• AI KATANA
  • Posts
  • Apple will spend more than $500 billion in the U.S. over the next four years

Apple will spend more than $500 billion in the U.S. over the next four years

Also: DBS CEO sees job cuts of 4,000 temp staff as AI replaces roles

Author

AI KATANA
February 24, 2025

Hello!

Today’s newsletter brings exciting developments from major tech companies significantly amplifying their investments in AI and technology infrastructure. Apple has announced an unprecedented investment exceeding $500 billion over the next four years in the U.S., notably establishing an advanced manufacturing facility and enhancing AI and silicon engineering capabilities. Similarly ambitious, Alibaba plans to allocate more than $52 billion towards cloud computing and AI, positioning itself prominently in China’s AI sector race. Google continues to innovate, testing new AI video capabilities within its Gemini platform, promising accessible high-quality video generation. In workforce trends, DBS reveals intentions to phase out 4,000 temporary roles, replaced by AI technology. Qatar partners with Scale AI, aiming to revolutionize government services through advanced AI applications. Meanwhile, a notable industry shift emerges towards Small Language Models (SLMs), which offer specialized, efficient, and cost-effective AI solutions. Lastly, DeepSeek introduces an open-source initiative and the FlashMLA decoding kernel, while Journify successfully raises $4 million to enhance AI-powered, privacy-focused marketing solutions.

Sliced just for you:

  • 👩🏻‍💻Apple will spend more than $500 billion in the U.S. over the next four years

  • 💰 Alibaba to invest more than $52 billion in AI over next 3 years

  • 🎬 New Google AI Leak Reveals Powerful Gemini Video Upgrade

  • 🏦 DBS CEO sees job cuts of 4,000 temp staff as AI replaces roles

  • 🇶🇦 Qatar signs deal with Scale AI to use AI to boost government services

  • 🔬 Small Language Models Could Redefine The AI Race

Apple has committed to investing over $500 billion in the United States within the next four years, marking its largest spending initiative ever undertaken. This significant investment will cover a wide array of projects including the construction of a new advanced manufacturing facility in Houston, Texas, intended to produce servers crucial for supporting Apple’s personal intelligence system, Apple Intelligence, and its secure Private Cloud Compute platform. The company plans to double its U.S. Advanced Manufacturing Fund to $10 billion, enhancing production capabilities, particularly in advanced silicon through a significant partnership with TSMC’s facility in Arizona. Additionally, Apple will establish a Manufacturing Academy in Detroit to support small- and medium-sized businesses in adopting AI-driven manufacturing technologies and to provide comprehensive skills training. The investment further encompasses significant expansions of Apple’s R&D operations nationwide, especially in AI, silicon engineering, and software development, with an expected creation of around 20,000 jobs predominantly focused in these high-tech areas. This extensive commitment is anticipated to bolster Apple’s already substantial contributions to U.S. employment and economic growth, further solidifying its role as a major taxpayer and employer in the country.

Alibaba has announced plans to significantly expand its investment in AI, dedicating over $52 billion (approximately 380 billion yuan) to its cloud computing and AI infrastructure within the next three years. This ambitious investment marks a considerable escalation compared to the company’s cumulative spending on AI and cloud services over the past decade. The announcement aligns with Alibaba’s impressive start to 2025, driven by strategic initiatives and partnerships that have led to substantial investor interest, reflected in a stock price increase exceeding 68% this year. Additionally, this major financial commitment positions Alibaba prominently in China’s intensifying AI competition, mirroring similar strategic investments by other leading Chinese tech companies, such as ByteDance, which has earmarked significant resources toward AI-related capital expenditure.

Google is currently testing powerful new AI video creation capabilities within its Gemini platform, which could soon enable users to generate realistic videos directly from text prompts. This exciting development includes potential integration with Google’s advanced text-to-video AI model, Veo 2, developed by Google DeepMind. Veo 2, already accessible in limited form as VideoFX via Google Labs and more broadly through Freepik’s creative suite, can produce highly realistic videos based solely on text or image inputs provided by users. Such integration with Gemini represents a significant step forward in expanding AI’s multimedia generation capabilities, potentially making advanced video creation more accessible to the general public and professionals alike. However, due to the substantial computational resources required for high-quality AI video generation, these advanced tools are expected to be limited primarily to users with paid subscriptions.

