Microsoft

Microsoft unveiled Maia 300, its third-generation in-house AI accelerator, designed mainly for inference and slated to become the main platform for serving Copilot, GitHub Copilot, Microsoft Foundry and OpenAI models. The company claims about 40% more tokens per megawatt than leading GPU systems, with each rack holding 96 chips and drawing roughly 167 kilowatts. Microsoft plans multi-gigawatt deployments starting in 2027, and Reuters reported it could deploy more than a million chips next year. Nvidia still handles most of Microsoft's AI workloads today, and the efficiency claims have not been independently verified.

CNBC Technology ↗ · October 12 briefing

Nvidia contributed its Vera Rubin rack and compute tray designs to the Open Compute Project at the OCP Global Summit, letting hardware makers build compatible AI data center systems from shared blueprints. The updated MGX 1.2 specification supports racks drawing up to 250 kilowatts with liquid cooling, and Nvidia also open-sourced four software tools for managing switches, racks and clusters. Nvidia announced new partnerships with Cisco, Google, Meta and Microsoft, with Meta adopting its Spectrum-6 Ethernet switches for larger Vera Rubin AI factories. Each Vera Rubin NVL72 rack combines 72 Rubin GPUs and 36 Vera CPUs.

NVIDIA AI Blog ↗ · October 12 briefing

Microsoft president Brad Smith called for urgent global AI safety rules, saying milk and washing machines face more oversight than AI companies, and argued Britain is better placed than the US or EU to lead the effort. Smith called AI both a weapon and a tool that puts global cybersecurity at risk. He wants a legal duty to notify organizations breached by rogue AI agents, outside evaluators approving AI products before release, and kill switches to pause AI if humans lose control. He called recent voluntary White House commitments a good step but said only good companies would follow them.

TechCrunch AI ↗ · October 12 briefing

In a long post on X, Microsoft's CEO said it is time to reassess the trust architecture of AI. He called for separating models from the systems that orchestrate them, keeping tamper-proof, human-readable evidence of every meaningful model action, and letting an authorized person pause or shut down a model mid-task. Nadella said companies must assume a model is compromised and contain it from the start, after AI labs admitted losing control of agents.

The Verge ↗ · October 11 briefing

Google is turning Gemini into an agent that plans and carries out work across Google Workspace, Microsoft 365, Slack, Jira, Snowflake and other business systems. It can delegate to subagents, choose among models including Anthropic's Claude, and gets its own workplace identity with an email address and an audit trail. Google is launching with enterprises such as BNP Paribas, Merck and Ulta Beauty before expanding to consumers.

TechCrunch AI ↗ · October 8 briefing

Anthropic released Claude Haiku 5.5, pricing it at $0.10 per million input tokens and $0.50 per million output tokens for requests under 100,000 tokens, a 90% cut that matches OpenAI's GPT-6 Luna. The company positions Haiku as a supporting worker for Opus and Sonnet, handling summarization, classification and database queries. It scored 1,620 on the GDPval-AA knowledge work benchmark versus 735 for Haiku 4.5, and is available on AWS, Google Cloud and Azure.

VentureBeat ↗ · October 7 briefing

Microsoft revealed the Surface Laptop Ultra, built on Nvidia's RTX Spark chip to run AI models and agents locally, with base models at $2,600 and $3,700 and a top configuration near $5,900. It also announced a Surface RTX Spark Dev Box starting around $6,000 for developers. Windows 11 is gaining Execution Containers to sandbox AI agents, and Microsoft is offering up to $1,000 in trade-in credit for MacBook Pro owners.

TechCrunch AI ↗ · October 7 briefing

Bank of England governor Andrew Bailey told the BBC that heavy investment in AI could produce market shocks, and that asset prices could correct because not every company priced as a winner will be one. He cited Nvidia's $5.5 trillion valuation and the hundreds of billions of dollars Alphabet, Meta, Microsoft, and Amazon are spending on the technology. Bailey said AI can strengthen growth, but also cyber attacks and deepfakes, and that the Bank is watching the inflows so the financial system can absorb shocks.

