🌅 Tekin Morning May 2: OpenAI Chips, White House Cyber Warning, RTX 5070 & Forza 6

blog-optimized-image-container">

🤖 OpenAI + MediaTek/Qualcomm: The AI Agent Smartphone Revolution by 2028

In one of the biggest shifts in the tech industry, OpenAI — the creator of ChatGPT — is collaborating with MediaTek and Qualcomm to design and manufacture custom chips for AI Agent smartphones. According to renowned analyst Ming-Chi Kuo from TF Securities, the project is still in its infancy, but mass production is targeted for 2028.

This news signals that OpenAI is no longer just a software company — it's transforming into a full-stack player in the hardware and software ecosystem. This move directly competes with Apple's strategy (which maintains complete control over hardware and software) and could fundamentally reshape the future of smartphones. The implications are profound: we're witnessing the birth of a new category of devices where AI isn't just an app — it's the operating system itself.

According to Kuo's report, the processor design will prioritize power efficiency, memory management, and on-device AI capabilities, while more complex tasks would be offloaded to cloud infrastructure. Specifications and supply chain partners are expected to be finalized by late 2026 or early 2027. This hybrid approach — local processing for privacy-sensitive tasks, cloud for heavy lifting — represents a pragmatic middle ground between pure edge computing and cloud dependency.

But why is OpenAI moving into hardware? The answer lies in the concept of AI Agents. Instead of users navigating between multiple apps, a unified system directly completes tasks. To deliver a seamless AI Agent experience, OpenAI believes that control over both hardware and software ecosystems is essential. This isn't just about making better phones — it's about fundamentally reimagining how humans interact with computing devices.

The concept of an AI Agent is a driving force behind OpenAI's latest initiative, which could potentially bring about a shift in how users interact with their devices. Imagine a phone that doesn't just respond to commands, but anticipates your needs, manages your digital life autonomously, and acts as a true personal assistant rather than a collection of disconnected apps. This is the vision OpenAI is pursuing — and it requires silicon designed from the ground up for agentic AI workloads.

OpenAI has already signaled its interest in developing chips and other AI-centric hardware. The company is believed to be building an AI device in partnership with Jony Ive (former Chief Design Officer at Apple). OpenAI also announced a multi-year strategic collaboration with Broadcom in 2025 — a move that would see the two companies design and develop chips and systems to power the growing demands of AI compute. These aren't isolated experiments — they're pieces of a coherent strategy to own the full stack.

The analyst further suggests that OpenAI views the tighter integration of software, operating systems, and physical hardware as essential to its long-term strategy. This vertical integration approach has historically been the domain of Apple, but OpenAI's deep AI expertise and massive user base (ChatGPT has over 300 million weekly active users as of early 2026) give it a unique advantage. The question isn't whether OpenAI can build a phone — it's whether it can build an ecosystem compelling enough to make users switch from iOS or Android.

💻 Qualcomm: Dedicated CPU for Agentic Experiences — Entering the Data Center Market

Qualcomm, the mobile chip giant, is expanding its territory into the data center market. During the Q2 2026 earnings call, Cristiano Amon (Qualcomm CEO) announced that the company has quietly entered the hyperscale custom silicon market and is working on data center CPUs and high-performance AI inference accelerators.

But the most intriguing part is this: Amon revealed that Qualcomm has built a "dedicated CPU for agentic experiences in the data center." He explained: "When you think about agents, CPU becomes very important." This signals that the industry is entering a new phase where generating demand for tokens to power agentic AI is critical. This isn't just about faster inference — it's about fundamentally different compute architectures optimized for the unique demands of autonomous agents.

In Amon's telling, AI kicked off with GPUs for training, dedicated inferencing hardware was the next necessity, but the market is now entering a new phase in which it is important to "generate demand for tokens" to power agentic AI. This is a subtle but crucial distinction: traditional AI inference is about answering questions quickly. Agentic AI is about sustained, multi-step reasoning and action — which requires different compute characteristics. CPUs, with their flexibility and low-latency memory access, are better suited for these workloads than GPUs.

Amon also teased the advent of what he calls "agentic smartphones," citing recent products from Chinese handset makers as examples. He mentioned a ZTE phone that includes the Doubao personal assistant developed by ByteDance, and Xiaomi's miclaw — an AI-powered assistant that's integrated with the OS kernel and divines smartphone users' intent and then drives third-party tools to make it happen. These aren't just voice assistants — they're autonomous agents that can book appointments, manage emails, and coordinate complex tasks across multiple apps without user intervention.

