AI Data Center & Build-out

ALT Telecom Public Company Limited, or ALT, disclosed that it is in negotiations for additional deals with global-tier customers, the Global Hyperscalers, in the fourth quarter of 2026, after signing network and digital infrastructure lease contracts worth a total of approximately 2 billion baht in the third quarter of 2026, with a contract term of 20 years. Preeyaporn Tangpaosak, Managing Director, told the Stock Vision news team that the company targets double-digit growth in revenue in the second half of 2026, driven by recurring revenue recognition from its existing network. The expansion of its large-customer base has raised the value of work awaiting revenue recognition, or backlog, for the group from 5.115 billion baht as of the second quarter of 2026 to approximately 6.9 billion baht in September 2026. The company is also proceeding with bids for smart grid power network systems and AMI smart meter installation work for the Provincial Electricity Authority and the Metropolitan Electricity Authority, as well as government work in security and digital fields, with a combined value of several billion baht. It is also pushing to connect networks in the Eastern Economic Corridor, or EEC, with data centers in Bangkok and overseas, setting a target internal rate of return, or IRR, for new projects of no less than 10 to 15 percent. The company sees three key drivers supporting revenue and profit growth over the next 12 to 18 months: the data center and AI trend, growth in smart meter and smart grid work, and margin expansion from operating leverage.

Blackfuel announced in late September 2026 that it has chosen Digital Realty's BCN1 data center in Barcelona to host and scale its liquid-cooled, AMD GPU-powered AI inference platform, tightly integrated with the ServiceFabric interconnection network for low-latency, private connectivity. The move comes shortly after Digital Realty's plan to add a new cable landing station at its LAX12 facility, which is slated to support subsea cables from 2028, tying subsea routes into the company's interconnection fabric. Together, the two projects support Digital Realty's near-term interconnection and AI demand catalyst, including the conversion of a US$1.9b lease backlog and US$410m of recent hyperscale signings into revenue. The company's narrative projects $9.7 billion in revenue and $1.4 billion in earnings by 2029, requiring 12.7% yearly revenue growth and an earnings increase of about $0.6 billion from $758.3 million today. The central debate remains Digital Realty's enlarged 1.4 GW, US$4.25b to US$4.75b 2026 CapEx plan and whether demand will match that build out, with three fair value estimates from the Simply Wall St Community spanning roughly US$223 to US$303 per share.

CIBC raised its price target on Enerflex to CA$30 from CA$27.50 after updating its model for a 450 MW behind-the-meter power generation award tied to a North American data center developer, while keeping a Neutral rating on the stock. The firm had already lifted its target to CA$30 in July 2026, and it cited strong Engineered Systems bookings and a modest EBITDA beat in the second quarter as positives supporting execution on the core business. CIBC noted that earlier weakness in the shares followed a lack of secured data center power generation bookings, which it believes pushed potential catalysts into later quarters. On the updated assumptions, Simply Wall St's fair value for Enerflex rose to CA$46.94 from CA$44.50, with revenue growth now 6.72% versus 3.39% previously, net profit margin at 8.64% versus 9.01%, a future P/E of 18.29x versus 17.87x, and a discount rate of 6.83% versus 6.68%.

Schneider Electric, alongside partners Wärtsilä and Stanley Consultants, introduced a "Generator-to-Chip" power approach and unveiled software-defined medium-voltage switchgear and advanced cooling solutions for AI-driven data centers in late September 2026. The company also launched cybersecurity-focused automation for water utilities and a new decarbonization allyship with Lenovo's 360 Circle community. The software-defined medium-voltage switchgear is being deployed in AI-oriented data centers, including an Equinix pilot. Schneider Electric's narrative projects €56.4 billion revenue and €8.3 billion earnings by 2029, requiring 10.3% yearly revenue growth and a €3.6 billion earnings increase from €4.7 billion today. The forecasts yield a €325.04 fair value, a 7% upside to its current price.

