AI Infrastructure
Ongoing coverage on AI Infrastructure.
Latest coverage
OpenAI head Sam Altman says 'people hate data centers,' but the company still plans to invest $50 million more in AI infrastructure this year alone
Despite public backlash and negative sentiments towards data centers and AI, OpenAI plans to invest an additional $50 million in AI infrastructure this year. CEO Sam Altman acknowledges the public's negativity but emphasizes the continued need for compute power, even exploring renting out capacity for profitability. The article highlights community opposition to data centers, citing examples of lawsuits and environmental concerns.
Bitcoin Mining Meets the AI Infrastructure Boom: How SHRMiner Brings Cloud Computing Power to Crypto Users
Bitcoin mining companies like IREN are increasingly shifting towards AI cloud services, with IREN reporting an eightfold revenue increase in this sector. The SHRMiner cloud mining platform offers users access to mining infrastructure through a rental model, simplifying participation in cryptocurrency mining without the need for physical equipment. This trend highlights the convergence of Bitcoin mining and AI computing due to shared infrastructure needs.
SSD price increases might be with us for the long haul, as Phison's CEO says NAND supply won't catch up with…
Phison CEO Pua Khein-Seng predicts that NAND flash prices will continue to rise, potentially through 2030, due to a significant imbalance between supply and demand. Increased demand from AI infrastructure is a major factor, and new production capacity will take years to come online. Phison is shifting focus from the retail market to enterprise and AI solutions.
Sandisk (SNDK) Q4 FY26 earnings results beat revenue & EPS expectations
Sandisk exceeded Wall Street's revenue and earnings-per-share expectations for its fourth fiscal quarter of 2026, reporting $8.97 billion in revenue and $39.25 EPS. The company's performance was driven by its Solid State Drive products and its role in providing infrastructure hardware for the AI market. Despite beating estimates, Sandisk stock saw a decline in after-hours trading.
NAND and HBM have been merged into a dorky megazord for datacenters by Sandisk and SK hynix
SK hynix and Sandisk have unveiled High Bandwidth Flash (HBF), a new memory technology merging High Bandwidth Memory (HBM) and NAND capabilities for data centers. HBF aims to provide high-speed data transfer with expanded capacity, leveraging an open standard via the Open Compute Project and the Universal Chiplet Interconnect Express (UCIe) interface. This development comes amidst a growing demand for AI infrastructure and ongoing memory supply shortages.
SpaceXAI pats itself on the back for removing temporary turbines in Memphis under agreed order over the next year: "We care deeply about being good neighbors"
xAI and SpaceX have agreed to remove 69 mobile turbines powering their Mississippi data centers over the next 11 months, following an agreement with the Mississippi Department of Environmental Quality. The company plans to switch to power from a new 1.2 GW plant under construction and will invest in emission control technologies for the turbines. This action comes amid ongoing environmental and ethical disputes regarding the turbines and the rapid expansion of AI infrastructure.
Microsoft (MSFT) agrees to utilize AMD Helios AI infrastructure in its data centers
Microsoft has agreed to utilize AMD's new Helios AI infrastructure platform in its upcoming data center expansions. This deal positions AMD's Helios system as a competitor to NVIDIA's platforms and brings Microsoft in line with other major clients like Meta, OpenAI, and Oracle. AMD's stock saw a jump following the announcement of this agreement.
AI Yi-Yi!
NVIDIA and LG Group are collaborating to build an AI factory focused on advancing physical AI, mobility, and AI infrastructure. Separately, OpenAI is reportedly preparing a significant overhaul of its ChatGPT service, with internal discussions suggesting a shift away from the current chat interface.
A single Nvidia Vera Rubin rack is estimated to cost $7,803,148 with over $2 million of that figure spent on memory…
A Morgan Stanley research report estimates that a single Nvidia Vera Rubin VR200 NVL72 rack costs approximately $7.8 million, with over $2 million attributed to memory. These costs are representative of what cloud service providers would pay, indicating a significant markup from Nvidia's internal build costs. Nvidia CEO Jensen Huang predicts Vera Rubin will be more successful than their previous Grace Blackwell AI GPUs.
AMD is investing $10 billion+ into Taiwan tech to 'accelerate next-gen AI infrastructure', with some fancy new…
AMD announced an investment of over $10 billion into Taiwan's semiconductor and AI ecosystem to expand strategic partnerships and scale advanced packaging facilities for AI infrastructure. The company also highlighted production milestones for its Venice Epyc CPUs on TSMC's 2nm technology and plans for its Helios rack-scale platform.
Nvidia says 'demand for AI infrastructure continues to expand at an unprecedented pace' but after speaking to…
Nvidia reports unprecedented demand for AI infrastructure, with its revenue soaring due to AI-focused hardware. However, experts offer mixed opinions on the future of AI growth, with some predicting a slowdown by late 2026 due to market saturation and pushback against data centers, while others foresee continued broad adoption across major companies driving GPU demand.
Evening Tech Bits
Meta's significant investment in Graviton technology underscores the ongoing CPU shortages impacting AI infrastructure, according to Tom's Hardware. The article also provides further details on Intel's revenue streams.
A lot of you panic-bought PCs to avoid RAMaggedon 2026
Global PC shipments saw a modest 3.2% year-over-year growth in Q1 2026, driven by pre-emptive purchases before anticipated memory price hikes and the end of Windows 10 support. However, rising AI infrastructure investment is increasing component costs, with analysts predicting significant negative impacts on PC market growth and retail prices throughout 2026.
Shoe retailer AllBirds (BIRD) sees 700 percent stock surge following pivot to AI infrastructure
Allbirds, a shoe retailer, experienced a 700 percent surge in its stock price after announcing a strategic pivot from selling sustainable footwear to providing AI compute infrastructure. The company's stock rose from under $3 to over $17 following this announcement.
An American shoe company has taken a $50 million investment to pivot to, of all things, AI infrastructure
Allbirds, a shoe company, has announced a significant pivot to AI infrastructure, securing a $50 million investment to become a GPU-as-a-Service and AI-native cloud solutions provider. The company plans to rebrand as "NewBird AI" and will lease high-performance GPUs to customers requiring dedicated AI compute capacity. This strategic shift has already led to a substantial increase in its stock price.
Dell's CEO reckons that the total memory demand from the entire AI market in 2028 will be 625x bigger than it was…
Dell's CEO Michael Dell predicts that the total memory demand for the AI market will increase by approximately 625 times by 2028. This surge is driven by expanding memory per accelerator and system scale in AI infrastructure. The article highlights that current memory manufacturers like SK hynix, Samsung, and Micron may struggle to meet this future demand, potentially impacting the availability and cost of components like HBM4, LPDDR5x, and NAND flash storage.
First came RAMageddon and storage inflation, now a CPU shortage could be on the cards as PC makers report limited…
PC makers are reporting limited CPU supply, potentially leading to shortages and price inflation similar to recent RAM and storage issues. This is partly driven by AI infrastructure demands consuming manufacturing capacity. Companies like Intel and AMD have previously warned customers about server CPU shipment delays.
How a surge in memory prices will affect the game industry
A surge in demand for DRAM and NAND memory, driven by AI infrastructure expansion, is significantly increasing prices for components used in gaming hardware and development. This trend is expected to squeeze hardware margins for consoles like the PlayStation 5 and Nintendo Switch, potentially leading to price adjustments for subscriptions and peripherals. PC gamers may delay upgrades, while developers face rising server costs, necessitating smarter infrastructure strategies.
