What's happening: Liquibase launches Secure 5.1 to tighten data platform security and AI readiness
Liquibase introduced Secure 5.1, extending its change‑control framework to Snowflake’s control plane and treating access, sharing, and cost‑control changes as auditable objects. The update adds policy enforcement, drift detection, automated rollback, and support for platforms such as Databricks, MongoDB, Couchbase, AWS Keyspaces, and DataStax Enterprise.
Also developing:

Amazon Web Services announced a $200 billion capital‑expenditure plan through 2026, with networking as a primary focus to support its AI push. The company is trialing hollow‑core fiber in a handful of locations to extend latency‑critical inter‑zone links, while a newly designed control plane delivers sub‑second recovery across hundreds of thousands of links. AWS continues to design its own ASIC‑based networking hardware, enabling uniform updates and tighter integration with compute and storage. These moves aim to keep the network invisible to customers as AI workloads surge.

UK chip startup Fractile announced a £100 million expansion across its Bristol and London sites, creating a national industrial hardware engineering hub and adding 40 new roles to its 70‑person team. The company claims its in‑memory compute architecture can run AI...

Harvey, the legal‑AI startup, is reportedly negotiating a $200 million raise that would push its valuation to $11 billion, led by Sequoia and Singapore’s GIC. This follows a rapid funding series: a $300 million Series D at $3 billion, a $300 million Series E at $5 billion, and...

Oracle Corp.’s shares jumped up to 12% on Monday, marking the stock’s strongest intraday gain since September 10. The rally was sparked by news that U.S. tech giants, led by Amazon’s $200 billion capex commitment for data centers and chips, are accelerating...

CyrusOne announced a 760 MW data‑center campus in Freestone County, Texas, partnering with Constellation’s Calpine unit. The project secures two 380 MW power agreements—Phase I and Phase II—leveraging the adjacent natural‑gas Freestone Energy Center. Construction is underway with an expected operational date in Q4 2026....

Deutsche Telekom has opened an Nvidia‑powered AI factory at the Polarise data centre in Munich’s Tucherpark, installing roughly 10,000 Blackwell GPUs including DGX B200 and RTX PRO systems. The facility, branded the Industrial AI Cloud, underpins the ‘Germany Stack’ – a sovereign...

Meta has acquired an extra 1,400 acres adjacent to its Hyperion data‑center campus in Louisiana, expanding the footprint of a project originally planned on 2,250 acres. The Hyperion complex is designed to deliver 2 GW of compute power, with the potential...

Intersect Power, a clean‑energy firm being acquired by Alphabet, has filed a Texas Department of Licensing and Regulation application to build a 761,000‑square‑foot data center near Miami, Texas. The Project Pumpkin 2A development will cost roughly $400 million and is scheduled for...

Edge provider Pulsant has completed a £10 million expansion of its Milton Keynes data centre, adding a 1.2 MW data hall optimized for high‑density AI workloads. The new facility is part of the company’s platformEDGE network, which now spans 14 edge sites...

The article warns that excessive caution in data‑center design leads to chronic oversizing. Studies show many facilities operate at only 20‑60% of installed capacity, inflating capital, energy and maintenance costs by up to 30%. Oversized power, cooling and backup systems...
I now constantly get questions about the SAAS meltdown, role of AI, system of records etc. I don't have an answer to all these. But I do know that we saw an acceleration in our business in Q2, Q3, and...
Job titles that have emerged in Tech, Data and AI the last 5 years: 👇

Estonian supercapacitor maker Skeleton Technologies opened its first U.S. engineering facility in Houston, Texas, to support AI data‑center customers. The graphene‑based devices can smooth power spikes and claim up to 45% energy savings for high‑performance computing workloads. Skeleton already has...

Aerospike released version 8.1.1, introducing native Dynamic Data Masking (DDM) for its high‑performance NoSQL database. The feature lets administrators define masking rules that hide personally identifiable information at the database layer, automatically applying to all users and machines except those...
Geotab’s telematics analysis of 22,700 EVs shows ultra‑fast charging above 100 kW roughly doubles battery degradation, reaching about 2.5 % capacity loss per year versus 1.2‑1.5 % with slower Level 2 charging. The effect intensifies when more than 12 % of sessions use high‑power chargers,...

Johnson Controls unveiled the York YDAM air‑cooled magnetic‑bearing centrifugal chiller, delivering 3.5 MW of cooling and capable of operating with 45 °C warm‑water. The company also previewed the two‑stage YK‑HT economizer that can produce 44 °F chilled water and 140 °F hot water simultaneously....

