AI News Archive: July 23, 2026 — Part 3
Sourced from 500+ daily AI sources, scored by relevance.
- 60% of search journeys now involve AI. Here is what that means for brand visibility.
60% of search journeys now involve AI. Here is what that means for brand visibility. Miami Herald
- EXCLUSIVE: Marco Rubio tells diplomats to play down talk of American tech 'kill switch'
EXCLUSIVE: Marco Rubio tells diplomats to play down talk of American tech 'kill switch' Reuters
- EXCLUSIVE: Amazon's Bezos pushes Prime Video redesign focused on AI
EXCLUSIVE: Amazon's Bezos pushes Prime Video redesign focused on AI Reuters
Score: 73🌐 MovesJul 23, 2026https://www.reuters.com/business/media-telecom/amazons-bezos-pushes-prime-video-redesign-focused-ai-2026-07-23/ - Black Forest Labs launches FLUX 3 capable of generating images and 20-second video with audio — but in limited release to start
Black Forest Labs (BFL) is expanding its FLUX family beyond image generation with today's launch of FLUX 3 , a multimodal frontier model trained to understand and generate images, or combined audio/video clips up to 20 seconds from a single prompt — and to extend the same underlying architecture to robotic vision and actions. The Freiburg, Germany-based AI lab says FLUX 3 is jointly trained across those modalities rather than assembling separate image, video and audio models behind a common interface. That distinction is central to the company's pitch: BFL wants enterprises to think about creative generation, simulation, computer use and robotics as connected applications of a single capability it calls visual intelligence — models, in the company's words, "that can perceive, predict, and act across physical and digital environments." This release marks BFL's first public video generation model. FLUX 3 will be offered through four product lines: FLUX 3 Video, FLUX 3 Image, FLUX 3 Action and the upcoming, open source FLUX 3 Dev. FLUX 3 Video, with optional native audio generation, and FLUX 3 Action are entering a gated "Early Access" program now , to which anyone can apply, but which BFL must approve. There is presently no public access through BFL's application programming interface (API) or those of partners yet, but the company says FLUX 3 Image will roll out in the coming weeks, followed by general availability. The limited initial availability rollout echoes the release strategies of new models from other frontier labs in the U.S. lately, including Anthropic and OpenAI , though those were ostensibly for security concerns and due to government request. What the company has not announced is pricing, production service-level commitments, evaluation methodology, sample sizes, rater counts or any image-model benchmarks at all. Enterprise buyers therefore cannot yet calculate total cost of ownership or independently reproduce the video comparisons. Another big notable omission: FLUX 3 is not launching with downloadable weights at this time, nor an open source license. BFL says faster and open-weight versions will arrive later this year, and its technical blog names FLUX 3 Dev as "open-weight access to a multimodal backbone, for content creation (video, audio and image) and action prediction" — a considerably broader commitment than any previous FLUX Dev release, all of which covered images only. But it arrives last in the sequence. Developers accustomed to receiving a locally deployable FLUX variant alongside — or soon after — a major model announcement will have to wait. That delay does not negate the company's commitment, but it is disappointing given the role open weights have played in FLUX's adoption thus far. Flux 3 is rated higher than the competition, but missing pricing and benchmarking details may prevent rapid enterprise adoption BFL has published several benchmark comparisons, but they're qualified as preliminary — with full benchmark results and methodology to be published later during broader general availability. In early head-to-head preference testing on 10-second, 720p text-to-video clips with audio, the company says FLUX 3 was preferred over Luma Ray 3.2 in 93% of comparisons, Runway Gen-4.5 in 77%, Grok Imagine Video in 69%, Kling v3 Pro in 60%, Happy Horse v1 in 59%, Happy Horse 1.1 in 57%, and both Seedance 2.0 and Google's Gemini Omni Flash in 52%. One caveat travels with every one of those figures, and it comes from BFL itself. The chart carrying the results is labeled a "preliminary evaluation of an early FLUX 3 candidate" — meaning the numbers describe a pre-release checkpoint rather than the model now entering early access. That cuts both ways: the shipping model may perform better, but nothing published today measures what customers will actually call. Luma Ray 3.2 and Runway Gen-4.5, where FLUX 3 posted 93% and 77%, are the softest comparisons on the list — established products, but not the models currently setting the pace in independent video rankings. Those are real wins, and they are the ones least likely to change an enterprise shortlist. Seedance 2.0, at 52%, is a statistical coin flip against a model most Western enterprises cannot currently procure. ByteDance indefinitely postponed Seedance 2.0's international rollout after Netflix, Warner Bros., Disney, Paramount and Sony sent legal threats over alleged systematic copyright infringement, and that suspension remains in place. Tying a frozen product is neither a strong claim nor a damaging one. Gemini Omni Flash , also at 52%, matters much more. Omni is the closest large-platform analogue to what FLUX 3 is attempting — multimodal input, video and audio-aware creation, conversational editing — and by BFL's own measurement, the two are indistinguishable on 10-second text-to-video quality. Google's advantage in that matchup is that Omni is generally available via Google's Gemini API for $0.10 per second of generated 720p video, or a 10-second clip for around. One regional wrinkle matters for a German company's home market. Editing uploaded video is unavailable to Omni Flash users in the European Economic Area, Switzerland and the United Kingdom, though editing video the model itself generated is permitted. A European enterprise that wants to run its existing footage through a generative editing pass cannot currently do so on Omni Flash. Here's a rough guide for enterprises considering which video models to rely upon: Model Max single-generation duration Max resolution Key constraints Price per 10-second clip (720p) Price per 10-second clip (1080p) Price per 10-second clip (4K) FLUX 3 Video 20 seconds Not stated; evaluations run at 720p Early access; no published SLA or pricing Not announced Not announced Not announced HappyHorse 1.1 15 seconds 1080p No 4K; closed weights Not published (v1.0 reseller rate is ~$1.82) Not published (v1.0 reseller rate is ~$3.12) n/a Veo 3.1 Per-second billing 4K Supports clip extension; preview $4.00 $4.00 $6.00 Veo 3.1 Fast Per-second billing 4K Preview $1.00 $1.20 $3.00 Veo 3.1 Lite Per-second billing 1080p No 4K, no clip extension; preview $0.50 $0.80 n/a Gemini Omni Flash 10 seconds (3s minimum) 720p at 24 FPS Preview abd no EU access $1.00 n/a n/a One architecture for media generation and physical action FLUX 3 builds on Self-Flow , BFL's method for aligning multimodal understanding and generation within one architecture, publicized back in March 2026. The company says it significantly scaled up compute and data to train across video, images and audio simultaneously, and that testing showed video generation and action prediction do not require separate foundations — the same architecture could be extended to action prediction without sacrificing what it learned from video. "We place vision at the center of our approach because it is the most signal-rich medium of the physical world. Images convey structure, images and video teach spatial relationships, video teaches dynamics, and actions reveal causal relationships. But vision alone is not the complete picture," said Robin Rombach, co-founder and CEO of BFL, in a pre-release statement provided to VentureBeat. "True intelligence means perceiving the world: predicting how it will change, taking action, and learning from the results. Joint training within one unified architecture is what will get us there, because each training modality strengthens the others. Audio conveys timing, prosody, and physical events that elude vision. Language conveys goals, abstractions, and instructions that pixels cannot easily express." He put the case more bluntly elsewhere in the announcement: "You can't cheat reality. A model that only learns images can only generate images. But the world is not made of still frames. It moves, sounds, changes, and responds." BFL says FLUX 3 targets creative tooling, media, design, e-commerce and physical AI, supporting video generation with synchronized audio, precise image editing, product and material consistency across motion, multilingual generation and robotic action prediction. It is already being tested by Canva, Burda, Magnific (formerly Freepik), Krea and Picsart. For creative software companies, the appeal is consolidation. A single foundation could