In partnership with
In Today’s Issue:
🤑 Nvidia adds an $80B buyback while AI revenue keeps climbing
💵 SpaceX's IPO story gets complicated by losses and AI expectations
📈 A University of Tokyo chip experiment points to cooler future compute
📉 Trump's AI order whiplash shows the politics of model oversight
✨ And more AI goodness…
⚡ The Signal
Today's issue starts with the cleanest signal in AI right now: demand for compute is still outrunning almost every normal business comparison.
Nvidia just posted $81.6 billion in quarterly revenue, with data-center revenue alone hitting $75.2 billion. The company is now framing its future around AI factories, hyperscalers, enterprise AI clouds, and edge computing, while Washington is still arguing over how closely the government should inspect frontier models before release.
All the best,
Kim Isenberg
🧊 A 40-Picosecond Chip Teases Cooler AI Compute
Researchers at the University of Tokyo have built a non-volatile quantum switching element that processed one bit in 40 picoseconds in lab tests. TechRadar reports the device uses magnetic electron properties rather than continuous electrical current, promising much faster switching without the same heat penalty.
👉 tl;dr: This is not a product yet. The prototype target is around 2030, but the direction matters because AI's next bottleneck is as much energy and heat as raw speed.
🚀 SpaceX's IPO Math Gets Messier
Axios says SpaceX's newly filed IPO prospectus shows a much messier business than the valuation hype suggests. The filing reportedly shows a $4.9 billion net loss on $18.67 billion in 2025 revenue, while the AI unit containing X and xAI generated $818 million in Q1 2026.
👉 tl;dr: SpaceX may still become the largest IPO ever, but the pitch increasingly depends on future AI and compute upside, not just today's rocket-and-Starlink fundamentals.
📊 ChatGPT Moves Into PowerPoint
OpenAI's ChatGPT for PowerPoint is now in beta, letting users create slides, update existing decks, and turn source material into presentation-ready content inside PowerPoint. The beta is available globally across paid workplace and education plans as well as Free, Go, Pro, and Plus users.
👉 tl;dr: AI is moving from separate chat windows into the actual work surface. The catch is still review: OpenAI says users should check formatting, claims, and numbers before sharing.
When you read an AI earnings report, ask your assistant to separate demand from dependency: what customers are buying now, what capacity is still constrained, and what assumptions the next quarter depends on.
Why it helps: Nvidia's quarter looks enormous, but the real intelligence is in the second layer: supply assumptions, China exposure, data-center mix, and how much future growth is already priced into the story.
Try this: Paste an earnings release and ask: "Split this into current demand, future guidance, supply constraints, geographic risks, and the one metric that would change the story next quarter."
🎬 Watch This
Demis Hassabis explores the rapid evolution of artificial intelligence, discussing its potential to transform science and medicine. The conversation examines the path toward artificial general intelligence, the societal implications of this technological shift, and how current research is being integrated into everyday tools and global problem-solving initiatives.
"the largest infrastructure expansion in human history"
– Jensen Huang, Nvidia founder and CEO, in Nvidia's Q1 FY2027 earnings release
The day's dustup comes from SpaceX's IPO prospectus, which Axios says undercuts the simple story that Elon Musk's space company is already a financial monster. The filing reportedly shows big losses, a valuation target that would place SpaceX among the world's most valuable companies, and a surprisingly small Q1 revenue contribution from the AI unit containing X and xAI. The bullish twist is that SpaceX says Anthropic has agreed to pay $1.25 billion per month for compute, turning AI infrastructure contracts into a key part of the IPO narrative.
Nvidia's $81.6B Quarter Shows the AI Factory Era Is Still Accelerating
The Takeaway
👉 Nvidia reported record Q1 FY2027 revenue of $81.6 billion, up 85% from a year ago.
👉 Data Center revenue hit $75.2 billion, up 92% year over year, making AI infrastructure the center of the company.
👉 Nvidia guided for roughly $91 billion in Q2 revenue and said it assumes no Data Center compute revenue from China in that outlook.
👉 The company also added an $80 billion share-repurchase authorization and raised its quarterly dividend from $0.01 to $0.25.
Nvidia's latest quarter is less an earnings report than a map of the AI economy's center of gravity. The company reported $81.6 billion in Q1 FY2027 revenue, up 20% sequentially and 85% from a year ago, while Data Center revenue reached $75.2 billion. In plain English: the AI boom is no longer just a model story. It is an infrastructure story, and Nvidia is still collecting the tolls.
