Jabil Guides AI Revenue to $22.1B by 2027

Key Takeaways

• <ul><li>Jabil guides AI revenue to $22.1 billion by fiscal 2027, up from about $9 billion</li><li>SpaceX reports AI revenue of $7.6 billion in Q3 2026, an 830% increase representing 52% of total revenue</li><li>Blacksmith raises $66 million in Series C round, reaching $1 billion valuation</li><li>Meta develops Muse Charm, a wearable AI device planned for release later this year</li><li>Visa launches stablecoin processing pilot using USDC through Visa Direct</li><li>Mastercard introduces Global Money Platform with BVNK and Agent Pay for Machines</li><li>The Pensions Regulator warns pension trustees about hidden AI investment risks</li><li>An Oregon PAC faces scrutiny under 2024 AI disclosure law for posting an AI-altered campaign photo</li><li>RS offers Omron AI data center products including power supplies and liquid leakage sensors</li><li>Trainocate Malaysia launches a seven-level AI training roadmap with over 100 training hours</li></ul>

3 AI Infrastructure Stocks To Watch in October 2026

Lumentum Holdings makes optical modules and photonic parts that move data in AI data centers. Jabil builds cloud data center server platforms and rack systems for hyperscalers with AI revenue growing from about $9 billion to $22.1 billion. Super Micro Computer builds GPU-heavy servers but faces legal charges over alleged illegal sales of AI servers to restricted customers.

Adtran Networks and 2 Other AI Infrastructure Stocks

Adtran Networks develops optical and Ethernet networking gear for AI data center connections with a market value of about €1.1 billion. Hamamatsu Photonics supplies photodiodes and image sensors used in data center optical transceivers with a value of roughly ¥613.4 billion. Jabil assembles rack-scale servers and chassis for hyperscalers with AI revenue guided to reach $22.1 billion by fiscal 2027.

3 AI Hardware Stocks With Strong Earnings Growth

Micron Technology supplies DRAM, NAND and high bandwidth memory for AI accelerators with a market value of about $1,014 billion. Super Micro Computer builds GPU-heavy servers earning about $39.1 billion but depends heavily on one customer representing 63% of revenue. Penguin Solutions builds AI data centers and makes SMART Modular memory and storage products with a market value of about $2.9 billion.

SpaceX Launches Super Intelligence Partner Network

SpaceX created a partner network for companies building products with Grok for joint customers. The network includes Transformation Partners who deploy Grok across enterprises and Marketplace Partners who build plugins and integrations. More than 130 companies joined the early-access phase and partners serve customers worldwide from startups to Fortune 20 companies.

SpaceX Stock Falls Despite New AI Partner Network

SpaceX shares dropped 7% to about $148 despite launching its super intelligence partner network for Grok to compete with Google in enterprise AI. The company reported AI revenue of $7.6 billion in Q3 2026, an 830% increase representing 52% of total revenue. Wall Street maintains a Moderate Buy rating with an average price target of $240 suggesting about 62% upside.

RS brings Omron AI data center control panel products

RS is offering Omron Automation products built for AI data centers. The lineup includes single-phase and three-phase power supplies that are compact and efficient. It also includes motor condition monitors and liquid leakage sensors. These tools help operators improve reliability and prevent downtime.

TPR warns pension schemes about AI investment risks

The Pensions Regulator has warned pension trustees to carefully assess their exposure to AI investments. It says AI could create hidden risks even in diversified portfolios. The regulator calls for thorough due diligence and better understanding of AI tools. It also warns that AI use may reduce transparency and accountability.

Oregon AI disclosure law tested in Multnomah County campaign

A political action committee called Don't Rank Julia faces scrutiny for posting an AI-altered photo of Multnomah County Commissioner Julia Brim-Edwards. A 2024 Oregon law requires disclaimers on campaign materials when AI changes create a misleading impression. The secretary of state's office is considering whether to investigate. The case could help clarify how the new law works.

Meta develops wearable AI device called Muse Charm

Meta is moving forward with dedicated AI hardware despite a shift toward cloud-based services. Mark Zuckerberg described the Muse Charm, a small wearable device that lets users talk to AI using voice commands. The device is designed for personal, conversational AI interactions. It is still in development and planned for release later this year.

Visa and Mastercard launch AI and crypto tools

Visa and Mastercard are rolling out new products focused on AI and digital assets ahead of earnings reports. Visa released tools for supplier spending analytics, virtual-card fraud prevention, and AI agent payments. It also launched a stablecoin processing pilot using USDC through Visa Direct. Mastercard introduced Global Money Platform with BVNK, Stablecoin Security, and Agent Pay for Machines to support AI-driven payments.

Blacksmith reaches $1 billion valuation with $66 million funding round

Blacksmith, an AI code-testing startup, reached a $1 billion valuation after raising $66 million in its Series C round led by Sequoia Capital. The company has raised about $124.5 million total and now serves over 10,000 companies. Blacksmith buys its own servers instead of renting from cloud providers like AWS or Google Cloud. The company plans to grow from 40 to about 75 employees by year end.

Trainocate Malaysia launches 7-level AI training roadmap for businesses

Trainocate Malaysia launched the AI Training Roadmap 2026, a seven-level framework with over 100 training hours. The programs cover skills from basic AI literacy to AI ethics and governance. Each level targets different groups from employees to executives and uses measurable KPIs. The training is HRD Corp claimable and customizable for 17 or more industries.

AI Loops 101 explores what happens after AI models are trained

This article continues a series on AI loops by examining what happens after an AI model is trained. It references data from The New York Times about the math behind OpenAI GPT models. The training compute grew from 18 quintillion calculations for GPT-1 in 2018 to nearly an octillion for GPT-6 Astra. The piece notes a detour into Anthropic's draft S-1 filing before returning to the main topic.

Sources

• 3 Great AI Infrastructure Stocks To Own In October 2026
• Adtran Networks And 2 Other AI Infrastructure Stocks To Watch
• 3 AI Hardware Stocks With Earnings Growth Over 18%
• Introducing the SpaceX Super Intelligence Partner Network
• SpaceX Stock (SPCX) Sinks Despite New Network to Challenge Google, Others in Enterprise AI
• RS enables next-generation AI data centers with Omron Automation's industrial control panel products
• TPR warns pension schemes over hidden AI investment risks
• OPB’s First Look: A test for Oregon’s AI disclosure law
• Meta Isn't Giving Up On Dedicated AI Hardware, Apparently
• Visa, Mastercard bet on AI, crypto ahead of earnings
• Exclusive: AI code-testing startup Blacksmith hits $1 billion valuation as AI-written code surges
• Trainocate Malaysia Launches AI Training Roadmap 2026, a 7-Level Framework to Close the Enterprise AI Skills Gap
• AI: AI Loops 101 (Part 3). Beyond AI Training. AI-RTZ #1247