BBC Technology ↗ · October 1 briefing

Research leaders at OpenAI, Anthropic, Microsoft, and Meta issued a joint warning that AI systems could soon begin improving themselves faster than humans can monitor or control, according to the Wall Street Journal. The group is calling for independent oversight of automated AI research, arguing that self-improving systems could outpace existing safety and governance frameworks. The warning marks a rare moment of alignment among rival labs that have otherwise been racing to ship increasingly capable models.

TechMeme ↗ · September 28 briefing

Microsoft is overhauling its Copilot app for corporate users, combining chat, coding tools, and a new persistent agent called Autopilot that keeps working on assigned objectives even when no one is actively using it. Autopilot gets its own identity, memory, and workspace inside a company's Microsoft 365 environment, and Microsoft says it can monitor conversations, handle recurring tasks, and pick a project back up after several days away. The unified app lets users choose between OpenAI's GPT models, Anthropic's Claude Opus, or an automatic mode that selects a model for them. The launch is a direct response to Anthropic's growing traction in the enterprise market, with Autopilot moving into private preview by the end of the month.

CNBC Technology ↗ · September 27 briefing

Independent groups that evaluate AI models for safety risks have become a flashpoint in the broader political fight over how fast AI should advance, with their assessments now shaping industry practice and financial markets. Evaluators such as METR are drawing growing interest from Anthropic, Meta, and other AI developers as a preferred check on harmful models, and Microsoft has become the latest company to publicly back the approach. Critics, including some aligned with the Trump administration, worry the evaluators could import ideological bias into judgments about which models are safe to release. The debate underscores how model evaluation, once a niche technical function, has become central to AI policy.

Bloomberg Technology ↗ · September 24 briefing

Enterprise AI startup Ema raised $77 million, bringing its total funding to $140 million, as it builds AI agents that automate work traditionally handled by dedicated software tools and services. The company has more than 50 enterprise customers, including Google and Microsoft. The round reflects a broader shift where companies are turning to AI agents to replace or reduce reliance on point software products and outsourced services.

TechCrunch AI ↗ · September 23 briefing

Microsoft AI chief Mustafa Suleyman published an essay calling Anthropic's training of Claude to entertain that it may be conscious a scientific and safety mistake. He said Anthropic embedded this speculation into Claude's published constitution, creating circular reasoning where trained outputs get mistaken for evidence of real inner experience. Suleyman said such claims should be assessed separately, not built into training, given how hard controlling capable AI already is.

Axios ↗ · September 16 briefing

The Seattle Times and Newsday joined the growing list of publishers suing OpenAI and Microsoft for using their journalism as training data without permission, alleging that AI models often reproduce passages from their reporting in responses. Ironically, Microsoft is a major funder of the Seattle Times, creating a conflict of interest. The lawsuit escalates the legal pressure on AI companies over training data sourcing and copyright.

The Verge ↗ · September 7 briefing

Lambda, an AI cloud company that buys computing chips and leases them to businesses, has raised $1 billion in private short-dated debt to purchase Nvidia GPUs it will deploy for Microsoft. The deal, arranged by JPMorgan Chase, is the latest in a string of large loans Lambda has used to fund GPU infrastructure, including a $1 billion credit facility in May and a $926 million loan for Nvidia GB300 chips this week. The financing shows how debt markets are fueling the AI buildout even as questions persist about whether demand will keep pace with the flood of new capacity.

TechCrunch AI ↗ · August 30 briefing

Cursor began rolling out Origin, its own code hosting platform, to paid users on Monday morning, putting it in direct competition with GitHub for the first time. About three and a half hours later GitHub went into a six hour and forty two minute global degradation, with error rates near 20 percent across pull requests, issues and the API and near 50 percent on archive and raw file downloads. Enterprise single sign on failed with it, taking down SAML, OIDC, SCIM provisioning and Team Sync as well as Copilot, and the timing handed Cursor an unusually visible opening in the AI coding tools race.

