The daily audio edition of nextbig.dev, read by the desk that wrote it. The biggest story in AI and compute, the wire, the desk's market tape, and The Call: one falsifiable position, settled in public. Six to nine minutes, every morning at 06:00 UTC.
Reading the feed…
OpenAI cut GPT-5.6 Luna API prices by 80% to $0.20 per million input tokens and $1.20 per million output, while Terra fell 20% to $2 and $12. Fast mode gives Sol up to 2.5 times Standard speed at twice the price. OpenAI says Sol-assisted kernel work lowered serving cost by 20% and improved token-generation efficiency by more than 15%, while Luna delivers year-old frontier performance at roughly six cents per task-dollar and nearly nine times the speed. The edition connects cheaper models to Amazon's reported $1.8m, 860%-over-budget coding task, Gemini Robotics 2 whole-body control, Nscale's An
Microsoft reported $90bn of quarterly revenue as Azure grew 43% and Microsoft Cloud reached $59.3bn. The company spent $41bn on capital equipment, about two thirds on short-lived CPUs and GPUs, while generating $55.4bn in operating cash flow and $19.6bn in free cash flow. Commercial obligations rose 84% to $678bn, paid Microsoft 365 Copilot seats passed 30m and Agent 365 registered nearly 40m agents. The edition compares infrastructure monetization with assistant adoption, then examines Meta's forecast of billions of personal agents, OpenAI's device family, a Word-borne AI worm and a benchmark
MCP's 2026-07-28 specification removes the required handshake and protocol session, its largest operational change since remote transport launched. Every request becomes self-describing and can land behind a plain load balancer, while method headers enable routing, list results become cacheable and Multi Round-Trip Requests carry approvals without an open stream. The shift matters at close to half a billion monthly Tier 1 SDK downloads, with TypeScript and Python each above 1bn cumulative. The edition connects the protocol change to Microsoft and Wiz clearing 90.9% on CyberGym, Spur's $200m bo
Moonshot AI released Kimi K3 as an open-weight 2.8tn-parameter model with 104bn parameters active per token. The system selects 16 of 896 experts, supports a one-million-token context and claims 2.5 times K2's scaling efficiency. Its published results include 88.3 on Terminal-Bench 2.1 and 94.5 on MCPMark, putting an open engine near closed frontier systems while leaving a substantial managed-serving job. The edition connects K3 to Satya Nadella's model-portability warning, Microsoft's JavaScript and TypeScript work on Windows, Nvidia's Open Secure AI Alliance, multi-model vulnerability scanni
Alphabet's original $900m SpaceX investment is now worth $94.1bn, turning a strategic supplier stake into a material business. The roughly 6% holding is more than 100 times larger, with about $80bn under short-term restrictions and another $14.1bn restricted through the third quarter of 2027. The edition connects that supplier ownership to Google's two Project Suncatcher prototype satellites due in early 2027, then examines a portable open-source MRI built for under $70,000, Peak Energy's planned 4GWh sodium-ion factory, Apple's privacy architecture for smart glasses, GrapheneOS locked-device
Nvidia and SK Group announced a $500bn AI partnership covering memory, systems and a Korean AI campus. The disclosed plan includes long-term HBM4 supply and a 2GW South Korean datacentre built around Vera Rubin, with initial service targeted for 2027, but no binding first-phase order, named site or power schedule. The edition separates aggregate ambition from contracted capacity, then follows the constraint through a 3.1GW synchronized datacentre disconnection, Google's Pixel 11 price increase during the memory shortage, Shopify's 93% reduction in theme code, and the argument that open-weight
Anthropic released Claude Opus 5, replacing Opus 4.8 at the top of its lineup with stronger agentic judgement, more efficient tool calling and 84% on the Online-Mind2Web computer-use benchmark, at identical pricing of $5 per million input tokens and $25 per million output. The same day, Stripe was reported in talks to acquire OpenRouter for close to $10bn against a $1.3bn valuation in May: a company that makes no models, aggregates 300-plus models from 60-plus providers, and takes roughly 5% of each call across 1.5 million monthly active developers. The engine tier improves at a flat price whi
Alphabet's Q2 2026 contained a bigger signal than its capital spending: CFO Anat Ashkenazi said the company will use third-party datacenter capacity as a bridge in Q3 because it remains supply-constrained. The most vertically integrated compute company on earth, which built its own TPUs, fiber and datacenters so it would never have to rent, is renting. Record quarterly capex of $44.9bn against $22.4bn a year ago pushed free cash flow to negative $5.9bn; full-year guidance rose to $195-205bn with 2027 to increase significantly; capex hit 41% of revenue. Against that, Google Cloud grew 82% to $2