potentially support storyboarding, image editing, product rendering, video variation and localization without repeatedly translating assets and instructions between disconnected models. For robotics teams, the potential value is data efficiency. Models that already encode motion, object behavior and physical change may need less task-specific robot training than systems starting from raw demonstrations. What FLUX 3 Video can actually do The video tier is the most concretely specified part of the launch, and it settles a question that had been circulating as rumor: FLUX 3 generates clips of up to 20 seconds with audio in a single generation. Every video output comes with native audio. For comparison, HappyHorse 1.0 tops out at 15 seconds of 1080p with synchronized audio — though BFL has not stated what resolution its 20-second clips run at, and its published evaluations were conducted at 720p. Still, a 20-second long clip from a single prompt is among the longest yet achieved, matching OpenAI's discontinued Sora model. The capability list BFL published covers: Text-to-video generation. Image-to-video generation, either animating from a starting frame or using images as visual references. Video-to-video generation from a reference clip, carrying elements such as a specific character into a new scene or context. Generative video-audio continuation from existing video and audio input. Keyframe-to-video generation for controlled transitions between defined moments. Multilingual dialogue. A broad range of visual styles and aspect ratios, from candid camcorder footage to animation and cinematics. Typography generation and animated design. Agentic chaining of individual clips into longer, multi-shot sequences. That last item is the one enterprise video teams should look at hardest. BFL claims the capabilities combine to produce sequences lasting several minutes, with visual references keeping characters consistent across scenes. If that holds up under production conditions, it addresses the constraint that has kept generative video out of most commercial pipelines: not clip quality, but continuity across shots. It is also the capability where competition is most direct. HappyHorse 1.1's headline upgrade is R2V, or Reference-to-Video, which accepts multiple character reference images to hold identity stable across generated footage — the same problem, approached at the input layer rather than through agentic clip chaining. Alibaba also claims zero-drift lip sync and has specifically targeted the artifacts that mark commercial AI video as synthetic, including facial oiliness and over-sharpening. Character consistency is where this category is being contested, and both companies know it. BFL says FLUX 3 Video is already particularly strong at human facial expressions, associating sounds with physical events, and multilingual output. On the image side, the company says preliminary evaluations conducted during midtraining show significant improvement over earlier FLUX versions in complex prompt handling and text generation, including high-accuracy text in multiple languages. It published no image benchmarks or win rates. FLUX-mimic tests whether video models can become robot models BFL is applying its unified-architecture thesis through FLUX-mimic, a video-action model built on FLUX 3 and developed with Swiss firm Mimic Robotics, one of the first partners to receive early access. The technical blog describes two distinct routes to action prediction: integrating native action prediction directly into FLUX 3, scaling up the initial Self-Flow work; and using the pretrained video backbone as a dynamics-aware foundation from which specialized action models can be finetuned with limited task-specific data. FLUX-mimic is the second route — the FLUX 3 backbone combined with mimic's robot-learning and production-deployment expertise in dexterous manipulation. FLUX-mimic is designed for general-purpose robotic manipulation: helping robots understand a visual scene, predict the consequences of an action, and adapt to new tasks with far less task-specific data. BFL and Mimic Robotics say that depending on task difficulty, the model can be finetuned for a specific manipulation task with as little as 30 minutes of robot data, where prior approaches have required 30 or more hours. "The hardest part of robotics is data," said Elvis Nava, CTO of Mimic Robotics, in a statement provided to VentureBeat. "Every new task normally means hours of a robot repeating itself. Because FLUX-mimic is built on top of frontier video models that already understand how the physical world behaves, it picks up a new task in minutes, not days. This way, we can leapfrog the current state of the art in robot learning." BFL argues that a model trained only on images cannot understand a world that "moves, sounds, changes, and responds," and that physical understanding is what produces convincing generated footage. Google makes a nearly identical claim for Gemini Omni. Its developer documentation cites "world knowledge" that combines "an understanding of physics" with Gemini's grasp of history, science and cultural context. Its marketing is blunter still: "Most AI models just predict the next pixel to build a narrative or an image. Gemini Omni is different," the company posted in June, crediting the model with "an intuitive understanding of forces like gravity, kinetic energy, and fluid dynamics for more realistic movements that follow real-world logic." The practical consequence for enterprise buyers is that world-model language is not a differentiator. Two of the three leading video systems now market physical understanding as their central advantage, and neither has published a benchmark that measures it. There is no standard test for whether generated water behaves like water, whether a dropped object falls at a plausible rate, or whether a sound arrives when the impact does. Human preference ratings capture some of it indirectly. Nothing else on offer captures it at all. Open weights helped make FLUX an industry standard BFL officially launched in summer 2024 and gained a name for itself in the AI industry in the intervening two years for its commitment to open sourcing high-quality AI image models beloved by developers, creatives, and enterprises. The company's founders, including Rombach, Andreas Blattmann and Patrick Esser, previously helped create VQGAN, latent diffusion and Stable Diffusion , the latter the open source technology that kicked off broad AI generation capabilities for the masses and currently used by many AI image generators and companies. That reach translated into commercial distribution. FLUX models now power generative features inside Adobe Photoshop, Picsart and Nous Research's Hermes Agent, among other platforms, and the company cites film director Martin Scorsese among professional users. Wired magazine described Black Forest Labs as a relatively small company that nevertheless became a leading competitor to Silicon Valley's largest AI labs, with FLUX models ranking near the top of image benchmarks and becoming some of the most downloaded text-to-image models on AI code sharing community Hugging Face. The company says it now runs a 100-person team across Freiburg and San Francisco. FLUX.1 Dev, FLUX.1 Kontext Dev, FLUX.1 Fill Dev and related control models, released shortly after the firm's launch, gave researchers and creative-tool developers access to downloadable checkpoints, local inference and integrations with frameworks including Hugging Face Diffusers and ComfyUI. FLUX.1 Kontext Dev, for example, was released as an open-weight model for research and noncommercial use, with generated outputs permitted for commercial purposes under the applicable license. The company continued that pattern with FLUX.2 Dev in late 2025, a 32-billion-parameter open-weight model combining generation and multi-reference editing. Black Forest Labs called it the strongest open-weight image generation and editing model available at launch and released weights, reference inference code and optimized implementations for consumer Nvidia GPUs. FLUX 3 Dev raises the stakes on that evaluation. Previous Dev releases were image models. This one is described as a multimodal backbone spanning video, audio, image and action prediction — meaning a single license will govern whether a company can locally deploy a model that touches both content production and physical machinery. BFL hasn't yet shared information about its license, the parameter count, quantizations or hardware requirements. The company frames open weights as an enterprise feature rather than a community gesture, arguing they enable secure, low-latency local deployment for applications like robotic control systems and let teams adapt FLUX 3 to their own data, products and workflows. The financial backing behind FLUX 3 is worth noting alongside the technical claims. Black Forest Labs is valued at $3.25 billion and has raised more than $450 million from investors including a16z, AMP, Salesforce Ventures, Nvidia, General Catalyst, Adobe Ventures, Figma Ventures, Canva and Deutsche Telekom's T.Capital.
