In partnership with
In Today’s Issue:
💻 Google splits its custom silicon strategy into two specialized architectures
🤖 AI-driven efficiency sparks major workforce cuts at Meta
🇨🇳 China blocks Meta's $2 billion acquisition of Manus in a major escalation of the tech rivalry
⚡ A proposed hyperscale data center in Utah could consume more than double the entire state's current electricity output
⚖️ Elon Musk's high-stakes lawsuit against OpenAI goes to trial with a star-studded witness list
✨ And more AI goodness…
Dear Readers,
Google just split its AI chip strategy in two, and the customer list tells you everything you need to know: OpenAI, Anthropic, and Meta are all buying TPU capacity, marking the first real crack in NVIDIA's dominance over frontier AI training. That's our featured story today, and it's a signal that the $400 billion AI accelerator market is about to get a lot more competitive. But hardware isn't the only thing being restructured this week. Meta and Microsoft are cutting over 20,000 jobs as AI-driven efficiency moves from corporate talking point to workforce reality, S&P 500 companies shed 400,000 positions last year alone, and the uncomfortable question of who actually benefits from the AI boom is getting harder to dodge.
Meanwhile, China just blocked Meta's $2 billion acquisition of Manus in a dramatic escalation of the US-China tech rivalry, Elon Musk's lawsuit against OpenAI finally goes to trial in Oakland with a witness list that reads like a tech billionaire reunion, and a proposed Utah data center could consume more than double the state's current electricity output. Grab your coffee, this one's packed.
All the best,
Kim Isenberg
🤖 AI Layoffs Shake Tech Workforce
Meta and Microsoft’s planned cuts, alongside broader reductions at Amazon, Snap, Oracle and Nike, suggest AI-driven efficiency is moving from theory into workforce reality. The piece frames this not as a routine tech correction, but as a deeper shift: companies are pouring billions into AI infrastructure while simultaneously reducing roles, especially in entry-level, generalized IT, support and tech operations.
The sharper tension is that AI may still create new jobs, but the article suggests the transition is already uneven, with leaner startups, smaller teams and “50-person unicorns” becoming the new ideal. That leaves many workers caught in an anxious middle ground, participating in the AI boom while wondering whether it is quietly redesigning their own jobs out of existence.
China Blocks Meta-Manus Acquisition
China has moved to cancel Meta’s $2 billion acquisition of Manus, a striking intervention after Beijing faced criticism over possible AI technology leakage to the US. The decision shows a tougher stance on foreign investment in strategic AI firms, turning what was once seen as a model global startup exit into a flashpoint in the wider US-China tech rivalry. This could further inflame the US-China tech rivalry, coming just days after White House science advisor Michael Kratsios accused Chinese firms of running "industrial-scale" distillation campaigns to copy American AI models, naming the practice a systematic effort to undermine US research and development.
🖥️ Utah Data Center Advances Quickly
Kevin O’Leary’s hyperscale data center project in rural Box Elder County is nearing final approval after Utah’s military development authority backed major tax incentives to help attract top cloud companies. The proposal is striking for its scale: at full buildout, it could generate and consume 9 gigawatts of power, more than double Utah’s current average electricity use, while supporters frame it as a boost for AI infrastructure, military modernization, jobs, and Great Salt Lake water benefits. Still, the speed of approval and sweeping tax rebates leave a notable question hanging: whether local officials and residents have had enough time to weigh the trade-offs.
Sam Altman and Greg Brockman use their first joint podcast appearance to unpack OpenAI’s past conflicts, current product bets (like GPT-5.5 and agents), and future vision for personal AGI, while addressing legal battles, safety debates, and the global AI race.
TPU 8 Splits Training From Inference
The Takeaway
👉 Google introduced TPU 8t (training) and TPU 8i (inference) at Cloud Next 2026, splitting its custom silicon for the first time into two specialized architectures with 3x training performance and 80% better inference cost-efficiency over the previous generation.
👉 OpenAI is now booking Google TPU capacity alongside Anthropic and Meta, marking the first significant shift of frontier AI workloads away from an exclusively NVIDIA-dependent supply chain.
