Stocks.

·AI infrastructure

NVIDIA Corporation

$1142.36
+2.41%

7 of 22

Prev close $1,115.48H $1,142.36L $980.00Start of last 10 sessionsLatest
Interactive close chart for NVDA. Move the pointer across the chart to read the close at each session; the dashed reference line marks the previous session’s close.

Designs the GPUs and accelerators behind most AI training stacks.

$2.81T

3 of 22

62.1

4 of 5in sector

+126%

1 of 5in sector

75.0%

2 of 7in sector

1.2%

1 of 5in sector

More angles · 6

The headline metrics above sit at the top because they're what most readers glance at first. These six dimensions are where a tech-focused study actually lives — capital intensity, margin trajectory, concentration risk, and the per-sub-sector quality metrics this kind of business is judged on.

9.4%

leverage on scale — single-digit despite >$100B rev base

48.2%
3.1%

asset-light — TSMC fabs the wafers

5.1%
~46% of DC rev

top 4 hyperscalers — single-customer concentration

3yr rev CAGR
+94%

Stands out in coverage

Where NVDA leads — and where it trails — the rest of coverage

2 callouts

Leads

  • Best Rev YoY+126%in Semiconductors vs 4 peers
  • Lowest Short interest1.2%in Semiconductors vs 4 peers

Trails

NVDA doesn't trail on any tracked row.

Fundamental rows (P/E, margins, growth) rank NVDA against same-sector peers that disclose the metric — a high P/E means different things in different sectors, so the comparison stays within one. Sectors with only one ticker in coverage fall back to the full catalog so the chip isn't silently dropped. Market cap and today's move stay coverage-wide — those are positional summaries, not quality reads. Direction (higher- or lower-is-better) is editorial; rows we don't have a confident direction on are skipped. Structural ranking, not individualised investment advice.

What changed on the call

How management's posture shifted from FY25-Q4 to FY26-Q1

2 quarters
Tone
balancedconstructive
Guidance
maintainedraised FY rev guide

This quarter · FY26-Q1

Called 2026-05-21

Management framed data-center revenue as still supply-constrained at the high end, with Blackwell ramp ahead of the prior quarter's internal plan. Software and services were called out separately for the first time as a material recurring line, and sovereign-AI engagements were named as the largest new pipeline driver.

  • Inference attach rate on Blackwell systems running higher than Hopper at the same maturity, supporting the unit-economics framing.
  • Sovereign-AI deals named explicitly across three regions; revenue cadence described as multi-quarter rather than lumpy.
  • Gaming addressed only briefly; management declined to re-segment the quarterly softness call from the prior cycle.
  • Networking attach to data-center systems above prior commentary, with Spectrum-X cited as the differentiator.

Prior · FY25-Q4

Called 2026-02-19

Management held the FY guide and characterised data-center demand as still exceeding supply on the new generation. The tone on gaming softened materially, with weaker channel inventory and a mix shift into lower-ASP SKUs. China cadence was called "variable" rather than recovering.

  • Data-center demand described as continuing to outstrip supply; Blackwell shipments still gated by HBM availability.
  • Gaming flagged as softer; management framed it as a near-term setup, not a structural change.
  • China revenue characterised as variable rather than recovering — explicitly declined to upgrade the framing.
  • Capex commentary unchanged from prior call; no new commitment beyond what hyperscalers had already signposted.
New this quarter
inference attach ratesovereign-AI pipelinenetworking attachsoftware & services recurring
Dropped from prior
data-center capex pacegaming softnesschina variabilityhbm supply

Publisher summary. Structured editorial commentary, not a verbatim transcript.