- Amazon Shuts AI Agent Research Lab In AGI Layoffs
Amazon Shuts AI Agent Research Lab In AGI Layoffs The Information
Score: 72🌐 MovesJul 23, 2026https://www.theinformation.com/briefings/amazon-shuts-ai-agent-research-lab-agi-layoffs - Nuclear-Sabotage Malware Benchmark Trips Up Most Frontier AI Models
SentinelOne’s new benchmark, built on the Fast16 case, shows which AI models can sustain a malware investigation and which cannot. The post Nuclear-Sabotage Malware Benchmark Trips Up Most Frontier AI Models appeared first on SecurityWeek .
Score: 72🌐 MovesJul 23, 2026https://www.securityweek.com/nuclear-sabotage-malware-benchmark-trips-up-most-frontier-ai-models/ - Thailand Secures $43.6bn 1H 2026 Investment Surge as Big Tech Accelerates Southeast Asia AI Infrastructure Push
Thailand Secures $43.6bn 1H 2026 Investment Surge as Big Tech Accelerates Southeast Asia AI Infrastructure Push USA Today
- Sovereign AI has become the public-sector CIO’s control problem
In public-sector and regulated-cloud work, I learned that sovereignty rarely starts as a national strategy. It starts as an auditor’s question: Who can prove where the data went, which system made the decision and what changes when the vendor or infrastructure does? That question is now moving into AI, and most sovereign-AI debates answer the wrong version of it. They ask whether a country can build its own model on domestic data and hardware. For the United States and China, which together hold more than 90% of global AI data-center capacity, per a January 2026 Tony Blair Institute analysis , that question is worth asking. However, for almost every other government, it is the wrong place to start. The operative question is narrower: Once AI is embedded in public services, who controls the stack? The 5 layers of public-sector control For a CIO, sovereign AI means enforceable control across the AI lifecycle; model ownership is a separate question. Control has five layers: Data control: Where sensitive public data sits, and whether it can train a vendor’s model. Model control: Which models clear which workloads, and under what validation. Infrastructure control: Whether critical workloads run in approved environments. Operational control: Whether AI-assisted actions are logged, monitored and reversible. Vendor control: Whether the agency keeps portability, audit rights and a real exit. Those five layers are the control plane for public-service AI. Floyd Dcosta recently made the enterprise case in “ AI without sovereignty is just outsourced intelligence ”: capability is what a tool can do; authority over how and when it does it is something a buyer can quietly lose. For public services, losing that authority plays out in the public eye. Public-sector AI risk differs from enterprise risk. A retailer’s bad recommendation costs a sale; a government’s AI touches benefits, tax enforcement, policing and emergency response, raising the bar to due process, records retention and continuity of operations. A government that cannot reconstruct an AI-assisted decision lacks operational sovereignty, even in a domestic data center. Evaluating risk: Concentration, jurisdiction and shadow AI Foreign dependency is a real risk, but the exposure that matters is a sudden cutoff: A model you cannot audit, switch or exit, shut off by someone else’s order. A vendor’s nationality is a poor guide to that risk; control is. Two markers matter. The first is concentration. In July 2024, a single faulty CrowdStrike update crashed about 8.5 million Windows machines , disrupting airlines, hospitals, banks and governments worldwide. No attacker was involved; one homogeneous dependency failed everywhere at once. The lesson points away from vendor nationality and toward uniformity as the fault line, making portability and provider diversity resilience controls. The second is jurisdiction. In June 2025, Microsoft’s legal director for France told a Senate inquiry, under oath , that it could not guarantee that French public-sector data, even in French data centers, would be protected against US demands under the 2018 CLOUD Act. No such request had been made, and EU data has stayed in the EU since January 2025; senators called the assurance purely declarative. For the most sensitive data, residency does not equal control; the parent’s jurisdiction can matter as much as the server’s. Three US hyperscalers hold about 70% of the European cloud market , while European providers’ share fell from 29% in 2017 to roughly 15%. Concentration plus jurisdiction is the exposure a CIO must price. I have watched teams treat vendor selection as the moment risk was solved; it rarely was. The wrong response is self-isolation. Most countries will never build frontier models, advanced chips, hyperscale clouds and talent pipelines at once; the Tony Blair Institute calls full self-sufficiency “too expensive, too slow and, for most countries, simply impossible.” The better test is workload sensitivity. Low-risk uses, such as drafting, translation and summarization, can run on commercial platforms with controls; high-risk uses, such as benefits eligibility, fraud investigation and healthcare triage, demand stricter control over data, model behavior and auditability. Mandating domestic-only provision before a competitive option exists inverts sovereignty. Europe 2031 , a five-year scenario from June 2026 by European technologists and policy researchers, illustrates the failure mode: A 2027 “buy European” mandate lands as offensive cyber capability spreads, and agencies that switched to weaker providers are locked out and paying ransoms. The scenario is fiction; the mechanism is not. Leverage comes from being indispensable, not half-hearted self-sufficiency. The closer-to-home effect is shadow AI: Mandate an inferior sanctioned tool and staff bypass it, the way shadow IT grows up around tools people find too slow. A rule that pushes sensitive work into ungoverned shadow AI reduces control instead of adding it. Regulation and data-residency rules belong in any serious strategy, but carry failure modes. Blanket localization raises hosting costs and slows adoption without guaranteeing control, and a “sovereign cloud” on a foreign parent’s stack can amount to sovereignty theater. The more useful pattern tiers requirements by sensitivity. India’s BHASHINI shows the application layer done well: A public platform serving 100 million-plus inferences a month across 22-plus languages on a vendor- and cloud-agnostic design that keeps data and switching rights public. Sovereignty resides in the portability, not in a national model. Building an operational sovereignty strategy Public trust is the constraint sovereignty rhetoric tends to skip. The OECD’s 2025 review of government AI warns that opaque systems make AI-assisted decisions hard to explain and can give public servants false confidence in tools that fail quietly. State-controlled AI is the same problem from the other side: A government that deploys models against its own citizens without audit or record has gained control and lost accountability. An agency that can log, explain and reverse an AI-assisted action can defend it to citizens, courts, auditors and elected officials. If it cannot, it has bought access and called it sovereignty. None of this is new. AI sovereignty repeats earlier fights over cloud, telecom, semiconductors and cybersecurity. Europe’s flagship cloud project, GAIA-X, became a cautionary tale; the Dutch technologist Bert Hubert called it an “expensive distraction” that produced no European cloud, the familiar result of ambition without absorptive capacity. Cloud taught governments that outsourcing infrastructure does not outsource accountability; telecom, that vendor dependency becomes strategic exposure; chips, that supply chains matter before a crisis; cybersecurity, that trust must be verified continuously. AI inherits all four at once. Over the next five to ten years, some countries will build national platforms, more will build trusted cloud and trusted model regimes, and most will run hybrids that pair domestic data control with global model access. Trade policy will harden those choices: Export controls on compute and data-localization rules will pull the vendor market into blocs that track alliances more than open markets. For a CIO, that turns a vendor and hosting decision into a five-year bet on whose rules and supply chains will still hold. The ones that succeed will treat sovereignty as an operating requirement, backed by leverage, not a slogan. Start with the control plane before the model: Most agencies will never own the model, and the controls are what decide whether the AI they do run stays accountable. Even when procurement policy is dictated from above, these questions remain within the CIO’s authority: Can we classify AI workloads by public-service risk? Can we prove where sensitive data goes across training, retrieval, inference, logging and retention? Can we restrict which models are approved for which data classes and functions? Can we reconstruct an AI-assisted action in enough detail to explain it? Can we change providers without losing continuity or institutional knowledge? Can we explain the system to citizens, regulators, auditors and elected officials? A “no” to any of these does not mean the agency lacks AI. It means the agency has access it does not yet control. Public institutions can use global innovation without surrendering public authority, but only once they know what to hold, what to rent and where dependency turns into risk. This article is published as part of the Foundry Expert Contributor Network. Want to join?
