By Simon Ben Gemmill | 23 September 2026
Artificial intelligence has become one of the central stories in global technology and financial markets. In reporting on 22–23 September 2026, the technology-heavy Nasdaq Composite reached an intraday record of 27,212.68, while AI demand and semiconductor shares remained important drivers of investor attention. Reuters reported that AMD’s market value had also moved above $1 trillion and that the semiconductor index had recovered strongly. Reuters market report.
Why AI Is Driving Wall Street in 2026
The AI boom is not based on one product. It is an expanding technology chain involving semiconductor designers, chip manufacturers, cloud companies, data centres, networking equipment, software developers and businesses buying AI services.
- AI chips: powerful processors are needed to train and run increasingly capable models.
- Data centres: AI requires large amounts of computing capacity, electricity, cooling and networking.
- Cloud computing: businesses can access AI infrastructure without owning all the hardware themselves.
- AI software: companies are attempting to turn large language models and AI agents into commercial products.
- Productivity: investors are watching for evidence that AI spending can translate into higher revenue, lower costs or new businesses.
The Nasdaq and the AI Investment Cycle
Technology companies have become closely associated with the AI investment cycle. The current market story includes demand for processors, memory, networking and data-centre infrastructure as well as enthusiasm for applications such as AI assistants and agents.
However, markets can move in both directions. Reuters noted that the Nasdaq had previously fallen more than 10% from its late-July intraday high amid concerns that very large AI investments might take longer than expected to produce returns. The September recovery therefore illustrates both sides of the story: strong demand can support technology valuations, while questions about profitability can quickly affect sentiment.
Why AI Chips Matter So Much
AI models require specialised computing. Graphics processing units (GPUs), AI accelerators, memory, advanced packaging and high-speed networking are all part of the infrastructure needed to train and operate modern AI systems.
This has turned semiconductor supply into an economic and national-security issue. The US government said in January 2026 that the United States consumes roughly one quarter of the world’s semiconductors but fully manufactures only about 10% of the chips it requires. Its policy response includes measures intended to strengthen domestic semiconductor supply chains, including attention to AI-related chips.
US semiconductor policy information – White House
AI Chips Are Also a Cybersecurity Issue
The AI race is creating a second race: protecting the systems that make AI possible. A modern AI system can contain valuable model weights, sensitive training data, proprietary software and connections to other business systems. Attackers may try to steal models, manipulate data, disrupt services or compromise autonomous AI agents.
The UK government has specifically identified theft of AI model weights, compromise of sensitive data, modification of system behaviour and attacks through deployed autonomous agents as security concerns. The UK published a Secure AI Infrastructure call for information in January 2026.
UK Government – Secure AI Infrastructure
AI Can Help Hackers – and Defenders
AI can potentially make cyber attacks faster by helping malicious actors analyse information, find weaknesses and automate parts of an operation. At the same time, defenders can use AI to identify suspicious activity, analyse vulnerabilities, improve monitoring and respond more quickly.
The US National Institute of Standards and Technology (NIST) says AI security involves both familiar cybersecurity risks and AI-specific risks, including model extraction, evasion, data-related attacks and attacks against the availability of AI systems. NIST is developing security guidance for AI models, AI agents and AI developers.
NIST – AI Security and Resilience
A New Warning About AI Model Theft
In September 2026, the NSA, FBI and CISA published a cybersecurity advisory warning that China-based AI companies were conducting what the agencies described as industrial-scale distillation campaigns aimed at extracting capabilities from US frontier AI models. The agencies said the activity was intended to help train other AI systems.
This is important because the competition is no longer only about who has the fastest chip. It also concerns intellectual property, model security, data protection, compute infrastructure and the ability to defend AI systems from sophisticated cyber operations.
NSA, FBI and CISA cybersecurity advisory
Are Governments Doing Enough?
There is no single measure that can make AI or the semiconductor industry secure. Governments are responding through several different approaches: supporting domestic chip production, protecting supply chains, funding research, controlling sensitive technology exports, improving cybersecurity standards and encouraging secure AI development.
United States
The US has combined semiconductor industrial policy with AI security measures. A June 2026 White House executive order directed agencies to strengthen federal cyber defence, expand access to AI-enabled cybersecurity tools and establish an AI cybersecurity clearinghouse for identifying and remediating software vulnerabilities at scale.
White House – AI Innovation and Security
United Kingdom
The UK published its AI Hardware Plan in June 2026. It focuses on innovation, skills, procurement and investment and includes an £18 million Hardware Security R&D Programme. The plan also highlights secure-by-design hardware and the need to reduce critical supply-chain dependencies.
European Union
The European Commission published an Action Plan on Cybersecurity and Artificial Intelligence in July 2026. It recognises that advanced AI can improve cyber defence but can also be used to automate attacks and identify vulnerabilities. The plan includes AI evaluation, cybersecurity research and investment in European AI capabilities.
European Commission – AI and Cybersecurity Action Plan
China
China is also expanding domestic AI-chip capability. Reuters reported in September 2026 that Huawei was accelerating development of new Ascend AI processors and that demand for its AI computing systems in China was exceeding current production capacity. China has also been pursuing greater self-sufficiency in advanced computing as US restrictions limit access to some advanced foreign chips.
Reuters – China and Huawei AI chips
Producing More AI Chips Will Not Stop Cyber Attacks by Itself
Building more domestic AI chips can reduce dependence on overseas supply chains and can provide governments and companies with greater control over critical hardware. But chip production alone cannot prevent cyber attacks.
- AI models need secure software as well as secure processors.
- Data centres need strong identity, access and network controls.
- Model weights and sensitive training data need protection.
- AI agents need limits on what they can access and change.
- Semiconductor supply chains need monitoring and resilience.
- Governments need cooperation between intelligence, cybersecurity, technology and industry organisations.
- Companies still have responsibility for patching systems, managing credentials and responding to incidents.
The Bigger Wall Street Question
The long-term market question is whether today’s enormous AI infrastructure spending will generate equally large and durable economic returns. Investors are therefore watching more than chip sales. They are looking at cloud revenue, AI subscriptions, enterprise adoption, productivity gains, energy costs, data-centre investment and the ability of AI companies to turn research breakthroughs into sustainable businesses.
The September 2026 Nasdaq record shows how important AI has become to the market narrative. But the same technology creates major challenges around cybersecurity, energy, semiconductor supply, regulation and international competition. The AI boom is therefore both a technology story and an infrastructure story – and increasingly a security story.
Sources and Further Information
- Nasdaq
- NIST AI Security and Resilience
- NSA/FBI/CISA AI cybersecurity advisory
- UK AI Hardware Plan
- European Commission AI Cybersecurity Plan
Image credit: Carlos Delgado, “Wall Street – New York Stock Exchange”, Wikimedia Commons, licensed under CC BY-SA 3.0. Image used with attribution.
Market information is provided for general information and should not be taken as personal investment advice. Market prices can change rapidly.
