+353-1-416-8900REST OF WORLD
+44-20-3973-8888REST OF WORLD
1-917-300-0470EAST COAST U.S
1-800-526-8630U.S. (TOLL FREE)

Thematic Intelligence: Deep Dive into the Semis-AI-ESG Trinity

  • PDF Icon

    Report

  • 36 Pages
  • June 2024
  • Region: Global
  • GlobalData
  • ID: 5983393
Artificial intelligence (AI) is evolving rapidly from three perspectives: software, hardware, and regulation. This makes both IT and financial commitments highly risky. Specifically, AI algorithms are still evolving rapidly, which limits options for hardware acceleration to either i) use case or workload-specific chip development by Big Tech, who are the only ones technically capable and who can afford it, or ii) more generic optimizations using off-the-shelf graphics processing units (GPUs). Big Tech’s advantage will eventually vanish when merchant AI silicon (i.e., commercial AI chips) emerges, which may be less than five years away.

Key Highlights

  • Hardware processing improvements are not keeping up with the increase in AI model sizes. So, barring a semiconductor breakthrough, demand for raw compute capacity in data centers is bound to dramatically increase, increasing AI’s contribution to carbon emissions.
  • As the carbon footprint impact of large language models (LLMs) becomes more transparent, organizations must consider it when selecting an AI delivery model and in real-time orchestration of AI-enabled services. Scope 3 emissions guidance will be needed, and AI vendors must step up disclosures. An LLM’s carbon footprint and its transparency will become a competitive differentiator.

Scope

  • This Deep Dive report looks at the relationship between AI, semiconductors, and environmental, social, and governance (ESG). The environmental cost of AI is becoming a competitive factor. If this development unfolds, the sector will shift away from increasingly larger models and the optimization of AI chips for raw processing performance. Instead, it would focus on smaller models, including small language models (SLMs), somehow reducing the scale advantage of LLM vendors providing extremely large models, and focusing on performance to power. As a result, open-source LLMs would become a more compelling option, as a model’s sheer size and training would not be a barrier to entry any longer, democratizing access to competitive AI technology.

Reasons to Buy

  • Understand the long-term scenarios for the AI market.
  • Discover the likely winners and losers in each of these scenarios.
  • Get recommendations for adopting generative AI in your organization.

Table of Contents

  • Executive Summary
  • AI Technology is in Flux and Evolving Rapidly
  • Power Efficiency in AI Enjoys Large Economies of Scale
  • The AI Market Will Change Dramatically as Technology Matures
  • Glossary
  • Further Reading
  • Thematic Research Methodology

Companies Mentioned (Partial List)

A selection of companies mentioned in this report includes, but is not limited to:

  • 3dfx
  • Alibaba
  • Alphabet
  • Amazon
  • AMD
  • Anthropic
  • Arista
  • ATI
  • Baidu
  • Barefoot
  • Bentley Systems
  • Biren Technology
  • Boeing
  • Broadcom
  • Cambricon
  • Cerebras
  • Cisco
  • Citroen
  • Eleuther.AI
  • Graphcore
  • Groq
  • Huawei
  • Hugging Face
  • IBM
  • Integraph
  • Intel
  • Juniper
  • Marvell
  • Mellanox
  • Meta
  • Microsoft
  • MIT
  • Nvidia
  • OpenAI
  • Renault
  • Salesforce.com
  • Seiko
  • Silicon Graphics
  • Stability.ai
  • United Computing
  • Zuken