When asked what agentic smartphones will mean, Amon said: "We see interesting associations now starting to form between smartphones and AI companies. We're starting to see some very interesting dynamics there, which is changing the nature of designs." Those dynamics mean smartphone designs are "moving towards products [that] have much more capable CPU." They may also need more memory, which is currently in short supply. Amon noted: "We see new memory players coming and building capacity. So we're going to have to monitor the situation and see what happens in 2027."

The CEO is already keenly aware of the memory shortage, because it's hurting Qualcomm as smartphone manufacturers — especially Chinese companies — decide to build fewer units. Amon and CFO Akash Palkhiwala both predicted demand will bottom out in Q3, then rebound. This cyclical downturn is creating strategic opportunities: Qualcomm expects to win 70 percent of Samsung's SoC business this year and next, up from its usual 50 percent. Samsung is trying to improve its Exynos SoCs and make more of them, but clearly the Korean giant isn't yet ready to stand on its own two feet.

🛡️ White House: Warning About AI Cyberattacks — Anthropic Mythos at the Center

On April 30, 2026, the White House convened an urgent meeting with a group of tech companies to discuss how to ward off digital attacks that frontier artificial intelligence tools could soon enable. The meeting was largely prompted by concerns about Anthropic's newest AI model, Claude Mythos, and its ability to unearth hidden software flaws and outpace the world's top hackers at certain tasks.

Anthropic has, for now, restricted access to Mythos to a small group of security researchers and tech companies through what it has dubbed Project Glasswing. Anthropic CEO Dario Amodei recently visited the White House to discuss Mythos with administration officials, including Treasury Secretary Scott Bessent, chief of staff Susie Wiles, and National Cyber Director Sean Cairncross. The urgency of these discussions reflects a sobering reality: AI has crossed a threshold where it can autonomously discover and exploit vulnerabilities faster than human defenders can patch them.

One list of questions sent by the White House to some tech and cyber firms, obtained by POLITICO, covers a range of technical and policy considerations. Four people said some industry representatives were confused by the questions they received, several of which were seen as vague. Two of those people said other questions addressed internal security practices some of these representatives did not feel comfortable disclosing to the government without a clear justification.

The White House is also weighing executive action on AI, which two of the people noted was discussed during Tuesday's meeting. One U.S. official familiar with the discussion, who was not at the Tuesday meeting, said the draft executive order has undergone interagency review at the deputies' level, though some resistance to the plan remains. The political complexity is significant: balancing national security imperatives with industry innovation, civil liberties concerns, and international competitiveness is a delicate act.

The White House has been taking steps to defuse a monthslong legal battle with Anthropic over the company's efforts to set ethical limits on government use of AI — a fight that led President Donald Trump in February to ban all federal agencies from using the AI company's software. Since then, growing awareness of Mythos' cyber prowess — as well as concerns that unauthorized users might be commandeering technology — has agencies clamoring for access to the tool.

Competitors, including OpenAI, have also recently begun testing their own advanced cybersecurity-focused AI models, creating additional pressure on the government to find a solution. The arms race dynamic is unmistakable: as one company develops offensive cyber AI, others must follow or risk being left defenseless. This creates a dangerous escalation spiral where the technology proliferates faster than governance frameworks can adapt.

🎮 Nvidia RTX 5070 Laptop: RAM Upgrade from 8GB to 12GB — Responding to Criticism

In an unexpected move, Nvidia has upgraded the laptop version of the GeForce RTX 5070 with a memory increase from 8GB to 12GB GDDR7 — a 50 percent boost that should reduce some performance bottlenecks and generally future-proof the GPU. This decision comes after widespread criticism from the gaming community about the 8GB limitation in an era where games and AI applications demand more memory.

The RTX 5070 Laptop with 12GB VRAM can now run AAA games at Ultra settings in 1440p and even 4K (with DLSS 5) without hitting memory bottlenecks. This is also beneficial for professional workloads such as 4K video editing, 3D rendering, and local AI model training. The upgrade addresses a fundamental issue: modern game engines and AI frameworks are increasingly memory-hungry, and 8GB has become a limiting factor for high-end experiences.

This update comes at a time when the industry is grappling with a global memory crisis. AI demand for HBM (High-Bandwidth Memory) has caused manufacturers like Samsung and SK Hynix to allocate all their capacity to data center orders, which in turn has limited the supply of GDDR7 for consumer GPUs. The fact that Nvidia managed to secure enough GDDR7 for the 12GB upgrade suggests the company prioritized this SKU — likely in response to competitive pressure from AMD's upcoming RDNA 4 mobile GPUs, which are rumored to feature 16GB of VRAM.