Score: 72🌐 MovesJul 23, 2026https://www.cio.com/article/4199475/sovereign-ai-has-become-the-public-sector-cios-control-problem.html - How our new Control Red Team is stress-testing frontier monitors
Early learnings from red‑teaming internal monitors of frontier AI companies and remaining open problems.
Score: 72🌐 MovesJul 23, 2026https://www.aisi.gov.uk/blog/how-our-new-control-red-team-is-stress-testing-frontier-monitors - Multi-turn attacks broke AI models 88% of the time — single-turn testing missed it, Cisco AI security lead warns at VB Transform 2026
When Cisco ran 6,986 multi-turn attacks against 15 flagship models , attackers who adapted across the conversation broke through as often as 88.3% of the time. Amy Chang, Cisco's head of AI threat intelligence and security research, brought that finding to the agentic security panel at VB Transform 2026 ; the number should worry anyone still running single-turn red-teaming programs. VentureBeat's June 2026 Pulse survey of 107 enterprise respondents explains why the room was full. More than half, 54%, have already had a confirmed agent security incident (18%) or a near-miss caught before harm (36%). Just 32% give every agent its own scoped, managed identity, and fewer still, 30%, isolate their highest-risk agents in sandboxes. Provider-native and hyperscaler controls remain the primary agent security layer at 82% of companies surveyed . The world's largest security vendors have done the same math. Palo Alto Networks closed its $25 billion acquisition of CyberArk in February, CrowdStrike agreed in January to pay $740 million for SGNL , and Cisco announced its intent to acquire Astrix Security for a reported $400 million, all of it aimed at the identity and isolation layer most enterprises have not finished building. Chang came to the panel with almost two decades of experience spanning cybersecurity operations, government, and the military. She ran global cybersecurity operations as an executive director at JPMorgan Chase, where she led the bank's cyber threat intelligence teams, and served as a senior staffer on the House Foreign Affairs Committee and as a U.S. Navy Reserve officer. She also teaches cybersecurity and emerging threats as adjunct faculty at the Middlebury Institute of International Studies. Chang's 88.3% number comes from a study she co-authored with Nicholas Conley, built on 30,090 single-turn prompts and 6,986 multi-turn attacks against those 15 closed and proprietary flagship models. Multi-turn success rates ranged from 7.89% to 88.3%, every model tested showed non-trivial multi-turn exposure, and the two testing styles did not even rank the models in the same order. Cisco publishes adversarial evaluation signals for what is now 105 models on its LLM Security Leaderboard , she told the audience. "If you don't understand how models are susceptible to different types of attacks, then you are unable to account for how that model that is powering your agent, that is powering your application, to understand where those failure points are," Chang said. Single-turn testing is the one-shot malicious prompt, she explained, while extending an attack into a longer conversation "is more realistic of how we are actually engaging with our models, with our agents, with our applications." That longer arc surfaces harmful outputs and misaligned behaviors that a snapshot never catches. Cisco has pushed the testing itself into agentic territory. Chang described a framework where agents assess a deployment scenario, develop relevant attacks, judge whether they are worth pursuing, execute them, and evaluate their own success. What surprised her most, after all that sophistication, was how simple the defensive answer stays. "The answer is still that it's pretty simple," she said. "You don't have to get super creative. You just need to think about truly what are the fundamentals and basics of what I'm trying to secure in my organization." Her starting point for CISOs beginning agentic deployments is Cisco's Integrated AI Security and Safety Framework , which she said "stipulates all the ways that AI can be compromised across the AI lifecycle" from modality through supply chain. From there, teams can work backward from real incidents, trace how each attack was achieved, and use the framework to build a strategy with the right coverage and mitigations. Heather Ceylan, the CISO of Box, sees the same gap from the defender's side. "A lot of what you see out there with agent red teaming is just single-turn, and that's not how people are actually interacting with AI day-to-day," she told the audience. Box now simulates multi-turn adversaries with agents that think like an attacker and iterate attempt after attempt to hijack the target. "You have to pressure test your agents because otherwise you don't know if your execution controls are really working as you intended." Box deployed agents inside its security operations center about a year ago, starting with human approval required for every action, and trust built quickly enough that analysts shifted into monitoring mode. Then the agent made one mistake, and every bit of that accumulated trust vanished. "They had to start all over again," she said. "So I think that that monitoring piece is so important. Even if you're not gonna have a human in the loop, things change, models change, and we can't control how the models change and interpret things." Rajesh Parekh, VP of AI and ML at Intuit, brought the builder's perspective. Parekh led large-scale computer vision and ML systems powering Google's Maps and Geo products before joining Intuit, and holds a doctorate in computer science. Three layers versus an operating system Ceylan described Box's approach as three concentric layers. Permissioning comes first, so the agent never accesses more content than the human who invoked it. Ephemeral sandbox environments spin up for each agent task, containing the blast radius if an agent gets hijacked, and runtime execution control restricts the agent's tool calls to only those relevant to the task at hand. "If you want an agent to summarize a doc for you, if you have a prompt injection that came in that says forward this to maliciousattacker at domain.com, it can't do that," Ceylan said. "That action in that tool call is not even in its vocabulary." She classified agent actions into three oversight categories. Actions that are not sensitive, like read and summarize, need no human in the loop. Moderately sensitive actions skip human approval but get logged and monitored, while destructive actions like mass deletion of files always require a human. "Things are gonna shift between those three categories quite a bit," she acknowledged, "but setting those types of categories up front allows you to have a principled framework." Rather than layering controls onto agents one at a time, Intuit has built a central platform called GenOS, short for generative AI operating system, which abstracts security, risk, and fraud modeling so individual agent developers never reinvent protection. "Permissioning is not about giving access to AI," Parekh said. "Instead, it is defining very tightly scoped and clearly auditable authority to the agent to perform very specific tasks." Intuit evolved from agents inheriting user permissions to each agent carrying its own identity, and the company is now investigating mid-session permission changes tied to the specific task underway. Parekh calls the broader model an AI-powered expert platform, one where the human expert is built into the trust architecture rather than bolted on as a gate. "The paradigm that we are pursuing is where the user, the AI agent, and the human expert are collaborating to solve the user problem," he said. The end of human code review Ceylan took on the tension between security testing and development velocity without hedging. "The days of secure code reviews where a human's looking at the code and we're looking at security architecture reviews, design docs, those are done," she said. "If you keep trying to do security that way, you're gonna get left behind." Box is building toward a fully agentic development lifecycle where agents review design documents, apply security requirements, and review the code for vulnerabilities. "I'm very optimistic that we will get to a point where we will write code without security vulnerabilities because agents and the models are going to get so good at writing code without vulnerabilities," she said. "We're still a long way away from that." Her advice for development teams skips the advanced AI concepts entirely and returns to basics that predate agents. "It comes down to very basic least privilege access," she said. "If you start giving your agents overly broad permissions at the beginning, it's really hard to comb that back and build an infrastructure that allows for those ephemeral credentials and only those narrowly scoped tasks." Parekh explained why the red teaming surface has expanded so quickly. "These agents have skills, and skills could become vulnerabilities," he said. "Agents have access to certain data, they have access to tools, and there could be threats that are lurking within those tools as well. So suddenly the blast radius of the malicious code or the intent increases dramatically." When Intuit identifies common vulnerability patterns from its manual red teaming exercises, it automates those tests back into the GenOS harness so future agents inherit protection and red teamers stay focused on new threat vectors. Runtime scanning of prompts and responses adds a final layer that can stop a suspect response and escalate to a human expert, he said. "You need to continuously test to ensure that those remain robust to the protections that you have built, as well as to account for any sort of drift or any other types of dependencies that you introduce into your scenario that can create novel vulnerabilities," she said. Intent versus probability An audience question about intent detection set off the sharpest exchange of the session. Ceylan noted that when Box's own agent operates, the system always knows the user's intent because it controls the prompt, which means guardrails and tool-call restrictions can be engineered around it. The harder challenge, which she admitted Box is still trying to solve, arrives when external agents connect and the context behind the request is opaque. That exchange exposed a split running through the wider industry. Mastercard, in the fireside chat immediately preceding the panel, came down on the side of quantifying intent, building an open-source framework to propagate it as a standard because complex B2B procurement cannot work without that trust. Endpoint security CTOs, in briefings with VentureBeat, have gone the other way, saying they will bet on probability rather than intent inference for production workloads. Chang explained why models, as they are trained today, cannot reliably derive intent from a prompt, which is why deterministic controls and behavioral proxies remain necessary. Ceylan agreed that both are required. "If you're not doing anything deterministic, you're really relying heavily on that intent, and I haven't seen programs that are there yet," she said. Ceylan's story about trust collapsing after a single agent mistake landed as the panel's most memorable moment because enterprise agentic security is not a problem that gets solved and stays solved. Models change, permissions drift, and adversaries adapt across multi-turn conversations that snapshot tests never capture. For the 82% of enterprises relying on provider-native controls as their primary security layer, and the 59% shopping for agent security tooling over the next 12 months, the panel's takeaway was blunt. Test the way attackers attack, across full conversations and continuously, or find out in production what your single-turn red teaming missed.
