Tech Express | July 23, 2026 Samsung Unveils Galaxy Z Fold 8 Series with First Ultra Foldable

---

📰 Samsung Galaxy Unpacked 2026: Redesigned Z Fold 8, First Ultra Foldable Debut

Samsung unveiled its latest foldable lineup at the summer Galaxy Unpacked event in London. The highlight was the Galaxy Z Fold 8, featuring a completely redesigned, shorter and wider form factor with a 7.6-inch inner display and 5.5-inch outer screen—finally shedding the awkwardly narrow remote-control shape. The new device introduces Flex Titanium display technology that, while not eliminating the crease entirely, makes it significantly less noticeable than on previous models. Powered by the Snapdragon 8 Elite Gen 5 for Galaxy with up to 16GB of RAM and 1TB of storage, the Z Fold 8 starts at $1,899.

Samsung also introduced its first "Ultra" foldable—the Galaxy Z Fold 8 Ultra. It inherits the Z Fold 7's design language but packs a 200MP main camera, a 5,000mAh battery, and 45W wired charging with 20W wireless charging, starting at $2,099. The Z Flip 8 received a spec bump with the Snapdragon 8 Elite Gen 5 chip and 12GB of RAM, though its starting price rose $100 to $1,199.

Samsung also refreshed its wearable lineup with the Galaxy Watch 9 and Watch Ultra 2. The Ultra 2 packs an impressive 800mAh battery, up from 590mAh, and introduces four new preventative health metrics: Vitals, Heart Health Score, Daily Cardio Load, and Fitness Index. Samsung confirmed the new devices will launch in Canada on August 7.

— The Verge

---

📰 Alphabet and Tesla Q2 Earnings: Soaring AI Spending Rattles Investors

Tech earnings season kicked off with Alphabet and Tesla, and the picture was sobering for Wall Street. Alphabet posted better-than-expected Q2 revenue, but raised its 2026 capex forecast from $180-190 billion to $195-205 billion while warning of even higher spending in 2027. The company reported negative free cash flow, sending its stock down more than 3% in after-hours trading.

Tesla's results were more concerning. While revenue showed recovery, profits still missed analyst estimates as AI-related spending continued to climb. Tesla shares dropped 4% after the report, bringing July's cumulative decline to over 11%. Both reports highlight a central question: when will Big Tech's massive AI infrastructure investments begin generating meaningful returns?

Analysts note that the rise of cheaper open-source models from China and growing frugality among enterprise AI customers are eroding confidence in AI investment returns. With Meta, Microsoft, Amazon, and Apple all reporting next week, the market's scrutiny of AI spending versus profitability will only intensify.

— CNBC

---

📰 Amazon Cuts Jobs in AGI Unit, Sharpens Focus on Business AI

Amazon confirmed layoffs within its artificial general intelligence (AGI) organization, the group behind its Nova family of models. According to LinkedIn posts from affected employees, roles in model customization and post-training optimization were primarily targeted. Amazon did not disclose the scale of the cuts but characterized them as a strategic realignment toward customer-facing AI products and tools.

The move reflects Amazon's pragmatic pivot in AI strategy. Rather than competing head-to-head with OpenAI and Anthropic on frontier models, Amazon is leaning into its cloud computing and enterprise services strengths, building deployable AI solutions for AWS customers. This approach aligns with the $1 billion forward-deployed engineering initiative Amazon announced just two days earlier.

Amazon's AGI team remains operational—this is a strategic restructuring rather than a full retreat. As the enterprise AI market shifts from model bragging rights to practical implementation, every major tech company is reassessing how to allocate R&D resources for maximum business impact.

— CNBC

---

📰 Microsoft Commits $2.5 Billion and 6,000 Employees to AI Implementation Unit

Microsoft announced a $2.5 billion investment in a new business unit called Microsoft Frontier Co., deploying 6,000 employees directly into customer organizations to help implement artificial intelligence. Led by President Rodrigo Kede Lima, formerly Microsoft's commercial business lead in Asia, the unit brings together forward-deployed engineers, technical consultants, support staff, and industry-focused salespeople to help enterprises move from AI pilots to full-scale production.

Frontier Co. will work alongside existing consulting partners like Accenture and EY while supporting multiple AI models, data sources, and open systems. Judson Althoff, CEO of Microsoft's commercial business, noted that customers remain "in very different places right now" with AI deployment, and that the strongest results have come from close collaboration on intelligence platforms that protect proprietary data.

The launch comes at a critical moment—Microsoft's stock is down 21% this year amid uneven Copilot adoption, and investors are closely watching whether the company's heavy AI spending can translate into broad enterprise adoption. The Frontier Co. represents Microsoft's bet that services revenue can bridge the gap between AI investment and commercial returns.

— The AI & Software Report

---

📰 WAIC 2026: China's Full Domestic Chip Lineup Debuts as AI Enters the Industrial Phase

The 2026 World Artificial Intelligence Conference (WAIC) opened in Shanghai on July 17 under the theme "Intelligent Partners, Co-creating the Future." For the first time, the conference spanned three districts—Zhangjiang, the Expo area, and Xuhui—covering the complete AI supply chain from chip computing power to consumer AI devices. This year's defining shift was unmistakable: AI is no longer about model capability demonstrations but about real industrial deployment.

