Top AI Stories – June 25, 2026

The artificial intelligence landscape continues to evolve at breakneck speed. This Wednesday, June 24, 2026, brought a cascade of major developments — from OpenAI’s first custom silicon to a landmark AI chip IPO, a geopolitical clash over semiconductor export controls, and the largest known AI model distillation attack on record. Here are the five most important AI stories you need to know.

1. OpenAI Unveils “Jalapeño” — Its First Custom AI Chip, Built with Broadcom

OpenAI took a major step toward vertical integration on Wednesday, unveiling its first custom-designed AI chip, named “Jalapeño,” developed in partnership with Broadcom. The inference processor is purpose-built for running OpenAI’s AI models and is currently in testing, with early results showing “significantly better performance-per-watt than current state-of-the-art alternatives,” according to the company.

OpenAI President Greg Brockman described the chip’s development on the company’s in-house podcast: “We have a deep understanding of the workload. We’ve really been looking for specific workloads that are underserved, [and asking] how can we build something that will be able to accelerate what’s possible?”

The move represents a strategic effort to reduce OpenAI’s dependence on Nvidia GPUs, following in the footsteps of Google (TPU) and Amazon (Trainium), who have built their own custom AI accelerators. Jalapeño is designed specifically for inference — running pre-built models in response to user commands — rather than the computationally intensive pre-training phase, which will continue to rely on Nvidia hardware. Even marginal reductions in inference costs could meaningfully improve OpenAI’s bottom line as it scales its agentic products like Codex.

OpenAI’s announcement emphasized a full-stack approach: “OpenAI is not only developing frontier models or building products on top of them; it is designing the infrastructure underneath them: chip architecture, kernels, memory systems, networking, scheduling, deployment systems, and product experience.” The chip was first rumored in February 2025, with the Broadcom partnership formally announced in October 2025.

2. Qualcomm to Acquire Modular for .9 Billion in AI Software Push

Qualcomm announced on Wednesday it would acquire Modular Inc. — an AI-native software platform company — in an all-stock deal valued at approximately .92 billion (19.2 million Qualcomm shares). The acquisition represents a bold bet on AI infrastructure software and a direct challenge to Nvidia’s dominant CUDA ecosystem.

Modular was founded in 2022 by Chris Lattner — the legendary engineer who co-invented the LLVM compiler infrastructure (foundational to Swift, Rust, and Clang) and worked at Apple, Tesla, and Google — alongside Tim Davis. The company built a platform that enables developers to deploy AI models efficiently across diverse hardware architectures — CPUs, GPUs, NPUs, and custom ASICs — without rewriting code for each processor. In effect, Modular provides a hardware-agnostic compute layer that competes directly with Nvidia’s CUDA lock-in.

Qualcomm CEO Cristiano Amon framed the deal as essential for the coming era of agentic AI: “As agentic AI scales across data centers and edge environments, the industry is moving toward disaggregated, multi-vendor architectures that demand a more open and modern software foundation.” Lattner added: “Joining Qualcomm gives us the scale and platform reach to accelerate that mission.”

The deal is expected to close in the second half of 2026, subject to regulatory approvals. It arrives ahead of Qualcomm’s investor day, where the company is expected to reveal a major data-center chip customer and next-gen processor plans. Qualcomm shares have risen 57% over the past three months amid growing enthusiasm for its AI chip strategy.

3. Anthropic Accuses Alibaba of Massive Claude AI Distillation Campaign

Anthropic sent a letter to U.S. Senators Tim Scott (R-S.C.) and Elizabeth Warren (D-Mass.) on June 10, accusing Chinese tech giant Alibaba of waging what it calls the “largest known distillation attack on Anthropic to date.” According to the letter — viewed by CNBC and first reported by Bloomberg — Alibaba-linked operators conducted 28.8 million exchanges with Claude using approximately 25,000 fraudulent accounts between April 22 and June 5, 2026.

Distillation is an AI training technique where outputs from a stronger, frontier model are used to train a smaller, cheaper model — effectively allowing competitors to replicate capabilities without investing in original research. Anthropic states that the attackers targeted Claude’s most valuable capabilities, including its ability to handle longer and more complex tasks and its decision-making approach.

The campaign comes just two months after the White House Office of Science and Technology Policy issued an April 2026 memorandum pledging to help AI companies detect and coordinate against industrial-scale distillation. Anthropic stated that Alibaba “ignored the Trump Administration’s warnings.” An Anthropic spokesperson said: “We believe combating the threat of illicit distillation requires coordinated action between government and industry, and we will continue working with Congress and the Administration to maintain American AI leadership.”

