DeepSeek's new V4.1-Flash-1017 model is reshaping investor interest in AI infrastructure, spotlighting companies like Zeta Global, Everpure, and Procore Technologies, which is acquiring DroneDeploy for US$845 million. Other infrastructure picks include Sandisk, Cerebras, and Seagate, with Cerebras integrating its CS-3 systems with AWS in March 2026. Lite-On, Wiwynn, and 2CRSI also benefit as Broadcom reports US$16.7 billion in sales, underscoring strong demand for data center hardware.
IREN, which operates data centers with 816 NVIDIA H100 GPUs, saw a 21% revenue increase in June 2024, while Nebius Group earns about US$1.3 billion from its AI Cloud. Dell Technologies has gathered more than US$130 billion in AI server orders over the past 12 months. Apple, with US$466.8 billion in revenue, is embedding AI features across its iPhone ecosystem, and Workday and Fortinet round out a US tech watchlist focused on cloud software and AI-centric network security.
UK small cap AI companies such as Beeks Financial Cloud, Softcat, and TPXimpact are drawing attention after Meta's earnings, though each faces questions about margins and execution. Microsoft's cancellation of its anchor-tenant AI data center deal with Nscale has shifted focus toward TTM Technologies, Trimble, and Octave Intelligence, which serve long-term infrastructure needs. Meanwhile, Google and Synaptics unveiled a low-power edge AI chip that reportedly cuts energy use by 40%, set for formal debut on October 20 in San Jose.
Anthropic is lobbying Congress for an AI safety bill that would permit government intervention only when a model poses a foreseeable risk of over US$1 billion in damages or injury to more than 50 people. Independent evaluators would assess labs, and a federal judge would rule within 72 hours of a report. The proposal follows several AI agent incidents this summer. In a separate development, Resy banned an Instinct agent that made 200 calls per hour, prompting industry leaders to push for per-call runtime authorization through frameworks like Keycard's OAuth 2.0-based architecture.
AI is also influencing politics and consumer safety. A University of Tennessee at Chattanooga professor noted the growing use of AI in campaigns after Tennessee candidate Marsha Blackburn posted an AI-generated video mimicking a Kendrick Lamar performance. National Trading Standards warns that UK criminals now use AI for voice cloning, deepfake videos, and personalised phishing, making traditional scam warning signs unreliable. About one in five US midterm voters use AI chatbots for ballot research, though experts say this can reinforce existing views. AI labs are also acquiring employee work emails and Slack data for training, often without workers' knowledge or compensation.
Wall Street warns of a potential AI stock bubble, but current valuations remain well below dot-com era levels, with risks tied to concentrated market leadership and heavy borrowing. IQVIA has adopted Veeva Systems' clinical tools as enterprise standards, pairing them with Veeva Falcon and Vault AI platforms, and projects US$4.9 billion in revenue and US$1.5 billion in earnings by 2029.
Key Takeaways
- <ul><li>DeepSeek's V4.1-Flash-1017 model is driving investor interest in AI infrastructure stocks including Zeta Global, Everpure, and Procore Technologies</li><li>Procore Technologies plans to acquire DroneDeploy for US$845 million</li><li>Cerebras integrated its CS-3 systems with AWS in March 2026</li><li>Broadcom reports US$16.7 billion in sales, reflecting strong AI chip demand</li><li>Dell Technologies has received more than US$130 billion in AI server orders over the past 12 months</li><li>Apple generates US$466.8 billion in revenue and is focusing on AI features for its iPhone ecosystem</li><li>Anthropic's proposed AI safety bill sets a high threshold, requiring over US$1 billion in damages or injury to more than 50 people before government intervention</li><li>An Instinct agent made 200 calls per hour to Resy.com and was banned, prompting calls for per-call runtime authorization</li><li>About one in five US midterm voters use AI chatbots for ballot research</li><li>Google and Synaptics previewed a low-power edge AI chip that cuts energy use by 40%</li></ul>
DeepSeek model boost highlights three AI infrastructure stocks
New competition from DeepSeek's V4.1-Flash-1017 model is changing how investors view AI infrastructure. This shift benefits three specific companies from a global cloud screener. Zeta Global Holdings uses its data cloud for marketing tools and has a market cap of about US$8.0 billion. Everpure provides storage devices for AI workloads and recently secured its first EXA-scale customer. Procore Technologies manages construction projects via the cloud and plans to acquire DroneDeploy for US$845 million.
