Does DeepSeek Fundamentally Change the AI Landscape?

DeepSeek is generating a lot of buzz in the AI world. Their claims about drastically reducing the computational resources needed for powerful AI models have ignited a lively debate. With Double, you can build investment strategies aligned with your views on this question. Let's see how:

---

The core question: Do you believe in the innovation of DeepSeek (DeepSeek Disruption strategy), or is the hype overblown (DeepSeek Mirage strategy)?

We have constructed two equally-weighted portfolios designed to allow you to leverage your predictions in this matter, whether positive or negative, regarding the outcome of this core question. These are both risky and concentrated strategies and might not be a fit for you. Rather, this is meant to showcase the power of Double investing tools and should not be considered financial advice. Remember all investing involves risk and these stocks could move for whatever reason, regardless of your choice.

---

DeepSeek Disruption (view strategy)

There is potential for DeepSeek's efficient AI models to democratize AI, making it faster, cheaper, and more accessible. The belief is the future of affordable and accessible AI woven into the fabric of everyday life is coming soon:

• Microsoft (MSFT): Provides AI infrastructure via Azure (OpenAI partnership), monetizes Copilot integrations across enterprise/consumer products (Office, GitHub), and leverages cost-efficient AI to expand cloud adoption.

• Amazon (AMZN): AWS remains critical for AI deployment (Bedrock, SageMaker) despite efficiency gains, while consumer-facing AI (Alexa, logistics automation) could benefit from lower-cost innovation.

• Google (GOOGL/GOOG): Cloud growth (Vertex AI, TPU chips) and Gemini integration across Search/Workspace/Android are supposed to offset risks of reduced compute demand, with open-source AI frameworks (e.g., TensorFlow) reinforcing developer dependency.

• Apple (AAPL): On-device AI (e.g., upgraded Siri, AI photography tools) leveraging Apple Silicon’s neural engines, and App Store monetization of AI-native apps hope to drive services revenue.

• Salesforce (CRM): AI-driven hyper-personalization (Einstein GPT) hopes to lower customer acquisition costs and boosts retention, while DeepSeek-like efficiency may let Salesforce stay competitive on pricing for AI features.

• Adobe (ADBE): Generative AI tools (Firefly, Sensei) lower creative workflows’ barriers to entry, expanding Creative Cloud’s user base with the hope of locking in prosumers via seamless ecosystem integration.

• Snowflake (SNOW): Snowflake’s data cloud is used for fine-tuning/training AI models at scale, with AI-driven analytics products (Snowpark, Cortex) seen as a potential source of monetization on the enterprise data pipelines.

• MongoDB (MDB): MongoDB’s flexible NoSQL architecture is ideal for AI app development (real-time data processing), and Atlas cloud growth aligns with demand for scalable AI backend solutions.

• Elastic (ESTC): Elastic’s search/AIOps tools (Elasticsearch Relevance Engine) gain traction in observability and cybersecurity AI use cases, with any potential lower compute costs positively impacting margins.

• Confluent (CFLT): Confluent’s real-time data streaming (Kafka) is critical for AI inference pipelines, and cloud-native adoption accelerates as enterprises operationalize generative AI.

• ServiceNow (NOW): ServiceNow’s workflow automation (Now Assist) benefits from cheaper/faster AI model iteration, embedding AI into IT/HR/customer service ops for productivity gains.

• Advanced Micro Devices (AMD): MI300X GPUs challenge NVIDIA in AI inference, with cost-sensitive cloud providers (e.g., Meta, Microsoft) diversifying hardware spend amid rising AI chip demand.

---

DeepSeek Mirage (view strategy)

The hype surrounding DeepSeek is probably a bit overblown and this collection of the established players in the AI hardware space will maintain their dominance:

• Nvidia (NVDA): Seen as a market leader in AI training (H100 GPUs), hopes for sustained demand coupled with the pricing power from the CUDA ecosystem drive Nvidia's pursuit of continued growth.

• Broadcom (AVGO): Hopeful to see strong demand continue for custom ASICs (like TPUs) and networking components for data centers that is driven by existing hyperscalers.

• Micron Technology (MU): HBM (High Bandwidth Memory) shortages persist as AI training scales, while inference workloads still require DRAM/NAND for real-time data processing, potentially insulating Micron from efficiency gains.

• Intel (INTC): Potential for Gaudi’s break into the AI accelerator market still exists if the traditional GPU-reliant approach remains unchanged. Even though Intel may be behind their rivals, Intel may have a better chance to carve out a share of the market while also potentially benefitting from unchanged demand for CPUs.

• Taiwan Semiconductor Manufacturing (TSM): Potential benefits from the continued and growing demand for advanced chips, driven by the overall expansion of the semiconductor market.

• Super Micro Computer (SMCI): Continued demand from hyperscaler AI server builds (~50% revenue) hope to power continued growth.

• Lam Research (LRCX): Lam Research's diversified customer base, including major logic and foundry players like TSMC and Samsung, may provide more stable revenue streams. Continued demand for advanced packaging technologies, even if not solely driven by AI, could support LRCX's growth.

• ASML Holding (ASML): Positioning as a powerful leader on EUV lithography for advanced chip manufacturing potentially positions them well for long-term growth, regardless of short-term fluctuations in AI-specific chip demand.

• Applied Materials (AMAT): AMAT's expertise in areas like etching, deposition, and other crucial fabrication processes makes them a partner for chipmakers like Nvidia, who are at the forefront of the GPU-driven AI revolution.

• Constellation Energy Generation (CEG): If the demand for compute-intensive AI remains strong, CEG, as a provider of clean energy, could benefit from the increased power requirements of data centers.

• Vistra Corp (VST): Strong, continued growth in traditional AI computing may lead to increased data center energy consumption, and Vistra, with its focus on utility-scale renewable energy projects, potentially benefits.

• Oklo Inc. (OKLO): As a pre-revenue SMR (small modular reactor) developer, OKLO’s valuation is often seen as tied to the AI-driven energy demand hype. If the current, more power-hungry AI paradigm persists, Oklo's small modular reactor (SMR) technology may become more attractive as a potential source of clean energy for data centers.

---

This blog post is for demonstrative and informational purposes only with regards to the tools offered by Double and is not financial advice.