Oracle and We Energies announced a nuclear power subscription deal on October 2, 2026, under which Oracle will take 10-20% of the output from the Point Beach Nuclear Plant, or 125-250 megawatts, for the $15 billion Lighthouse Campus AI data center in Port Washington, Wisconsin. The campus, co-developed by Oracle, OpenAI, and Vantage Data Centers as part of the broader Stargate initiative, is designed to house close to 1 gigawatt of AI capacity within a 1.3 gigawatt total electrical footprint, with completion targeted for 2028. Oracle has committed to fully funding the energy costs for its portion, and the deal is the primary driver of a proposed $176 million electric rate hike for We Energies customers in 2027, accounting for roughly 20% of that increase, while the utility projects the arrangement will save customers approximately $300 million in fuel costs between 2027 and 2033. The subscription, a first-of-its-kind model in Wisconsin, requires approval from the Wisconsin Public Service Commission, which is weighing whether to require tech companies to cover 100% of new power plant costs. The deal is part of an industry-wide pivot in which Big Tech has contracted over 10 gigawatts of new nuclear capacity in the United States over the past year, including Microsoft's 20-year power purchase agreement for the 835-megawatt restart of Three Mile Island, Amazon's acquisition of a nuclear-adjacent Pennsylvania campus for over $650 million and its $500 million investment in X-energy small modular reactors, and Google's order of 500 megawatts from Kairos Power. Data center power demand reached 29.6 gigawatts by late 2025, equivalent to the peak demand of New York state, and the International Energy Agency projects it will rise by 130% by 2030, even as U.S. nuclear output stayed largely flat between 2020 and 2025 and no small modular reactors are under construction in the country.

Amazon has established an $8 billion special purpose vehicle to buy and lease back Nvidia Grace Blackwell chips to AWS clients, an off balance sheet financing tool for AI hardware as cloud capital demands rise. Morgan Stanley has reinstated Nvidia as a Top Pick, citing its view on AI data center demand and capacity expansion. The Amazon vehicle and the Morgan Stanley call both sit inside a wider Nvidia AI infrastructure story, alongside Nvidia's own US$500 billion ecosystem financing plans. The structure lets AWS keep building AI capacity without stacking all the hardware on its own balance sheet, pointing to new funding structures that can support large orders as AI factories become more capital intensive across cloud and sovereign projects. The clearest early signal on whether the model scales would be other hyperscalers or sovereign AI buyers setting up similar chip vehicles that explicitly name Nvidia hardware, with disclosed sizes and timelines.

AI-related investment is generating unusually large spillover effects across the U.S. economy, according to UBS research, with every 1 percentage-point contribution to growth from AI-related investment generating about 1.46 percentage points of investment growth in the following quarter. UBS economist Arend Kapteyn said in a Sept. 25 research note that AI-related technology investment is growing about 30% year over year, while other business fixed investment has risen just 0.8% and residential investment has contracted 3.8%. Of the 1.46 percentage points, 0.46 percentage points comes from continued investment within AI-related categories, 0.30 percentage points through software and research and development, and 0.69 percentage points through other sectors. UBS defines its AI investment proxy across five national accounts categories — electrical transmission and distribution equipment, special industry machinery, computers and peripherals, communications equipment and data centres — which together account for roughly 18% of non-residential fixed investment and about 2.5% of GDP. The spillover is particularly visible in industries supplying data-centre infrastructure, including gas turbines, electricity infrastructure and utilities, and UBS also pointed to less obvious beneficiaries such as Japan's TOTO, whose electrostatic chucks for semiconductor production now account for more than half of the company's profits. The multiplier effect has strengthened over time, with the estimated contribution from a 1 percentage-point increase in AI-related capital expenditure rising from 88 basis points in 1980 to 106 basis points in 2000 and 146 basis points by June 2026.

Dell Technologies has joined a planned $15 billion AI data center initiative in Japan aimed at supporting large-scale AI workloads by pairing new data facilities with dedicated power supply capacity. The company also introduced rugged laptops with on-device AI features and added security aimed at field and mission-critical use cases. Dell's participation plugs directly into its existing hardware, software and support services business, tying its gear more tightly to power-hungry enterprise workloads and edge environments. The move fits a narrative centered on converting a US$95b AI backlog and a large pipeline into deployed systems, while the rugged laptops push AI closer to field workers in sectors such as public safety and energy. Dell still relies heavily on hardware in segments where HP and Lenovo also compete on price and features.

