NVIDIA (NVDA): FY25 Q4 earnings call, Feb 26, 2025
On its Feb 26, 2025 earnings call (FY25 Q4), NVIDIA said 9 of the 23 words Kalshi and Polymarket listed; Parrot’s count agreed with 22 of 23 settlements. Quotes below are at most 25 words each.
UpdatedMarkets
23 words · 9 settled yes · our count agreed on 22 of 23- Kalshi: company speakers
Kalshi - Polymarket: named speaker only
Polymarket
backtest The probabilities on settled rows are a walk-forward backtest: the model as of before each event, computed after Parrot went live; they were not published at the time.
| Word | Result | Our count | All speakers | Agrees | Probability | Market |
|---|---|---|---|---|---|---|
| Blackwell | Yes | 49 | 52 | ✓ | 91%transcripts from the transcript model | – |
| Inference | Yes | 26 | 28 | ✓ | 95%transcripts from the transcript model | – |
| China | Yes | 4 | 5 | ✓ | 88%transcripts from the transcript model | – |
| Microsoft | Yes | 3 | 3 | ✓ | 81%transcripts from the transcript model | – |
| Grok | Yes | 3 | 3 | ✓ | 15%transcripts from the transcript model | – |
| CUDA | Yes | 2 | 2 | ✓ | 82%transcripts from the transcript model | – |
| Supply Chain | Yes | 2 | 3 | ✓ | 82%transcripts from the transcript model | – |
| Perplexity | Yes | 1 | 1 | ✓ | 41%transcripts from the transcript model | – |
| DeepSeek | Yes | 0 | 0 | ✗ | 15%transcripts from the transcript model | – |
| H100 | No | 0 | 0 | ✓ | 82%transcripts from the transcript model | – |
| H200 | No | 0 | 0 | ✓ | 82%transcripts from the transcript model | – |
| Oracle | No | 0 | 0 | ✓ | 64%transcripts from the transcript model | – |
| Tesla | No | 0 | 0 | ✓ | 38%transcripts from the transcript model | – |
| Amazon | No | 0 | 0 | ✓ | 36%transcripts from the transcript model | – |
| TSMC/Taiwan Semiconductor Manufacturing Company | No | 0 | 0 | ✓ | 35%transcripts from the transcript model | – |
| Apple | No | 0 | 0 | ✓ | 28%transcripts from the transcript model | – |
| Taiwan | No | 0 | 0 | ✓ | 28%transcripts from the transcript model | – |
| Satya/Nadella | No | 0 | 0 | ✓ | 25%transcripts from the transcript model | – |
| AMD/Intel | No | 0 | 0 | ✓ | 15%transcripts from the transcript model | – |
| Buyback | No | 0 | 0 | ✓ | 15%transcripts from the transcript model | – |
| Elon/Musk | No | 0 | 0 | ✓ | 15%transcripts from the transcript model | – |
| RTX 5080 | No | 0 | 0 | ✓ | 15%transcripts from the transcript model | – |
| Trump | No | 0 | 0 | ✓ | 15%transcripts from the transcript model | – |
Market prices are mid quotes from Kalshi and Polymarket, shown for context, may be delayed by several minutes and are not a signal. Quotes as of Oct 21, 2026 04:22 UTC.
Transcript
quotes of at most 25 wordsSpeaker-labelled transcript · Source and credits · Used to count words; quotes of at most 25 words.
8 matches for “Hopper”; 5 quoted, at least 25 words apart.
…sequentially and 93% year on year. As the Blackwell ramp commenced, and Hopper 200 continued to contribute growth, In Q4, Blackwell sales exceeded our expectations.…
…with up to 25x higher token throughput and 20x lower cost versus Hopper 100. It is revolutionary. Transformer engine is built for LLM. And mixer…
…AI models. Meta's cutting-edge Andromeda, advertising engine runs on NVIDIA Corporation's Grace Hopper Superchip. Serving vast quantities of ads across Instagram, Facebook applications. Andromeda…
…system architecture is exactly the same. It's a lot harder going from Hopper to Blackwell because we went from an NVLink 8 system to a…
…I'm curious as we now approach the two-year anniversary of really the Hopper inflection that you saw in 2023 in Gen AI in general. We…
Where the listed words were said
…of cutting-edge models and unlock the next level of AI capabilities. With Blackwell, it will be common for these clusters to start with 100,000 GPUs…
…and can collectively demand orders of magnitude more compute than pretraining. Our inference demand is accelerating. Driven by test time scaling and new reasoning models.…
…as a percentage of total data center revenue, data center sales in China remained well below levels seen onset of export controls. China shipments absent…
…inference costs 3x with NVIDIA Corporation Triton inference server and TensorRT LLM. Microsoft Bing achieved a 5x speedup at major TCO savings for visual search…
…the more the model thinks the smarter the answer. Models like OpenAI's Grok 3, DeepSeq R1, are reasoning models that apply inference time scale. Reasoning…
…And the reason for that is because they're always things that because CUDA is so programmable, you could use it right well, one of the…
…across a range of applications, including customer support, fraud detection, and product supply chain and inventory management. Leading AI agent platform providers, including SAP and…
Agents: this page as TSV at /api/v1/appearances/nvidia:2025-02-26:earnings_call · words: /api/v1/brief?appearance=nvidia:2025-02-26:earnings_call · columnar JSON in #parrot-data · llms.txt · MCP