Apple just shook up its entire chip supply chain. Markets hit record highs. And AI legislation is spreading across US states faster than anyone is paying attention to.
May 8, 2026 · Global Tech
Friday delivered one of the biggest chip industry stories of the year. Here is everything that happened.
It was a Friday that moved fast. The Apple and Intel story alone would have been enough to carry the week. But there was also a wave of AI legislation passing through American state governments that deserves more attention than it is getting, a major infrastructure announcement from Anthropic, and markets setting new records across the board. Let's go through all of it properly.
Apple and Intel just made a deal. The chip industry will not look the same after this.
The Wall Street Journal reported today that Apple and Intel have reached an agreement for Intel to supply chips to Apple. The market reaction was immediate and dramatic. Intel stock jumped nearly 14%. Taiwan Semiconductor, which currently produces Apple's chips and has done so for years, fell. Broadcom also dropped over 3% on related concerns about losing Apple business.
To understand why this matters you need to understand what Apple gave up to get here. Apple spent the better part of a decade building its own silicon, the M-series chips, specifically to reduce dependency on outside suppliers. The fact that it is now bringing Intel back into the picture tells you something about the pressure the AI era is placing on even the most disciplined hardware supply chains. Demand for AI-capable chips has created shortages that even Apple, with all its leverage, cannot simply wait out.
Intel on the other hand needed this badly. The company has been fighting for relevance for several years, losing ground to Nvidia, AMD and the custom silicon push from the hyperscalers. A deal with Apple is not just a revenue story. It is a credibility story. It tells the industry that Intel's manufacturing capabilities are good enough for the most demanding customer in consumer electronics.
Intel's stock surged nearly 14% on the news. Taiwan Semiconductor fell. The chip supply chain just shifted.
Markets hit record highs today. Chip stocks led the way.
The Nasdaq closed up 1.71% on Friday, setting a new all-time record. The S&P 500 and Russell 2000 also hit new highs. The Apple-Intel story drove chip stocks broadly, with Micron up 13.7%, Oracle up 13.56% and SanDisk up 14.27%. AMD also posted double-digit gains. A stronger than expected jobs report added fuel, with the US adding 115,000 jobs in April against forecasts of 65,000. The S&P 500 technology sector is now up nearly 35% since late April alone, which is an extraordinary run in a very short window of time.
OpenAI wants to build its own AI chips. But it needs Microsoft to make it work.
OpenAI is in talks with Broadcom to produce its own custom AI processors. The catch is that the deal only goes through if Microsoft agrees to buy 40% of the chips, which it would install in its own data centres and rent back to OpenAI. If Microsoft walks away, OpenAI has to find other buyers to cover that commitment. The reasoning is straightforward. OpenAI is trying to reduce its dependence on Nvidia's expensive hardware. It has already been turning to AMD and Amazon's custom Trainium chips as alternatives. Building its own silicon would give it more control over cost and supply, which is becoming increasingly critical as compute demand outpaces what any one supplier can reliably provide.
Anthropic tapped SpaceX's Memphis data centre after growing 80 times over in one quarter.
Anthropic announced a deal to access all available compute at SpaceX's Colossus 1 data centre in Memphis, securing over 300 megawatts of capacity. The move came directly out of necessity. The company grew 80 times over in the first quarter of 2026, which created real reliability issues for Claude Pro and Max users. When your product is growing that fast, infrastructure stops being a background concern and becomes the most urgent thing on the table. The partnership also includes exploratory conversations about developing gigawatts of orbital compute infrastructure, which is the kind of sentence that would have sounded fictional five years ago.
Google killed Project Mariner and folded its features into existing products.
Google shut down Project Mariner, the experimental AI browser agent that could handle web tasks like booking hotels or organising emails. The company said the technology has moved into Gemini Agent and AI Mode instead. This is less a retreat than a consolidation. Google is clearly moving away from standalone AI experiments and toward embedding agentic capabilities inside the products people already use every day. Whether that makes the technology more useful or just less visible is an interesting question. Either way, the era of Google shipping separate AI agent demos appears to be ending.
Arm's phone business is slowing down. Its AI data centre business is not.
