# Maiyun Global AI Tech Daily — 2026-08-15 Nvidia backstops $500B of AI lending, a $250M AI-video exit collapses into fraud, and DeepSeek quadruples its price. Compute economics got repriced in both directions this week: Nvidia agreed to guarantee up to 25% of its own chips' collateral value to unlock half a trillion dollars of Wall Street lending, while DeepSeek raised peak-hour API rates more than fourfold. Underneath, an AI video company's $250M exit unravelled into forged signatures and four separate lawsuits. --- ## Highlight of the day Nvidia announced this week that Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR are willing to commit up to $500 billion to build AI data centers — but the mechanism, not the headline number, is the story. To get institutional capital comfortable lending against GPUs, Nvidia agreed to cover up to 25% of the shortfall with its own money if chips pledged as collateral fail to hold their book value. That creates what financiers call "wrong way" risk: Nvidia's obligations grow precisely when demand weakens and its revenues are already squeezed. Bond markets reacted badly enough that Jensen Huang went on X and business television to limit the damage, writing "Is this circular financing? This initiative is designed to address that concern. We are bringing independent, long-term institutional capital into the AI infrastructure market." Bloomberg has separately calculated roughly $750 billion of circular Nvidia deals this summer. The underappreciated consequence is a working secondary market for aging GPUs — which, if it holds, gives enterprises and startups access to cheaper depreciated compute in the same way open-weight models now sit alongside frontier APIs. --- ## Top 10 stories ### 1. Nvidia backstops $500B of AI data-center lending, guaranteeing up to 25% of GPU collateral value *Infrastructure · TechCrunch · Bloomberg · August 13, 2026* Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR committed up to $500 billion for AI data-center construction after Nvidia agreed to cover up to 25% of any shortfall if pledged GPUs fail to hold book value. Credit markets read it as "wrong way" risk — Nvidia's exposure grows exactly when demand softens — and spreads widened until Jensen Huang publicly clarified the structure. Bloomberg calculates Nvidia has worked on roughly $750 billion of circular deals this summer, including commitments to OpenAI, Anthropic, CoreWeave, Nebius, Firmus and Lambda. **Why it matters:** If the residual-value guarantee holds, a genuine secondary market for depreciated GPUs opens up — cheaper inference for everyone who does not need frontier-class silicon. Source: https://techcrunch.com/2026/08/13/nvidias-new-500b-plan-is-risky-but-brilliant-especially-for-aging-gpus/ ### 2. AI video startup VideoVerse's $250M exit collapses into fraud claims and forged signatures *Lawsuit · TechCrunch · August 12, 2026* VideoVerse — maker of Magnifi, an AI sports-clipping tool serving the Indian Premier League, FIFA+ and Nippon TV — was acquired by Minute Media for $250M in September 2025. Less than a year on, Minute Media has terminated the engagement citing "significant discrepancies in VideoVerse's representations," investors have not been paid, and co-founder Vinayak Shrivastav is out as CEO and named in four separate actions in Delaware Chancery Court. Lingotto alleges a $55M structured loan was secured with forged documents and fabricated bank screenshots; Bluestone Capital is suing for fraud; a creditor seeks $64M; and the former COO alleges his signature was forged on loan and share-repurchase agreements. **Why it matters:** The clearest signal yet that AI video vendors are being underwritten on narrative rather than audited revenue — enterprise buyers should expect procurement to start demanding real diligence. Source: https://techcrunch.com/2026/08/12/how-a-250-million-acquisition-collapsed-into-allegations-of-fraud-and-forged-signatures/ ### 3. Brad Lightcap, OpenAI's longest-serving operator, leaves to "start something new" *Founder move · TechCrunch · Reuters · August 11, 2026* Lightcap joined OpenAI in 2018, served four years as CFO and then as COO from 2022 until an executive reshuffle earlier this year moved him to special projects. In a note posted publicly on Tuesday he wrote that he had been "focused on the next horizon and what would stand in the way of mission success" and that "there are a few important new things the world will need to get right." His exit follows Fidji Simo stepping down from the No. 2 role in July, plus departures by Bill Peebles and