Sandisk Cerebras and Seagate lead AI infrastructure picks
Investors are looking at physical infrastructure stocks as AI demand grows for chips and models. Sandisk sells solid-state drives for data centers and has a market cap of US$190.2 billion. Cerebras Systems designs wafer scale AI chips and integrated its CS-3 systems with AWS in March 2026. Seagate Technology supplies hard drives and SSDs with a market cap of US$145.3 billion to store training data.
Lite-On Wiwynn and 2CRSI ride data center growth wave
AI infrastructure chips are now generating hard revenue as Broadcom reports US$16.7 billion in sales. Three specific stocks from a semiconductor screener are highlighted for their exposure to this trend. Lite-On Technology supplies power and optical components for data centers and expects AI revenue to reach 25% by 2026. Wiwynn builds cloud servers and racks with a market cap of NT$1.03 trillion. 2CRSI designs high performance servers with a market cap of €592 million.
IREN Nebius and Dell attract retail investor attention
Volatile headlines around big tech companies are making investors watch AI heavy business models closely. Three stocks exposed to these storylines are being watched by retail investors. IREN operates data centers with 816 NVIDIA H100 GPUs and saw a 21% revenue increase in June 2024. Nebius Group supplies GPU cloud infrastructure with about US$1.3 billion in revenue from its AI Cloud. Dell Technologies supplies AI servers and has received more than US$130 billion in orders in the past 12 months.
Apple Workday and Fortinet shape US tech watchlists
New ideas like Innovation Accounts are directing money toward AI and tech stocks. This trend highlights three companies from a US technology screener. Workday provides cloud software for finance and HR with a market cap of US$46.0 billion. Apple generates US$466.8 billion in revenue and is focusing on AI features for its iPhone ecosystem. Fortinet secures enterprise networks and launched new appliances designed for AI-centric workloads.
Beeks Financial Cloud and UK AI Stocks to Watch
Investor focus on Meta's earnings and global AI trends has highlighted three UK small cap AI companies. Beeks Financial Cloud Group uses AI in its Market Edge Intelligence platform for trading, with a market value near £120 million. Softcat integrates AI into its £2.0 billion IT reseller business. TPXimpact applies AI and data analytics in public and private sector projects with a market value near £89 million. Each company shows potential but faces key questions about margins, execution, and pricing power.
3 Stocks to Watch After Microsoft AI Data Center Pullback
Microsoft cancelled its anchor-tenant AI data center deal with Nscale, shaking confidence in hyperscale infrastructure building. This shift highlights three established cloud players. TTM Technologies makes circuit boards and RF components for AI-ready data center hardware with a US$9.9 billion market cap. Trimble provides cloud-based workflow tools for construction, agriculture, and transportation with a US$13.9 billion market cap. Octave Intelligence develops software connecting operational data for critical infrastructure operators with a US$5.4 billion market cap. All three benefit from long-term AI infrastructure demand rather than single project deals.
Anthropic's AI Safety Bill Sets High Threshold for Government Action
Anthropic is lobbying for an AI safety bill that would allow the government to block a model only if it risks over US$1 billion in damages or injuring more than 50 people. An independent third party registered with the Department of Commerce would evaluate labs and models periodically. The bill covers bioweapons, cyberattacks, and loss of control over AI models. Only the Commerce Department could order a halt to model releases. A federal judge in Washington D.C. would make the final decision after reviewing evidence within 72 hours of a report.
Anthropic Pitches Its Own AI Safety Bill to Congress
Anthropic has proposed an AI safety bill based on its Advanced AI Framework, which it has promoted to lawmakers since June. The draft sets a high bar, requiring a foreseeable risk of over US$1 billion in damages or serious injury to more than 50 people before action can be taken. The danger must come from a bioweapon, a model slipping out of control, or a cyberattack. Outside evaluators would assess labs, and Anthropic may pay them as long as independence is preserved. The bill follows several reported AI agent incidents this summer, though none reached the scale the bill describes.