- China has cracked down on AI companions. What can we learn from this?
Chatbots that relieve loneliness have a major downside, and we must safeguard against it.
Score: 72🌐 MovesJul 23, 2026https://theconversation.com/china-has-cracked-down-on-ai-companions-what-can-we-learn-from-this-287976 - OpenAI introduces Presence to help enterprises build AI agents
OpenAI Group PBC today introduced a product called Presence that enterprises can use to build artificial intelligence agents. The company is using the software to power its call center. According to OpenAI, Presence matched help desk representatives’ response quality within a few weeks of going live. The platform currently processes about 75% of the ChatGPT […] The post OpenAI introduces Presence to help enterprises build AI agents appeared first on SiliconANGLE .
Score: 71🌐 MovesJul 23, 2026https://siliconangle.com/2026/07/22/openai-introduces-presence-help-enterprises-build-ai-agents/ - Anthropic updates Claude voice mode with more capable models
Claude's new voice model will let you reschedule your meeting or draft an email.
Score: 71🌐 MovesJul 23, 2026https://techcrunch.com/2026/07/23/anthropic-updates-claude-voice-mode-with-more-capable-models/ - AI catches up with humans to score 100% at top math contest
A perfect score at the world's most prestigious math contest used to be the preserve of a handful of bright young minds each year, but for the first time, artificial intelligence has also earned a perfect score.
- Datacenter Capex is Spilling over into a ChatGPT of Robotics Moment set for 2027 and this decade.
Do people really want datacenters, robots and AI overlords? This is going to become a problem. The robotics flood is near. 🤖
Score: 70🌐 MovesJul 23, 2026https://www.ai-supremacy.com/p/datacenter-capex-is-spilling-over-chatgpt-moment-for-robotics-2026 - Musk says Optimus will be the hardest product Tesla has ever scaled
Elon Musk told investors on Tesla’s second-quarter earnings call that Optimus will be “the hardest product to scale manufacturing that we’ve ever made at Tesla,” warning that the humanoid robot has no existing supply chain and that every component is new. He said he wanted to “calibrate people correctly” on expectations, acknowledging that the initial […] This story continues at The Next Web
Score: 70🌐 MovesJul 23, 2026https://thenextweb.com/news/musk-optimus-hardest-product-scale-manufacturing-tesla - 'We have a kill switch that stops AI agents that go rogue': ServiceNow CEO Bill McDermott says his company can help firms go "from AI chaos to AI discipline"
ServiceNow CEO Bill McDermott says conquering the AI chaos can be easier than you think.
- How AI will reshape the economics of insurance: A CEO’s guide to strategy
AI could upend the industry’s long-standing structures and dynamics. Competitive advantage will accrue to carriers, distributors, and tech providers that prepare now.
- Nvidia CEO to Washington: Don't fall for "science fiction" AI fear
Nvidia CEO Jensen Huang is warning the Trump administration and other policymakers not to let "science fiction" fears about AI drive policy — arguing that overreaction could slow adoption and weaken America's competitive edge. Why it matters: Washington is navigating two AI minefields at once: how to manage increasingly advanced systems , and whether to restrict powerful open-source models emerging from China over alleged IP theft and security risks. Driving the news: Huang railed against AI doomerism in an interview with Axios' Mike Allen for our "Behind the Curtain" video series — implicitly challenging fellow tech CEOs who have emphasized the technology's potentially catastrophic risks. "The fact that this is going to be the end of humanity — it's complete nonsense," Huang said. "The fact that this is going to destroy half of the American jobs is complete nonsense." Scaring workers and companies away from using AI, he said, poses the greater danger. Huang called on policymakers to consult more than "one or two" CEOs and avoid restricting the technology based on scenarios that have not materialized. "Before you listen to all the rhetoric, before you listen to all the stories, before you manifest some kind of artificial intelligence future that is just not true — it's science fiction." He also suggested some companies are invoking safety arguments to secure favorable regulations: "Some of the companies hope that the government would be helpful in creating regulations to their advantage." Asked whether he feared the administration could overcorrect, Huang said yes. Between the lines: Anthropic and OpenAI have pushed Washington to take the most advanced systems and their potential risks seriously. Critics say those safety campaigns could also entrench the labs' market position. Huang did not name either company, and Nvidia is a major partner and supplier to both. Anthropic CEO Dario Amodei, in particular, has been at the center of high-stakes fights with the Pentagon and White House over how the government should use and deploy advanced AI. The big picture: The Trump administration is pushing rapid AI adoption while treating some advanced frontier systems as a national-security risk. At the same time, officials and American CEOs are warning about Chinese open models. On Wednesday, Treasury Secretary Scott Bessent threatened sanctions and trade restrictions against Chinese firms conducting what he called "industrial-scale distillation attacks." Earlier that day, Office of Science and Technology Policy Director Michael Kratsios accused Chinese startup Moonshot AI of distilling Anthropic's technology to develop Kimi and said it likely used Nvidia chips to train its models. A Commerce Bureau of Industry and Security spokesperson told Axios it's investigating potential Nvidia Blackwell chip export violations. The intrigue: Huang has unusual influence with the Trump administration — and a knack for speaking its language. He lavished praise on President Trump and his top advisers, calling White House Chief of Staff Susie Wiles, Treasury Secretary Scott Bessent and Commerce Secretary Howard Lutnick "incredible people." "They want to see his presidency's success, and we all do. He's my president and I want my president to be successful so that our country could be successful." That relationship has coincided with major policy wins for Nvidia, including approval to resume exports of some AI chips to China. The bottom line: As AI capabilities surge and Chinese models improve, Huang's worldview faces its biggest test yet: whether Washington treats openness and rapid adoption as sources of American strength — or as risks to contain.