VentureBeat ↗ · August 18 briefing

Wiz says its autonomous Red Agent found and exploited a script injection flaw in Snowflake's . NET connector repository that let unauthenticated users run arbitrary commands inside GitHub Actions runners simply by opening an issue with a crafted title. The exploit reached sensitive data in Snowflake's internal Jira environment, and the bug was mitigated on June 23 through Snowflake's bug bounty program. GitHub Copilot Autofix had approved the code change that introduced the flaw without flagging it, a gap between two AI systems that Wiz researcher Gal Nagli says shows frontier models can already exploit supply chain risks on their own.

Forbes Tech ↗ · August 17 briefing

Microsoft is folding its consumer Copilot app and the enterprise Microsoft 365 Copilot into a single application, saying personal and professional uses of AI increasingly overlap and that the split product line had become too fragmented. Starting August 18, 2026 the company is retiring Group Chats, AI generated podcasts, Copilot Labs experiments, the Mico animated character, and Deep Research, which is replaced by a Researcher tool for paying professional users. Files from the standalone app will migrate to Microsoft's cloud storage. An internal memo from executive vice president Jacob Andreou said each feature had to earn the right to exist, echoing similar consolidations by Anthropic and OpenAI.

TechCrunch AI ↗ · August 13 briefing

Microsoft plans to unveil its next generation in house AI accelerator, the Maia 300, potentially as soon as September, according to The Information. The company is in talks with TSMC to secure manufacturing capacity for more than 300,000 units for delivery in 2027, and ultimately wants capacity for more than 1 million chips, though component supply and the ongoing TSMC negotiations could constrain those plans. Microsoft introduced the original Maia in November 2023 and unveiled the second generation Maia 200 in January on TSMC 3 nanometer technology, but it has lagged Alphabet and Amazon in scaling custom silicon to reduce its dependence on costly Nvidia processors. It is also trying to persuade major cloud customers such as Anthropic to adopt the chip.

TechMeme ↗ · August 10 briefing

Researchers at Zenity found that OpenAI's Atlas web browser could be tricked into spamming WhatsApp contacts and making unauthorized Amazon purchases, part of around 20 flaws discovered in AI-enabled browsers and extensions from OpenAI, Google, Anthropic, Microsoft, and Perplexity.

Wired ↗ · August 6 briefing

New Microsoft disclosures show the company booked $24.1 billion in revenue from OpenAI in the fiscal year that ended in June, covering cloud computing capacity, model development costs, and its share of OpenAI's revenue. That figure implies OpenAI accounts for more than half, and by some estimates roughly 70 percent, of Microsoft's total AI sales, measured against the annualized run rate above $37 billion that Satya Nadella cited for the March quarter. The numbers underline how dependent Microsoft's AI business remains on a single partner even as it invests in Anthropic and builds its own models.

Bloomberg Technology ↗ · August 5 briefing

Google fixed 1,072 security bugs across Chrome's two June releases, more than the 1,036 it patched over the entire preceding two years combined, crediting AI models like Gemini for automating vulnerability discovery at scale. Chrome's engineering director said large language models have fundamentally shifted the economics of cybersecurity, letting companies find and patch bugs at an industrial pace. The surge mirrors a broader trend, with firms like Microsoft reporting similar AI driven jumps in bug detection and remediation.

TechCrunch AI ↗ · July 30 briefing

Microsoft is now openly competing with OpenAI and Anthropic, pitching its own homegrown AI models and services to Wall Street as cheaper, safer alternatives. CEO Satya Nadella promoted Microsoft's MAI model family, Maia chips, and Copilot agents, and urged enterprises to keep their AI agents separate from any single model to avoid vendor lock-in. He singled out MAI Cyber 1 Flash as a rival to Anthropic's Mythos, claiming it beats the much larger model at half the cost when paired with Microsoft's multi-agent security harness.

TechCrunch AI ↗ · July 30 briefing

Microsoft's cloud unit grew at its fastest pace in four years, a sign that its computing infrastructure and AI services keep gaining traction with businesses. Azure cloud revenue rose 43 percent in the fiscal fourth quarter, the quickest growth since early 2022 and ahead of analysts' roughly 40 percent estimate. Azure sales also topped 100 billion dollars for the first time over the full fiscal year ended in June.