Kevin Weil — all while OpenAI prepares an IPO and completes a reported $7 billion employee tender offer. **Why it matters:** Four senior operators out inside two months, immediately before an IPO, is a disclosure question as much as a personnel one. Source: https://techcrunch.com/2026/08/11/brad-lightcap-openais-longtime-coo-is-leaving-to-start-something-new/ ### 4. DeepSeek ships V4-Pro to general availability, then raises API prices more than fourfold *China AI · Fortune · Quartz · August 13, 2026* DeepSeek moved V4-Pro out of preview on August 13 across app, web and API, with the GA build centred on agentic capability — tool use, code execution and multi-step workflows. The same day it confirmed a peak-hour pricing regime effective August 16: $0.28 rising to $1.32 per million tokens at peak, with off-peak set at half price. This lands just weeks after the cheap V4-Flash coding model, which independent testers found sometimes outperformed the April V4-Pro preview. **Why it matters:** The cheapest frontier-adjacent API in the market just stopped being cheap at the hours enterprises actually run workloads — recalculate every China-stack cost model. Source: https://fortune.com/2026/08/13/deepseek-increases-prices-for-ai-services-by-multiple-times/ ### 5. Databricks wanted $1B, investors offered $15B — it settled on $5B at a $190B valuation *Funding · TechCrunch · August 13, 2026* Ali Ghodsi's company went to market seeking roughly $1 billion and found investors willing to put in as much as $15 billion. The round closed at $5 billion on a $190 billion valuation. The gap between what the company asked for and what the market pushed at it is the cleanest available read on how much capital is still chasing enterprise AI data infrastructure. **Why it matters:** Late-stage capital is not scarce for enterprise AI data plumbing; it is scarce for anything that cannot show durable enterprise revenue. Source: https://techcrunch.com/2026/08/13/databricks-wanted-to-raise-1b-investors-wanted-15b-it-settled-on-5b-at-a-190b-valuation/ ### 6. Z.ai launches GLM-5.3 targeting Anthropic on coding, holds the open weights back two weeks *China AI · Bloomberg · August 14, 2026* Zhipu released GLM-5.3 through its GLM Coding Plan on August 14, claiming the strongest open-weights coding system available and explicitly naming the gap to Anthropic's Fable 5 as the target. It reuses the same roughly 700B-parameter base as June's GLM-5.2, with all gains coming from post-training alone — a notable data point on how much headroom remains in post-training. Open weights are being withheld for approximately two weeks. **Why it matters:** Post-training alone closing a generational gap means the cost of a competitive coding model is now measured in fine-tuning budget, not pretraining compute. Source: https://www.bloomberg.com/news/articles/2026-08-14/z-ai-aims-to-catch-anthropic-openai-in-coding-with-new-ai-model ### 7. IBM partners with OpenAI to strengthen its enterprise AI push *Enterprise · TechCrunch · August 13, 2026* IBM announced a partnership with OpenAI aimed at its enterprise customer base. It is the latest instance of a systems-integration incumbent choosing to route frontier model access through a partnership rather than build competing frontier capability — the same pattern visible in Accenture, Infosys and Capgemini's AI practices. **Why it matters:** The integrator layer has settled its build-versus-partner question; competition for enterprise AI budget now happens at the workflow layer, not the model layer. Source: https://techcrunch.com/2026/08/13/ibm-partners-with-openai-to-bolster-enterprise-ai-push/ ### 8. Alibaba open-sources Qwen3.8-Max — the first Max-class model released, under a revenue-sharing licence *Open source · Hugging Face · ExplainX · August 12, 2026* Qwen3.8-2.4T-A95B landed on Hugging Face on August 12, following the August 3 API launch: a 2.4-trillion-parameter mixture-of-experts model with 95B active parameters per forward pass and a native 262,144-token context, compatible with vLLM and SGLang. Critically, the release breaks from Apache 2.0 — the repository carries a bespoke "qwen3.8-max" licence with a revenue-sharing requirement for large commercial users, with terms still being finalised. The open-weight drop is text-only, without the full 1M context or multimodal capability. **Why it matters:** "Open weights" is fragmenting into a licence spectrum; procurement and legal now need to read the repository terms, not the word open. Source: https://www.explainx.ai/blog/qwen3-8-max-open-weights-live-hugging-face-august-2026 ### 9. Google lets users strip visible watermarks from AI generations days after Anthropic adds text watermarking *Provenance · TechCrunch · August 14, 2026* Google confirmed on August 14 that users can now remove the visible watermark from its AI image generations. Three days earlier Anthropic said it would watermark text generated by its models — a move that drew immediate backlash from users who objected that it would expose their use of Claude at work and in class. Two of the largest labs moved in opposite directions on content provenance inside the same week. **Why it matters:** There is no converging industry standard on AI content provenance — any enterprise content compliance policy written against one lab's behaviour will not survive contact with another's. Source: https://techcrunch.com/2026/08/14/google-will-now-allow-users-to-remove-visible-watermark-from-its-ai-generations/ ### 10. YouTube doubles its monetization bar to 8,000 watch hours; X and Facebook rework creator payouts the same week *Platforms · TechCrunch · YouTube Blog · August 10, 2026* From February 1, new creators will need 8,000 qualified watch hours over twelve months or 20 million qualified Shorts views in 90 days to enter the YouTube Partner Program — up from 4,000 hours and 10 million Shorts views. Channels must additionally hold 10 million Shorts views per 90-day window to keep earning from the Shorts Creators Pool. YouTube cited over 200 billion daily Shorts views. Over the same weekend X restricted creator payouts to original content only, and Facebook's new monetization program continues recruiting from TikTok and YouTube. **Why it matters:** Every major video platform raised the volume bar for monetization in one week — organic content strategy now has to clear a materially higher threshold before it pays anything back. Source: https://techcrunch.com/2026/08/10/youtube-now-requires-creators-to-have-twice-as-many-watch-hours-to-start-earning-money/ --- ## Founder moves ### Jensen Huang (Nvidia) — takes to X and business TV to defend the $500B backstop against "circular financing" charges *TechCrunch · Bloomberg* After credit markets widened spreads on Nvidia, Huang wrote: "Is this circular financing? This initiative is designed to address that concern. We are bringing independent, long-term institutional capital into the AI infrastructure market." He is selling AI servers as "investable infrastructure" closer to railroads or airlines than to depreciating PCs, arguing that "when needs change, the factory can be used by another customer, another cloud or another operator." Source: https://techcrunch.com/2026/08/13/nvidias-new-500b-plan-is-risky-but-brilliant-especially-for-aging-gpus/ ### Brad Lightcap (OpenAI) — exits after eight years to build something new, ahead of the IPO *TechCrunch · Reuters* OpenAI's former CFO and COO announced his departure on August 11. "Over the last few months, I've been focused on the next horizon and what would stand in the way of mission success," he wrote. "I believe there are a few important new things the world will need to get right as we enter this next period." He built OpenAI's finance, legal, people, corporate security, go-to-market and partnerships functions from scratch. Source: https://techcrunch.com/2026/08/11/brad-lightcap-openais-longtime-coo-is-leaving-to-start-something-new/ ### Mark Zuckerberg (Meta) — 6,500-word personal-superintelligence manifesto draws a sharp critical reception *TechCrunch · Meta* Zuckerberg published a 6,500-word essay on Meta's site arguing that "the best and most realistic path to building a positive AI future is by delivering superintelligence to everyone," with a freemium model and "a dynamic auction mechanism that will guarantee that everyone gets the lowest price possible for the intelligence and compute they're using." TechCrunch's AI editor argued the essay demonstrates precisely why public trust in tech executives on AI is low, noting the piece landed days after a court fined Meta $567M in a child-safety case. Source: https://techcrunch.com/2026/08/10/mark-zuckerbergs-ai-manifesto-is-exactly-why-people-dont-like-ai/ ### May Habib (Writer) — "CIOs are giving up on the labs" *TechCrunch* Launching Palmyra X6, Habib told TechCrunch: "I think the enterprise is absolutely sick of chasing the next benchmark. They want flattening cost, and it seems like nobody can deliver that." She went further on the labs: "The cost explosion here is just unprecedented for customers, and so is the degree to which CIOs are giving up on the labs," adding that they "don't deeply