Nvidia Shatters Records With $96.2B Quarterly Revenue — AI Infrastructure Spending Reaches Unprecedented Heights
August 28, 2026 — By Vito Ruocco
Introduction: The Age of AI Infrastructure Has Arrived
There was a time not so long ago when a $10 billion quarter was considered a milestone for a semiconductor company. Then came the AI boom, and the old rules went out the window. Today, Nvidia — the undisputed king of AI chips — has reported a staggering $96.2 billion in quarterly revenue, with a forecast of $108 billion for the coming quarter. That number alone would make it one of the highest-grossing companies in any industry, rivaling the quarterly hauls of Amazon, Apple, and Alphabet.
But this isnt just a story about one company financial success. Its a window into the sheer scale of the AI revolution that is reshaping every corner of the global economy. From hyperscale cloud providers committing millions of GPUs to government AI factories and a historic cross-industry cybersecurity coalition, the week of August 28, 2026, marks a defining moment in how the world builds, deploys, and secures artificial intelligence.
Let dive into the numbers, the deals, and the strategic moves that define this pivotal moment.
1. Nvidia Quarter: $96.2 Billion and Climbing
Nvidia second-quarter fiscal 2027 earnings report landed like a thunderbolt. The company brought in $96.2 billion in revenue — a jump of over $10 billion from the previous quarter and more than double the $45 billion it reported in the same period last year. Profits more than doubled to $59.7 billion, a margin that would make any CEO envious.
The lion share of this revenue came from the data center segment, which alone generated a record $89 billion — up more than 100% year-over-year. This is the engine of the AI era: the H100, B100, and next-generation Blackwell chips that power virtually every major large language model in production today.
Nvidia edge computing category, which includes its consumer gaming business, accounted for just $7.2 billion — a modest 27% year-over-year gain, but a reminder that Nvidia has fundamentally transformed from a gaming graphics company into an AI infrastructure powerhouse. The company acknowledged slower consumer PC sales tempered by elevated memory and system prices, but the core story is unmistakable: Nvidia future is the data center, and the data center is Nvidia.
CEO Jensen Huang, in his typical leather-jacket style, summed it up: The world is racing to build AI infrastructure at a scale we never seen, and Nvidia is the fulcrum. He not wrong. With a market cap now flirting with $4 trillion, Nvidia is the largest pure-play AI bet in the world — and it paying off.
2. AWS and Nvidia: 2 Million More GPUs, Vera CPUs, and AI Factories
Hot on the heels of Nvidia earnings, Amazon Web Services (AWS) and Nvidia announced a massive expansion of their 16-year partnership. The headline: AWS will deploy 2 million additional Nvidia GPUs across its global infrastructure in 2027-2028, on top of the 1 million it had already committed to at Nvidia GTC 2026.
AWS CEO Matt Garman said customer demand exceeded expectations, and the new expanded commitment reflects the breakneck pace at which enterprises, governments, and AI labs are consuming compute power. This brings AWS total Nvidia GPU commitment to 3 million units — an almost unimaginable number considering that just two years ago, the entire world AI compute cluster capacity was a fraction of that figure.
Nvidia Vera CPUs Coming to AWS
Perhaps the most strategic announcement in the deal is the introduction of Nvidia Vera CPU-based infrastructure to AWS. Vera is Nvidia custom ARM-based CPU, designed specifically for the intensive compute demands of agentic AI workloads — code execution, tool orchestration, sandboxing, analytics, and reinforcement learning pipelines.
This is a significant expansion of the Nvidia-AWS relationship beyond GPUs. Vera will serve both as a host CPU for accelerated GPU systems and as a standalone CPU for AI factory workloads. By bringing Vera to AWS, Nvidia is positioning itself as a full-stack computing company — not just an AI accelerator provider.
AI Factories for the U.S. Government
In a move that underscores the geopolitical importance of AI infrastructure, AWS and Nvidia will also build AI factories for the U.S. government, including clusters of 100,000 GPUs running on secure AWS infrastructure. This echoes the growing recognition that AI compute is a national security asset, on par with nuclear reactors or aircraft carriers in strategic importance.
3. A Historic Cyber Defense Coalition: Over 100 Companies Sign On
On a different but equally consequential front, OpenAI, Anthropic, Google, and over 100 other companies and organizations released an unprecedented open letter calling for a global surge in cyber defense. Signatories include AWS, Microsoft, IBM, CrowdStrike, Cloudflare, Palo Alto Networks, Visa, Mastercard, Uber, Shopify, Hugging Face, Arm, AMD, Cisco, and dozens more across every sector of the economy.
The core message: AI-enabled cyber attacks will become far more widespread and sophisticated as models around the world become increasingly capable. The window for action is limited — measured in months, not years.
The Three Principles
The coalition proposed three foundational principles for collective action:
- Recognize that status quo security wont be enough. Longstanding bugs, excessive permissions, misconfigurations, insecure unpatched software, weak authentication, and technical debt in legacy systems have left infrastructure dangerously exposed. Security teams, particularly for critical infrastructure (hospitals, water treatment, internet infrastructure), have been historically under-resourced.
- Empower more defenders with cyber-capable AI. AI brings specialist skills to more defenders and makes core security tasks faster, cheaper, and better. The coalition calls for sharing tools, practical knowledge, and verified fixes so one organization work can protect many others.
- Mobilize a collective response. Cyber capabilities are advancing worldwide, and no single company should control the future of defense. A global response is necessary, requiring new partnerships to raise security standards and find new solutions to emerging cyber threats.