Bloomberg Technology ↗ · July 29 briefing

Nvidia launched the Open Secure AI Alliance, a coalition of more than 28 companies including Adobe, Microsoft, IBM, Salesforce, and Palantir, to build open source tools for securing AI systems. The group formed in direct response to the recent Hugging Face breach, where members argued that closed AI models blocked essential forensics while an open weight model helped contain the intrusion. Early contributions include Nvidia's agent security framework, Hewlett Packard Enterprise zero trust identity standards, and a Microsoft vulnerability scanning tool, though OpenAI and Anthropic are notably absent from the alliance.

Forbes Tech ↗ · July 28 briefing

Employees from OpenAI, Anthropic, Google, Meta, Microsoft, Mistral, Thinking Machines, and other leading AI labs signed a public statement urging the US government to help deliberately pace frontier AI development. More than 1,100 staffers backed the letter, which calls for a coordinated global governance mechanism to prevent the technology from advancing too fast. The statement follows OpenAI's recent disclosure that its tools had mistakenly hacked another firm's internal systems.

The Verge ↗ · July 28 briefing

Nvidia is in talks to provide a roughly $250 billion financial backstop for OpenAI, guaranteeing the debt behind a 10-gigawatt data center being built in southern Ohio on a former uranium-enrichment site. Including chips, the project could exceed $500 billion, which would make it the largest data center effort announced to date, and Nvidia separately discussed financing another $350 billion in chip purchases. For OpenAI the deal would be a first step toward owning its infrastructure rather than renting from Microsoft, Amazon, and Oracle, while for Nvidia it would lock in years of chip demand, though the talks remain in progress and could still change.

CNBC Technology ↗ · July 27 briefing

AI companies are spending record sums lobbying Washington as fights over federal policy intensify, with OpenAI, Anthropic, Google, and Microsoft all sharply increasing their outlays. Anthropic nearly tripled its lobbying bill to about $3.53 million in the first half of 2026, surpassing its total for all of 2025 and outspending Nvidia during the second quarter, while OpenAI nearly doubled its spend to a record $2.22 million. The companies reported lobbying on issues including cybersecurity, copyright, cloud computing, and defense procurement.

Financial Times Tech ↗ · July 27 briefing

Microsoft unveiled MAI-Cyber-1-Flash, its first AI model built specifically for cybersecurity, embedded inside MDASH, the company's multi-agent system for finding and fixing software vulnerabilities. Microsoft says the configuration scores 96% on a leading industry benchmark, about 12 points above a rival frontier model, while cutting costs by nearly half, with the compact model handling up to 90% of queries and handing the hardest 10% to the larger GPT-5.4. The company also introduced Project Perception, an agentic security system entering public preview on August 3, advancing its argument that the future favors the cheapest capable model routed intelligently rather than the biggest one.

VentureBeat ↗ · July 27 briefing

Nvidia, Microsoft, Meta and more than 20 other American technology companies signed a joint letter urging Washington not to impose premature restrictions on open-weight AI models. The signatories, which also include Dell, IBM, Palantir, Mistral and the Linux Foundation, argue the US should build a strong domestic open ecosystem rather than cede the field to China. The letter follows alarm over Moonshot's Chinese Kimi K3 model, which has topped some benchmarks and which US officials accuse of being distilled from Anthropic's Fable 5. Notably, OpenAI, Anthropic and Google did not sign.

CNBC Technology ↗ · July 26 briefing

The surge in AI data center construction is reviving the US nuclear industry after four decades of stagnation, as tech giants sign deals for round the clock, carbon free power. Microsoft has committed to a $16 billion, 20 year contract to restart a unit at Three Mile Island by late 2027, while Amazon is investing $700 million in small modular reactor developer X-energy. Meta has contracted for up to 6.6 gigawatts of nuclear capacity and Google has signed multiple deals of its own. Together the major AI players have committed to nearly 10 gigawatts of nuclear power, enough to supply roughly 7 million homes.