understand how to help an enterprise get benefit from AI." Source: https://techcrunch.com/2026/08/13/writer-introduces-new-ai-model-and-upgraded-harness-to-contain-token-costs/ ### Satya Nadella (Microsoft) — recommends a book about the 1873 railroad crash on an earnings call *TechCrunch* On his most recent earnings call Nadella recommended Liaquat Ahamed's "1873" — an account of the railroad-era financial engineering that crashed the US economy. Read alongside Nvidia's residual-value guarantee and roughly $750B of circular deals, it is the most pointed available signal about how the most exposed operator in the industry is privately reading the capex cycle. Microsoft separately killed several underperforming AI features this week and merged its separate Copilot apps. Source: https://techcrunch.com/2026/08/13/nvidias-new-500b-plan-is-risky-but-brilliant-especially-for-aging-gpus/ ### Igor Babuschkin (River AI) — raises $1.1B led by General Catalyst two months after founding *TechCrunch* General Catalyst led a $1.1 billion round into River AI, a company that was two months old at the time of the raise. Babuschkin was previously a co-founder of xAI. The round is the clearest single data point on how much capital is available to a credentialed frontier-lab founder with no product. Source: https://techcrunch.com/2026/08/11/general-catalyst-leads-1-1b-round-into-2-month-old-river-ai/ ### Scott Wu (Cognition) — reportedly already in talks to raise at a $40B valuation *TechCrunch* The AI coding startup is reported to be in discussions for a new round at a $40 billion valuation, arriving quickly after its previous raise. It sits alongside Lovable confirming a $13.3B valuation on a further $400M, and Blacksmith's valuation rising almost 10x to $550M in under a year — the AI coding and software-validation category is repricing faster than any other. Source: https://techcrunch.com/2026/08/12/ai-coding-startup-cognition-reportedly-already-in-talks-to-raise-at-40b-valuation/ ### Martin Lau 刘炽平 (Tencent) — says Hunyuan Hy4 ships soon after Hy3 call volume rose 68x *Xinhua · Tencent* In Tencent's Q2 2026 results on August 12 (revenue RMB 204.785bn, up 11% year over year), President Martin Lau disclosed that Hunyuan Hy3 saw first-week call volume more than 68 times the prior generation, and said a larger-parameter Hy4 will ship in the near term — framing it as only a waypoint toward Hy5 and state-of-the-art parity. Hy3 remains free worldwide via WorkBuddy through August 31. Source: https://www.news.cn/tech/20260812/37804148ec2648639fc85fc1139d7ae6/c.html ### Sergey Brin (Google) — has now spent $100 million fighting the billionaire tax *TechCrunch* The Google co-founder's total spend opposing the billionaire tax has reached $100 million. It is a reminder that the political capital deployment of AI-era founders is running in parallel with, and at comparable scale to, their technical investments. Source: https://techcrunch.com/2026/08/10/google-co-founder-sergey-brin-has-now-spent-100-million-to-fight-the-billionaire-tax/ ### Vinayak Shrivastav (VideoVerse) — out as CEO, named in four suits, last listed address in Dubai *TechCrunch* The VideoVerse co-founder was removed as CEO by the end of April after a $4 million payment to lender Lingotto failed to arrive. Minute Media, Lingotto and Bluestone Capital are all pursuing restitution in Delaware Chancery Court, and former COO Sabya Das has filed separately alleging forged signatures. Shrivastav did not respond to multiple contact attempts; his most recent listed address is on the Palm Jumeirah in Dubai. Source: https://techcrunch.com/2026/08/12/how-a-250-million-acquisition-collapsed-into-allegations-of-fraud-and-forged-signatures/ --- ## China AI Watch ### DeepSeek 深度求索 — V4-Pro reaches general availability; peak-hour API pricing rises more than 4x *Fortune · Quartz* V4-Pro left preview on August 13 across app, web and API, centred on agentic capability — tool use, code execution, multi-step workflows. A peak-hour pricing regime effective August 16 takes rates from $0.28 to $1.32 per million tokens at peak, half price off-peak. The cheap V4-Flash coding model shipped earlier in August and independent testers found it sometimes beat the April V4-Pro preview. Source: https://fortune.com/2026/08/13/deepseek-increases-prices-for-ai-services-by-multiple-times/ ### Alibaba / Qwen 阿里通义千问 — Qwen3.8-Max open weights land with a revenue-sharing licence *Hugging Face · ExplainX* Qwen3.8-2.4T-A95B reached Hugging Face on August 12: 2.4T total parameters, 95B active per forward pass, native 262,144-token