Nvidia is preparing to ship its DSX Pod AI Factory starting in early 2027, a system that packs seven Vera Rubin AI racks into a container weighing roughly 28,000 kilograms. The modular design arrives pre-assembled and ready to plug into power, cooling, and networking, cutting deployment time from six months down to about 14 days. Partners including Foxconn, Lenovo, HPE, and Google Cloud are involved in the rollout, and the system uses liquid cooling that Nvidia says can save up to 54 million liters of water per day.
At Nvidia GTC Berlin, Jensen Huang highlighted that deploying a gigawatt of data center capacity can currently take four to six months after shipping, and the DSX POD approach targets that bottleneck. Strong demand is coming from companies like SpaceX, Core Weave, and Nebius. Analysts estimate a single Vera Rubin rack costs over $7 million, underscoring the scale of investment involved.
Moonshot AI, the Chinese startup behind the Kimi K3 model, is expanding into Singapore. The company posted a new sales role on LinkedIn seeking candidates with experience in sales, venture capital, consulting, product, or engineering. Kimi K3 launched in July and is claimed to be the world's largest open-weight AI system. Backers include Tencent Holdings, Meituan, and China Mobile, though Moonshot has not commented publicly on its Singapore plans.
Microsoft introduced a workload-based Sovereign Cloud at Ignite, featuring a Sovereign Control Panel now in public preview inside the Azure portal. New capabilities include Just-in-Time Access, Customer Lockbox, Data Guardian, and Production Touch Ledger. Azure Key Vault Managed HSM reached general availability for external key management, and tools like Foundry Local and GitHub Enterprise Local also became generally available.
ServiceNow released research showing a sharp rise in ungoverned AI exposure at enterprises, with most organizations lacking formal AI governance programs. In response, the company launched AI governance tools including AI Passport and real-time agent control, along with an Autonomous Security offering. The findings were tied to NYSE:NOW earnings and revenue growth as of October 2026.
OpenAI President Greg Brockman is supporting a political group that opposes AI restrictions, which is spending millions of dollars in congressional races to influence legislation around AI regulation. His backing lends significant weight to the effort.
3 E Network signed a roughly $5 million storage hardware sales contract through its Hong Kong subsidiary, covering memory modules and data center-grade storage controllers. Revenue from the deal is expected to fund its Mikkeli AI compute center in Finland and support proprietary chip development. The 9 MVA European center recently began hardware evaluation.
Hewlett Packard Enterprise will send its Spaceborne Computer-4 to the Moon on Astrolab's FLIP rover. The module tests large language models, LiDAR data processing, and autonomous monitoring on the lunar surface. NASA will run a neural network trained on Earth, Moon, and ISS data, while Deloitte will test its Silent Shield cybersecurity system. It is the first HPE space computer designed to operate in the vacuum of space, using a radiator instead of fans.
New York City proposed a rule requiring companies to disclose when customers are speaking with AI. Mayor Zohran Mamdani supports the proposal, which targets frustrating "doom loops" where customers cannot reach a human. Violations could carry fines starting at $525 per incident, with the Department of Consumer and Worker Protection leading the effort.
Researchers have built a mathematical framework identifying when humans and AI collaborate better than either alone. Experiments across seven clinical cases showed that strong AI performance can sometimes reduce human input, yet synergy still holds when human and AI errors are independent. The findings aim to help designers create better AI decision support systems.
Key Takeaways
- Nvidia's DSX Pod AI Factory ships in early 2027, cutting AI data center deployment from six months to 14 days.
- Each Vera Rubin rack is estimated to cost over $7 million, with the full container system weighing 28,000 kilograms.
- Moonshot AI is expanding into Singapore with a new sales role, backed by Tencent, Meituan, and China Mobile.
- Microsoft launched a Sovereign Cloud with Just-in-Time Access, Customer Lockbox, and Data Guardian controls at Ignite.
- ServiceNow reports most enterprises lack formal AI governance programs and launched AI Passport and real-time agent control tools.
- OpenAI President Greg Brockman supports a political group spending millions to oppose AI restrictions in congressional races.