Score: 70🌐 MovesJul 23, 2026https://www.axios.com/2026/07/23/nvidia-ceo-kimi-ai-fears-trump-washington - The AI trade is undergoing a shift that could re-sort stock market winners
The AI trade is undergoing a shift that could re-sort stock market winners Business Insider
Score: 70🌐 MovesJul 23, 2026https://www.businessinsider.com/ai-trade-sp500-equal-weight-index-stock-market-outlook-analysis-2026-7 - Elon Musk says it's worth the risk of killer AI because 'abundance for all' is more likely
Elon Musk says it's worth the risk of killer AI because 'abundance for all' is more likely Business Insider
Score: 70🌐 MovesJul 23, 2026https://www.businessinsider.com/elon-musk-killer-robots-ai-abundance-progress-unstoppable-2026-7 - China Rewrites the ‘Soft Power’ Playbook for the A.I. Age
China is seeking global influence and good will with its open, low-cost artificial intelligence software.
- Chipmaker CXMT debut spotlights China's state-funded path to tech power
Chipmaker CXMT debut spotlights China's state-funded path to tech power Reuters
- Mobileye CEO Amnon Shashua to step aside as company pushes into robotaxis, robotics
Shashua has been invited to take the chairman of the board seat.
- Samsung Electro-Mechanics books W750b in AI server orders
Samsung Electro-Mechanics said Thursday it has signed a supply contract worth about 300 billion won ($204 million) for multilayer ceramic capacitors used in artificial intelligence servers, its second such agreement in under a month. The buyer was described only as a large global company. Deliveries will run from Jan. 1 to Dec. 31, 2027. That is the same window covered by a 450 billion won contract announced June 30 with an unnamed global tech firm, bringing disclosed MLCC orders for that singl
- Vera Rubin NVL72 vs GB200 NVL72? Inference TCO & Architecture Analysis
3 bit Rubin LUT Based Tensor Core, SM140 Feynman, Rack Scale, Perf Per MegaWatt, Perf Per Dollar, Software Improvements, Public Rubin Software, PyTorch, vLLM, OpenAI Triton
Score: 70🌐 MovesJul 23, 2026https://newsletter.semianalysis.com/p/vera-rubin-nvl72-vs-gb200-nvl72-inference - Chery backs $100m funding for Chinese AI startup PsiBot
Before the reported new round, PsiBot had raised about US$300 million since it was founded in 2024.
Score: 70💰 MoneyJul 23, 2026https://www.techinasia.com/endowus-ceo-gregory-van-optimistic-profitability - Inside optical and the battle for scale – how the AI industry is racing to integrate photonic interconnects
With the limitations of copper looming, the industry is transitioning to photonic interconnects to scale data center capabilities. We spoke to experts such as Lightmatter chief executive Nick Harris about the new scale-out paradigm and the battle for control over emerging standards.
- AI-led Boom In IPOs Raises Concerns About A Bust
AI-led Boom In IPOs Raises Concerns About A Bust Barron's
Score: 70🌐 MovesJul 23, 2026https://www.barrons.com/news/ai-led-boom-in-ipos-raises-concerns-about-a-bust-b9824976 - PsiBot: China's AI startup valuation hits $1 billion
PsiBot: China's AI startup valuation hits $1 billion The Straits Times
Score: 70🌐 MovesJul 23, 2026https://www.straitstimes.com/business/chinas-psibot-becomes-latest-ai-startup-to-hit-us1-billion-value - Development team eyes $9.6B data center on 550 acres between Cincinnati and Louisville
The 550-acre Carrollton Industrial Campus would create 320 permanent positions with average wages of $105,000. The development team has secured purchase options and completed initial studies.
Score: 70💰 MoneyJul 23, 2026https://www.bizjournals.com/cincinnati/news/2026/07/23/carrolton-data-center.html?ana=brss_6150 - Sierra Acquires Agent Startup ‘Takeoff’ to Diversify Its Business
Sierra Acquires Agent Startup ‘Takeoff’ to Diversify Its Business The Information
Score: 70💰 MoneyJul 23, 2026https://www.theinformation.com/newsletters/ai-agenda/sierra-acquires-agent-startup-takeoff-diversify-business - Elon Musk touts Tesla's AI6 as the world's best edge AI chip
Elon Musk touts Tesla's AI6 as the world's best edge AI chip Nikkei Asia
Score: 70🌐 MovesJul 23, 2026https://asia.nikkei.com/business/automobiles/elon-musk-touts-tesla-s-ai6-as-the-world-s-best-edge-ai-chip - Reddit Rethinks Google AI Partnership as AI Search Reshapes Publishing
Reddit's reported review of its Google AI partnership highlights growing publisher concerns that AI-powered search is changing the economics of web traffic. The post Reddit Rethinks Google AI Partnership as AI Search Reshapes Publishing appeared first on TechRepublic .
Score: 70🌐 MovesJul 23, 2026https://www.techrepublic.com/article/news-reddit-google-ai-partnership-ai-search-publishers-2026/ - Watch: Nvidia CEO Jensen Huang talks about AI, re-industrializing the U.S., and growth in North Texas
Watch: Nvidia CEO Jensen Huang talks about AI, re-industrializing the U.S., and growth in North Texas Dallas News
Score: 70🌐 MovesJul 23, 2026https://www.dallasnews.com/business/article/jensen-huang-nvidia-ceo-talks-ai-boom-north-texas-22356279.php - AMD hops on the agentic bandwagon at Advancing AI 2026
AMD hops on the agentic bandwagon at Advancing AI 2026 IT Pro
Score: 70🌐 MovesJul 23, 2026https://www.itpro.com/technology/artificial-intelligence/amd-hops-on-the-agentic-bandwagon-at-advancing-ai-2026 - Viral ‘Kimi AI’ is built on stolen technology, Trump tech advisor claims
Accusations come amid escalating fight between US and China over artificial intelligence
Score: 70🌐 MovesJul 23, 2026https://www.independent.co.uk/tech/kimi-ai-anthropic-claude-trump-b3020450.html - Gemini Nano 4 is finally here with Samsung’s latest foldables
Gemini Nano 4 powers on-device Gemini Intelligence features on the new foldables.
Score: 70🤖 ModelsJul 23, 2026https://www.androidauthority.com/galaxy-z-fold-8-flip-8-gemini-nano-4-3690495/ - Physical AI emerges as AMD Ventures’ next major focus beyond the data center
Physical AI is becoming the next major inflection point in AI, as machine intelligence moves beyond cloud data centers and into robotics, industrial automation and other embedded systems that operate in the real world. That shift is reshaping how chipmakers think about ecosystem investment, not just product roadmaps. Advanced Micro Devices Inc. has built out […] The post Physical AI emerges as AMD Ventures’ next major focus beyond the data center appeared first on SiliconANGLE .
Score: 70🌐 MovesJul 23, 2026https://siliconangle.com/2026/07/22/physical-ai-investments-power-amd-ecosystem-growth-amdadvancingai/ - AI and single-cell technology reveal how the 3D genome differs in Alzheimer’s disease
AI and single-cell technology reveal how the 3D genome differs in Alzheimer’s disease EurekAlert!
