AI Infrastructure News

Madison Air, a provider of ventilation and filtration systems, raised $2.23 billion in its initial public offering, according to Bloomberg. The company's stock surged following the IPO, with President and CEO Jill Wyant appearing on Bloomberg Tech to discuss the offering and the company's growth outlook. Wyant highlighted the expanding opportunity in data center infrastructure, specifically referencing the buildout of Terafabs as a key growth driver. The interview was conducted by Bloomberg's Ed Ludlow and published on April 16, 2026. Madison Air's positioning centers on supplying critical physical infrastructure components — ventilation and filtration systems — to support the rapid expansion of data center facilities.

Allbirds, the footwear company that went public at a $4 billion valuation in 2021 but never turned a profit, announced a dramatic pivot away from shoes toward artificial intelligence, causing its stock to surge approximately 600 percent. The company had experienced a sharp decline in sales of nearly 50 percent between 2022 and 2025, and had agreed to sell its name and remaining assets to American Exchange for $39 million after closing its physical stores. CEO Joe Vernachio announced a plan to raise $50 million from an undisclosed investor to transform the company's shell listing into a new entity called NewBird AI, which the company describes as 'a fully integrated GPU-as-a-Service (GPUaaS) and AI-native cloud solutions provider,' according to reporting by The Verge citing the Financial Times. The strategic shift leverages the remaining public shell company structure — a listed entity with no core operating business — as a vehicle to enter the AI infrastructure space without undergoing a traditional IPO process.

The article, sourced from AI Business, covers Orbital, a startup founded in 2026, which is among a growing number of companies pursuing the development of AI-powered data centers in space. The company has received funding intended to lay the groundwork for this orbital infrastructure initiative. Beyond these high-level details, the provided article content does not include specific funding amounts, investor names, launch timelines, technical specifications, or other quantifiable data points. The concept positions Orbital within an emerging segment of space-based computing infrastructure. Due to the limited content provided, a comprehensive factual summary with specific figures cannot be fully constructed from the available text.

Allbirds, once a trendy direct-to-consumer shoe brand, is pivoting to AI infrastructure, according to a report from AI Business. The company is financing this dramatic strategic shift through a $50 million convertible financing facility agreement. The move represents a full rebranding of the company, moving away from its origins as a sustainable footwear brand. The limited details available from the source indicate this is a significant corporate transformation, though specific operational details of the AI infrastructure business have not been fully disclosed in the article.

Parasail has raised $32 million in a Series A funding round, according to TechCrunch, positioning the startup around a concept called 'tokenmaxxing' — a strategy focused on maximizing token throughput for AI developers. The funding signals investor appetite for infrastructure plays that sit at the intersection of model deployment and compute optimization. The company is betting that as AI models proliferate and fragment across providers, developers will need specialized infrastructure to efficiently manage and scale token generation workloads. The Series A raise suggests early-stage confidence in Parasail's thesis that compute demand driven by tokenmaxxing could underpin a next-generation compute giant. The TechCrunch report frames the funding as a broader signal of a 'fractured future' of AI models and compute, implying the market is moving away from monolithic infrastructure toward more distributed, workload-specific solutions.

Allbirds, the sustainable footwear company, is pivoting away from its core shoe business and rebranding as NewBird AI, according to a TechChrunch report dated April 15, 2026. The company has sold its shoe business and secured a $50 million convertible financing facility to fund its transition into the AI sector, specifically targeting AI servers. The rebrand represents a complete strategic overhaul from Allbirds' original identity as a sustainable wool sneaker brand. The $50 million convertible financing structure suggests the company is seeking capital while potentially deferring equity dilution, a common approach for companies undergoing significant business model transformations. Specific details regarding the acquirer of the shoe business, the terms of the sale, or the identity of the financing providers were not disclosed in the available content.

JPMorgan Chase & Co. and Mitsubishi UFJ Financial Group underwrote a record-breaking $38 billion loan package tied to Oracle Corp. data center projects in Texas and Wisconsin, according to Bloomberg. The debt package represents an exceptionally large financing arrangement, described as record-shattering in scale. The distribution of this debt across the broader market has taken several months, indicating the significant challenge of syndicating such a large loan to institutional investors and lenders. The financing is directly linked to Oracle's infrastructure expansion efforts, specifically supporting new data center developments in two U.S. states. The prolonged distribution timeline highlights the complexity and scale of placing $38 billion in debt with market participants, even for major financial institutions like JPMorgan and MUFG.

