
Why It Matters
Affordable models are rapidly gaining capabilities once limited to premium AI systems.
Vision enables AI agents to interact more effectively with websites, software interfaces, and documents.
Multimodal capability is becoming a standard expectation rather than a premium feature.
Businesses will have more options for balancing AI performance with deployment cost.
4 AI Tools & Community Workflows
1. AI Research Agent
Use an AI research assistant to collect sources, compare competitors, summarize findings, and turn large amounts of information into structured reports.
2. Screenshot Analysis Workflow
Give a vision-enabled model screenshots of dashboards, websites, or interfaces to identify problems, extract information, and suggest improvements.
3. Visual Browser Agent
Combine vision models with browser automation so an AI agent can understand page layouts, buttons, charts, and other visual elements instead of relying only on text.
4. Model Cost Comparison
Test the same workflow across premium and lower-cost multimodal models. Compare accuracy, latency, and cost before deciding which model should handle production workloads.
Ready to level up your AI skills?
Explore Our Courses →That’s it for today.
The AI space doesn’t slow down - and neither should your thinking.
See you in the next drop.