◉Tech Investing RadarBY MARC WINN
← Historical patterns

SIMILAR CONDITIONS. DIFFERENT OUTCOMES.

Compare the pattern.
Keep the differences.

Find the shared questions across two episodes, then inspect the evidence that makes each one different.

SHARED RESEARCH HYPOTHESES

2 shared pattern features

Adoption acceleratesThe economics improve

These labels describe retrospective research hypotheses. They are not a similarity probability, forecast or evidence that either case predicts the other.

Important differences

  • The documented business outcomes differ; business progress must be assessed separately from shareholder return.
  • The episodes use different time horizons and cannot establish the same investment window.
  • The episodes occurred in different calendar periods; market conditions are not controlled for.

Missing evidence

  • AMD: annual endpoints do not resolve a two-to-eight-week investment window.
  • AMD: no comparable measured historical sentiment series.
  • NVIDIA: annual endpoints do not resolve a two-to-eight-week investment window.
  • NVIDIA: no comparable measured historical sentiment series.

AMD · 2017–2019

AMD: renewed expansion as the product roadmap reaches customers

Verified rise
Annual context 2014–2019Episode 2017–2019 shaded

Prior belief

Hypothesis: technical competitiveness still needed repeated commercial execution.

New possibility

Hypothesis: a multi-generation competitive position and better economics were plausible.

Alternative explanations

  • Semiconductor-sector valuation changes
  • Other AMD product lines and market share shifts

Evidence gaps

  • Selected retrospectively, rather than discovered by a predeclared stock-universe screen.
  • Belief labels are interpretations, not a measured sentiment series or evidence of market consensus.
  • Daily adjusted prices, liquidity, transaction costs and exact entry or exit boundaries are not verified.
  • Related episodes can overlap and are not independent statistical observations.
  • The February 2020 filing describes prior-year outcomes. It must not be replayed as information available during 2018 or early 2019.
Inspect the full dated evidence →

NVDA · 2023–2024

NVIDIA: generative AI changes the scale of demand

Verified rise
Annual context 2020–2025Episode 2023–2024 shaded

Prior belief

Hypothesis: a cyclical GPU supplier with promising data-centre exposure.

New possibility

Hypothesis: a central infrastructure supplier for a major computing transition.

Alternative explanations

  • Market-wide technology recovery
  • Supply scarcity
  • Customer capital expenditure concentration

Evidence gaps

  • Selected retrospectively, rather than discovered by a predeclared stock-universe screen.
  • Belief labels are interpretations, not a measured sentiment series or evidence of market consensus.
  • Daily adjusted prices, liquidity, transaction costs and exact entry or exit boundaries are not verified.
  • Related episodes can overlap and are not independent statistical observations.
Inspect the full dated evidence →

Belief descriptions are reconstructed unless the case explicitly identifies contemporary sentiment evidence. Annual charts may cover different dates and use different benchmarks. They must not be read as matched trade results.