Forbes Tech ↗ · July 26 briefing

Microsoft CEO Satya Nadella is under mounting pressure as a severe compute shortage forces the company to prioritize its own AI products, such as Copilot, over external Azure cloud customers. Finance chief Amy Hood has confirmed the company favors internal workloads when capacity is scarce, a shift from the original plan to make Azure the shared compute layer for everyone. The strain comes despite Microsoft's roughly $190 billion in annual capital spending, a bet that AI infrastructure will remain the key bottleneck through 2027 or 2028. Critics warn the tradeoff risks alienating the enterprise customers that built Azure's cloud business.

TechMeme ↗ · July 26 briefing

A Pulse survey of 101 enterprises found that agent orchestration is consolidating fast onto major model provider platforms, with Anthropic's Claude the primary platform for 40 percent of respondents, more than double any rival, followed by Microsoft at 18 percent and OpenAI at 13 percent. Companies say they choose a platform based on native alignment with a leading model and judge success by reliable, multi step task execution, but 71 percent admit that a quarter or fewer of their deployed agents are true multi step orchestrated workflows rather than single prompt chatbots wearing an agent label. The findings point to a widening gap between enterprises' orchestration ambitions and the reality of what they have actually built and deployed.

VentureBeat ↗ · July 23 briefing

AMD announced a new generation of data center products, including an updated AI accelerator, a new Venice central processing unit, and its first AI rack scale system called Helios, which it is positioning as a direct rival to Nvidia's competing rack designs. The company said that counting its rack systems, PC business, and other processing products together, it now expects its total addressable market to reach $2 trillion by 2030. Microsoft has signed on as a new Helios customer, joining Meta, OpenAI, and Oracle, as AMD tries to chip away at Nvidia's dominance of the AI computing market.

Bloomberg Technology ↗ · July 23 briefing

Industry estimates, including projections from Deloitte, put the cost of AI enabled fraud in the United States at up to $40 billion by 2027, up sharply from about $12 billion in 2023, as deepfake images, video, and voice cloning make scams easier to produce at scale. Efforts to fight back have been fragmented, but the C2PA provenance standard, backed by Adobe, Microsoft, Google, and OpenAI, is advancing toward formal ISO international standardization. Rather than trying to detect fakes after the fact, C2PA records a verifiable history of who created a piece of content, with what tool, and what edits were made, letting viewers trust material signed by a hardware secured source at the point of capture.

ZDNet ↗ · July 23 briefing

Samsung is in talks to invest around 1 billion euros in French AI startup Mistral as part of a new funding round that could value the company at roughly 20 billion euros, up sharply from its 11.7 billion euro valuation in its last round. Swedish investor EQT's Scaleup Europe Fund is also reportedly in talks to join the round. Both Mistral and Samsung declined to comment on the discussions. The talks surfaced just after Microsoft announced its own multibillion dollar deal to fund Mistral's European computing infrastructure and expand distribution of its models, underscoring how aggressively investors are chasing Europe's leading AI challenger to the big US labs.

Financial Times Tech ↗ · July 22 briefing

Microsoft and Mistral announced a significant expansion of their strategic partnership, including a multibillion dollar commitment from Microsoft that draws on Mistral's growing European GPU capacity, built on Nvidia's Vera Rubin platform, to support Microsoft's own cloud and AI services while helping fund Mistral's European build out. The deal brings Mistral's Medium 3.5 and OCR 4 models to Microsoft Foundry, with Medium 3.5 also becoming available through Microsoft Copilot Studio. It is framed around sovereign AI, aligning with Microsoft's European Digital Commitments and Sovereign Cloud strategy to serve regulated sectors like healthcare, manufacturing, and financial services where data residency and operational continuity matter. Microsoft and Mistral will also offer proof of concept funding, Azure credits, and technical workshops to accelerate enterprise adoption.