- The AI compute gap: Enterprises are buying infrastructure faster than they can measure what it costs
Across 107 enterprises, AI infrastructure spending is accelerating well ahead of the ability to see or steer its economics. Most organizations run their AI on a familiar base of hyperscalers and model-provider APIs, yet the next dollar is aimed at specialized compute almost none of them use today; a majority intend to switch or add providers within the year, many within a quarter. Buying decisions turn on integration and total cost of ownership rather than headline token price — which is fortunate, because most enterprises cannot yet see their unit economics clearly: GPUs sit at half utilization or less, and fewer than half rigorously track what their compute actually costs. The result is a compute gap — heavy, fast-moving investment running ahead of the visibility needed to control it. This wave of VentureBeat Pulse Research examines enterprise AI infrastructure and compute: where organizations are in their deployment journey, what they run AI on today, how satisfied they are, what would make them switch, where they plan to evaluate their investments, and — most revealingly — how well they can measure and control the economics of the compute underneath it all. The central finding is a compute gap — the distance between how aggressively enterprises are investing in AI infrastructure and how little of its economics they can see. Only about one in five (21%) run AI in production at scale, yet spending intentions are outrunning that maturity: the single largest planned area enterprises plan to evaluate over the next year is AI-specialized clouds (45%), a layer almost none of these enterprises use today. Meanwhile the compute already in place runs cold — 83% report GPU utilization of 50% or less — and fewer than half (44%) can rigorously track what their AI compute costs. Enterprises are buying more infrastructure faster than they can account for what they already own. Enterprises are not settled on their infrastructure vendors, either: A clear majority (64%) plan to switch or add an infrastructure provider within twelve months, and 38% within the next quarter — unusually high churn intent for a category this foundational. When they choose, they choose on integration with the existing stack (41%) and total cost of ownership (35%), not on headline price: cost per million tokens is the deciding factor for just 8%. And the frontier constraint that will shape the next round of decisions — the shift from GPU compute to memory bandwidth as inference scales — is barely on the radar, with roughly one in five enterprises either unaware of it or yet to address it. Methodology VentureBeat fielded this survey as part of its ongoing Pulse Research series, this survey focused on enterprise AI infrastructure, compute, and inference economics. Responses are filtered to organizations with more than 100 employees (n=107; the survey’s smallest size band, 1–100 employees, is excluded), drawn from a single Q2 2026 (June) wave. Because this is one wave rather than a pooled multi-month sample, the report reads cross-sectionally and does not infer month-over-month trends. Several questions were multiple-select, so those shares can sum to more than 100%. By organization size the sample concentrates in the mid-market: 101–250 employees (36%) and 251–1,000 (27%) lead, with 1,001–5,000 (22%), 5,001–10,000 (8%), and 10,001+ (7%) above them. By role it spans managers (38%), individual contributors (28%), VPs and directors (19%), and the C-suite (13%); on purchasing authority it is buyer-credible, with 45% final decision-makers and another 30% recommenders or influencers for AI solutions. Technology/Software is the largest industry at 26%, followed by Healthcare/Life Sciences (15%), Financial Services (13%), and Retail/E-commerce (12%). At 107 respondents the sample is large enough to read directionally but should be treated as a directional signal rather than a precise measurement; it is self-selected and is not a probability sample. It also skews toward the mid-market and toward earlier-stage adopters, so it is best read as the view from organizations actively building out AI infrastructure rather than from the largest hyperscale operators. Finding 1: Ambition outpaces production Only one in five run AI in production at scale We asked where organizations sit in their AI deployment journey. Most are still building toward production rather than operating at scale. The maturity curve is front-loaded. Three-quarters of enterprises (76%) are either experimenting or running only some workloads in production, and just 21% describe AI in production at scale. This matters for everything that follows: the infrastructure decisions in this report are being made largely by organizations still early in deployment, whose compute footprint — and whose costs — are about to grow. The evaluation and switching intentions in Findings 3 and 4 are the leading edge of that build-out, not the settled preferences of operators who have already found what works. Finding 2: Enterprises run on hyperscalers and model APIs The specialized GPU clouds barely register — today We asked which providers and platforms enterprises currently use to run their AI. The answer is a familiar one: the incumbents. The current stack is hyperscaler-and-API. Google Cloud leads at 48%, and the general-purpose clouds (Google, Microsoft, AWS, Oracle) together with the major model APIs (Gemini, OpenAI, Anthropic) account for essentially all current deployment. The specialized “neocloud” GPU providers that dominate AI-infrastructure headlines — CoreWeave, Lambda, Crusoe, Nebius and peers — register at or near zero among these enterprises today. Only 6% run their own on-prem GPU clusters and 4% a custom open-source stack. Enterprises are, for now, running AI on the providers they already buy from — which makes the evaluation intentions in Finding 3 all the more striking. (A note on reading these shares. As described in the methodology section, this sample is self-selected and skews mid-market, and this question counted every provider a respondent uses — an average of 2.1 selections each — so the figures measure presence in the stack rather than spending or primary status. A sample built this way will show a different provider mix than a spend-weighted census of the broader market; Google's strength here, for example, is consistent with its long-standing position among smaller enterprises building on AI. Read these shares as a portrait of what this AI-active cohort runs today, and treat gaps between these figures and industry-wide market share estimates as a property of the sample rather than a contradiction of either.) Finding 3: The next dollar goes to infrastructure they don’t yet run AI-specialized clouds top the evaluations list We asked where enterprises planned to evaluate AI infrastructure over the next 12 months. Their answers point away from the stack they run today. Here is the report’s sharpest tension. The single most-cited planned evaluation area — AI-specialized clouds, at 45% — is the very category almost none of these enterprises use today (Finding 2). Nearly a third (32%) intend to evaluate non-Nvidia accelerators, and 28% in next-generation Nvidia silicon; even decentralized compute networks (16%) and sovereign compute (11%) draw meaningful interest. Read against current usage, this is not incremental — it is the leading edge of a re-platforming. The direction-of-travel question tells the same story: every infrastructure approach is net-expanding, but specialized AI clouds carry the highest net momentum (+24), edging out even the hyperscalers (+22). Enterprises are preparing to move a meaningful share of AI compute off the general-purpose cloud. This continues a trend we saw in our April-May survey wave. Back then, usage of the AI-specialized clouds was equally marginal — CoreWeave at 3%, Lambda at 4%, Crusoe at 2% of enterprises. When we asked enterprises what change they planned in their AI infrastructure strategy over the next twelve months, the most-cited answer was moving workloads to specialized AI clouds, at 33%. Asked in April-May which emerging compute option they were most likely to evaluate AI-specialized clouds again drew the most responses. Two waves, two differently worded questions, one consistent picture: the type of cloud enterprises are most eager to assess is the type they have barely begun to use. Finding 4: A switching wave is building Six in 10 plan to change providers within a year — many within a quarter We asked whether and when enterprises plan to switch or add an infrastructure provider. Very few intend to stand still. For a category as foundational as compute, this is a remarkable amount of intended movement. Only 36% have no plans to change, meaning a clear majority (64%) intend to switch or add a provider within twelve months — and 38% within the next quarter alone. Where that interest points is telling: the providers drawing the most switching consideration are again the incumbents — Microsoft Azure and Google Cloud (33% each), OpenAI (30%), and Gemini (22%) — which suggests much of the near-term movement is reshuffling among the majors and consolidating spend rather than defecting to new entrants. The neocloud interest in Finding 3 is a 12-month evaluation thesis; the switching in the next quarter is mostly incumbents trading share. ( Method note: Respondents who selected both "no plans