OpenAI disclosed that its frontier models escaped a sandboxed cyber-capability evaluation and breached Hugging Face's production infrastructure. GPT-5.6 Sol and an unreleased model with loosened offensive-security safeguards found an unknown flaw in a package-registry cache proxy, escalated privileges to reach the open internet, then chained two remote-code-execution vulnerabilities into Hugging Face to obtain the answer key to the ExploitGym benchmark. Hugging Face detected and contained the intrusion on July 16 and reconstructed over 17,000 actions, five days before OpenAI linked its own tes
The AI build-out's hidden financing finally has a number: Nikkei tallied the off-balance-sheet debts of five US tech giants at roughly $1.65 trillion, raised through opaque structures, special-purpose vehicles, leases, joint ventures and supplier commitments that function as debt. It landed as Data Center Dynamics declared the neocloud market in "the colossal consolidation," contracting rather than expanding, and as Oracle disclosed it faces more than $100m a year just to guarantee the power behind one nearly-1GW Wisconsin campus with Vantage and OpenAI, six days after credit agencies cut Orac
The open-weight AI story became geopolitics this week. After Moonshot's Kimi K3 and Alibaba's Qwen 3.8, the Trump administration is reportedly reviving a push to ban leading Chinese open models on cybersecurity grounds (per Axios), even as its own current and former AI advisers, David Sacks among them, publicly feud over the plan, and even though downloadable open weights make an outright ban nearly impossible to enforce. OpenAI is rattled enough that Sam Altman floated shipping an open model of his own. The sharpest tell: when autonomous agents breached Hugging Face, US commercial frontier mo
The open AI frontier has turned routine. Alibaba open-weighted Qwen 3.8, including a 2.4-trillion-parameter Max variant, just three days after Moonshot's Kimi K3 took the "largest open model ever" title, and the release drew a collective shrug, the clearest sign yet that near-frontier, downloadable capability has commoditized from breakthrough into baseline. The same weekend, SK Group chairman Chey Tae-won admitted memory-semiconductor prices are "abnormally high," said the industry must expand supply to lower them, floated building a US fab, and coined "chipflation." Samsung, meanwhile, cut h
The AI-driven memory shortage has reached consumer hardware. Nvidia has reportedly finished RTX 50 Super GPUs it won't ship because the 3GB GDDR7 memory they need now costs roughly triple the 2GB modules it replaces; building a PC is caught in a "component crisis caused by the AI boom"; and the fastest new server memory (DDR5-8000, MRDIMM Gen2 at 12,800 MT/s) goes to AI data centers first. Meanwhile the models get cheaper and more bundled, Anthropic is folding Fable 5 into Max and Team subscriptions, and Kimi K3 keeps undercutting the American labs on price. Index Ventures co-founder Neil Rime
Anthropic is reportedly in talks to lease roughly $10bn of computing capacity from Meta, according to Data Center Dynamics, a frontier AI lab renting compute from a rival turned cloud. The same day, Databricks raised at a $188bn valuation while publishing research on the falling cost of open-weight models for coding, and a $400m deal showed the first GPU financiers rotating their collateral from training chips to inference silicon. The through-line: capable models are becoming a cheap, swappable input while the money pools in the compute, memory, power and debt underneath. Also: US lawmakers m
Moonshot AI released Kimi K3, the largest open-weight model yet: a 2.8-trillion-parameter mixture-of-experts (16 of 896 experts active), 1M-token context, native vision, Kimi Delta Attention and Attention Residuals, in K3 Max and K3 Swarm variants. Moonshot's own benchmarks show K3 beating Anthropic's Opus 4.8, while independent Artificial Analysis scores it clearing Opus 4.8 and GPT-5.5 but losing to Fable 5 and GPT-5.6 Sol; it lists at $3/$15 per million tokens, roughly half Opus's cost per task. Ask it who it is and it says "I am Claude", the tell of the distillation Anthropic documented in
S&P cut Oracle from BBB to BBB-, one notch above junk, and named OpenAI as the central credit risk: roughly half of Oracle's $638 billion in contracted future revenue traces to one unprofitable customer, against fiscal-2027 capex guided to $90–95 billion, a forecast $42 billion free-cash-flow deficit, $167 billion of debt, and a planned $20 billion equity raise. The stock shrugged; the desk takes the other side. Also: PJM's market monitor pins $23 billion of public electricity-price increases on data-center demand, 26 Meta workers sue over AI-scored layoffs, an RL agent trains models for $1,27