Nordic data‑center operator atNorth has filed a permit request to add three new buildings to its FIN04 campus in Kouvola, Finland. The expansion would raise the site’s planned capacity to 430 MW across a 45‑hectare footprint, up from the original 60 MW...
Swiss‑based NexaVM, backed by Alibaba, delivers an all‑in‑one VMware replacement aimed at mid‑market cloud service providers, sovereign‑cloud operators, and hardware‑agnostic OEMs. The platform bundles a production‑grade KVM hypervisor, Ceph‑based hyper‑converged storage, integrated Kubernetes‑as‑a‑Service, and multi‑tenant management under a subscription licence....

Global AI, in partnership with Saudi AI firm Humain, is converting a former Kodak‑Carestream site near Windsor, Colorado into a high‑density AI data center. After purchasing 438 acres for $15.6 million, the company plans to launch an 18‑24 MW facility by the...

Corning and Meta have signed a multiyear agreement worth up to $6 billion to accelerate the construction of advanced data centers in the United States. Under the deal, Corning will provide Meta with its latest optical fiber, cable and connectivity solutions...

Red Hat announced an expanded partnership with NVIDIA, introducing Red Hat Enterprise Linux for NVIDIA—a specialized RHEL edition tuned for the NVIDIA Rubin platform. The collaboration delivers Day 0 support for the Vera CPU, Rubin GPUs and BlueField‑4 DPU across Red...

IBM has issued a global request for proposals for the next cohort of its Impact Accelerator, targeting AI-driven solutions in education and workforce development. The program invites nonprofits, government entities, and academic institutions to develop tools that bridge the widening...
Today, I dig into the details of StarRocks and how it is gaining traction in the real-time database world. DWH-like joins and fast retrieval from a #Lakehouse-native data architecture, without additional data engineering work to persist and ingest data. https://www.ssp.sh/blog/starrocks-lakehouse-native-joins/

US‑based Vantage Data Centers was denied permission by the Groß‑Gerau city council to build a 174 MW, €2.5 billion data‑center campus on a 14‑hectare site outside Frankfurt. The council voted 18‑14 against contract negotiations, citing concerns over limited job creation, visual impact,...

Singapore telecom giant Singtel’s data‑centre arm Nxera has opened DC Tuas, an 120,000‑sq‑ft, eight‑storey facility delivering 58 MW of power, the highest‑capacity data centre in the city‑state. More than 90% of its multi‑tenant capacity was pre‑leased, and the site features advanced...

Embedded analytics moves data insights from isolated dashboards into the applications where decisions happen, cutting context‑switching delays. By delivering real‑time metrics within workflows, it accelerates decision cycles, boosts operational efficiency, and encourages higher user adoption. The approach also fuels customer...
Global semiconductor revenue is projected to breach the $1 trillion mark, propelled by explosive AI demand. The Semiconductor Industry Association reported $791.7 billion in sales for 2025, a 25.6% year‑over‑year increase. Tech giants are pouring capital into AI‑focused data‑center infrastructure, accelerating the...
What's changed since Kimball wrote The Data Warehouse Toolkit: 1. Surrogate keys are less necessary with better databases 2. Denormalization for performance matters less with modern engines 3. Snapshotting dimensions beats complex SCD2 logic 4. Collaboration requirements mean looser conformance Kimball's principles still matter. But...
In this episode, Viktor Gamov interviews Jeremy Custenborder of Confluent about his journey from a paper boy to a leader in large‑scale systems, focusing on his experience keeping MySpace operational at massive pre‑cloud scale. Jeremy explains how he built custom...
Most AI programmes fail for the same reason sports teams lose: weak fundamentals. 🏅 In this video, I share 6 lessons from elite sport that make AI adoption work when it moves from pilot to daily workflow: 1️⃣ Clear roles and ownership,...

In episode #345, DataFramed hosts Adel Nehme and Richie Cotton sit down with Brian Solis, Head of Global Innovation at ServiceNow, to explore how organizations can foster a culture of continuous innovation. Solis emphasizes the importance of aligning innovation with...

Nigeria’s credit penetration has topped 40%, signaling a rapid shift toward broader financial inclusion. CRC Credit Bureau, the country’s largest licensed bureau, has built the most comprehensive credit data ecosystem by pulling information from banks, fintechs, utilities, telcos, and digital...
![[Industry News] Yodo1 Unveils IPVerse to Eliminate High-Stakes Guesswork in Global Game Licensing](/cdn-cgi/image/width=3840,quality=75,format=auto,fit=cover/https://mcvuk.com/wp-content/uploads/IPVerse-Logo.jpeg)
Yodo1 Games has launched IPVerse, a free data‑intelligence platform that consolidates global IP and licensing information for in‑game collaborations. The service draws from over 80,000 events, 20,000 brands and 15,000 games, offering AI‑driven matching, popularity indices and regional insights, including...

