VC
Value Add VC
⚡HomePulse⚡Helpful Apps📝Blog🤝Partner
Illustration for: AI Chip Stocks Rally as Anthropic's Blowout Quarter Lands
Value Add VC/Pulse/AIDEEP DIVE

AI Chip Stocks Rally as Anthropic's Blowout Quarter Lands

Micron and Sandisk led premarket gains Monday after Anthropic's Q2 revenue surge bolstered the market's view that AI infrastructure spending will keep climbing rather than plateau.

By the Numbers

+0.5%
Nasdaq 100 futures
+0.49%
QQQ premarket
+0.1%
S&P 500 futures
$11.5B+
Anthropic Q2 revenue
TC
By the AI Desk
Edited by Trace Cohen · Early-stage VC & angel · Founder, New York Venture Partners
August 17, 2026
2 min read
ShareXLinkedInEmail

THE RUNDOWN

1

Nasdaq 100 futures rose 0.5% and the QQQ gained about 0.49% in Monday premarket trading, per [Bloomberg](https://www.bloomberg.com/news/articles/2026-08-16/stock-market-today-dow-s-p-live-updates)

2

Storage and semiconductor makers including Sandisk and Micron Technology rallied on the read-through from Anthropic's Q2 numbers

3

The move suggests investors are treating Anthropic's revenue, not just Nvidia's or OpenAI's, as a bellwether for whether AI infrastructure spending is durable

4

No major earnings were scheduled for Monday itself, making the rally a pure sentiment reaction rather than a reaction to new company-specific data

TC

The VC Read · Trace's Take

Trace Cohen

The diligence item: how much of Micron and Sandisk's rally is genuine demand signal versus a levered bet on Anthropic's unaudited, preliminary numbers holding up through an S-1. I'd want every LP asking their infra-fund GPs whether they're underwriting to audited revenue or press-release revenue right now, because those two things are about to diverge hard once IPO filings force real disclosure. Watch memory pricing data, not stock charts, for the real tell.

AI Valuations Tracker →

Analysis

Technology stocks opened the week higher, with memory and storage names leading the move after Anthropic's blowout second quarter reinforced the market's bet that AI infrastructure spending has further to run. Bloomberg's markets desk reported Nasdaq 100 futures up 0.5% and the Invesco QQQ Trust gaining roughly 0.49% in premarket trading Monday, with Sandisk and Micron Technology among the session's biggest gainers.

Why a private company's earnings move public stocks

Anthropic isn't publicly listed, but its preliminary Q2 revenue of more than $11.5 billion functions as a proxy data point for the entire AI compute-demand story.

“- Databricks' $190B round -- $5B raised on 80%-plus year-over-year growth and a $7B revenue run-rate.”

Revenue growing more than 14-fold year over year, while the company posts its first positive adjusted operating income, validates the thesis underlying every chip, memory, and data-center-infrastructure stock: that enterprises are actually paying for inference at scale, not just experimenting with pilots that never convert.

The mechanism, not the vibes

Memory and storage names specifically benefited because AI inference workloads are increasingly memory-bound, not just compute-bound -- serving larger context windows and more concurrent users requires more DRAM and high-bandwidth memory per GPU, which is exactly the bottleneck Micron and Sandisk sell into. That's also why CXMT's rise to China's most valuable listed company has tracked the same story on the other side of the Pacific: memory is where AI capex increasingly lands.

Competitive and historical context

The same read-through has played out repeatedly this month, each time treated by public markets as confirmation that AI infrastructure demand is real revenue rather than speculative capex:

  • Nvidia's SpaceX and Intel stakes -- disclosed at roughly $21B and $30B respectively, the first time Nvidia put a number on either position.
  • Databricks' $190B round -- $5B raised on 80%-plus year-over-year growth and a $7B revenue run-rate.
  • Anthropic's Q2 revenue -- Monday's own catalyst, more than $11.5B and rising over 14x year over year.

The risk is that markets are now trading on a relatively small number of private companies' self-reported, preliminary figures as if they were audited public disclosures.

The counterweight

What the rally glosses over: Anthropic's figures are preliminary and could still change, the company remains privately held with no audited public filings yet, and a single quarter of accelerating revenue doesn't resolve the open question of whether AI infrastructure spending is being paid for by durable enterprise revenue or by continued venture and sovereign-wealth capital chasing the same narrative. Critics have flagged that circular capital flows -- chipmakers investing in the labs that buy their chips -- make it hard to separate genuine demand from vendor-financed demand.

What to watch

Anthropic's actual audited numbers, expected as part of its IPO filing process this fall, will be the real test of whether Monday's rally was justified. Until then, expect chip and memory stocks to keep trading as a leveraged bet on private AI labs' self-reported growth.

ShareXLinkedInEmail

More on

Anthropic →

Reported by Bloomberg · Analysis by Value Add Pulse.

← Back to Pulse

THE WIRE in your inbox— Tech, startup & VC news with Trace's take. Free, no spam.

Read Next

AI· Aug 16, 2026

DeepSeek Raises V4 Prices Hours After Topping Agent Tests

Illustration for: DeepSeek Raises V4 Prices Hours After Topping Agent Tests
AI$0.44/$1.32 per 1M tokens

DeepSeek Raises V4 Prices Hours After Topping Agent Tests

DeepSeek moved its V4 models to peak and off-peak pricing on August 16, raising rates across every tier, days after V4-Flash beat the company's own flagship on nine agent benchmarks at three cents per task.

AI· Aug 16, 2026

ChatGPT Can Now Log Every Click and Keystroke on Your Mac

Illustration for: ChatGPT Can Now Log Every Click and Keystroke on Your Mac
AI

ChatGPT Can Now Log Every Click and Keystroke on Your Mac

OpenAI's new Computer History feature records mouse clicks, typing and app switches on macOS to build a searchable timeline ChatGPT can reference, stored locally as unencrypted plain text and off by default.

AI· Aug 15, 2026

Eval Harness Finds Models Most Confident When Wrong

Illustration for: Eval Harness Finds Models Most Confident When Wrong
AI

Eval Harness Finds Models Most Confident When Wrong

A systematic evaluation harness surfaced what qualitative review missed: model confidence rises on the outputs that turn out to be wrong, inverting the signal teams rely on for routing and human escalation.

@Trace_Cohen·t@nyvp.com