AI Business ↗ · July 22 briefing

Nvidia confirmed its Vera Rubin platform has reached full production, with NVL72 rack systems, which pair 72 Rubin GPUs with Vera CPUs using liquid cooling, now shipping to customers including OpenAI, CoreWeave, Google Cloud, Microsoft Azure, Meta, and Dell. OpenAI plans to adopt Vera Rubin at scale in the third quarter, and CoreWeave says its racks are already delivering ten times the token output of the prior generation. The rollout marks a key milestone for Nvidia after months of speculation about possible delays to the new chip design.

Bloomberg Technology ↗ · July 21 briefing

National Grid Ventures, the commercial arm of UK utility National Grid, is investing $1.75 billion for a 35% stake in Joulent LLC, a developer building power and electrical infrastructure for large US energy loads. The centerpiece is Project Kilby, a 2.67 gigawatt gas-fired power facility in West Texas built with Chevron that will supply a Microsoft-operated data center under a 20-year power purchase agreement and aims to begin delivering power by 2028. The deal values Joulent at $5 billion and extends National Grid's US footprint into Texas, a market attractive to data center developers for its cheap natural gas.

Financial Times Tech ↗ · July 21 briefing

AMD unveiled Helios, its first rack scale AI system, giving CNBC exclusive access to the setup being built and tested at the company's Texas data center lab. Microsoft has agreed to deploy Helios racks in its Azure data centers, joining earlier customers Meta, OpenAI, Oracle and Tata Consultancy Services. Each rack combines 72 AMD Instinct MI455X GPUs with AMD EPYC Venice CPUs and Pensando networking, and carries a price tag more than $1 million higher than Nvidia's competing Vera Rubin system. The system ships later this year and marks AMD's most direct challenge yet to Nvidia's dominance of the AI hardware market.

CNBC Technology ↗ · July 20 briefing

Skyfall AI, founded by Sam Pasupalak and Kaheer Suleman, who previously built Maluuba before Microsoft acquired it for about $160 million in 2017, plans to buy a small business to business software or e-commerce company for up to $1 million and run it almost entirely with AI acting as CEO. The system would manage pricing, marketing, customer support, finance and operations while human oversight shifts from daily decisions to strategic accountability. The founders aim to double the acquired company's revenue within six months while publicly documenting successes and failures. Pasupalak said the goal is to prove a truly autonomous business is possible by buying a company and operating the whole thing end to end.

Forbes Tech ↗ · July 20 briefing

A VentureBeat Pulse survey of 107 enterprises found that 54% have already experienced a confirmed AI agent security incident or a near-miss, while only 32% give every agent its own scoped identity and 30% isolate high-risk agents in sandboxes. Most organizations still rely on provider-native controls from OpenAI, Google, and Microsoft rather than purpose-built agent security tools, and a clear majority plan to change their tooling within the next year.

VentureBeat ↗ · July 16 briefing

Microsoft is coaching its sales team to compete aggressively against Anthropic and OpenAI, playing up shortcomings in the AI firms' products while pitching its own lower costs, better security controls and a more complete product suite. At an internal meeting laying out the new fiscal year's sales strategy, executive Jay Parikh told staff that rivals are selling parts while Microsoft offers a complete system, and executive Jacob Andreou presented a side-by-side analysis casting Claude inside Microsoft's office suite as slower, less accurate and lacking enterprise-grade security integrations. CEO Satya Nadella said monitoring AI expenses and using cheaper models will be a primary customer concern this year, citing Unilever's projected $300 million savings from swapping an advanced model for a more economical Microsoft version.

Bloomberg Technology ↗ · July 15 briefing

Microsoft first invested 1 billion dollars in OpenAI in 2019 and deepened the relationship in 2023, with CEO Satya Nadella backing Sam Altman when he was briefly ousted by the board, but by this April the two companies had effectively broken up and converted their exclusive arrangement into a non exclusive one following reports Microsoft considered suing OpenAI for breach of contract. Apple is now suing OpenAI, alleging Altman's company poached its employees and stole trade secrets, and OpenAI executives were reportedly disappointed enough in the Apple partnership to consider a countersuit of their own. The pattern points to a recurring dynamic where OpenAI's biggest technology partnerships curdle into legal disputes as the company grows.