Recent insider activity

6 insider sales on the tape

· last 6 weeks
Net 30-day flow
−$7.3M

Sells outweigh buys

Distinct insiders
6

0 buying · 6 selling

Of which planned
0%

share of selling

Latest filings

  • Priya Ramesh · Director

    Filed May 8, 2026

    $1.4M

    Sell11.0K sh

  • Marcus Whitfield · Director

    Filed May 4, 2026

    $1.5M

    Sell11.1K sh

  • Sofia Alvarez · EVP Engineering

    Filed May 1, 2026

    $1.5M

    Sell11.2K sh

  • Elena Petrova · Chief Product Officer

    Filed Apr 25, 2026

    $1.5M

    Sell11.2K sh

  • Hiroshi Tanaka · SVP Sales

    Filed Apr 20, 2026

    $1.5M

    Sell11.3K sh

  • Michael Chen · CEO

    Filed Apr 12, 2026

    $1.5M

    Sell11.4K sh

Public Form 4 record. Plan-scheduled sales () are flagged so they don't read as conviction sales.

& borrow

Crowded short setup on NVDA — squeeze risk in both directions

Crowded short

on the borrow

Short interest is 19.3% of float with 73% utilization and rising borrow cost — the kind of crowded-short profile that historically reads as elevated volatility around NVDA. Cuts both ways: a rally forces covering, a fundamental miss feeds the thesis.

19.32%

Elevated

6.3d

Constrained

24.3%

risingExpensive

73%

Available

Public short-interest disclosure (Reg SHO) and securities-lending read. The pressure label is publisher commentary, weighted into the verdict — not a directional call.

Government activity

NVDA has a notable U.S. government footprint

Federal presence

footprint

Cumulative U.S. government awards on record total $420M — an meaningful federal contract footprint that reads as structural recurring revenue (security clearance, FedRAMP authorisation, and multi-year IDIQ vehicles are all moats that take years to replicate).

$420M

Notable

Net congressional flow
+$933K

Net buying

Disclosures
3

3 buys · 0 sells

Recent STOCK Act disclosures

  • Rep. Sample Member-H · House

    Traded Apr 27, 2026

    +$100,001–$250,000

    Buy

  • Sen. Sample Member-E · Senate

    Traded Apr 20, 2026

    +$500,001–$1,000,000

    Buy

  • Sen. Sample Member-F · Senate

    Traded Apr 14, 2026

    +$1,001–$15,000

    Buy

Public-record STOCK Act disclosures and federal contract awards. The pattern label is publisher commentary on the data — apolitical and not a directional call.

Quiver ↗

& positioning

Options tape on NVDA is quiet

Quiet tape

No notable options flow on record for NVDA. The tape is quiet — no large premium prints in the recent window.

Public OPRA prints filtered to large, premium-weighted sweeps. The conviction label is publisher commentary — a confirmation read, not a directional call.

change

Moderate edits in NVDA's latest 10-K

Moderate edit

Moderate language change

NVDA’s latest 10-K reads 13.9% different from the prior comparable filing (0001045810-25-000023) — edits concentrated in Item 7 MD&A, Item 1A Risk Factors. The research finds material rewrites of risk-factor and MD&A language historically predict underperformance, on the read that management quietly tweaks the wording before the market reprices.

Form
10-K

Annual report

Filed
Feb 25, 2026

3 months ago

Δ vs prior
13.9%

Material edits

Sections moved
2

2 sections

Per-section delta

  • Item 7 MD&A

    MD&A — narrative discussion of results and forward posture

    Δ 38.0%

    Heavy

  • Item 1A Risk Factors

    Risk factors — management's enumerated risks to the business

    Δ 6.5%

    Light

Year-over-year shingle-cosine diff of / Risk Factors and MD&A. Material edits historically predict underperformance — publisher commentary on a public filing, not a directional call.