The next signal is how Nvidia is reorganizing the way it reports the business. The company says it will now split its platforms into Data Center and Edge Computing, with Data Center further broken into Hyperscale and ACIE: AI Clouds, Industrial, and Enterprise. That matters because it points to a broader market than just the biggest cloud providers. Nvidia wants investors to see AI factories spreading across countries, companies, and specialized data centers.
The guidance is just as important as the quarter. Nvidia expects about $91 billion in Q2 revenue, plus or minus 2%, while assuming no Data Center compute revenue from China. That makes the story unusually sharp: demand is strong enough that Nvidia can guide higher even while leaving a major geopolitical market out of the compute forecast.
Why it matters: AI's bottleneck is increasingly physical: chips, power, networking, memory, cooling, and capital. Nvidia's numbers suggest that the buildout is still accelerating, even as politics and export controls make the next phase harder to forecast.
Sources:
🔗 https://www.cnbc.com/2026/05/20/nvidia-nvda-earnings-report-q1-2027.html
🔗 https://investor.nvidia.com/news/press-release-details/2026/NVIDIA-Announces-Financial-Results-for-First-Quarter-Fiscal-2027/default.aspx
🔗 https://www.axios.com/2026/05/20/nvidia-earnings-shows-ai-demand-is-still-roaring
The IT strategy every team needs for 2026
2026 will redefine IT as a strategic driver of global growth. Automation, AI-driven support, unified platforms, and zero-trust security are becoming standard, especially for distributed teams. This toolkit helps IT and HR leaders assess readiness, define goals, and build a scalable, audit-ready IT strategy for the year ahead. Learn what’s changing and how to prepare.
The chart: Qwen3.7 Max scores 56.6 on the Artificial Analysis Intelligence Index, a 4.8-point jump over Qwen3.6 Max Preview. Alibaba still trails the leading models from OpenAI, Anthropic, and Google, but this is its closest approach yet to the current frontier.
The lesson: Alibaba is narrowing the gap. The strongest gains are not broad across every benchmark, but concentrated in the areas that matter most for advanced AI systems: scientific reasoning, coding, agentic capability, and reduced hallucination. Qwen3.7 Max looks less like a minor refresh and more like a serious push toward frontier-level proprietary models.
The caveat: The headline score needs context. A meaningful part of the improvement comes from the model abstaining more often instead of answering more accurately, especially on AA-Omniscience. It also used significantly more output tokens than its predecessor, so this is not just a story of raw intelligence, but also of changed behavior, higher compute usage, and better hallucination management.
The AI Policy Fight Is Now Speed vs. Scrutiny
⚡ Bottom line: Trump called off a planned AI executive order hours before an expected signing, reportedly over concern that pre-release government review could slow America's AI lead.
💡 Why it matters: The fight is no longer abstract. Frontier models are becoming powerful enough that cybersecurity officials want early access, while political leaders worry that oversight could become a competitive drag.
🔎 What it means: AI policy is splitting into two tracks: Washington is hesitating over frontier-model review, while California is moving ahead on worker disruption, safety, transparency, and economic preparation.
The most interesting AI politics story this week is not a new law. It is a canceled signing ceremony. AP reports that President Trump pulled back a planned AI executive order on Thursday after deciding the text could interfere with America's lead over China. The draft would have created a voluntary framework for government review of advanced AI systems before public release, according to people familiar with the deliberations.
Axios had described the draft as a plan for AI labs to share certain frontier models with the government at least 90 days before public release, with a focus on cyber-capable systems and national-security review. That idea sounds modest compared with hard regulation, but it hits the deepest tradeoff in AI politics: the same models that create security risk are also the models governments want their domestic companies to ship first.
California is taking a different route. Governor Gavin Newsom signed an executive order aimed at preparing workers, small businesses, and state agencies for AI-driven disruption, including early warning signals for labor-market shocks and new policy ideas around transition support. The contrast is useful: federal AI politics is wrestling with speed and national advantage, while state-level politics is already shifting toward jobs, services, and who shares in the gains.
Your ads ran overnight. Nobody was watching. Except Viktor.
One brand built 30+ landing pages through Viktor without a single developer.
Each page mapped to a specific ad group. All deployed within hours. Viktor wrote the code and shipped every one from a Slack message.
That same team has Viktor monitoring ad accounts across the portfolio and posting performance briefs before the day starts. One colleague. Always on. Across every account.