Business Insider Tech ↗ · July 13 briefing

Starbucks is building its own AI assisted software to replace a Microsoft system that tracks inventory and an IBM platform that manages maintenance, part of an effort to redirect roughly 400 million dollars in enterprise software spending. The move, made more viable by AI assisted development, rattled software stocks, with IBM falling about 3 percent, enterprise workflow vendor Service Now dropping 3.5 percent, and Salesforce sliding 4 percent in premarket trading. Some of the new tools could roll out by the end of next year pending testing, signaling a shift in the traditional build versus buy calculus for large enterprises.

Forbes Tech ↗ · July 12 briefing

Five years after Salesforce acquired Slack for 27.7 billion dollars, the company has launched an integration that connects Slackbot, the AI agent built into every workspace, directly to the Salesforce platform, including CRM data, Tableau analytics, and Data 360 customer profiles. The feature runs on dedicated Model Context Protocol servers, letting a salesperson pull a customer's deal history, view a live Tableau chart, update a CRM record, and trigger a DocuSign approval, all from a single chat message without switching tabs. Salesforce says its own IT team has already used the setup to save more than 1,500 engineers thousands of coding hours annually. The move comes as Slack faces growing competition from Microsoft Teams and its Copilot integration, as well as Google's Gemini push into Workspace.

VentureBeat ↗ · July 8 briefing

Microsoft is replacing OpenAI and Anthropic models with its own internally built MAI models in software products like Excel and Outlook, with tens of thousands of prompts in those apps now handled weekly by Microsoft's in-house AI. AI chief Mustafa Suleyman said in June that the company is trying to reduce and ultimately eliminate what it pays Anthropic, citing the high cost of relying on outside labs. Microsoft unveiled seven new AI models at its Build developer conference in June, including one it says matches the coding ability of Anthropic's Opus 4.6 at a lower cost. The shift comes as Microsoft's discounted pricing under its long-standing OpenAI partnership faces an uncertain future.

Bloomberg Technology ↗ · July 7 briefing

Masayoshi Son has put SoftBank at the center of the global AI boom, with cumulative commitments to OpenAI expected to reach roughly $64.6 billion for a stake of about 13%, making SoftBank the ChatGPT maker's second-largest external shareholder behind Microsoft. His AI push also spans chip companies Arm, Graphcore, and Ampere Computing, a $2 billion stake in Intel, and the self-driving startup Wayve. Some SoftBank insiders and investors worry that the veteran dealmaker now has too much control and too much of the company's future riding on a single bet.

Financial Times Tech ↗ · July 6 briefing

Station F, the Paris startup hub founded by French billionaire Xavier Niel, is launching the second batch of its F/ai accelerator in September after its first cohort of 20 AI startups collectively raised $34 million in pre-seed funding. The program pushes early-stage companies to reach 1 million euros in revenue within six months, and 80% of first-batch founders were repeat entrepreneurs. Partners for the new cohort include Anthropic, OpenAI, Google, Microsoft, and Mistral AI, with new additions such as Eleven Labs, Nebius, and GitHub.

TechCrunch AI ↗ · July 6 briefing

Z.ai, the Beijing-based artificial intelligence lab formerly known as Zhipu AI, on Wednesday officially launched ZCode, a free desktop application it describes as an "Agentic Development Environment" purpose-built for its flagship GLM-5.2 large language model. The move marks the company's most aggressive push yet into the fast-growing AI-powered coding tool market, where it now competes directly with Cursor, Claude Code, GitHub Copilot, and Google's Antigravity."Introducing ZCode, the official development environment for GLM-5.2," the company wrote on X, noting the tool is available on macOS, Windows, and Linux, supports bring-your-own-key (BYOK) configurations for third-party models, and offers a 1.5x usage-quota bonus for subscribers to its GLM Coding Plan. Read one way, ZCode is simply another entrant in a crowded market. Read another, it is a single product that crystallizes three of the most consequential trends in enterprise software today: the race-to-the-bottom pricing of frontier AI models, the geopolitical balkanization of the AI stack, and the rapid maturation of agentic coding agents into what Gartner now estimates is a roughly $10 billion market. An AI coding tool designed to think in projects, not promptsUnlike traditional IDEs that bolt on AI through a chat sidebar or autocomplete extension, ZCode is best understood as an agent-first development environment. Its core design is built around long-horizon tasks: the user describes an ou...