Arm reported quarterly royalty revenue of $671 million, below analyst estimates of $693 million. The shortfall came from weaker demand for lower-end smartphones, partly driven by higher memory costs making budget devices less attractive. But the AI data centre side of the business told a completely different story. Arm said its CPU chips now represent about 50% of compute among the world's top cloud companies. The company upgraded its AI outlook accordingly. It is a useful reminder that the same company can be losing ground in one market while dominating another at the same time.
78 AI-related bills are currently alive in 27 US states. This is no longer a niche conversation.
Worth sitting with
AI Laws Are Passing Across America. Nobody Is Watching Closely Enough.
While the chip stories dominated headlines today, something else was happening in state legislatures across the United States that deserves more attention than it is getting. Connecticut passed one of the most comprehensive AI regulation bills in the country. Iowa's governor signed a chatbot safety bill into law. Utah's governor signed nine separate AI bills in a single legislative session. Colorado moved bills covering chatbot safety, therapy bots and algorithmic pricing closer to passage.
"78 chatbot bills are currently alive in 27 US states. This is no longer a niche regulatory conversation. It is becoming everyday law."
Utah's package included bills on deepfake protections, online age verification and a ban on students using smartphones or smartwatches during school hours. One bill declared AI systems legally non-sentient and prohibited them from obtaining legal personhood. Another would regulate how AI is used by licensed therapists. These are not abstract policy debates. They are becoming law right now.
The reason this matters beyond America is precedent. US state-level regulation tends to travel. California's data privacy law became a reference point for legislation across dozens of countries. The same pattern is likely to play out with AI safety and chatbot regulation. Whatever framework American states establish over the next twelve months will shape how governments everywhere else think about the same questions. This is worth watching closely regardless of where you are.
May 9, 2026 · Update
Big Tech is spending $725 billion on AI this year. The people paying for it are its own employees.
The four largest technology companies in the world, Google, Amazon, Microsoft and Meta, have collectively committed $725 billion in capital expenditure for 2026. That is a 77% jump from last year's already record-breaking $410 billion. Almost all of it is going into data centres, custom chips, GPUs and AI infrastructure. To understand where the money is coming from, look at the other column. Meta announced 8,000 layoffs effective May 20. Amazon has cut approximately 16,000 corporate roles in Q1 alone. Microsoft offered voluntary buyouts to around 8,750 US employees. Oracle eliminated up to 30,000 positions. Snap cut 16% of its workforce. In total, over 92,000 tech workers have lost their jobs so far in 2026.
Meta's CFO was unusually candid about the connection on the Q1 earnings call. The layoffs are not happening because the work no longer needs doing. They are happening because the budget for GPUs is bigger than the budget for people, and something had to give. Mark Zuckerberg put it plainly to employees: the May cuts are a direct consequence of the AI infrastructure budget. He was not speaking in metaphor. The machines are not replacing the workers in the sci-fi sense. The capital allocation decision replaced them first.
Elon Musk filed plans for a $55 billion chip factory in Texas. The full buildout could cost $119 billion.
SpaceX filed a public hearing notice in Grimes County, Texas this week outlining plans for Terafab, a semiconductor manufacturing facility that would cost at least $55 billion in its first phase and up to $119 billion if all planned phases go ahead. The project is seeking a property tax abatement agreement from the county, with a public hearing scheduled for June 3. Musk announced Terafab in March, framing it as a response to future chip shortages and an attempt to bring logic, memory and advanced packaging manufacturing under one roof. The chips would serve Tesla, SpaceX and xAI. If completed at full scale, Terafab would be one of the largest single manufacturing investments in American history. It also represents Musk's most direct move yet into the chip production business, an area that sits at the centre of the AI infrastructure race and is currently dominated by TSMC, Samsung and Intel.
Terafab in Grimes County, Texas. A $55 billion first phase. A $119 billion ceiling. Musk's most ambitious manufacturing bet yet.
Eight stories across two days. The thread connecting almost all of them is the same: the race to own the physical infrastructure of AI is now the defining competition in the technology industry, and every other decision, who gets hired, who gets cut, who supplies whom, is downstream of that race.
Which of these hit closest to home? Drop it in the comments.
0 Comments