Allbirds Inc., a wool sneaker company that was once valued at over $4 billion, announced a pivot to AI computing infrastructure just days before the company was set to permanently close, according to Bloomberg. Following the rebrand announcement, shares in the company soared 373%, representing a dramatic reversal for a business that had been on the verge of shutting down entirely. The company, which had been known for its eco-friendly footwear products, did not disclose specific details about its AI infrastructure plans in the initial announcement. The sudden strategic shift came as a last-ditch effort to salvage the struggling business, with the AI pivot announced only days before the firm's anticipated closure.

Data centers linked to Alphabet Inc.'s Google are pursuing a $5.7 billion junk-bond sale, which would represent the largest deal of its kind, according to Bloomberg. The financing is intended to support the ongoing artificial intelligence infrastructure buildout boom. The bond issuance is classified as high-yield, or 'junk,' debt, indicating it carries a credit rating below investment grade. The scale of the transaction underscores the significant capital requirements associated with expanding AI-capable data center capacity. Bloomberg reported the deal as record-setting within the data center financing space, reflecting accelerating private and institutional investment in AI infrastructure.

The provided article content from AIBusiness.com is insufficient to produce a fully detailed summary. The article, titled 'Oracle Expands AI Infrastructure Drive with Bloom Deal' and sourced from aibusiness.com with a relevance score of 72/100, references a deal between Oracle and what appears to be Bloom Energy, framed within the context of AI infrastructure expansion. The only substantive detail provided in the content is that the deal comes amid growing interest in onsite energy options, as AI compute vendors are struggling to meet power demands. No specific financial figures, deal terms, dates, executive names, capacity targets, or other concrete data points were included in the supplied article text. A full and accurate summary cannot be responsibly constructed from the available content alone.

Microsoft Corp. has agreed to rent data center capacity at a site in Norway that was originally designated for OpenAI and marketed under OpenAI's Stargate initiative, according to Bloomberg. The arrangement represents a transfer of the Norway data center capacity from OpenAI to Microsoft. The Stargate initiative had been positioned as a major AI infrastructure program associated with OpenAI. The Bloomberg report, dated April 14, 2026, indicates Microsoft will now take over the capacity at the Norwegian facility. Specific financial terms of the rental agreement, the total capacity of the site, and the timeline for Microsoft's operational takeover were not detailed in the available content. The development reflects ongoing shifts in how AI infrastructure assets are being allocated and managed between major players in the sector.

According to a Bloomberg report, Lumen Technologies Inc. CEO Kate Johnson stated that over half of global internet traffic is now generated by AI bots, marking a significant shift in the composition of internet activity. Johnson, who leads the enterprise network giant, indicated that this development is compelling executives across multiple sectors to fundamentally reassess how their organizations manage operations ranging from customer-service functions to network security threats. The scale of AI bot traffic — now representing a majority of all internet data flowing across infrastructure like Lumen's — underscores the accelerating pace at which automated AI systems are being deployed at scale. The remarks highlight both operational and security implications, as AI-driven traffic introduces new challenges around identifying and managing both legitimate and potentially malicious automated activity on networks.

According to Bloomberg, Chinese data center operators are increasingly utilizing a REIT-style asset-backed security (ABS) financing structure to raise capital, having collectively raised more than one billion dollars from yield-seeking investors. This financing wave reflects growing institutional appetite for infrastructure-backed securities in China's data center sector. The trend mirrors real estate investment trust (REIT) structures, allowing data center operators to monetize their physical assets while providing investors with yield-generating instruments. The Bloomberg report highlights this as a fast-growing segment of China's capital markets, with data center infrastructure serving as the underlying collateral for these securities.

According to The Decoder, SoftBank is leading an initiative to unite Japan's major industrial corporations — including steel giants, automakers, and banks — to develop a domestically built AI foundation for the country. The effort is framed as Japan's strategic response to the dominance of American and Chinese AI models, with the goal of reducing national dependence on foreign AI infrastructure. The article does not specify the names of all participating companies, the total capital committed, or a formal launch date based on the content provided. SoftBank, already a significant player in global AI investment through its Vision Fund, appears to be positioning itself as the central organizer of this national AI coalition.

According to The Decoder, the AI industry is experiencing a significant compute capacity crunch driven by surging demand for AI agents. Anthropic has been struggling with service outages, while OpenAI announced the discontinuation of Sora, both developments attributed in part to constrained infrastructure resources. GPU prices have jumped nearly 50 percent according to market data cited in the article, reflecting the acute supply-demand imbalance in AI hardware. The report describes conditions of rationing across the industry as compute availability fails to keep pace with accelerating AI workloads. The Decoder's relevance score for this story was rated at 92 out of 100, underscoring the significance of the compute shortage as a central issue for the AI sector.