This is not Anthropic’s first encounter with distillation campaigns. In February 2026, the company revealed it had detected three “industrial-scale” operations from Chinese AI labs DeepSeek, Moonshot, and MiniMax. The situation is further complicated by the Trump administration’s recent export control directive ordering Anthropic to suspend access to its latest Claude models (Fable 5 and Mythos 5) by foreign nationals — a restriction that has not yet been resolved despite high-level meetings in Washington. Alibaba has not commented on the allegations but is separately suing the U.S. government over its inclusion on a Pentagon blacklist of companies tied to the Chinese military, a designation the company denies.

4. SK Hynix Targets 9 Billion Nasdaq Listing, Second-Largest Share Sale in History

SK Hynix, the South Korean memory chip giant and key supplier of high-bandwidth memory (HBM) to Nvidia and Google, announced plans to raise up to 9.4 billion through a secondary listing on the Nasdaq. The deal would rank as the second-largest share sale in history, behind only SpaceX’s 5.7 billion IPO, and would surpass Saudi Aramco’s 5.6 billion IPO (2019) and Alibaba’s landmark 2014 offering.

The company plans to issue up to 17.79 million new shares via American Depositary Receipts (ADRs), with trading expected to begin as early as July 10, 2026. Proceeds will fund new chip factories in South Korea and purchases of advanced chipmaking equipment, including extreme ultraviolet (EUV) scanners from Dutch manufacturer ASML.

SK Hynix has become a case study in how the AI boom can transform a company’s fortunes. Two decades ago, it nearly collapsed under debt. Today, it is South Korea’s most valuable company with a market capitalization of approximately .2 trillion, having overtaken its cross-town rival Samsung Electronics this week. Its stock has surged 329% year-to-date, far outpacing Micron (+314%) and Samsung (+183%).

Analysts see the U.S. listing as a catalyst for further re-rating. Ryu Young-ho, senior analyst at NH Investment & Securities, noted: “The most attractive benefit for investors is that SK Hynix will trade on Nasdaq alongside rival Micron, giving the company an opportunity to be re-rated in the U.S. market.” The deal is underwritten by BofA Securities, Citigroup, Goldman Sachs, and J.P. Morgan.

5. Europe Pushes Back on Washington’s Chip War as Dutch Minister Lobbies Against MATCH Act

Dutch Trade Minister Sjoerd Sjoerdsma visited Washington this week to lobby against the MATCH Act (House Bill 8170), proposed U.S. legislation that would extend export restrictions on semiconductor equipment sold to China. Speaking to Congress, he said: “It’s exceptional that I’m coming here to broadly outline our concerns to Congress. The stakes for the Netherlands may be very high.”

At the center of the dispute is ASML, the Dutch company that is the world’s sole manufacturer of advanced lithography machines essential for producing cutting-edge AI chips. China is already barred from buying ASML’s most advanced extreme ultraviolet (EUV) tools, but can currently purchase older-generation deep ultraviolet (DUV) immersion machines. The MATCH Act would extend the ban to cover those older DUV systems as well.

ASML CEO Christophe Fouquet previously told TechCrunch that the DUV machines China can currently buy are already approximately a decade old, making the proposed extension particularly punitive. China accounts for 19% of ASML’s net system sales, giving the company — and by extension, the Netherlands — substantial economic exposure to the dispute.

Introduced in April 2026, the MATCH Act has not yet faced a full House or Senate vote and would likely need to be folded into a larger legislative package to pass. The pushback from Europe underscores growing transatlantic tensions over semiconductor policy, as allied nations seek to balance U.S. national security objectives against their own economic interests in the global AI chip supply chain. The outcome will have significant implications for the availability of AI hardware and the pace of China’s domestic AI ambitions.

The Week Ahead

Between OpenAI’s first custom silicon, Qualcomm’s billion-dollar software bet, the escalating distillation war between U.S. and Chinese AI labs, a record-setting chip IPO, and growing geopolitical fractures over semiconductor supply chains, the AI industry is experiencing a defining week. As investors await Micron earnings and the market digests the implications of these developments, one thing is clear: the infrastructure race underpinning the AI revolution is accelerating on every front — hardware, software, geopolitics, and capital markets alike.


This article was published on June 25, 2026 and covers developments from June 24, 2026.