- 3 E Network signed a $5 million storage hardware contract to fund its Mikkeli AI compute center in Finland.
- HPE will send its Spaceborne Computer-4 to the Moon on Astrolab's FLIP rover to test LLMs and LiDAR processing.
- New York City proposed fines starting at $525 for companies that fail to disclose AI customer service interactions.
- A new study finds human-AI synergy depends on each side being able to correct the other's mistakes.
Nvidia launches containerized DSX Pod AI Factory
Nvidia announced the DSX Pod AI Factory, a containerized system holding seven Vera Rubin AI racks. The system is pre-assembled at the factory and arrives ready to connect to power and cooling. Nvidia says it cuts AI data center deployment time from six months to 14 days. Deliveries are expected to begin in early 2027.
Nvidia reveals AI factory in a shipping container plan
Nvidia unveiled its DSX POD at GTC Berlin as a modular AI data center packed into a box the size of a shipping container. The Vera Rubin-powered system arrives ready to plug into power, cooling, and networking infrastructure. Nvidia says deployment drops from six months to 14 days. The design can also save up to 54 million liters of water per day.
Nvidia's Jensen Huang discusses AI factory shipping container plan
Jensen Huang discussed the DGX Pod at Nvidia GTC Berlin as an AI factory in a shipping container. He explained that deploying a gigawatt of data center can currently take four to six months after shipping. The new approach aims to shorten that timeline and reduce idle cash costs. Huang noted strong demand from companies like SpaceX, Core Weave, and Nebius.
Nvidia unveils DSX POD and secure AI agent framework
Nvidia introduced the DSX POD at GTC Berlin as a reference design for Vera Rubin AI factories built in modular containers. The system reduces deployment time from six months to 14 days and uses liquid cooling to save 54 million liters of water daily. Nvidia also announced OpenShell Reference Architecture to help secure AI agents. Partners include Foxconn, Lenovo, HPE, and Google Cloud.
Nvidia announces plug and play Vera Rubin racks
Nvidia's new DSX Pod AI Factory packages seven Vera Rubin AI racks inside a shipping container weighing about 28,000 kilograms. All cabling and liquid cooling are assembled at the factory before shipping by truck. Analysts estimate a single Vera Rubin rack costs over $7 million. Nvidia expects deliveries to start in early 2027 from partners including Foxconn and Lenovo.
Nvidia unveils DSX POD, an AI factory in a shipping container
Nvidia introduced the DSX POD at its GTC event in Berlin. The design packs an AI data center into a standard shipping container. Deployment drops from six months to 14 days. The Vera Rubin-powered system supports liquid cooling and saves up to 54 million liters of water per day. It does not replace traditional data centers but lets companies run AI applications faster.
China's Moonshot AI expands into Singapore
Chinese AI startup Moonshot is expanding into Singapore with a new sales role listed on LinkedIn. The posting seeks candidates with sales, venture capital, consulting, product or engineering experience. Moonshot released its Kimi K3 model in July, claiming it is the world's largest open-weight AI system. Backers include Tencent Holdings, Meituan and China Mobile. The company has a Singapore business entity incorporated in July 2023.
China's Moonshot AI lists Singapore sales role
Moonshot AI posted a new sales position on LinkedIn to expand operations in Singapore. The role requires experience in sales, venture capital, consulting, product or engineering. Moonshot's Kimi K3 model launched in July as the world's largest open-weight AI system. The company did not comment on its Singapore plans. Backers include Tencent Holdings, Meituan and China Mobile.
Microsoft launches Sovereign Cloud controls at Ignite
Microsoft introduced a workload-based Sovereign Cloud with new controls at Microsoft Ignite. The Sovereign Control Panel enters public preview inside the Azure portal. New features include Just-in-Time Access, Customer Lockbox, Data Guardian and Production Touch Ledger. Azure Key Vault Managed HSM is now generally available for external key management. Several tools including Foundry Local and GitHub Enterprise Local reached general availability.