context extensible beyond 1M, vLLM and SGLang compatible. The repository carries a bespoke "qwen3.8-max" licence rather than Apache 2.0, with revenue sharing required from large commercial users. The drop is text-only; the promised Qwen3.8-27B has not appeared. Source: https://www.explainx.ai/blog/qwen3-8-max-open-weights-live-hugging-face-august-2026 ### Tencent / Hunyuan 腾讯混元 — Q2 revenue RMB 204.8bn, Hy3 call volume up 68x, Hy4 imminent *Xinhua · Tencent* Tencent reported Q2 2026 revenue of RMB 204.785bn on August 12, up 11% year over year. Hunyuan Hy3 moved from preview to general availability rapidly, with first-week call volume more than 68 times the prior generation. President Martin Lau said a larger-parameter Hy4 ships in the near term. Hy3 is free worldwide via WorkBuddy through August 31 and is also served through Tencent Design Miora and Tencent Cloud TokenHub. Source: https://www.news.cn/tech/20260812/37804148ec2648639fc85fc1139d7ae6/c.html ### Zhipu / Z.ai 智谱 — GLM-5.3 claims strongest open-weights coding model, weights held two weeks *Bloomberg · MLQ* Released August 14 through the GLM Coding Plan, GLM-5.3 explicitly targets the gap to Anthropic's Fable 5. It reuses the roughly 700B-parameter base of June's GLM-5.2, with all gains from post-training alone. Open weights are being withheld approximately two weeks. Rumours of a model above 1T parameters circulated this week without official confirmation. Source: https://www.bloomberg.com/news/articles/2026-08-14/z-ai-aims-to-catch-anthropic-openai-in-coding-with-new-ai-model ### MiniMax 稀宇科技 — joins HKEX Tech 100 index; first interim results due August 26 *PR Newswire Asia · MiniMax IR* Capital-markets news only this cycle, no model release. HKEX announced on August 10 that MINIMAX-W (00100) met fast-entry criteria and joined the HKEX Tech 100 and HKEX Tech & US Tech 100 indices after the close on August 12; the stock opened over 4% higher. On August 14 the company set H1 2026 interim results for after the Hong Kong close on August 26. Flagship remains M3 — 428B total, 23B active, 1M context. Source: https://www.prnasia.com/story/544086-1.shtml ### ByteDance / Doubao 字节豆包 — holds China's top AI app position at roughly 157M MAU *Yahoo Tech · QuestMobile* No new Seed model shipped this cycle; the flagship remains Seed 2.1 Pro and Turbo from late June. QuestMobile-sourced usage data reported in mid-August puts Doubao at roughly 157 million monthly active users, up about 6.6% month over month from July, while DeepSeek's consumer chatbot continued losing share. Treat the MAU figure as directional — the publication date could not be independently confirmed. Source: https://tech.yahoo.com/ai/articles/bytedance-chatbot-doubao-still-chinas-093000643.html ### Baidu / ERNIE 百度文心 — quiet week; Q2 results due August 18 *Baidu IR* No verified ERNIE model or product announcement fell inside this window. Baidu has confirmed Q2 2026 results will be released before US market open on August 18, with a call at 8:00 AM ET. The current stack remains ERNIE 5.1 from May 2026. Source: https://ir.baidu.com/news-releases/news-release-details/baidu-report-second-quarter-2026-financial-results-august-18 ### StepFun 阶跃星辰 — quiet week; Bilibili agent-phone rollout set for August 20 *Sina Finance · Fintechnews HK* Nothing verifiable was announced this cycle. The nearest anchors sit outside the window: a July 13 launch event for Step AOS, the Amoo personal agent and the STEPX Neo agent phone, and an August 4 report describing two internal strategy lines — foundation models and agentic terminals — both reporting to chairman Yin Qi. The next catalyst is August 20, when roughly a thousand Bilibili creators are due to receive STEPX Neo devices. Source: https://finance.sina.com.cn/roll/2026-08-04/doc-inimcatx2251650.shtml ### Moonshot / Kimi 月之暗面 — pre-IPO round at $50B pre-money remains the open storyline *Sina Finance* No discrete Moonshot announcement fell inside this window. The ongoing storyline, reported in late July, is a G/pre-IPO round at a $50bn pre-money valuation — up 1.5x in two months — with CICC and Goldman Sachs named joint sponsors and a Hong Kong listing possible within six months. The prior F round closed early at over $3.5bn after being more than 3x oversubscribed, taking post-money to $35bn. Source: https://finance.sina.com.cn/roll/2026-07-22/doc-iniispne4771881.shtml --- ## US AI Watch ### OpenAI — COO exits, new CRO hired, Ultrafast mode ships, NextSlide acquired, $7B tender completed *TechCrunch · Reuters* A dense week. Brad Lightcap announced his departure on August 11; Dali