Kepler Communications has launched what it describes as the largest orbital compute cluster, consisting of 40 GPUs operating in Earth orbit, according to a TechCrunch report published April 13, 2026. The company has announced Sophia Space as its latest customer for this space-based computing infrastructure. The orbital cluster represents a significant step in the emerging market for in-space computing, bringing GPU-based processing capabilities beyond Earth's atmosphere. Specific financial terms of the Sophia Space customer agreement were not disclosed in the available reporting.

Oracle Corp. has agreed to purchase up to 2.8 gigawatts of fuel-cell generated power from Bloom Energy Corp. to supply its data centers supporting artificial intelligence workloads, according to Bloomberg. The deal represents a significant energy procurement arrangement between the two companies, with Oracle tapping Bloom Energy's fuel-cell technology as a power source for its AI infrastructure. Fuel cells offer an alternative to traditional grid power, potentially providing more reliable and lower-emission energy for the energy-intensive demands of AI data center operations. The agreement underscores the growing pressure on major cloud and AI infrastructure providers to secure large-scale, dedicated power sources as AI workloads drive substantial increases in data center electricity consumption.

According to The Decoder, OpenAI is actively pitching investors on the argument that its early-mover advantage in infrastructure development provides a decisive competitive edge over rival Anthropic. The company is positioning its established infrastructure buildout as a key differentiator in the intensifying competition among leading AI labs. However, OpenAI is simultaneously pausing its data center project in the United Kingdom, signaling a potential shift or delay in its international expansion plans. Meanwhile, competitor Anthropic is reportedly exploring the development of custom AI chips, a move that could reduce its dependence on third-party hardware suppliers and potentially close the infrastructure gap OpenAI is highlighting to investors.

CoreWeave has signed a multi-year cloud infrastructure deal with AI startup Anthropic, according to The Decoder. The agreement is intended to power Anthropic's Claude AI models, though specific financial terms, contract duration, and capacity details were not disclosed in the source article. CoreWeave, a specialized AI cloud computing provider that went public in March 2025, will provide the GPU-intensive infrastructure required to run and scale Anthropic's Claude large language models. Anthropic, the AI safety-focused company behind the Claude model family, has been expanding its infrastructure partnerships as demand for its models grows. The deal represents a notable customer win for CoreWeave as it competes with hyperscalers such as AWS, Google Cloud, and Microsoft Azure for AI workload contracts.

Intel has secured a new AI infrastructure deal with Google, according to AIBusiness.com, though the article provides limited detail on the financial terms, scope, or timeline of the agreement. The deal is described as part of a broader wave of AI infrastructure partnerships among major technology vendors that have emerged over recent months. No specific contract value, product lines, or deployment details were disclosed in the source material. The agreement signals continued demand for AI hardware and infrastructure solutions from hyperscale cloud providers such as Google. Intel, which has been working to strengthen its position in the AI chip and infrastructure market, adds Google to its roster of enterprise AI partners.

CoreWeave CEO Michael Intrator announced in an interview with Caroline Hyde on Bloomberg Tech that the AI infrastructure company has signed a multibillion-dollar contract with Anthropic, though exact financial terms were not disclosed by either company. This deal with Anthropic, the AI safety company behind the Claude models, comes during what Intrator described as a significant week for CoreWeave. Separately, CoreWeave also announced a $21 billion deal with Meta to provide computing power, marking two major enterprise agreements in close succession. The back-to-back announcements position CoreWeave as a key infrastructure provider across multiple leading AI developers. According to Bloomberg, Intrator confirmed both deals but declined to elaborate on the specific terms of the Anthropic contract.

Blackstone Inc. is considering an initial public offering that could raise $2 billion for a newly formed acquisition vehicle focused on purchasing data centers, according to people familiar with the matter as reported by Bloomberg. The IPO would create a publicly traded entity specifically designed to acquire data center assets, reflecting growing institutional interest in AI-driven infrastructure demand. The article, published April 10, 2026, does not specify a timeline for the offering or name the acquisition vehicle, but cites sources with knowledge of the plans. Blackstone, one of the world's largest alternative asset managers, has been an active investor in digital infrastructure, and this move would represent a significant expansion of its data center strategy into public markets. The $2 billion fundraising target underscores the scale of capital Blackstone is looking to deploy in the sector.

Chinese AI startup DeepSeek is actively recruiting for two data center engineering positions located in Inner Mongolia, according to Bloomberg. The job postings signal that DeepSeek is expanding its physical infrastructure in the region. Notably, Bloomberg reports that the Inner Mongolia facility is reportedly relying on Nvidia's Blackwell chips, which are banned from export to China under U.S. semiconductor export control regulations. The presence of restricted Nvidia Blackwell hardware at the site raises significant compliance and geopolitical questions surrounding DeepSeek's operations and supply chain.