How to Protect Yourself From AI-Powered Scams
National Trading Standards warns that criminals are using AI to supercharge scams across the UK. Fraudsters create fake retailer websites with AI-generated images, clone voices, and use deepfake videos to impersonate people. AI also powers personalised phishing emails and romance scams that can last for months. Old warning signs like typos and grammar errors are no longer reliable. To stay safe, never send money to people you have not met, verify shop addresses, and be wary of urgency and pressure to act quickly.
Resy ban highlights need for per-call agent authorization
An Instinct agent made 200 calls per hour to Resy.com and got its user banned. Leaders from Keycard, JPMorgan Chase and AOH pushed for per-call runtime authorization. Keycard checks policy on every token exchange using OAuth 2.0 Token Exchange RFC 8693. Its reference architecture for secure agent onboarding is published at agentbaseline.org.
US midterm voters turn to AI chatbots for ballot research
Voters like Adam Johnson in West Virginia are using ChatGPT to research midterm ballots. About one in five voters use AI chatbots for election news according to a September survey. Experts say chatbots may reinforce views users already hold. Some voters cross-check answers across multiple chatbots to reduce the risk of false information.
AI labs seek employee work emails and Slack data
AI labs are using Slack DMs and work emails as training data. They are offering business owners large payments to access this data. Employees rarely know or profit from their workplace data being sold. Some CEOs are surprised to learn their company data is being sold to the highest bidder.
IQVIA picks Veeva clinical tools as enterprise standard
IQVIA chose Veeva Systems' CTMS, eTMF, Study Startup and Site Connect as enterprise standards across clinical studies. Veeva is pairing this with AI platforms like Veeva Falcon and Vault AI. The company also offers pre configured Veeva Basics deployments for more than 175 emerging biotechs. Veeva projects US$4.9b revenue and US$1.5b earnings by 2029.
Wall Street warns of AI bubble but valuations look lower
Wall Street is worried about an AI stock bubble but current valuations are much lower than dot-com boom levels. Major risks include concentrated market leadership and heavy borrowing for AI investment. There is uncertainty whether current profits will support large ongoing investments. Elevated expectations could still disappoint shareholders even if AI succeeds.
UTC professor discusses AI use in political campaigns
A University of Tennessee at Chattanooga professor shared thoughts on the growing use of AI in politics. The discussion follows Tennessee governor candidate Marsha Blackburn posting an AI-generated video of herself and other Republicans mimicking Kendrick Lamar's halftime performance during the 2025 Super Bowl halftime show. The video was created using artificial intelligence tools.
Google and Synaptics reveal new edge AI chip
Google previewed a new low-power AI chip built with Synaptics at its Research@ event in Mountain View, California, on October 14, 2026. The chip combines Google's Coral NPU M2 with Synaptics' Torq NPU and Arm CPU cores. The companies say it cuts energy use by 40% compared with existing edge AI processors. Synaptics CEO Rahul Patel said the chip will move beyond lab demos, and the formal unveiling is set for Synaptics' Tech Day in San Jose on October 20.
Sources
- 3 AI Infrastructure Stocks for Data Center Demand After DeepSeek Shock
- Our Pick Of The Best AI Infrastructure Stocks In October 2026
- 3 AI Infrastructure Stocks Riding Data Center Demand
- Nvidia and 3 AI Infrastructure Stocks Retail Investors Are Watching Now
- Apple Stock And AI Demand Are Shaping US Tech Watchlists
- Beeks Financial Cloud Group And 2 British AI Stocks To Watch
- 3 Stocks to Watch After the Microsoft AI Data Center Pullback
- Anthropic’s proposed AI safety bill only kicks in if 50 deaths or $1 billion in damages are on the line — proposal would effectively make the Commerce Department the only agency capable of stopping an AI model
- Anthropic is pitching Congress its own AI safety bill, report says
- Don't fall for this - how to protect yourself from AI scams
- Resy ban shows why agents need per-call authorization
- In the US, some midterm voters are asking AI to research the ballot, and even pick candidates
- AI labs really want your work emails
- Did Enterprise Client Adoption Just Shift Veeva's (VEEV) Investment Narrative?
- Wall Street sees warning signs of an AI bubble, but valuations tell different story
- Campaign advancements: UTC professor weighs in on use of AI for political purposes
- Google Synaptics Edge AI Chip Cuts Power 40% [2026]
Comments
Please log in to post a comment.