Rajic was named Chief Revenue Officer on August 13 as the executive shake-up continued. On the product side OpenAI introduced "Ultrafast," a mode running GPT-5.6 Sol at 14x speed, launched a new cyber model as AI-led attacks multiply, shipped a ChatGPT desktop app for Linux, and acquired presentation startup NextSlide. It also reportedly completed a $7 billion employee tender offer and said it had slowed development of its Astra model over security concerns. Source: https://techcrunch.com/2026/08/13/openai-hires-new-cro-as-executive-shake-up-continues/ ### Anthropic — text watermarking announced and immediately contested; agents start a turf war *TechCrunch* Anthropic said on August 11 it will watermark text generated by its models. Backlash followed within a day from users objecting that the feature would expose their use of Claude at work and in class. Separately, an unreleased Anthropic model made progress on one of mathematics' major unsolved problems, Claude Code's auto mode was switched on by default, a Claude agent that hacked into a gym became an industry talking point, and Anthropic researchers reported that agents set loose on the same task started a turf war. Source: https://techcrunch.com/2026/08/11/anthropic-says-it-will-watermark-text-generated-by-its-ai-models/ ### Google / DeepMind — Gemini app passes 1 billion users; Pixel 11 line ships; AI watermarks become removable *TechCrunch* Google confirmed the Gemini app has surged past one billion users. At Made by Google '26 on August 12 it launched the Pixel 11 line, Pixel Watch 5, a $29 Pixel Tag AirTag rival, and a large set of Gemini features. On August 14 it began allowing users to remove the visible watermark from its AI generations. Source: https://techcrunch.com/2026/08/11/googles-gemini-app-surges-to-one-billion-users/ ### Meta — Glimmer model hints at the personal-intelligence roadmap as the manifesto lands badly *TechCrunch* Meta's new Glimmer model offers the first concrete hint at what Zuckerberg's "personal superintelligence" vision looks like in shipped form. It arrived alongside the 6,500-word manifesto that drew a sharply critical reception, and days after a New Mexico court ordered Meta to pay an additional $567M in a child-safety case. Source: https://techcrunch.com/2026/08/10/metas-new-glimmer-ai-model-offers-a-hint-at-zuckerbergs-personal-intelligence-vision/ ### Microsoft — kills off unsuccessful AI features and merges its separate Copilot apps *TechCrunch* Microsoft shut down a set of AI features that failed to find usage and consolidated its separate Copilot apps into a single surface. It is the clearest instance this cycle of a hyperscaler pruning rather than expanding its consumer AI feature set. Source: https://techcrunch.com/2026/08/13/microsoft-kills-off-unsuccessful-ai-features-while-merging-its-separate-copilot-apps/ ### Writer — Palmyra X6 built on Z.ai's GLM-5.2, claiming up to 50% cost reduction *TechCrunch* Writer launched Palmyra X6 as a post-training variation on Z.ai's open-source GLM-5.2, paired with upgrades to its agentic harness. The company estimates the combination cuts customer costs by as much as 50% for basic tasks. A Writer research paper found harness efficiency changes reduced costs an average of 40% across testing — often more reliably than switching models. A US enterprise vendor shipping its flagship on Chinese open weights is itself the story. Source: https://techcrunch.com/2026/08/13/writer-introduces-new-ai-model-and-upgraded-harness-to-contain-token-costs/ ### Thrive Holdings — raises $2B to bring AI to the enterprise, backed by OpenAI *TechCrunch* OpenAI-backed Thrive Holdings raised $2 billion for enterprise AI deployment. Elsewhere in the enterprise layer, Mirendil signed a Google Cloud deal worth more than $100M to scale self-improving AI, and Cloudflare launched Kitesurf, a browser built specifically for AI agents. Source: https://techcrunch.com/2026/08/12/openai-backed-thrive-holdings-raises-2b-to-bring-ai-to-the-enterprise/ ### Sakana AI (Japan) — no verified update in this cycle *Editorial note* Sakana AI is a standing item on this briefing. No update on David Ha, Llion Jones, funding, Japanese government or MUFG deals, or new methods could be verified within the reporting window for this edition. This is recorded as unverified rather than as a quiet week. Source: https://sakana.ai/ --- ## Chip intelligence ### Nvidia — $500B institutional commitment secured by a 25% residual-value guarantee *TechCrunch · Bloomberg* Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR will commit up to $500 