Tech Giants Scramble: Did DeepSeek R1 Just Burst the AI Hype Bubble?

DeepSeek R1 makes stocks tumble, Meta's $65B AI push, Perplexity proposes Government stake in TikTok

January 29, 2025

In partnership with

The AI race is reaching a fever pitch, with groundbreaking developments from every corner of the globe. DeepSeek R1, a low-cost Chinese model, is shaking up the U.S. tech industry, while Meta’s $65B investment signals its ambition to dominate the AI landscape.

Alibaba’s Qwen2.5 models are redefining long-context capabilities, and Perplexity AI’s bold proposal to merge with TikTok U.S. has the world watching. Dive into these stories and see how AI is reshaping industries worldwide.

This week in AI:

  • DeepSeek R1: The Game-Changing Chinese AI Model Shaking the Tech World

  • Meta's $65B AI Push: Aiming for Industry Leadership

  • Alibaba’s Qwen2.5 Models Set a New Benchmark

  • Perplexity AI Proposes U.S. Government Stake in TikTok

  • L’Oréal and IBM Partner to Revolutionize Sustainable Cosmetics with AI


DeepSeek, a year-old Chinese AI startup, has stunned the tech industry with its groundbreaking AI model, DeepSeek R1, rivaling OpenAI's GPT-4 and Google's Gemini at a fraction of the cost. Developed using just $5.6 million and less advanced AI chips, R1 defies expectations, even as the U.S. enforces chip export restrictions to China. The open-source model, now dominating app store charts with nearly 2 million downloads, raises questions about the financial sustainability of America’s AI dominance.

Wall Street is reeling, with major tech stocks tumbling in response to fears that U.S. AI leadership faces serious competition. While skeptics highlight the unverified low-cost claims and R1's focus on consumer-facing applications, the breakthrough underscores the potential for more cost-efficient innovation in AI development.

Source: Reuters

An entirely new way to present ideas

Gamma’s AI creates beautiful presentations, websites, and more. No design or coding skills required. Try it free today.


Meta CEO Mark Zuckerberg revealed a $60-65B AI infrastructure investment for 2025, including a 1GW datacenter and 1.3M GPUs, positioning Meta AI as the leading assistant and Llama 4 as the state-of-the-art model. The spending marks a 70% increase from 2024 and reflects Meta’s ambitions to reach 1 billion users this year.

This announcement comes amid fierce competition, following DeepSeek R1’s low-cost breakthrough and OpenAI’s $500B U.S. AI infrastructure plans. The AI race is heating up, with companies pouring billions into building the next generation of data centers to maintain dominance in the field.

Source: Mark Zuckerberg on Facebook


Alibaba’s Qwen team unveiled two open-source models, Qwen2.5-1M, with 7B and 14B parameters, supporting 1M token context lengths while delivering up to 7x faster processing using a custom vLLM framework. The models outperform rivals like Llama-3 and GPT-4 on complex long-text tasks and come with an upgraded Qwen Chat v0.2 offering web search, text-to-video generation, and enhanced image features.

This breakthrough highlights the rapid advancements in long-context AI, with models like Google’s Gemini and Flash 2.0 paving the way for massive input capabilities. Faster processing and extended context lengths unlock unprecedented possibilities for data analysis and complex use cases.

Source: Qwen

Perplexity AI revised its merger proposal for TikTok’s U.S. operations, suggesting the creation of ‘NewCo,’ a company potentially valued at $300B post-IPO. The plan includes up to 50% U.S. government ownership, addressing a key requirement of President Trump, while ByteDance retains the recommendation algorithm.

This bold move positions Perplexity as a rising AI powerhouse but faces competition from heavyweights like Elon Musk, Oracle, and Microsoft. With Trump granting TikTok a 75-day extension for negotiations, the proposal is ambitious but far from guaranteed.

Source: Getty Images

L’Oréal is teaming up with IBM to develop a groundbreaking AI foundation model tailored for cosmetic formulation. This bespoke system will accelerate L’Oréal’s R&D efforts, enabling the creation of eco-friendly products using renewable raw materials while reducing energy and material waste. Positioned as the first AI model of its kind in the cosmetics industry, this partnership underscores a commitment to innovation, sustainability, and personalization.

By leveraging IBM’s generative AI expertise and L’Oréal’s deep beauty science knowledge, the collaboration aims to revolutionize cosmetic development. Beyond creating new products, the AI will analyze massive datasets of formulas and ingredients to reformulate existing products sustainably. This initiative could set a new standard for AI-driven sustainability across industries.

Source: Pexels

Tools to Check Out this Week

Ghola - Get personalized conversations wit this AI chatbot

Ocoya - Create & schedule posts faster on social media

HereAfter AI - Preserve memories through interviews, stories, & cards for friends and family

SaaS Prompts - AI & ChatGPT prompts for SaaS entrepreneurs

This Week’s Jobs On insidersedge.io 

Lead Cryptographer & ML Engineer - Cerebellum Network

ML Egineer Intern - OP3N

AI Backend Engineer - Harmony

Senior Technical Lead - ChainGPT

Thanks for tuning into today’s edition!

How's your vibe with our newsletter?

Be brutally honest. DM me or email me back with any suggestions!