Running Haiku → Sonnet pipeline…

History · NVDA

6 items
Daily insightAlphabet Earnings Release · 2026-07-22
5d ago

Mag-7 Q2 capex wave confirms $725B hyperscaler AI buildout — NVDA, AMD, AVGO demand read-through

Alphabet's Q2 print — the first Magnificent Seven report of the season — confirms that the aggregate ~$725B hyperscaler capex cycle for 2026 remains on track, with Google Cloud backlog at $514B. MSFT, META, AAPL, and NVDA report in the coming weeks, with AI accelerator spend broadly concentrated in GPU procurement from NVIDIA, networking from Broadcom (AVGO), and memory from Micron (MU). The combined data point reinforces the demand backdrop across the AI semiconductor supply chain.

NewsNVIDIA Blog · 2026-07-22
5d ago

NVIDIA donates DGX GB300 Blackwell Ultra supercomputer to U.S. Naval Postgraduate School

NVIDIA CEO Jensen Huang personally commissioned the company's most powerful enterprise AI system — a DGX GB300 NVL72 packing 72 Blackwell Ultra GPUs — at the Naval Postgraduate School in Monterey, California on July 22. NVIDIA donated the system, valued well into seven figures, through the NPS Foundation under a cooperative research agreement signed in December 2024. The school plans to use it for weather prediction, cybersecurity, ocean modeling, and disaster-response research.

NewsReuters · 2026-07-21
6d ago

Wistron Opens $700M Texas AI Factory to Produce NVIDIA Superchips

Nvidia supplier Wistron inaugurated its first U.S. manufacturing facility — a 324,000-sq-ft plant in Fort Worth, Texas — on July 21, with NVIDIA CEO Jensen Huang attending the ceremony. The $700 million facility is already producing NVIDIA GB300 Grace Blackwell Ultra Superchips and is slated to add Vera Rubin Superchip production, with output expected to scale to tens of thousands of boards per month by year-end. NVIDIA said the plant is part of its $500 billion U.S. investment commitment announced in 2025.

NewsBloomberg · 2026-07-21
6d ago

Bloomberg: Nvidia Vera Rubin systems now being delivered to major AI customers

Nvidia said computing systems based on its new Vera Rubin technology have been delivered to major AI companies and are about to enter active use, according to Ian Buck, an Nvidia vice president and general manager. The update, reported by Bloomberg on July 21, signals that NVDA's next-generation GPU platform is on track to reach production scale in H2 2026.

NewsNVIDIA Newsroom · 2026-07-21
6d ago

NVIDIA Vera Rubin NVL72 Enters Gigascale Production at Major Cloud Partners

NVIDIA announced on July 21 that Vera Rubin NVL72 production is ramping up with racks already running at CoreWeave, Google Cloud, Microsoft Azure, and Oracle Cloud Infrastructure. The company described AI as having "entered the gigascale era," with advanced AI factories now combining hundreds of thousands of GPUs and CPUs across more than 350 factory sites in 30 countries.

NewsTom's Hardware · 2026-07-21
6d ago

TSMC finalizes 5–10% wafer price hike for 2027, covering all advanced nodes

Nikkei Asia and Tom's Hardware report that TSMC has completed negotiations with major customers — including Nvidia, AMD, Apple, and Qualcomm — on wafer price increases of 5%–10% across advanced and mature nodes effective 2027. Legacy nodes (28nm, 16nm, 12nm) are also included, broadening the cost impact across the chip supply chain. TSMC cited higher materials, production-equipment, and overseas fab-construction costs as drivers.

Peers in coverage

Other tickers in Semiconductors, then more from the coverage list.

Compare 4 side by side

Same impersonal verdict shown on each profile page, surfaced for discovery. The order does not account for your positions or concentration. Browse the full list on the screener.

What this profile tracks

  • PriceDaily move and a session .
  • Filings / text diffs and material .
  • Insider and sales.
  • Short side, , and — squeeze pressure flagged.
  • Government congressional disclosures and awards.
  • OptionsLarge prints, weighted by — call / put split and conviction tag.
  • SentimentAggregated, decayed .
  • NewsCurated wires and press releases.
  • AI takesAsk the assistant — it knows you're on NVDA.