Five stories this weekend. Two of them change the game for anyone building AI products. One puts a ticking clock on your inference costs. One proves that “public AI” is no longer just a slogan. And one is the most expensive marketing exercise in corporate history. Here’s the operator’s read.<\/p>\n\n\n\n
This is the biggest structural shift for AI distribution this year, and it landed on a Wednesday with almost no fanfare.<\/p>\n\n\n\n
The European Commission adopted binding specification measures under the Digital Markets Act<\/a> on July 16 ordering Google to open 11 Android feature groups to rival AI assistants. That means third-party assistants get the same voice activation, cross-app action capabilities, and system-level access that Google reserved exclusively for Gemini. Users will be able to say “Hey [your assistant]” and have it book a ride, draft a reply, or pull up context from another app, all without Google’s permission layer in between.<\/p>\n\n\n\n The second measure is arguably bigger: Google must share its search data with competing search engines and AI developers, anonymized and priced by a set formula, starting January 2027<\/a>. The Android interoperability changes follow by July 2027. Non-compliance carries fines up to 10% of global revenue.<\/p>\n\n\n\n Here is why this matters if you build with AI. The moat around Google’s distribution just got a regulatory hole punched through it. If you’re building an AI assistant, your EU go-to-market just changed: you can now compete for default status on two billion Android devices without building your own hardware. If you’re building on top of search data, the cost of training and grounding your models against real query patterns just dropped from “impossible without Google’s scale” to “available at a regulated price.” I’ve been building content systems that depend on search intelligence for two years. This is the single most important distribution ruling since the browser ballot.<\/p>\n\n\n\n Here is a sentence I didn’t expect to write: a nonprofit just raised $400 million to compete with OpenAI, Google, and Anthropic.<\/p>\n\n\n\n Current AI, led by CEO Ayah Bdeir (formerly Mozilla’s AI lead, founder of littleBits), secured commitments from the French government ($100M), the Ford Foundation, MacArthur Foundation, DeepMind, and Salesforce<\/a>. The five-year target is $2.5 billion. The thesis: if AI is going to transform every aspect of life, there has to be a public option that doesn’t answer to shareholders.<\/p>\n\n\n\n This is not a research lab publishing papers. Current AI already shipped Suno Sutra in February: a pocket-sized, offline AI device running 22 Indian languages with no internet required, built with India’s Bhashini language division. Last month they launched an open-source chatbot at the AI for Good Summit in Geneva.<\/p>\n\n\n\n I’m filing this under “matters” because the funding mix is the tell. When a sovereign government, two major foundations, AND two of the biggest AI companies all write checks to the same nonprofit, that’s a signal that even the incumbents think the current market structure is fragile. For builders, watch this space. If Current AI ships usable open infrastructure, the cost floor for multilingual AI just dropped to zero.<\/p>\n\n\n\n SK Group Chair Chey Tae-won told the Korea Chamber of Commerce’s Jeju Forum on July 19<\/a> that foreign governments now classify AI memory access as “economic security” and that customers are requesting 60 to 100 percent more AI memory capacity for 2027. His blunt assessment: “No company has meaningful new capacity” to meet that demand.<\/p>\n\n\n\n This is a breaks-your-stack story. SK Group controls SK Hynix, the world’s dominant supplier of HBM (High Bandwidth Memory) chips that power every major AI accelerator from Nvidia’s H200 to AMD’s MI350. When the person who actually manufactures the memory tells you supply can’t keep up with demand, that is not speculation. That is a production forecast.<\/p>\n\n\n\n For builders, the implication is direct. If you’re planning inference-heavy workloads for 2027, your GPU costs are going up because memory availability is the binding constraint, not compute. The companies that locked in long-term supply contracts (Microsoft, Google, Meta) will be fine. Everyone else is bidding on a shrinking pool. If you run your own infrastructure, start those conversations now. If you rely on cloud providers, expect pricing pressure by mid-2027 at the latest.<\/p>\n\n\n\n Stanford’s Digital Economy Lab published a statement on July 13<\/a> signed by more than 200 economists and AI researchers, including 16 Nobel Prize winners: Michael Spence, Daron Acemoglu, Simon Johnson, Joseph Stiglitz, and others. The core claim: “AI may become radically more powerful over the next 10 years. This could drive an unprecedented transformation of our economy, larger than the Industrial Revolution, but unfolding over a vastly shorter time frame.”<\/p>\n\n\n\n The letter calls on policymakers and technology leaders to deepen research on economic impacts and build the policies and institutions to ensure AI complements human capabilities rather than displacing them wholesale. As Al Jazeera reported<\/a>, several signatories who previously dismissed AI job fears have reversed their positions, particularly on white-collar displacement.<\/p>\n\n\n\n I’m putting this in the “matters” column because of who signed it, not what it says. The policy recommendations are broad. But when 16 Nobel economists collectively reverse their prior stance and say the timeline is shorter than anyone assumed, that is a leading indicator. If you’re building AI tools that automate knowledge work (and I am), this letter is the establishment catching up to what operators already know: the displacement is real, it’s white-collar, and it’s faster than the Industrial Revolution analogy suggests.<\/p>\n\n\n\n Anthropic confidentially filed for IPO on June 1<\/a> following a $65 billion Series H that valued the company at $965 billion. The target is an October Nasdaq listing with Goldman Sachs, JPMorgan, and Morgan Stanley leading an offering expected to raise more than $60 billion. OpenAI filed its own confidential S-1 in late May<\/a> at a private valuation between $730 billion and $850 billion, also with Goldman and Morgan Stanley, targeting September to November 2026.<\/p>\n\n\n\n Two companies that build the models most of us ship against are about to be publicly traded at the same time. Neither is profitable. Anthropic reportedly spends $19 billion per year on compute. OpenAI lost roughly $9 billion on $13.1 billion in 2025 revenue. Reuters reported in late June that OpenAI may push to 2027.<\/p>\n\n\n\n I’m filing this under marketing until they actually price. The valuations are staggering, but they’re private-market numbers that haven’t been tested by public investors yet. For builders, the practical question is whether going public changes the API pricing calculus. Public companies face quarterly earnings pressure. That could mean either aggressive price cuts to grow revenue (good for you) or margin protection that keeps prices high (bad for you). Watch the S-1s when they go public. The cost structure they disclose will tell you more about your future API bills than any benchmark ever will.<\/p>\n","protected":false},"excerpt":{"rendered":" Five stories this weekend. Two of them change the game for anyone building AI products. One puts a ticking clock on your inference costs. One proves\u2026<\/p>\n","protected":false},"author":1,"featured_media":65,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[10],"tags":[],"class_list":["post-66","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-news"],"_links":{"self":[{"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/posts\/66","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/comments?post=66"}],"version-history":[{"count":0,"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/posts\/66\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/media\/65"}],"wp:attachment":[{"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/media?parent=66"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/categories?post=66"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/tags?post=66"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}Current AI Raises $400M to Build a Public Alternative to Big Tech’s AI<\/h2>\n\n\n\n
SK Group Warns the AI Memory Wall Hits in 2027<\/h2>\n\n\n\n
200 Economists and 16 Nobel Laureates Say “We Must Act Now” on AI’s Economic Impact<\/h2>\n\n\n\n
Anthropic and OpenAI Are Both Filing for IPO at the Same Time<\/h2>\n\n\n\n