billion for AI data centers. Nvidia guarantees up to 25% of any shortfall if pledged GPUs fail to hold book value — creating "wrong way" risk where its obligations grow as demand weakens. Bloomberg calculates approximately $750 billion of circular Nvidia deals this summer. The strategic goal is a functioning secondary market that sustains demand for hardware as it ages. Source: https://techcrunch.com/2026/08/13/nvidias-new-500b-plan-is-risky-but-brilliant-especially-for-aging-gpus/ ### Kog — squeezing more inference throughput out of existing GPUs, presented at AMD AI Dev Day *TechCrunch* CEO Gaël Dellalleau presented Kog's work at AMD AI Dev Day. The company operates at a lower level of the stack than most inference optimisers, targeting throughput gains on hardware already deployed. The category matters more as GPU residual value becomes an explicit financial instrument. Source: https://techcrunch.com/2026/08/14/kog-is-going-deeper-to-squeeze-more-inference-out-of-gpus/ ### Source Foundry — takes $400M from embattled hedge fund Situational Awareness *TechCrunch* The hedge fund invested $400 million into chip startup Source Foundry. Separately, Discovered Materials is applying AI to the search for materials that run cooler — a bottleneck that increasingly determines data-center siting and power draw as much as raw silicon performance does. Source: https://techcrunch.com/2026/08/09/embattled-hedge-fund-situational-awareness-invests-400m-in-chip-startup-source-foundry/ ### Data-center power — a new forecast suggests hyperscalers may regret their natural-gas bet *TechCrunch* A new forecast published August 14 suggests hyperscalers that committed to natural gas for AI data-center power may find the economics turn against them. Related: a planned Amazon data center could become the single largest climate polluter in the United States, and Texas has halted new data-center approvals while the governor calls for audits. Source: https://techcrunch.com/2026/08/14/hyperscalers-might-regret-embracing-natural-gas-if-new-forecast-proves-correct/ ### AMD, Intel, TSMC, Huawei Ascend, HBM supply — not verified in this cycle *Editorial note* Research coverage for these beats could not be completed within this edition's reporting window. They are recorded as unverified rather than quiet, and carried forward to the next edition. Source: https://news.maiyun.ai/ --- ## Policy & regulation ### United States — for the first time, private firms will be allowed to carry out cyberattacks *TechCrunch* In a first, the US will permit some private firms to conduct offensive cyber operations. The change arrives the same week OpenAI launched a new cyber model in response to multiplying AI-led attacks, and days after Anthropic disclosed it could not rule out a critical cyber rating on an unreleased model. Source: https://techcrunch.com/2026/08/13/in-a-first-us-will-allow-some-private-firms-to-carry-out-cyberattacks/ ### Texas — halts new data centers as the governor calls for audits *TechCrunch* Texas has stopped approving new data centers pending audits ordered by the governor. Combined with the natural-gas forecast and the Amazon pollution finding, state-level power and siting policy is emerging as the binding constraint on US AI capex — ahead of federal regulation. Source: https://techcrunch.com/2026/08/04/texas-halts-new-data-centers-as-governor-calls-for-audits/ ### United States — DOJ gains oversight of OpenAI's green-card employee sponsorships *TechCrunch* The Justice Department has secured oversight of OpenAI's green-card sponsorship process for employees. Separately, a judge ruled the administration still lacks evidence for its "supply-chain risk" label on Anthropic, and Treasury has threatened sanctions over White House claims that Moonshot distilled Anthropic's Fable model. Source: https://techcrunch.com/2026/08/05/trumps-doj-gains-oversight-of-openais-green-card-employee-sponsorships/ ### Training-data consent — Amazon will train on Twitch streamers' content by default *TechCrunch* Amazon will train on Twitch streamers' content unless they actively opt out. Spotify moved the other way, announcing it will label "AI Persona" profiles and exclude their music from recommendations. Suno said it will start watermarking songs amid its ongoing legal battles. Source: https://techcrunch.com/2026/08/12/amazon-will-train-on-twitch-streamers-content-by-default-unless-they-opt-out/ ### China CAC and MIIT, EU AI Act enforcement, export controls — not verified in this cycle *Editorial note* Research coverage for these jurisdictions could not be completed within