The United Arab Emirates’ Technology Innovation Institute (TII) has released its Falcon family of large language models as open‑source, positioning the nation as a challenger to big‑tech AI dominance. Falcon models emphasize efficiency, with the 7‑billion‑parameter H1R delivering high‑performance reasoning...

StarRocks is attracting heavyweight users such as Coinbase, Pinterest and Fresha because it delivers sub‑second query latency on terabyte‑scale analytics while reading directly from lakehouse storage. The platform’s shared‑nothing architecture, colocated joins, caching layer and a cost‑based optimizer let it...
The PostGIS development team announced a series of bug‑fix patch releases covering versions 3.2 through 3.6, while designating 3.0.12 and 3.1.13 as end‑of‑life. New patches include 3.6.2, 3.5.5, 3.4.5, 3.3.9 and 3.2.9, each addressing stability and security issues. Users still...
VMware released Workstation Pro 25H2 (version 17.6.4), expanding hardware compatibility and adding a host of new guest operating systems. The update introduces USB 3.2, hardware version 22, and a new dictTool CLI for editing VM configuration files. Licensing changed in 2024, making Workstation Pro free...

At my work every month we look at customer retention. A single churn rate doesn't tell the whole story. You need to show when customers are leaving. Here's how you can do it with a cohort analysis heatmap in python

I've been thinking about this math a lot and weighing if that 2% chance of $50m is worth it. I've built DataExpert.io to $2M ARR over the last three years. I could exit for $8-10m now. In 3 years not...

Jan Kristof Nidzwetzki’s article breaks down PostgreSQL’s spinlock mechanism, detailing its low‑latency design and adaptive backoff strategy. It explains the four‑function API (Init, Acquire, Release, Free) and shows how contention is handled through incremental micro‑second sleeps. The piece also introduces the...

The signal-to-noise problem is about to become unsolvable for humans alone. When AI makes ALL content (texts, emails, PRs, job apps, grant apps, college apps, mail) high-quality and infinite in volume, every filtering process breaks. And when it can’t be filtered...

The European Commission disclosed that its central infrastructure for managing mobile devices was compromised, potentially exposing staff names and phone numbers. The breach aligns with a zero‑day exploit (CVE‑2026‑1281) in Ivanti Endpoint Manager Mobile (EPMM) disclosed a day earlier. Dutch...
A data engineering team discovered that a three‑year SCD Type 2 backfill executed day‑by‑day produced 47,000 overlapping records, timeline gaps, and unrecoverable errors. The author introduced "Healing Tables," a framework that separates change detection from period construction and rebuilds the dimension in...

Waymo has launched a new "World Model" built on Google’s Genie 3 AI, enabling the creation of hyper‑realistic 3D driving simulations from simple text or image prompts. The platform can generate extreme edge cases—tornadoes, flooded streets, rogue elephants—and run them at...

The article explains how to read buffer statistics from PostgreSQL's EXPLAIN output, a skill essential for diagnosing I/O bottlenecks. Starting with PostgreSQL 18, buffer metrics are included automatically, eliminating the need for the BUFFERS option. It breaks down shared, local, and...

Cambridge researchers introduced Latency‑Aware Group Scheduling (LAGS), a Linux kernel patch that reshapes how the scheduler handles Kubernetes workloads. By favoring short‑running tasks and cutting context‑switch overhead, LAGS lifts cluster throughput by roughly 10‑20 %. Benchmarks show the same 600‑container workload...
Manufacturers are investing heavily in IoT, AI, and robotics, but legacy Wi‑Fi and wired networks are creating hidden bottlenecks that limit data flow and scalability. Industry forecasts show the manufacturing IoT market could reach $674 billion by 2032, yet many digital...

Machine learning teams are increasingly overlooking pipeline efficiency, a hidden driver of productivity. Slow data I/O, redundant preprocessing, and mismatched compute inflate the iteration gap, limiting the number of hypotheses tested per week. The article outlines five audit areas—data ingestion,...

Voltage ride‑through (VRT) enables data‑center hardware to stay connected to the grid during brief outages or brownouts, preventing immediate shutdowns. By using UPS units that buffer power and can switch instantly to backup sources, VRT eliminates the latency of manual...
📊 The ETL Data Pipeline From raw sources (databases, APIs, files) → clean & transform (cleaning, joining, aggregating) → load into warehouse/analytics for BI, reports & ML. E → T → L: the backbone of modern data workflows.🚀