Anthropic PBC has agreed to utilize data center capacity from CoreWeave Inc. to support the growing demand for its AI services, according to Bloomberg. The partnership involves Anthropic renting AI infrastructure capacity from CoreWeave to power its Claude AI models. The deal reflects the increasing computational demands facing AI model developers as adoption of their services scales. Bloomberg reported the agreement on April 10, 2026, though specific financial terms of the arrangement were not disclosed in the available content.

Blackstone is considering an initial public offering for its Digital Infrastructure Trust, a vehicle designed to acquire data centers, according to people familiar with the matter as reported by Bloomberg. The potential IPO could raise approximately $2 billion, positioning the acquisition company to target data center assets. The story was reported by Bloomberg's Bailey Lipschultz on 'Bloomberg The Close' and is based on sources with knowledge of the plans rather than official company announcements. The filing represents Blackstone's move to tap public markets to fund infrastructure acquisitions in the digital and AI compute space. No pricing date or exchange listing details were disclosed in the Bloomberg report.

PJM Interconnection LLC has put forward an emergency proposal targeting 15 gigawatts of new power supply to address potential electricity shortages driven by the surge in AI-related data center demand, according to Bloomberg. PJM, which operates the largest electricity grid in the United States covering 13 states and the District of Columbia, is responding to what it characterizes as an urgent capacity shortfall. The emergency nature of the proposal underscores the accelerating pace at which AI infrastructure buildout is straining existing power grid resources. The 15 gigawatt target represents a significant addition to regional electricity capacity, reflecting the scale of power consumption associated with modern AI data centers and high-performance computing facilities.

Anthropic has launched a new product called 'Claude Managed Agents,' a hosted infrastructure platform designed for developers to build and run autonomous AI agents, according to The Decoder. The platform provides managed infrastructure, meaning Anthropic handles the underlying compute and operational complexity rather than requiring developers to self-host. Early adopters of the system include productivity software company Notion and Japanese e-commerce and technology conglomerate Rakuten, indicating enterprise-level uptake at launch. The product represents Anthropic's move beyond standalone AI model licensing into providing end-to-end agentic infrastructure as a service. Specific pricing, technical specifications, and the full scope of capabilities were not detailed in the source article.

Meta has entered into a $21 billion deal with CoreWeave to expand their AI infrastructure partnership, according to AIBusiness.com. The agreement significantly deepens the relationship between the two companies, with Meta securing substantial GPU compute capacity from CoreWeave to support its growing AI workloads and model development. CoreWeave, a specialized AI cloud provider that went public in March 2025, has been rapidly expanding its client base among major technology firms seeking dedicated high-performance computing resources. The deal represents one of the largest AI infrastructure commitments reported in the sector and underscores Meta's aggressive investment strategy in building out the computational backbone required for its generative AI ambitions. This partnership adds to Meta's already substantial capital expenditure plans, which the company has indicated will reach between $60 billion and $65 billion in 2025.

Amazon CEO Andy Jassy used his annual shareholder letter to defend the company's planned $200 billion in capital expenditure while taking direct aim at several major competitors including Nvidia, Intel, and Starlink, according to TechCrunch. The letter, characterized as unusually combative in tone, positions Amazon's investments as a strategic response to competitive pressures across multiple technology sectors. Jassy's comments targeting Nvidia and Intel suggest Amazon is increasingly confident in its own silicon and AI infrastructure capabilities as alternatives to dominant chip suppliers. The mention of Starlink indicates Amazon's Project Kuiper satellite internet venture is becoming a more explicit competitive priority. The $200 billion capex figure underscores the scale of Amazon's commitment to building out AI and cloud infrastructure. The shareholder letter serves as a public-facing defense of this spending level to investors who may have concerns about the magnitude of capital allocation.

CoreWeave Inc. has signed a significantly expanded $21 billion deal with Meta Platforms Inc. to supply AI computing power, with the agreement extending through 2032, according to Bloomberg. The deal represents a major deepening of the existing commercial relationship between the two companies, positioning CoreWeave as a critical infrastructure partner for Meta's AI ambitions. The agreement underscores Meta's accelerating push to build more powerful artificial intelligence models as it works to close the competitive gap with rivals in the AI space. The scale of the contract — $21 billion over the duration of the partnership — highlights the enormous capital commitments being made across the industry to secure dedicated AI computing capacity.