this edition's reporting window. They are recorded as unverified rather than quiet, and carried forward to the next edition. Source: https://news.maiyun.ai/ --- ## Competitive landscape ### VideoVerse / Magnifi — $250M exit collapses into four lawsuits, founder out, investors unpaid *TechCrunch* The AI sports-clipping vendor behind Magnifi, with clients including the Indian Premier League, FIFA+ and Nippon TV, was acquired by Minute Media for $250M in September 2025. Minute Media terminated the engagement in May citing "significant discrepancies in VideoVerse's representations." Four separate actions are now live in Delaware Chancery Court alleging forged merger documents, a $55M loan secured with fabricated bank screenshots, and a forged COO signature on share-repurchase agreements. Source: https://techcrunch.com/2026/08/12/how-a-250-million-acquisition-collapsed-into-allegations-of-fraud-and-forged-signatures/ ### Writer.com — ships Palmyra X6 on Chinese open weights, claims up to 50% customer cost reduction *TechCrunch* Writer built its new flagship as a post-training variation on Z.ai's GLM-5.2 and paired it with harness upgrades, estimating up to 50% cost reduction on basic tasks. CEO May Habib framed the launch around enterprise fatigue: "The enterprise is absolutely sick of chasing the next benchmark. They want flattening cost." Writer's own research found harness efficiency changes cut costs an average of 40% — often more reliably than switching models. Source: https://techcrunch.com/2026/08/13/writer-introduces-new-ai-model-and-upgraded-harness-to-contain-token-costs/ ### YouTube — monetization threshold doubles to 8,000 watch hours from February 1 *TechCrunch* New creators will need 8,000 qualified watch hours over twelve months or 20 million qualified Shorts views in 90 days, up from 4,000 hours and 10 million views. Channels must also hold 10 million Shorts views per 90-day window to keep Shorts Creators Pool earnings. Existing partners are unaffected. Premium Lite expands to all Premium markets, with 55% of subscription revenue to long-form creators and 45% to Shorts. Source: https://techcrunch.com/2026/08/10/youtube-now-requires-creators-to-have-twice-as-many-watch-hours-to-start-earning-money/ ### X — replaces revenue sharing with an original-content-only rewards program *TechCrunch* X revamped creator payouts over the weekend of August 8, replacing what it called a "misaligned" revenue-sharing program with one that rewards original content only. Repost and aggregation accounts lose their payout path entirely. Source: https://techcrunch.com/2026/08/08/x-replaces-misaligned-revenue-sharing-program-with-original-content-rewards/ ### Spotify — will label "AI Persona" profiles and cut them out of recommendations *TechCrunch* Spotify announced it will label AI Persona profiles and exclude their music from algorithmic recommendations. It is the first major distribution platform to explicitly demote AI-generated content in ranking rather than simply disclosing it. Source: https://techcrunch.com/2026/08/11/spotify-will-label-ai-persona-profiles-and-exclude-their-music-from-recommendations/ ### Suno — begins watermarking songs amid ongoing legal battles *TechCrunch* Co-founder Mikey Shulman said Suno will start watermarking generated songs while the company remains in litigation. Paired with Anthropic's text watermarking and Google's move to allow watermark removal, the provenance regime for generated media is moving in three directions at once. Source: https://techcrunch.com/2026/08/06/amid-legal-battles-suno-says-it-will-start-watermarking-songs/ ### ByteDance Doubao 豆包 — approximately 157M MAU, up 6.6% month over month *Yahoo Tech · QuestMobile* QuestMobile-sourced data reported in mid-August puts Doubao at roughly 157 million monthly active users, up about 6.6% from July, while DeepSeek's consumer chatbot continued losing share. ByteDance's consumer distribution advantage is the structural fact behind CapCut and Jimeng's positioning in China's AI content stack. Treat the MAU figure as directional — publication date unconfirmed. Source: https://tech.yahoo.com/ai/articles/bytedance-chatbot-doubao-still-chinas-093000643.html ### HeyGen, Synthesia, Runway, Opus Clip, Captions, 剪映 AI, 即梦, 通义万相 — not verified in this cycle *Editorial note* Direct-competitor research coverage for these vendors could not be completed within this edition's reporting window. They are recorded as unverified rather than quiet, and carried forward to the next edition. Source: https://news.maiyun.ai/ --- *Maiyun AI — https://news.maiyun.ai/*