ServiceNow flags rising ungoverned AI security risks
ServiceNow released internal research showing a sharp rise in ungoverned AI exposure at enterprises. Most organizations lack formal AI governance programs for internal and external AI agents. In response, ServiceNow launched AI governance tools including AI Passport and real-time agent control. The company promotes its Autonomous Security offering to address these growing risks. The research was tied to NYSE:NOW earnings and revenue growth as of October 2026.
New Study Reveals Conditions for Human-AI Teamwork Success
Researchers created a mathematical framework to find when humans and AI work together better than either alone. The study shows synergy depends on each side being able to fix the other's mistakes. Experiments with seven clinical cases found that strong AI performance can actually reduce human input. However, synergy can still work even with biases like automation bias if human and AI errors are independent. The findings help designers build better AI decision support systems.
New York City Proposes Rules on AI Customer Service Disclosures
New York City proposed a rule requiring companies to tell customers when they are speaking with AI. The goal is to stop frustrating "doom loops" where customers cannot get help. Mayor Zohran Mamdani supported the proposal. Companies that break the rule could face fines starting at $525 per violation. The Department of Consumer and Worker Protection led the effort.
HPE to Send AI Computer Module on Moon Mission
Hewlett Packard Enterprise will send its Spaceborne Computer-4 to the Moon on Astrolab's FLIP rover this year. The module will test large language models, LiDAR data processing, and autonomous monitoring on the lunar surface. NASA will run a neural network trained on Earth, Moon, and ISS data, while Deloitte will test its Silent Shield cybersecurity system. Computer-4 is the fourth HPE space computer but the first designed to operate in the vacuum of space. It uses a radiator instead of fans and can lower its power during radiation events.
OpenAI President Supports Group Fighting AI Restrictions
OpenAI President Greg Brockman has supported a political group that opposes AI restrictions. The group is spending millions of dollars in congressional races. The group works to influence legislation around AI regulation. Brockman's backing adds significant weight to the effort.
3 E Network Sells $5 Million Hardware for AI Data Center
3 E Network signed a roughly $5.0 million storage hardware sales contract through its Hong Kong subsidiary. The order covers memory modules and data center-grade storage controllers. The company expects the sales revenue to fund its Mikkeli AI compute center in Finland and support proprietary chip development. The 9 MVA European center recently began hardware evaluation. A previous million-dollar memory component order was also delivered.
AI Sales Agents Ended Cold Email What Works Now
AI sales agents made cold email useless because every startup sends similar AI-generated messages. Reply rates dropped as the cost of sending emails fell to nearly zero. Buyers now move to harder-to-fake channels like referrals, Slack communities, GitHub issues, and free trials. Teams use intent data to send one accurate message instead of generic flattery. Physical kits with scannable links also work because AI cannot mail packages. You should test physical kits against your best 20 accounts to see what books more meetings.
Sources
- Nvidia Launches Containerized DSX Pod System To Accelerate AI Factory Deployments
- Nvidia reveals plans for 'an AI factory in a shipping container' to build data centers in weeks, not months
- Nvidia's Jensen Huang wants to build AI factories in weeks, not months
- Nvidia unveils DSX POD to accelerate AI factory deployment, adds secure AI agent framework
- Nvidia announces plug and play Vera Rubin racks that can be shipped on a truck
- Nvidia reveals plans for 'an AI factory in a shipping container' to build factories in weeks, not months
- China's Moonshot AI to operate in Singapore, jobs post shows
- China’s Moonshot AI to operate in Singapore, jobs post shows
- Microsoft sells sovereignty without letting go
- ServiceNow (NOW) Flags Rising Ungoverned AI Security Risks
- When Do Humans and Machines Truly Complement Each Other? Formal Conditions and Empirical Evidence for Human-AI Synergy
- Mamdani Targets Companies With AI Customer Service ‘Doom Loops’
- Hewlett Packard Enterprise to Send AI Compute Module to the Moon
- OpenAI president backs political group opposing AI restrictions
- About $5 million in hardware sales is expected to help fund AI research and a data center in Finland.
- AI Sales Agents Broke Cold Email: Here is What Replaced It
Comments
Please log in to post a comment.