The AI Industry Has Reportedly Spent $1.4 Trillion and Still Isn’t Profitable A website called is tracking the economics of the AI boom to determine if it is profitable, and the answer is no. As of May 2026, the AI industry has spent roughly $1.4 trillion on model development, data centers, chips, networking, and other AI infrastructure. Over the same period, it has generated about $613 billion in revenue. The biggest losses belong to the leading companies: - Amazon: -$291 billion - Google: -$262 billion - Microsoft: -$235 billion - Meta: -$227 billion - Oracle: -$39 billion - OpenAI: -$27 billion - Anthropic: -$26.5 billion - xAI: -$19.2 billion Only one company is profitable: Nvidia. According to the dashboard, Nvidia has generated an estimated $478 billion in AI revenue against $225 billion in AI-related spending, for a profit of roughly $253 billion. The figures are compiled from public filings, earnings reports, analyst estimates, leaks, and industry reporting. The site’s creator describes the project as a best-effort snapshot rather than a formal audit and updates the numbers monthly. The estimates also exclude indirect benefits from AI, such as improved search, advertising, and software sales.

The most important data point in AI just dropped and it changes the entire conversation about where we are in this cycle (Save this). Jensen Huang took the stage at NVIDIA GTC Taipei 2026 and highlighted how GitHub commits, a universal measure of global software output, climbed from 300 million in 2023 to 400 million in 2024 and 500 million in 2025. In the first few months of 2026 alone, that number has nearly tripled and Jensen's conclusion was that "Agentic AI has arrived, useful AI has arrived." Then he did the math and the numbers are staggering. 30 to 40 million professional software developers represent approximately $3 trillion worth of GDP that is their combined annual salary, generating economic output across $100 trillion worth of global industry. That same $3 trillion in developer salaries is now producing nearly three times as much output. "It's effectively $9 trillion of productivity from $3 trillion of salaries. The difference is absolutely extraordinary. This is the potential. This is the promise of AI." People talk about AI killing jobs but Jensen called it complete nonsense. His logic is that if you can hire a software engineer and generate $9 trillion worth of productive work, why would you hire fewer engineers? The answer is you hire more and the data confirms it, with a new developer joining GitHub every single second as of early 2026. GitHub COO separately disclosed that 2026 commits are on pace for 13–14 billion, a 1,300% increase from 2025 with GitHub Actions compute minutes already at 2.1 billion per week, more than double the 2025 baseline. But Jensen did not stop at the productivity argument. He connected it directly to token economics, the investment thesis that matters most for everyone in this room. "Tokens are now profitable units of revenue. Because it is now profitable, AI companies want to build more tokens, generate more tokens, build more AI factories which is the reason why compute demand here in Taiwan has skyrocketed." Every agent, every automated code commit, every workflow that runs without a human prompt consumes tokens. Taiwan's own government just upgraded its GDP growth forecast to 9.64% for 2026, a 16-year high driven entirely by AI infrastructure exports. This is exactly why Milk Road has been so convicted on the AI infrastructure buildout because the productivity data is now arriving in real time at a scale that nobody modeled.

The slow yes is the wrong yes. If I can't tell within 20 seconds whether a creator will convert, the answer is no. Every time I went back and forth, asked the team, and said yes anyway, it was almost always a mistake. Influencer evaluation isn’t a spreadsheet. It’s taste. I can’t perfectly explain what makes a creator great. But I know it when I see it.

The head of AARO explains the problem with detecting military UFOs is largely one of insufficient data and cautions against jumping to premature characterizations. Do the work, then characterize based on a solid foundation. "I have found it surprising that, in this age of ubiquitous sensor coverage, it is still so difficult to get high-quality, actionable data suitable for resolving, or even just advancing our understanding of some of the more intriguing cases. That said, I have also found that many of these initially baffling reports are fully explainable once you apply a rigorous, scientific process. It is easy to look at a strange video and jump to a conclusion. But time and time again, when our team of analysts and scientists dig in, we find the answer. It has been a powerful reminder of how important it is to stick to the data and not let assumptions get ahead of the evidence, regardless of how compelling a good mystery can be. In spite of all the noise, I always try to stay focused on the cases that may demonstrate true anomalies. "

“The answer is not to fear the future. The answer is to guide it wisely.” At @CarnegieMellon 2026 commencement, our CEO Jensen Huang shared why AI calls for optimism, responsibility, and ambition.

The data you need to make a decision is never in one place. The forecast is in Salesforce. The context is buried in a Slack thread. The budget lives in a spreadsheet that was emailed three days ago. And the competitive intel? That's in a SharePoint folder you haven't opened in weeks. Most AI tools search one system at a time. Amazon Quick connects all of them, and reasons across your entire work landscape at once. It doesn't just find the answer. It builds the deliverable, updates the record, and schedules the follow-up. Quick learns your role, your priorities, and your relationships, so every answer is grounded in what actually matters to you. Not just what you asked for. What you need. Turn questions into outcomes at

The answer is in your data. The problem is your data is in 12 different places. Amazon Quick connects your CRM, email, spreadsheets, dashboards, and documents. Then it turns them into clear answers and ready-to-use deliverables in one conversation. Less hunting. More deciding.