🏦DBS CEO sees job cuts of 4,000 temp staff as AI replaces roles

DBS Group Holdings has announced it will reduce approximately 4,000 temporary and contract roles over the next three years as AI begins to assume responsibilities previously fulfilled by human employees. The planned workforce reduction will occur through natural attrition, meaning these positions will gradually phase out rather than through immediate layoffs. Importantly, permanent staff will remain unaffected by this transition. This decision aligns with broader industry trends, as financial institutions globally anticipate significant job reductions over the coming years due to AI adoption, with estimates suggesting potential cuts reaching up to 200,000 positions worldwide.

Qatar has entered into a strategic five-year partnership with Scale AI, aimed at integrating advanced AI tools to enhance government services. Through this collaboration, Scale AI will implement AI-driven solutions such as predictive analytics, automation, and sophisticated data analysis, significantly streamlining governmental operations. Over the span of the agreement, more than 50 potential AI applications will be developed specifically for the Qatari government. This initiative positions Qatar competitively within the region, aligning with its goal to emerge as a leader in AI amid growing competition from neighboring nations, including Saudi Arabia and the UAE. Scale AI, recognized for supplying accurately labelled datasets used in AI model training for prominent entities like OpenAI, Microsoft, and Morgan Stanley, sees this agreement as a potential global blueprint for similar government collaborations.

Recently, there’s been a notable shift in the AI sector from LLMs to smaller, domain-specific models known as Small Language Models (SLMs). These smaller models are tailored to specific industries and tasks, providing more precise, efficient, and cost-effective solutions compared to their larger counterparts. SLMs excel in real-world applications by consuming fewer computational resources and being able to operate on local devices, which enhances data security and reduces compliance risks associated with cloud-based models. Industry experts suggest that SLMs enable more autonomous and accurate decision-making, especially valuable in sectors like finance, healthcare, and manufacturing, where tailored AI solutions can actively manage tasks such as financial trading, patient diagnostics, and equipment maintenance. While challenges remain in curating high-quality, domain-specific training data, the overall trend indicates a growing preference for these specialized models, driven by their economic advantage and effectiveness in delivering tangible business outcomes.

🛠️ AI tools updates

DeepSeek has recently announced an exciting open-source initiative aimed at accelerating technology innovation by progressively releasing five code repositories over the span of a week. Emphasizing collaborative growth, DeepSeek positions itself as community-focused, highlighting incremental yet meaningful contributions to the broader technological ecosystem. Alongside this initiative, DeepSeek unveiled FlashMLA, a decoding kernel specifically optimized for Hopper GPUs. FlashMLA is designed to efficiently handle variable-length sequences and is already operational in production environments. This move underscores DeepSeek’s commitment to transparent, collaborative innovation, providing valuable tools that foster technological advancements and community development.

💵 Venture Capital updates

Journify has successfully secured $4 million in its latest funding round led by Silicon Badia, with additional participation from RZM and strategic investors. Since its launch in March 2024, Journify has swiftly reached a significant milestone, achieving $1 million in Annual Recurring Revenue (ARR) within just nine months. This funding highlights Journify’s position as a leader in AI-driven, privacy-first marketing platforms. Journify’s technology addresses key industry issues, including the sharp rise in customer acquisition costs—which have surged 200% over the past decade—by optimizing Return on Ad Spend (ROAS) and reducing the Cost Per Purchase for brands significantly. With strategic partnerships including major platforms like TikTok, Google, and Snap, Journify’s personalized campaigns now effectively reach over 30 million users in the MEMA region. The platform leverages real-time first-party data activation and AI innovations such as Predictive Models, Advanced Attribution, and Real-Time Data Health & Alerts, delivering up to 5 times higher match rates for precise targeting and enhanced marketing efficiency.

🫡 Meme of the day

⭐️ Generative AI image of the day