The Zhangjiang exhibition zone emerged as the conference's hardware powerhouse, gathering over 100 chip and infrastructure companies with more than 200 exhibits, including 67 domestic debuts. Over a dozen general-purpose high-compute chip companies appeared together, representing nearly every major Chinese GPU manufacturer. The standout was Dongfang Suanxin's global debut of the DF1000, the first large-scale AI chip following what industry observers call a "third path" distinct from both NVIDIA's CUDA ecosystem and traditional architectures.

The conference also showcased super-node computing clusters, humanoid robots entering real factory training environments, and AI agents evolving from chatbots into genuine productivity tools. From super-nodes replacing single chips as the new battleground for computing power, to consumer AI devices moving from concept demos to everyday experiences, WAIC 2026 painted a clear picture of an industrial AI era arriving in real time.

— 36Kr

---

📰 China's AI Model "Total War": DeepSeek V4 Goes Full Commercial, Xiaomi Slashes MiLM Prices by 99%

July 2026 may be remembered as a watershed moment for China's AI model industry. Within weeks, five leading independent AI companies made dramatic moves: Zhipu completed a HK$31.4 billion equity swap and pledged two years of no monetization pressure; MiniMax raised HK$16 billion through a share placement and convertible bond mix, with its founder pledging zero salary until AGI; Moonshot AI launched the top-tier open-source model Kimi K3 and accelerated its Hong Kong IPO preparations; DeepSeek completed a massive ¥50+ billion funding round with an IPO on the horizon; and StepFun unveiled a concept AI phone exploring hardware-software integration.

DeepSeek V4 officially went into full commercial availability on July 15—the Pro flagship edition featuring 1.6 trillion MoE parameters and the Flash lightweight edition at 284 billion, both supporting a 1-million-token context window. Most notably, DeepSeek introduced China's first "peak-valley time-of-use pricing" for AI compute, slashing off-peak rates by approximately 60% during nighttime hours, effectively bringing the electricity industry's peak-valley pricing model to AI computing.

Xiaomi joined the fray by permanently cutting MiLM API prices, with some models dropping by as much as 99% to match DeepSeek's aggressive pricing. Alibaba's Tongyi Qianwen announced Qwen3.8 with 2.4 trillion parameters entering preview. Internationally, OpenAI temporarily removed Codex usage limits, Anthropic adjusted Fable pricing four times in two weeks, and xAI released Grok 4.5 with a promise of one new model per month. A global AI model price war and performance arms race is now in full swing.

— 36Kr

---

📰 Canadian Healthtech Unicorn Jane App Acquires Clinic Software Startup Ginger Desk

Canadian healthtech unicorn Jane App (Jane Software) has completed its first public acquisition, scooping up Vancouver-based startup Ginger Desk. Founded by naturopathic doctor Julie Durnan, Ginger Desk provides trained virtual assistants—real people, not bots—who manage clinic workflows including patient scheduling and bookings. The acquisition strengthens Jane App's clinic administration capabilities within its all-in-one practice management platform.

Durnan revealed on LinkedIn that the acquisition actually closed three months ago, with all 53 Ginger Desk team members joining Jane App. Durnan now serves as Jane's Head of Services. Jane App is a leading Canadian SaaS platform that enables health clinics to operate entirely online, serving practitioners across multiple disciplines.

The acquisition marks Jane App's expansion from a pure software platform into a "software plus services" model. As telehealth and online clinic management continue to proliferate, integrating virtual administrative assistant capabilities could create deeper customer retention and open new growth avenues for the Vancouver-based unicorn.

— BetaKit

---

📰 SpaceX Starship Flight 13 Set for Launch

SpaceX is making final preparations for Starship's 13th test flight, with the launch window opening July 23 from the Starbase facility in South Texas. The mission will pair the Super Heavy booster Booster 20 with the Starship upper stage Ship 40. As with the previous flight, both the booster and upper stage will attempt controlled ocean splashdowns rather than recovery.

Flight 13 marks Starship's return to flight following a grounding period mandated by the FAA after a Super Heavy booster anomaly in a previous mission. As the largest rocket system ever built by humanity, every Starship test flight commands intense global attention—it is not only the key vehicle for NASA's Artemis lunar landing program but also the centerpiece of SpaceX's Mars colonization vision.

SpaceX is accelerating Starship's test cadence, with Elon Musk previously stating the company aims to complete multiple flights in 2026. If Flight 13 succeeds, SpaceX may attempt booster recovery on a subsequent mission—a critical milestone for achieving Starship's fully reusable design goal.

— Space.com

---

📰 iOS 27 Code Reveals Apple "Restricted Mode": iPhones Could Be Locked After Missed Payments

Code discovered in the latest iOS 27 beta reveals a new "Restricted Mode" feature that would reportedly allow Apple to remotely lock financed iPhones if missed payments are detected. Once triggered, users would lose access to nearly all apps except Phone, Settings, and the App Store. The code also includes a "Partner Finance Lock" designed to prevent users from erasing, reselling, or stripping a restricted device for parts while reportedly not sharing the device's location with financing partners.

The finding follows a Bloomberg report earlier this week claiming Apple is preparing to launch a new "Apple Upgrade" device leasing program in partnership with Klarna. Restricted Mode operates through a new "App Managed Features" system that allows financing partner apps to continuously verify a device's payment standing.

It's worth noting this code is from a beta version of iOS 27, meaning the feature could be modified or removed before the official release. Apple has not officially announced either the Restricted Mode or the Apple Upgrade program. Nevertheless, the discovery has sparked a broader debate about device ownership and consumer rights—when a phone can be remotely bricked, are users really buying or merely renting their devices?

— The Verge