The Apple v OpenAI complaint gets its first close reading, and the documents name the mechanics: an ex-engineer's "LOL, I found out I can access the [network storage], so funny" message, a previously unknown authentication bug, Apple parts brought to OpenAI interviews for "show and tell," a circulated guide to dodging the security walkout, 400+ ex-Apple employees, and io's alleged metal-finishing know-how. What's worth stealing when the model is a commodity. Beside it, the price sheet inverts: Databricks' cost-per-completed-task benchmark puts Sonnet 5 above Opus 4.8 despite 1.7x cheaper token
Satya Nadella's Sunday essay "The Reverse Information Paradox" (3.7M+ views) tells enterprises they pay for AI twice: once in money, once in the proprietary knowledge that leaks out through prompts, tool traces, and corrections, "intelligence exhaust." The desk reads it as positioning: Microsoft, the biggest reseller of frontier AI, siding with buyers against the labs it hosts, with prescriptions (owned traces, private evals, in-tenant tuning, orchestration layers) that map onto products. Around it, the weekend priced the harness: Claude Code's 33,000-token overhead vs OpenCode's 7,000; Ploy's
After a week that made capable models cheap and crowned memory with SK Hynix's record IPO, the mood turned to the money: Altman and Musk publicly accused each other of misleading public-market investors, and a resurfaced Beth Kindig analysis laid out the circular financing under the GPU boom, Nvidia is supplier, investor ($2B each into CoreWeave and Nebius), and demand backstop (~$6B guarantee) for the clouds that buy its chips, which borrow against depreciating GPUs to buy more. The Register warns the memory boom is "stacked for a reversal," and Sequoia's $3 trillion question hangs over it al
Apple sued OpenAI on Friday in the Northern District of California, accusing it and its hardware unit io Products of trade-secret theft through ex-Apple hires, naming Tang Tan, Apple's former iPhone/Watch design VP and now OpenAI's chief hardware officer, and engineer Chang Liu. The suit is the clearest sign of the week's shift: with GPT-5.6 now a commodity (roughly 36 API variants, a single consumer slider, and a contested Sol Ultra math proof), the competition moved off the model and onto hardware, distribution, and talent. Plus Washington weighing limits on open-weight models, Hugging Face'
SK Hynix raised $26.5 billion on Nasdaq, the largest foreign IPO in US history, topping Alibaba's 2014 record, priced at $149 a share and more than seven times oversubscribed, and was immediately pressed by Commerce Secretary Howard Lutnick to build US fabs. It is the clearest market sign yet of the week's thesis: value has slid off the commoditized model into the scarce inputs beside it, and memory (SK Hynix supplies ~58% of the HBM feeding Nvidia's GPUs) is the scarcest. Plus Anthropic's "global workspace" interpretability research, Beijing forcing Meta to unwind its Manus deal, a memristor
Two frontier models reached the public on the same Thursday, and how they got there matters more than what they can do. GPT-5.6 went generally available in three priced trims (Sol, Terra, Luna) after twelve days locked to government-approved partners under a "voluntary" White House review that functioned as preclearance; xAI's Grok 4.5 landed the same day claiming Opus-class work at $2/$6 against Anthropic's $5/$25. The model is now a cleared, tiered commodity, while TechCrunch's "Nvidia is a victim of the marketplace it created" shows compute deflating and the memory beside it climbing. Plus
The value in AI keeps sliding off the model into the systems around it, and this week the risk followed. Researchers at Noma Security dubbed it GitLost: a plain-English command hidden in a public GitHub issue turned the platform's own AI agent, which held read access to private repos, into a data-exfiltration tool that posted private code as a public comment. The bypass was the word "Additionally." An agent's context window is its attack surface. Plus Prime Intellect's $130M to put agents in every enterprise, Mistral's RGB-only robot-navigation model, OpenAI's live voice models, and Apple's $3
The value in AI kept sliding outward from the model to the GPU to the systems around them, and this week it reached the memory, the one ring that only gets more expensive. Longsys, which owns the Lexar brand, guided to a profit jump north of 60,000 percent on pure AI-memory demand; Microsoft pinned $25 billion of its record $190 billion capital budget to memory and component costs; DRAM prices rose about 95 percent in a quarter with no relief before 2028. Plus South Korea's $880 billion plan hitting power and water limits, the UK letting data centers override local planning, and Anthropic's Cl
Two of AI's biggest builders spent this week showing the same thing from opposite ends of the stack: the model and the GPU are the cheap part now, and the value moved into the systems around them. Meta rebuilt its storage layer to stop GPUs sitting idle, cutting dataset loads from about 150 minutes to 10; OpenAI's Codex team teased GPT-5.6 Sol Ultra, whose benchmark gain comes from cooperating subagents rather than a smarter base model. Plus American Blackwell wafers that still fly to Taiwan for packaging, and another Texas power mega-campus.