👉 Broadcom co-designed the training chip, MediaTek the inference chip, both manufactured by TSMC, creating a multi-vendor silicon pipeline that gives Google leverage against any single supplier.
👉 NVIDIA retains 81% market share and its CUDA ecosystem remains deeply entrenched, but the competitive narrative for 2027 and beyond has fundamentally shifted as hyperscalers build credible in-house alternatives at scale.
Google just broke a decade-long tradition. At Cloud Next 2026, the company unveiled not one, but two new AI chips, the TPU 8t for training and TPU 8i for inference. For the first time ever, Google is splitting its custom silicon into specialized architectures instead of relying on a one-size-fits-all design.
The TPU 8t superpod packs 9,600 liquid-cooled chips delivering 121 FP4 ExaFlops of peak compute, roughly a 3x leap over the previous generation. The TPU 8i delivers 80% better performance-per-dollar than its predecessor, with triple the on-chip memory and a new Boardfly topology that cuts network latency in half.
The signal that matters most is not the specs, it's the customer list: Anthropic, Meta, and now OpenAI are buying multi-gigawatt allocations of TPU capacity. OpenAI booking Google silicon is a first visible crack in NVIDIA's grip on frontier AI training. Broadcom co-designed the TPU 8t, while MediaTek handles the TPU 8i, both fabbed by TSMC. NVIDIA still holds 81% of the AI chip market, but the era of serious competition has officially begun.
Can Google's dual-chip bet reshape the $400 billion AI accelerator market, or will NVIDIA's CUDA ecosystem prove too deep to dislodge?
Why it matters: Google's decision to split training and inference into separate chips signals that the AI hardware market is moving beyond general-purpose GPUs. With OpenAI, Anthropic, and Meta all signing up for TPU capacity, NVIDIA faces its first credible large-scale competitive threat in AI infrastructure.
Sources:
🔗 https://blog.google/innovation-and-ai/infrastructure-and-cloud/google-cloud/eighth-generation-tpu-agentic-era/
🔗 cloud.google.com/blog/products/compute/tpu-8t-and-tpu-8i-technical-deep-dive
Gladly Connect Live '26. May 4–6 in Atlanta.
AI has everyone talking. Not everyone has answers. At Gladly Connect Live, CX leaders from Condé Nast, Smith Optics, and more share exactly how they moved AI from pilot to production, the timeline, the systems, the QA loops. 13+ sessions built for the moment we're all in. For CX and ecommerce leaders. Atlanta, May 4–6. Space is limited, secure your spot now.
White-collar employment is reversing course as S&P 500 headcount dropped by 400,000 in 2025, its first decline since 2016, driven by layoffs at companies like Amazon, Meta, and Microsoft, with deeper cuts expected in 2026 as firms aggressively reallocate capital toward AI.
Musk vs. OpenAI Goes to Trial
Jury selection kicked off today in what could become the most consequential tech trial in years. Elon Musk is taking OpenAI, Sam Altman, and Microsoft to court in Oakland, claiming the company betrayed its founding nonprofit mission to chase profits. The trial is expected to run four weeks, with a star-studded witness list including Musk, Altman, Microsoft CEO Satya Nadella, and OpenAI co-founder Ilya Sutskever.
Musk strategically dropped his fraud claims last week, narrowing the case to two charges: breach of charitable trust and unjust enrichment. The move lets him argue this isn't about personal gain but about forcing OpenAI to return to being a purely charitable foundation. The most damaging evidence could be Greg Brockman's 2017 diary entry, reportedly calling OpenAI's nonprofit commitment "a lie."
The stakes are enormous. OpenAI is now valued at $852 billion with a trillion-dollar IPO planned for late 2026. A ruling in Musk's favor could unwind the entire for-profit conversion. But with Musk running xAI as a direct competitor, the jury will also have to weigh whether this is altruism or competitive warfare.
Hiring in 8 countries shouldn't require 8 different processes
This guide from Deel breaks down how to build one global hiring system. You’ll learn about assessment frameworks that scale, how to do headcount planning across regions, and even intake processes that work everywhere. As HR pros know, hiring in one country is hard enough. So let this free global hiring guide give you the tools you need to avoid global hiring headaches.