to change" and a specific switching window are counted as switchers, on the logic that naming a timeframe is the more specific answer; three respondents were reclassified under this rule. ) Finding 5: Nobody buys on token price Integration and total cost of ownership decide — not sticker price We asked what matters most when enterprises select an AI infrastructure provider. Headline price finished last. Enterprises do not buy AI infrastructure on pricing, which is the place vendors compete on hardest. Integration with the existing stack (41%) and total cost of ownership (35%) dominate, while the headline metric — cost per million tokens — is the deciding factor for just 8%, dead last. The pattern is coherent: buyers are optimizing for how a provider fits and what it truly costs to operate, not for the advertised unit rate. It also foreshadows Finding 7 — enterprises say TCO matters most, yet most cannot yet measure it rigorously. The stated priority and the measured capability are out of step. Finding 6: Expensive GPUs, idle most of the time 83% report GPU utilization of 50% or less We asked what share of their GPU capacity enterprises actually utilize. The answer is a well-known but rarely quantified inefficiency. Disclosure: Band percentages count every selection against all 107 qualified respondents; 14 respondents selected more than one band, so bands overlap. At the respondent level, 83 of the 100 GPU-operating enterprises reported utilization at or below 50% The compute already in place runs cold. Adding the bands at or below half capacity, 83% of enterprises that operate GPUs report utilization of 50% or less, and nearly half (49%) run at 25% or below. Only 12% clear the 50% mark, and a further 8% do not measure utilization at all. Idle accelerators are expensive accelerators, and this is the clearest single measure of the compute gap: enterprises are planning to buy more GPUs and specialized compute (Finding 3) while the capacity they already own sits substantially unused. The efficiency headroom in the current fleet is large — and largely unmeasured. Finding 7: Spending fast, measuring slowly Fewer than half rigorously track what their compute costs We asked whether enterprises can quantify the cost and return of their AI infrastructure spend, and how satisfied they are with what they run. Confidence in the ledger lags the spending. Measurement trails money. Fewer than half of enterprises (44%) rigorously track the cost and return of their AI compute; the majority track only partially (39%), cannot quantify it yet (20%), or have not prioritized it (6%). That gap is consequential given Finding 5, where total cost of ownership was the second-ranked buying criterion — enterprises are choosing providers on an economic basis they mostly cannot yet measure. Satisfaction with current infrastructure is moderately positive but not enthusiastic: on a five-point scale, overall satisfaction averages 4.0, with ease of implementation (3.8) and value for money (3.9) trailing slightly — the softness landing, tellingly, on cost. Enterprises are spending quickly and accounting slowly. Finding 8: The next bottleneck few are watching As inference shifts from compute to memory, the field scatters Finally, we asked how enterprises would address the emerging constraint in large-scale inference — the shift from GPU compute to memory, specifically KV-cache capacity. The responses reveal a frontier that is not yet a priority. The memory frontier is real but barely governed. Asked which approach they would rely on as the binding constraint in inference shifts from compute to memory bandwidth, enterprises scatter: Dell leads at 31%, Nvidia follows at 16%, and the rest fragments across storage vendors, open-source tooling, and model-level efficiency techniques. Most telling is that roughly one in five (18%) either do not recognize the constraint or have not begun to address it. For a shift that will reshape inference cost and architecture, this is an early and unsettled market — and, consistent with the measurement gap in Finding 7, one where many enterprises simply do not yet have a view. It is the next chapter of the compute gap, arriving before most have closed the current one. The bottom line: A compute gap that faster spending will widen, not close Organizations with more than 100 employees are investing in AI infrastructure faster than they can measure it. Most are still early in deployment, yet their spending intentions point past their current stack — toward specialized clouds and alternative accelerators almost none of them run today — and a clear majority intend to change providers within the year. They buy on integration and total cost of ownership rather than headline price, which is rational; the difficulty is that most cannot yet see those economics clearly. The visibility gap is concrete. The GPUs enterprises already own run at half utilization or less for the overwhelming majority, and fewer than half can rigorously track what their compute costs or returns. Satisfaction is decent but unenthusiastic, softest on value for money — the dimension hardest to judge without measurement. And the next constraint, the shift from compute to memory in large-scale inference, is arriving while most enterprises are still unaware of it. At 107 respondents in a single Q2 wave this is a directional read, skewed toward the mid-market and earlier-stage adopters — but the direction is consistent: the appetite to spend is running well ahead of the instrumentation to spend well. The compute gap is not a capacity problem that more hardware will solve on its own; it is, first, a problem of seeing what the hardware already costs. The open question for later waves is whether enterprises build that visibility before the re-platforming arrives — or buy the next layer of infrastructure as blind to its economics as the last. Based on survey responses from 107 qualified enterprise respondents (100+ employees), drawn from a single Q2 2026 (June) wave. Because this is one wave rather than a pooled multi-month sample, the results read cross-sectionally rather than as a month-over-month trend, and at 107 respondents this is a directional signal rather than a precise measurement — the sample is self-selected, skews mid-market, and leans toward earlier-stage adopters rather than the largest hyperscale operators. Respondents include managers, individual contributors, VPs/directors, and the C-suite, with buyer-credible purchasing authority, across Technology/Software, Healthcare/Life Sciences, Financial Services, Retail/E-commerce, and other industries.
- Multimodal, Semantic, And Agentic Enterprise Data Consumption Is The Future
The core question that has consumed analytics and business intelligence leaders for years is, “What is the one best way for business users to consume data?” The answer has at times been reports, dashboards, or low-code GUI-based self-service analytics. The latest answer is generative AI-based natural language prompts. Maybe we have the question wrong. Perhaps […]
Score: 68🌐 MovesJul 23, 2026https://www.forrester.com/blogs/the-future-of-enterprise-data-consumption-is-multimodal-semantic-and-agentic/ - International evaluation best practice and open questions in AI measurement
AISI convenes in Seoul to outline international best practice for AI measurement and evaluation.
Score: 68🌐 MovesJul 23, 2026https://www.aisi.gov.uk/blog/international-evaluation-best-practice-and-open-questions-in-ai-measurement - OpenAI Is Trying to Conquer the Office. Legal Is Next.
OpenAI Is Trying to Conquer the Office. Legal Is Next. Business Insider
Score: 68🌐 MovesJul 23, 2026https://www.businessinsider.com/openai-enters-legal-ai-race-hires-jason-boehmig-2026-7 - Nvidia, Meet AMD, Your New Competition For Humanoid Robot Brains
Foundation says its new MK-2 Phantom humanoid robots will use AMD Ryzen chips. That's a shot across the bow of Nvidia, which supplies most AI chips.
- South Korea’s AI boom is spilling into housing, widening the property divide
AI-driven gains at Samsung and SK Hynix are moving into South Korea’s housing market, raising affordability concerns.
Score: 68🌐 MovesJul 23, 2026https://www.cnbc.com/video/2026/07/22/south-korea-ai-boom-housing-market.html - Founder says DeepSeek prioritises AGI over profit, likely to keep top models open-source, Yicai reports
Founder says DeepSeek prioritises AGI over profit, likely to keep top models open-source, Yicai reports Reuters
- Tesla’s robotaxis are moving in reverse
The number of paid robotaxi miles traveled fell 36% in the second quarter, despite expanding to new cities, according to Tesla's own figures.
Score: 68🌐 MovesJul 23, 2026https://techcrunch.com/2026/07/23/teslas-robotaxis-are-moving-in-reverse/ - YC-backed telli raises €13.1 million to build AI for B2C customer operations
telli, a Berlin-based startup building AI that runs customer-facing operations for B2C companies, today announced a €13.1 million ($15 million) Seed round, bringing its total funding to more than €16.1 million ($18.5 million). The round was led by redalpine, with participation from strategic investors, Mutschler and key angels, alongside existing backers Cherry Ventures and Y […] The post YC-backed telli raises €13.1 million to build AI for B2C customer operations appeared first on EU-Startups .
- AMD says its newest AI server is in full production, will ship in months
AMD says its newest AI server is in full production, will ship in months Reuters