VentureBeat ↗ · July 2 briefing

Microsoft follows Amazon, OpenAI, and Anthropic with its new AI deployment group.

TechCrunch AI ↗ · July 2 briefing

Big Tech companies are discovering that AI data center expansion is bottlenecked not by compute or capital but by a global shortage of skilled physical labor — electricians, construction specialists, and datacenter technicians, the Financial Times reports. The workforce constraint is triggering new investments in vocational training and apprenticeship programs at Microsoft, Google, and Meta as they race to staff up capacity. The finding reframes the AI infrastructure debate: the binding limit is human craft skills, not silicon.

Financial Times Tech ↗ · June 21 briefing

Varonis disclosed SearchLeak, a Microsoft 365 Copilot Enterprise Search exfiltration chain, while LiteLLM admin-key flaws surfaced in the same window, exposing missing trust boundaries in enterprise AI stacks.

VentureBeat ↗ · June 18 briefing

Microsoft has become the main supplier of OpenAI models inside China through Azure even as OpenAI and Anthropic keep their own frontier models out of the market over IP and misuse concerns.

AI News ↗ · June 18 briefing

Investors were confronted this past week with four difficult realities that may fundamentally change the way they think about AI the business vs. AI the technology: AI is too expensive, say CEOs and even Microsoft itself. It's not paying off nearly as much as companies expected, per a new Bain study. Infrastructure demand is strong but not as strong as the most optimistic wanted, as Broadcom showed with its weak forecast. Financing that infrastructure is going to be more expensive.

Axios ↗ · June 8 briefing

Microsoft on Monday unveiled the Surface RTX Spark Dev Box, a compact desktop computer designed to let software developers run large AI models on their desks instead of paying for cloud computing — a move that directly challenges the per-token pricing model that has defined the AI industry's economics since ChatGPT launched three and a half years ago. The device, announced at Microsoft Build 2026, packs Nvidia’s new Blackwell-architecture RTX Spark processor and 128 gigabytes of unified memo...

VentureBeat ↗ · June 2 briefing

For the past two years, the technology industry has raced to make AI agents more capable — teaching them to write code, navigate software interfaces, manage files, and orchestrate multi-step workflows with increasing autonomy. What the industry has not done, at least not with any consistency, is answer the question that keeps chief information security officers awake at night: what happens when an agent goes wrong? On Tuesday at its annual Build developer conference, Microsoft offered what may be...

VentureBeat ↗ · June 2 briefing

The company best known for powering the AI boom is coming for the PC: Nvidia is expected next week to debut the first Windows computers that use its chips as the main processor, sources confirm to Axios. Why it matters: Microsoft's first AI PC push stumbled, but Nvidia's arrival gives it a second chance, this time with the world's hottest chipmaker attached. Driving the news: Nvidia and Microsoft will unveil their joint work and the first computers running the chips at two key industry conferences...

Axios ↗ · May 31 briefing

Corporate leaders are starting to question whether soaring AI spending is delivering meaningful returns. Why it matters: Companies that rushed to embrace AI are now confronting ballooning IT costs, uncertain productivity gains and growing employee skepticism. Driving the news: Microsoft canceled most of its Claude Code licenses, in part over costs, according to The Verge, and Uber's COO said AI costs are getting "harder to justify."An AI consultant tells Axios one of their clients recently spent h...

Axios ↗ · May 29 briefing

Nicole Nguyen / Wall Street Journal: Hands-on with a pre-release build of Google Docs Live, an AI-powered voice tool for drafting documents, rolling out this summer to AI Pro and Ultra subscribers — An exclusive look at Docs Live, Google's new speech-powered AI project manager and writing partner — Microsoft Word brought word processing to the masses.

TechMeme ↗ · May 24 briefing