Amazon CEO Jassy says company could sell AI chips, raising stakes for Nvidia, AMD
Amazon CEO Andy Jassy signaled the company could begin selling its in-house AI chips (Trainium/Inferentia) externally, directly challenging Nvidia and AMD in the merchant silicon market. This raises the long-term question of whether AWS's captive chip build-out becomes a structural demand headwind for both NVDA and AMD's data center franchises.
What happened
Amazon CEO Andy Jassy signaled the company could begin selling its in-house AI chips (Trainium/Inferentia) externally, directly challenging Nvidia and AMD in the merchant silicon market.
The read · Apr 9
The question for NVDA and AMD is whether Amazon externalizing its AI chip business is a slow-burn competitive threat or an overblown headline that leaves Nvidia's 71% gross-margin moat intact.
Wrong if
Nvidia's next earnings print showing sustained data center demand above expectations — or a customer like AWS reaffirming large H100/B200 orders — would invalidate the competitive threat narrative and squeeze the short NVDA leg hard.
CoverageSource: Yahoo Finance · Published here THU, APR 9 · 10:28 AM ET · the only report in this recordHow this is decided →
The chart · NVDATradingView · third-party feed, not the Wire’s licensed closes
🔒 Click to interact · scroll moves the page
The story1 min read
Andy Jassy's comments suggest Amazon may move beyond using its custom Trainium and Inferentia chips solely for internal AWS workloads and begin offering them as a merchant product — effectively entering the AI accelerator market against Nvidia and AMD. Amazon's $716.9B revenue base and deep cloud distribution give it a credible path to monetize silicon at scale, but external chip sales remain nascent and unquantified. The more immediate read-through is competitive pressure on NVDA and AMD: Nvidia's data center segment is running at 71% gross margins on 65% YoY revenue growth, and any credible hyperscaler-as-chip-vendor narrative chips away at that premium multiple. Watch for AWS re:Invent or earnings calls for concrete chip-sales volume disclosures; until then, this is a directional signal, not a structural shift.
The read, in fullWhy it matters
Jassy's comments open a credible narrative that hyperscaler-built silicon could erode external accelerator demand over time — even if near-term impact is minimal. NVDA trades at a massive premium (55.6% net margins, 65% YoY growth) that is sensitive to any crack in the data center TAM story, while AMZN gets a re-rating tailwind as a potential new silicon revenue stream. A short NVDA / long AMZN pair captures the narrative shift without betting on a near-term volume hit that isn't yet quantified.
The arc● Open · no developments yet
· Published
The read above, as written. kept as written
Next · what would move this
4-8 weeks. Follow to be told when one lands.
▲ The case it holds
NVDA's 71% gross margins and 65% YoY revenue growth reflect a software-plus-hardware moat (CUDA ecosystem lock-in) that Amazon's nascent merchant chip ambitions cannot realistically threaten in the near-to-medium term, keeping the premium multiple defensible.
▼ The case it breaks
If Amazon successfully externalizes Trainium/Inferentia, it joins Google (TPU), Microsoft (Maia), and Meta (MTIA) in a hyperscaler chip coalition that collectively represents the largest buyer cohort for Nvidia silicon — pricing power and TAM expansion assumptions embedded in NVDA's multiple face a genuine structural challenge.
Kept as written · your side, if you take one, is graded privately against licensed closes after 10 trading days · nothing here is advice · How the Wire is made →
Shares a name with this story — discovery, not a connection.
The memory
This page is kept as it was written on Apr 9. Later coverage joins it only when the company and catalyst evidence match, and what the stock did is shown from licensed end-of-day closes — never re-graded, never backdated. The judgment is yours.