Nvidia also announced that DLSS 5 (Deep Learning Super Sampling fifth generation) will soon be released for the RTX 50 series. DLSS 5 claims to boost performance up to 4x while maintaining image quality — a game-changer for 4K gaming and VR. The technology leverages the RTX 50 series' enhanced Tensor Cores and introduces new AI-driven frame generation techniques that can interpolate multiple frames between rendered frames, dramatically increasing perceived smoothness.

🏎️ Forza Horizon 6: Racing in Japan with Cherry Blossoms — Xbox's First Big Exclusive of 2026

Xbox is launching its first AAA exclusive of 2026 in May: Forza Horizon 6, a lavish open-world racing game set in Japan. Developed by Playground Games, this is poised to be one of the biggest titles of the year — not just for Xbox, but for the entire gaming industry. The choice of Japan as the setting is both a love letter to car culture and a strategic play for the Asian market.

Forza Horizon 6 takes players to a fictionalized version of Japan featuring Tokyo, Kyoto, Hokkaido, and Mount Fuji. The game boasts over 600 cars, from classic JDM legends (like the Nissan Skyline GT-R and Toyota Supra) to modern hypercars (like the Bugatti Chiron and Koenigsegg Jesko). The environment includes winding mountain roads, bustling urban highways, and rural routes lined with cherry blossoms. The attention to detail is extraordinary: seasonal weather affects driving conditions, neon-lit Tokyo streets come alive at night, and traditional festivals provide dynamic backdrops for races.

One of the most exciting features of FH6 is Touge Racing mode — a Japanese racing style that takes place on narrow, winding mountain roads. This mode is inspired by Japanese street racing culture and famous anime like Initial D. Players must defeat rivals on dangerous courses with skill and precision. The physics engine has been tuned specifically for this mode, with enhanced tire modeling and weight transfer simulation that rewards smooth, precise driving over raw power.

Forza Horizon 6 is also the first Forza game fully optimized for Xbox Series X|S and PC (no Xbox One support). This allowed Playground Games to leverage the full power of next-gen hardware — including fast SSDs for instant loading, Ray Tracing for realistic reflections, and 3D audio for an immersive experience. The game also supports 120fps mode on Xbox Series X and high-end PCs, delivering buttery-smooth gameplay that racing enthusiasts demand.

💾 Global Memory Crisis: Why Smartphones and GPUs Are Getting More Expensive

One of the biggest challenges facing the tech industry in 2026 is the global memory crisis that has severely impacted the production of smartphones, GPUs, and consumer devices. This crisis stems from unprecedented demand from AI data centers for HBM (High-Bandwidth Memory), forcing manufacturers to allocate all their capacity to enterprise orders.

According to industry reports, global smartphone chipset shipments are expected to decline 7 percent in 2026, but industry revenue will still grow in double digits due to rising prices. AI-driven data center demand is pushing chipmakers toward high-margin HBM, driving up DRAM prices and squeezing phone makers. This creates a paradox: fewer units sold, but higher revenues — a clear sign of supply-constrained pricing power.

Qualcomm, whose Snapdragon processors are commonly found in premium smartphones, has forecasted an 18 percent revenue decline for its smartphone processors in the second quarter of 2026. The company generated over $6 billion in mobile chip sales during the first quarter, but this figure is expected to fall to approximately $4.9 billion. This dramatic drop reflects both reduced unit volumes and the cascading effects of memory shortages on device assembly.

Samsung also complained about the memory crisis in its earnings report, saying that limited supply of memory chips has slowed production of Galaxy S26 phones. The company announced it is investing in new production capacity, but this capacity won't come online until late 2026 or early 2027. The capital expenditure required is staggering: Samsung alone is investing over $30 billion in new fab capacity, but the lead time for semiconductor manufacturing means relief is still months away.

According to a CNBC report, memory chips and storage drives are becoming a major bottleneck in the artificial intelligence buildout, driving companies' capital expenditures higher. Chipmakers warn the supply crunch will get worse before it gets better, and Wall Street sees an opportunity. Memory manufacturers' stock prices have surged as investors bet on sustained pricing power and margin expansion. SK Hynix, the leading HBM supplier, has seen its stock price triple over the past 18 months.