/
Navigation
Chronicles
Browse all articles
Explore
Semantic exploration
Research
Entity momentum
Nexus
Correlations & relationships
Story Arc
Topic evolution
Drift Map
Semantic trajectory animation
Posts
Analysis & commentary
Pulse API
Tech news intelligence API
Browse
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
Concept Search
Semantic similarity search
High Impact Stories
Top coverage by position
Sentiment Analysis
Positive/negative coverage
Anomaly Detection
Unusual coverage patterns
Analysis
Rivalry Report
Compare two entities head-to-head
Semantic Pivots
Narrative discontinuities
Crisis Response
Event recovery patterns
Connected
Search: /
Command: ⌘K
Embeddings: large
TEXXR

Chronicles

The story behind the story

days · browse · Enter similar · o open

A look at the Nvidia A100, a ~$10K GPU that has become a critical tool for generative AI; Nvidia has an estimated 95% market share of machine learning GPUs

- Companies like Microsoft and Google are fighting to integrate cutting-edge AI into their search engines, as billion-dollar competitors …

CNBC Kif Leswing

Context & Ripple Effects

The A100 was launched as a datacenter chip in May 2020, with 54 billion transistors and roughly 20 times Volta's performance, and was later expanded with an 80GB variant for supercomputers unveiled in May 2020. At the time, analysts were already benchmarking it against Intel and GraphCore on performance and economics head-to-head against Intel and GraphCore.

Three years on, that same ~$10K part has become the critical tool for generative AI, with an estimated 95% share of machine learning GPUs — and Microsoft and Google are now fighting to integrate cutting-edge AI into their search engines, making A100 capacity a direct input to that race.

First-order effects

  • Microsoft and Google's search AI buildout runs on Nvidia silicon, so every new AI feature they ship translates into demand for ~$10K A100 units and pricing power for Nvidia.
  • At an estimated 95% share of machine learning GPUs, Nvidia effectively sets the supply and cost floor for any company training or deploying generative models right now.

Second-order effects

  • The dependency forces AMD, Intel, GraphCore, and in-house chip efforts at Amazon's and Google's cloud units to position as alternatives, a rival landscape analysts size at Nvidia holding 80%+ of the market Nvidia's AI chip rivals.
  • Cloud providers competing with Microsoft and Google in search-adjacent AI have an incentive to fund second-source silicon, since A100 supply constrains how fast they can scale their own AI offerings.

Third-order effects

  • If the pattern holds, GPU capacity becomes utility-like infrastructure for AI — a scarce input whose allocation shapes which companies can compete in generative products at all.
  • Sustained near-monopoly economics in ML accelerators invite structural responses: customer-funded alternative chips, and eventually regulatory scrutiny of a single-vendor bottleneck for a strategic technology.

The trend: Generative AI is converting Nvidia's accelerator dominance into a strategic chokepoint, pushing its biggest customers to fund alternatives while GPU capacity becomes the gating resource for AI competition.

Discussion

  • @sarbjeetjohal Sarbjeet Johal on x
    .@nvidia jumps 14% after stellar earnings and a rosy outlook. Around 60% of their revenue comes from datacenter, most of that is from #Cloud #hyperscalers. IMO Nvidia will keep benefitting from rise of #AI #MWC23 #Barcelona @dvellante @furrier https://www.cnbc.com/...
  • @stagefright__ Ilija on x
    8 high-powered graphic cards each at 300W to compute a single answer to your chatGPT quip. if true that's at least 2400 Watts thrown at every question. for context my Macbook's (M1 Max) chip draws around 65W at max load https://twitter.com/... https://twitter.com/...