What Comes Next
The letter calls on every organization to make cyber defense an immediate leadership priority. Specifically, it urges companies to raise security standards, fix the highest-risk weaknesses, verify results, and use capable lower-cost models for broad coverage while applying frontier capabilities to the hardest problems.
This is a landmark moment of cooperation among companies that are otherwise fierce competitors. It signals that the AI industry recognizes both the enormous potential and the enormous risk of the technology it is building.
4. Meta Gambit: Glimmer — AI for Everyone, With Caveats
Meta released Glimmer this week, an open-weight AI model that anyone can download and run on their own hardware — a direct contrast to Muse Spark, Meta more powerful model that remains locked behind proprietary APIs. The release came alongside a letter from Mark Zuckerberg arguing that AI should be for everyone rather than controlled by a handful of labs.
Zuckerberg letter is characteristically philosophical: The future of AI shouldnt be determined by a few companies in San Francisco and Seattle. It should be shaped by researchers, startups, and creators everywhere.
But as TechCrunch Equity podcast pointed out, the vision comes with asterisks. Glimmer, while open-weight, is not the most capable model Meta has built. The companys best models remain proprietary and monetized through cloud APIs. Critics argue this is a strategic move to build goodwill and ecosystem lock-in rather than genuine democratization. Still, Glimmer is a powerful addition to the open-weight ecosystem, joining the ranks of Llama, Mistral, and others.
Benchmarking reports suggest Glimmer performs competitively with GPT-4o-mini on several standard AI benchmarks, though it trails the frontier models on complex reasoning and multi-step agent tasks. For most developers and small businesses, however, it offers a compelling free alternative that runs locally — no API costs, no data leaving your premises.
5. Google AI Mode Adds Hotel Bookings and Flight Tracking
Google AI Mode, the company conversational search overlay, is getting more practical. Starting in the coming weeks, users in the U.S. will be able to book hotel stays directly through AI Mode, with support for Expedia, Booking.com, Hilton, Hotels.com, IHG, Marriott International, Priceline, and others.
Additionally, Google announced flight tracking alerts through AI Mode, as well as the ability to view flight costs in points or miles by linking airline rewards accounts. This represents Google most aggressive push yet to turn AI Mode from a curiosity into a genuine utility that replaces traditional search and booking workflows.
The move is significant because it represents AI-assisted commerce — the AI doesnt just answer questions; it completes transactions. This is a key battleground in the war between Google, ChatGPT, and Perplexity for who becomes the default AI interface for daily life.
6. ChatGPT Temporary Chats Now Saveable — OpenAI Agentic Push Continues
OpenAI added a subtle but important feature to ChatGPT this week: temporary chats can now be saved. Users can personalize a temporary chat with existing memories, custom instructions, and plugins, and then choose to save it to their sidebar.
OpenAI noted that temporary chats dont create new memories, and stay out of your sidebar unless you choose to save them. It a privacy-forward feature — users who want a clean, non-persistent session can still get one, but they now have the flexibility to keep something useful if they want.
In parallel, ChatGPT Work — OpenAI AI agent designed to get things done — can now work more autonomously. It can sign in to websites to book a DMV appointment, cancel a reservation, or fill out a job application without OpenAI ever seeing the user login credentials. This is a crucial security architecture: the agent operates with credential isolation, meaning OpenAI infrastructure never touches the actual passwords.
The combination of these features points to a future where ChatGPT is not just a chatbot but an autonomous digital assistant that manages tasks across the web on behalf of users.
7. Anthropic Adds Memory to Cowork — Claude Gets Persistent
On the competitor front, Anthropic added memory to Cowork — the same memory capability that Claude chatbot already uses. This means Cowork can now remember details from regular chats, making it easier to pick up tasks without re-explaining important information.
Memory is on by default but wont save sensitive topics like health or beliefs unless users manually enable that setting. This is a careful balance between utility and privacy, and it mirrors the approach both OpenAI and Google have taken with their own persistent memory features.
In other Anthropic news, the Wall Street Journal reported that Anthropic is likely to tell investors its potential revenue opportunities are above $30 trillion — a staggering figure that would top even Elon Musk SpaceX, which has claimed a $28.5 trillion addressable market. The comparison drew raised eyebrows across the industry, with The Verge TC Sottek quipping that Anthropic may be even more delusional than Elon Musk.
8. Perplexity Portable Computer: On-Device AI Agents
Perplexity announced a new feature called Portable Computer, which runs AI models entirely on device. Unlike the computer-controlling AI tool Perplexity launched earlier this year, Portable Computer runs AI models fully locally and asks for permission if it needs to access the cloud for more advanced research and reasoning.
It will be available first on Nvidia DGX Spark, the company personal AI supercomputer, before rolling out to PCs with compatible Nvidia RTX GPUs. This marks a trend: the major AI platforms are racing to deliver local, private AI execution alongside cloud-powered capabilities.
Conclusion: The Infrastructure Age
Looking across the week events, a clear picture emerges. The AI industry has moved past the hype or real? debate. The numbers answer that question definitively: $96 billion in a single quarter for one company chips alone, with $89 billion of that from data center AI compute. AWS alone plans to deploy 3 million Nvidia GPUs. Over 100 companies — including the fiercest rivals in tech — have agreed to cooperate on cyber defense because the threat is real and growing.
This is the infrastructure age. The models that captured headlines in 2023 and 2024 are now being deployed at planetary scale. The winners are not just the companies that build the best AI — but the ones that build the factories, the networks, and the security systems to support it.
For anyone watching the AI space, one thing is clear: the train is accelerating, not slowing down.
— Vito Ruocco, August 28, 2026