April 25, 2026 • 6 min read
ZYMP Tech News — April 25, 2026
Meta Inks Multibillion-Dollar Deal With Amazon for AI Chips
BIG TECH
Meta Platforms has signed a multiyear agreement with Amazon Web Services to use tens of millions of AWS Graviton cores to power its growing artificial intelligence workloads. The deal, announced Friday, gives Meta access to hundreds of thousands of Amazon’s custom ARM-based Graviton5 processors in a partnership expected to last at least three years. Unlike GPUs, which dominate AI model training, the Graviton chips are CPUs designed specifically for the compute-intensive demands of agentic AI — real-time reasoning, code generation, and multi-step task coordination.
The announcement was strategically timed as Google’s Cloud Next conference wrapped up, underscoring the intensifying rivalry among cloud providers for AI workloads. Amazon CEO Andy Jassy has publicly stated his intention to compete on price-performance for AI compute, taking direct aim at Nvidia. The Meta deal also positions AWS as a viable alternative to Nvidia’s forthcoming ARM-based Vera CPU for enterprise customers.
The partnership deepens Meta’s existing cloud relationships. The social media giant previously signed a six-year, $10 billion deal with Google Cloud last August, while also maintaining ties with Microsoft Azure. For Amazon, the agreement follows closely on the heels of Anthropic’s commitment to spend $100 billion over ten years on AWS infrastructure.
Musk vs. OpenAI: Landmark AI Trial Set to Begin
ARTIFICIAL INTELLIGENCE
Jury selection begins Monday in Oakland, California, in the high-stakes legal battle between Elon Musk and OpenAI. Musk is seeking up to $134 billion in damages, alleging that OpenAI co-founders Sam Altman and Greg Brockman deceived him about the company’s nonprofit mission when he invested approximately $38 million as a co-founder in 2015. Both Musk and Altman are on the witness list, along with Microsoft CEO Satya Nadella.
The case centres on whether OpenAI betrayed its founding mission to develop AI for the benefit of humanity when it transitioned to a for-profit structure. Court filings reveal that OpenAI established a commercial subsidiary in 2017, just months after Altman assured Musk of his commitment to the nonprofit model. Microsoft subsequently invested billions, with its stake now valued at approximately $135 billion. On Friday, a US judge dismissed Musk’s fraud claims, though the core breach-of-fiduciary-duty claims will proceed to trial.
OpenAI, now valued at $852 billion, counters that Musk’s lawsuit is motivated by a desire for control and that his departure was driven by his own ambitions rather than any betrayal. The outcome could reshape the governance structure of one of the world’s most valuable AI companies.
Apple’s Incoming CEO John Ternus Charts Hardware-First Strategy
CONSUMER ELECTRONICS
Apple announced this week that John Ternus will succeed Tim Cook as CEO later this year, bringing a distinctly hardware-centric perspective to the role. Ternus, who joined Apple in 2001, has been instrumental in developing products including AirPods, Apple Watch, and Vision Pro. His appointment signals a renewed focus on devices at a time when Apple faces mounting pressure to define its AI strategy.
Rather than competing directly with companies building large AI models, Ternus is expected to prioritise AI-powered hardware — including rumoured smart glasses, wearable pendants with cameras, and AirPods with integrated AI capabilities. A foldable iPhone, long rumoured and reportedly on track for September, will be among the first major launches under his leadership. Ternus has also shown interest in robotics, having built assistive devices for quadriplegics during his college years.
The leadership transition comes amid significant challenges, including memory chip shortages, shifting US tariff policies, and Apple’s heavy reliance on Chinese manufacturing. The company recently accelerated production shifts to India and Vietnam to mitigate trade risks.
Nuclear Startup X-Energy Raises $1 Billion in Record IPO
STARTUPS
Amazon-backed nuclear reactor developer X-Energy raised $1.02 billion in its initial public offering on Thursday, selling 44.3 million shares at $23 each — a significant premium above the initially projected range of $16 to $19. The offering represents the largest nuclear public listing on record, and shares surged 27% to $29.20 on their first day of trading on the Nasdaq.
The strong demand reflects growing investor interest in nuclear energy as a power source for AI data centres. Major technology companies including Amazon, Microsoft, and Google have all signed agreements to procure nuclear-generated electricity to fuel their expanding AI infrastructure. X-Energy specialises in advanced small modular reactors, which promise safer and more flexible deployment than traditional nuclear plants.
The successful IPO underscores the convergence of two of the decade’s defining trends: the AI computing boom and the revival of nuclear energy as a clean, reliable power source for energy-intensive infrastructure.
Google Unveils Eighth-Generation TPU Chips at Cloud Next
INFRASTRUCTURE
Google announced its eighth generation of custom AI chips at the Cloud Next ’26 conference in Las Vegas this week. The new processors, designated TPU 8t and TPU 8i, are purpose-built for the agentic AI era — one optimised for training and the other for inference. Google says the TPU 8i delivers up to 80% improvement in performance per dollar over the previous Ironwood generation, particularly at low-latency targets.
The dual-chip approach reflects the industry’s growing recognition that training and inference have fundamentally different compute requirements. As AI agents become more sophisticated — handling multi-step reasoning, code generation, and real-time search — the inference workload has grown dramatically, demanding chips specifically tuned for sustained, low-latency operations.
The announcement intensifies the battle for AI infrastructure supremacy, with Google, Amazon, and Nvidia all vying for the massive compute budgets of enterprises deploying AI at scale. Unlike Nvidia, which sells chips directly, Google makes its TPUs available exclusively through its cloud platform, positioning them as a differentiated offering for enterprises already invested in the Google Cloud ecosystem.
ZY Media Productions
IT • Music • Technology