OpenAI unveils its trump card, GPT-6 Astra! AI trading sentiment is reignited. Which of the two major camps is worth betting on?

OpenAI has hit the accelerator again in the AI arms race.

OpenAI's latest flagship modelGPT-6 Astrarelease has reignited market enthusiasm for AI computing infrastructure, with memory chips—previously sold off by the market—being the first to be repriced by capital.

Last Friday, the US semiconductor sector strengthened collectively, $Micron Technology (MU.US)$ Rose by more than 6%, $SanDisk (SNDK.US)$ Gained nearly 12%, $SK hynix (SKHY.US)$ Surged over 8%; this week's rally further extended to Asian markets, $SK Hynix (000660.KR)$ Up approximately 8%, $Samsung Electronics (005930.KR)$ up more than 5%, $Kioxia Holdings (285A.JP)$ Rose by about 10%.

Symbols

Meanwhile,The market performance of the OpenAI-related ecosystem also significantly outperformed.Looking at two ETFs that track the Anthropic and OpenAI industrial ecosystems respectively, recently $OpenAI Lab Ecosystem ETF (OAIW.US)$ significantly outperformed $Anthropic AI Lab Ecosystem ETF (ANTW.US)$ 。

A new model—why has it once again shaken up the entire AI supply chain?

What the market is clearly trading on is no longer just "how much smarter Astra is than the previous generation," but rather a more important signal released behind it:The capabilities of frontier models are still rapidly improving, and to support increasingly complex AI tasks, the demand for computing power is far from peaking.

Why has Astra reignited AI trading?

ChatGPT-6 Astra was rolled out in two batches last Thursday and Friday. OpenAI President Greg Brockman described it as the company's "most intelligent and best-aligned" model to date.

Unlike previous generations that focused solely on chat and reasoning capabilities, new-generation models are increasingly oriented toward complex reasoning, software development, AI agents, scientific research, and enterprise-level tasks.This means AI is evolving from merely "answering questions" to actually "getting work done."

Rich Privorotsky, Head of Delta One Trading at Goldman Sachs, believes Astra represents a "major" breakthrough, with capability enhancements surpassing typical model iterations. This could well bethe catalyst AI bulls have been waiting for.。

The underlying reason is not complicated. Over the past year, one of the biggest debates in the market regarding AI has been whether the hundreds of billions of dollars in annual capital expenditure by tech giants can ultimately translate into sufficient commercial returns.

If models merely become cheaper and faster, AI may become more widespread, but it might not justify such massive infrastructure investment. However, if model capabilities truly leap forward, enabling entry into broader scenarios such as software development, financial analysis, scientific research, legal services, and enterprise process automation, then the potential market size for AI will be reopened.

This creates a more complete positive feedback loop:Enhanced model capabilities → Increased AI applications → Rising demand for tokens and computing power → Continued expansion of data centers → Synchronous growth in demand for GPUs, storage, networking, and power.

This is also why, following the release of Astra, the market initially shifted its focus back to AI infrastructure.

The AI competition has escalated from a battle of models into an ecosystem arms race.

In fact, the current AI competition is no longer just a model showdown between OpenAI and Anthropic. As training costs, inference demands, and infrastructure scales continue to grow, both AI labs have already formed massive industrial ecosystems behind them.

From the $Anthropic AI Lab Ecosystem ETF (ANTW.US)$ and $OpenAI Lab Ecosystem ETF (OAIW.US)$ Their holdings make this point very clear.

The two major camps share many common holdings, including $NVIDIA (NVDA.US)$、 $Taiwan Semiconductor (TSM.US)$、 $Advanced Micro Devices (AMD.US)$、 $Broadcom (AVGO.US)$、 $SK hynix (SKHY.US)$、 $Microsoft (MSFT.US)$、 $Amazon (AMZN.US)$、 $Corning (GLW.US)$、 $Blackstone (BX.US)$, Inventec, and others.

This means that regardless of whether OpenAI or Anthropic ultimately leads in the model competition, both will require the most advanced GPUs, HBM, chip manufacturing, networking equipment, servers, and data centers.

Therefore, rather than betting on which model company will emerge as the ultimate winner, a more certain theme in the capital markets is actually the AI "shovel sellers" relied upon by both camps.

Symbols

However, the unique holdings on each side also reflect different ecosystem characteristics.

The Anthropic campincluding $Alphabet-A (GOOGL.US)$ 、 $Micron Technology (MU.US)$ 、 $Marvell Technology (MRVL.US)$ 、 $Astera Labs (ALAB.US)$ 、 $Vertiv Holdings (VRT.US)$ 、 $American Electric Power (AEP.US)$ 、 $Salesforce (CRM.US)$ 、 $Zoom Communications (ZM.US)$ and other companies, overall leaning more towardscloud computing, enterprise software, semiconductors, and data center infrastructure。

andThe OpenAI campincludes $SoftBank Group (9984.JP)$ 、 $Oracle (ORCL.US)$ 、 $CoreWeave (CRWV.US)$ 、 $Arm Holdings (ARM.US)$ 、 $GE Vernova (GEV.US)$ 、 $Bloom Energy (BE.US)$ 、 $Celestica (CLS.US)$ 、 $Flex Ltd (FLEX.US)$ 、 $Lam Research (LRCX.US)$ among others, clearly leaning more towardscomputing power, servers, chip manufacturing, and power infrastructure。

In other words, the competition between OpenAI and Anthropic has evolved from a mere "model war" into a full-blown arms race between two AI industrial ecosystems. With every step forward in model development, the underlying infrastructure must expand in tandem.

Why is storage the first sector to be repriced in this cycle?

Some market analysts point out thatAstra's core innovation lies in its ability to directly operate computers and software, handling complex tasks for users such as programming, scientific research, 3D modeling, and creating presentations and spreadsheets. This capability will drive substantial demand for memory chips and flash storage.

This also explains why the market initially favored memory chip stocks following Astra's launch.

Charu Channa, Chief Investment Strategist at Saxo Bank, pointed out that Astra's release further supports the continued growth in AI spending. As frontier models increasingly demand computing power,bottlenecks in storage and networking are also widening, which bodes well for memory chip manufacturers like SK Hynix and Samsung Electronics.

Analysts at Goldman Sachs and Morgan Stanley project that global AI capital expenditure will reach $1.3 trillion to $1.5 trillion by 2027, with more than half allocated to memory infrastructure. If this trend persists, memory products such as HBM and DRAM are poised to continue benefiting from the expansion of AI computing capabilities.

Therefore, the significance of Astra for memory stocks is not merely that 'OpenAI released a new model,' but rather it reinforces a more critical industry trend:The bottleneck in AI is gradually shifting from standalone GPUs to the entire computing infrastructure system.

Summary

In summary, the true implication of Astra is not just that OpenAI has launched another stronger model, but that it once again validates how advancing AI capabilities continue to drive up infrastructure demand across the industry.

As model competition evolves from a simple race in parameters and performance toward Agents, enterprise-level applications, and complex workflows, the underlying support required extends beyond just GPUs to a comprehensive computing ecosystem encompassing storage, networking, servers, data centers, and power supply.

For investors, rather than betting on whether OpenAI or Anthropic will emerge as the ultimate winner, it may be more prudent to focus on the 'shovel sellers'—companies relied upon by both camps that benefit from the continuous expansion of AI capital expenditure. This could represent a more certain investment theme in the next phase of the AI arms race.

2.6M Views