$NVDA

Nvidia Is Letting Rival AI Chips Into Its Racks. Astera Labs Could Be the Quiet Winner

NVIDIA (NVDA) is integrating rival AI chips into its data-center racks using NVLink Fusion, partnering with d-Matrix and Astera Labs (ALAB). Systems are expected in 2027. Nvidia aims to control system standards, while Astera benefits from connectivity solutions for complex AI racks. Hedge funds increased positions in both companies, with higher short interest in Astera.

Original reporting
Published Sep 11, 2026, 11:45 PM UTC
Analysis
AlphAI AI DeskAI-generated
Added to AlphAI Sep 12, 2026, 12:21 AM UTC. Informational, not investment advice.
How this was made
AlphAI summarizes source reporting and applies a structured AI analysis for relevance, timing, sentiment and ticker impact. Always verify material claims with the original publisher.
Nvidia Is Letting Rival AI Chips Into Its Racks. Astera Labs Could Be the Quiet Winner — source image
Decision brief

The 30-second read

$NVDANeutralMed
01

Why it matters

The partnership could reshape AI data‑center architecture, influencing both hardware vendors and cloud providers.

02

Market read

First‑report disclosure of a cross‑vendor AI rack strategy that may affect GPU and connectivity markets.

03

What to watch

Adoption depends on hyperscalers' willingness to integrate non‑Nvidia chips, and d‑Matrix's ability to deliver competitive performance.

Relevance 6/10Novelty 7/10Timing: today

Background

Nvidia's NVLink Fusion is being extended to third‑party AI accelerators, marking a strategic shift from a closed ecosystem to a more open standard.

Company-level read

Ticker impact

$NVDANeutralMedium confidence
Context

Nvidia announced it will allow rival AI accelerators, like d‑Matrix, to connect via its NVLink Fusion technology in data‑center racks.

Expected impact

Short‑term modest downside risk; medium‑term upside if ecosystem gains traction.

Evidence & confidence

The partnership does not guarantee immediate revenue; impact depends on adoption by hyperscalers.

$ALABBullishMedium confidence
Context

Astera Labs is working with d‑Matrix to provide custom connectivity solutions for the new heterogeneous AI rack architecture.

Expected impact

Potential upside as investors price in higher AI‑infrastructure exposure.

Evidence & confidence

Astera benefits from the added complexity of multi‑accelerator racks, but execution risk remains.

Market effects

The move could spur broader AI‑infrastructure spending and benefit other connectivity and interconnect vendors.

U.S. AI hardware ecosystem may see increased collaboration, with limited immediate global ripple.

Highlights a shift toward heterogeneous AI compute, relevant for global data‑center operators.

Counterpoint

Nvidia's openness may erode its GPU margin leadership if rivals capture significant share of accelerator sales.

Key entities

  • Nvidia Corporation

    Designer of NVLink Fusion technology.

  • Astera Labs, Inc.

    Provider of high‑speed connectivity solutions.

  • d‑Matrix

    Startup building inference AI accelerators.

Related articles

$NVDAMed

Nvidia Stock Moves Lower as Einride Scales Physical AI

Nvidia (NVDA) shares fell to $224.80 Friday after partnering with Einride to supply AI tools for driverless trucks. Einride expects demand for 1,500-2,000 vehicles by 2028, but revenue impact for Nvidia is unclear. Nvidia's GF Score is 96/100, with profitability and growth strengths noted.

$NVDAMed

Goldman Sachs: AI Infrastructure Valuations Fall as Spending Climbs Toward $1.1 Trillion

Goldman Sachs reports AI infrastructure stocks' valuations have dropped to 22x earnings, down from 32x in April, amid spending estimates reaching $1.1T by 2027. Memory chip stocks trade at 4x earnings, while chip designers show more stable margins. Goldman maintains Buy ratings on Nvidia (NVDA) and Broadcom (AVGO), Neutral on Micron (MU). Cloud providers' spending may reach $800B in 2024, requiring $300B in annual AI revenue to break even by 2026-2027.

$NVDAMedAI 8/10

What Does NVIDIA Offer That Texas Instruments Does Not?

NVIDIA (NVDA) and Texas Instruments (TXN) both sell chips for data centers, but NVIDIA's revenue is more concentrated in this area. NVIDIA's Q2 2027 data center revenue was $89B, with Q3 guidance at $108B. TI's Q2 revenue was $5.5B, with Q3 guidance between $5.65B and $6.15B. NVIDIA faces supply constraints, while TI has spare capacity. NVIDIA's gross margin is expected to decline to 71-72% by Q4. NVIDIA leads in growth, profit margins, and debt metrics.

$NVDAMedAI 8/10

Nvidia Stock Rises as Elon Musk Reveals Huge New AI Chip Plans

Nvidia NVDA shares rose 1% after SpaceX announced plans to add 660,000 Nvidia GB300 processors to its Colossus 2 facility by year-end. SpaceX also has a deal with Anthropic for 325,000 Nvidia GPUs, with $1.25B monthly payments post-ramp. The expansion highlights AI computing infrastructure growth across Musk's businesses.

$NVDAMed

Elon Musk Makes Massive Nvidia Move at SpaceX

Elon Musk's xAI plans to more than double its Nvidia chip capacity by year-end, with Colossus 2 potentially reaching 1.21 million GB200 and GB300 chips. Current expansion includes 220,000 GB300s operational next week, followed by another 220,000 in November, and possibly 220,000 more by December. This highlights strong demand for Nvidia's latest AI hardware.