Dell Technologies has booked more than $130 billion of AI server orders over the past year and ended its latest quarter with a $95 billion AI backlog, with its AI customer base expanding to more than 6,500 customers. The company says some AI engagements require more than 50 unique designs as customers optimize systems around workload performance, power, cooling and data-center constraints. Dell's stock has risen more than fourfold since the end of February and trades at just 21.3 times forward earnings, with revenue expected to surge 69% this year before growing another 20% next year. Hedge fund sentiment strengthened in the second quarter, as the number of funds in Insider Monkey's database holding the stock rose from 72 to 77 and the value of those positions more than doubled from about $1.7 billion to $3.6 billion. Dell acknowledges component costs are rising, particularly for memory and storage, and that supply remains constrained across much of the AI infrastructure chain.

Super Micro has started shipping server racks built for NVIDIA's next-generation Vera Rubin AI architecture, sending its shares up 4.3% in the afternoon session. The start of shipments reinforces the server solutions provider's role in the AI infrastructure supply chain, as Vera Rubin-ready racks house and interconnect the high-density systems data centers need for advanced AI workloads. The stock was trading at $43.69, up 4.4% from the previous close. Super Micro is up 41.1% since the beginning of the year, but at $43.69 per share it remains 25.6% below its 52-week high of $58.68 from October 2025. The company's shares are extremely volatile, having logged 69 moves greater than 5% over the last year.

Hewlett Packard Enterprise received a US$1.2 billion order from Vultr for its new AMD Helios AI servers, marking the first commercial deployment of the Helios racks that combine AMD chips with HPE Networking hardware. The deal is an early revenue milestone for HPE's integrated AI data center platform as Helios moves from launch to customer rollout, and it gives the company a concrete reference customer for model training and inferencing workloads. The US$1.2 billion Vultr order is only one piece of the broader Hewlett Packard Enterprise story, and the company's Helios rollout plugs into its wider role supplying infrastructure for AI workloads at cloud providers and enterprises worldwide. The contract lines up with HPE's existing narrative, which leans on AI infrastructure, Juniper networking and higher margin recurring services as key catalysts. Investors will watch whether HPE converts this first Helios deployment into a visible pipeline of similar rack scale AI deals and repeat orders from Vultr, along with disclosed Helios contract wins and networking order intake linked to AI racks.

Nscale, the artificial intelligence infrastructure company, has hired long-time Meta Platforms executive Justin Osofsky as its chief operating officer ahead of its upcoming initial public offering. Osofsky spent 18 years at Meta in various executive roles and was most recently its chief partnerships officer; he will report to Chief Executive Officer Josh Payne and oversee global operations. Nscale, a London-based developer of AI data centers that counts Nvidia and Microsoft among its partners, filed in September for an IPO in New York and is seeking to raise as much as $3 billion, Bloomberg News has reported. Spun off from a cryptocurrency mining operation in early 2024, the company became one of the most prominent data center newcomers powering the AI boom as a so-called neo-cloud operator, renting out computing resources to companies developing AI. Osofsky joins a group of Meta veterans helping lead the company, including former Meta Chief Operating Officer Sheryl Sandberg, who sits on Nscale's board and helped recruit former head of global affairs Nick Clegg as well as former OpenAI executive Fidji Simo.

SpaceX shares have stabilized after notching an all-time low of $104.83 in early August, with the company citing fresh catalysts that could drive the stock higher into year-end and into 2027. Starship, SpaceX's two-stage rocket and the largest human-made flying object, recently reached orbit for the first time, and the recent flight deployed the first 26 next-generation Starlink V3 satellites. Anthropic expanded its computing agreements with SpaceX, committing up to $84.5 billion through May 2029 for NVIDIA-based AI infrastructure, while Zacks Consensus Estimates suggest SpaceX will grow revenue by 147% in 2027. The odds of a SpaceX and Tesla merger by the end of next year have spiked to 63% on Polymarket. The company also pointed to Planet Labs and Google's launch of the first AI satellite as proof of concept for its ambitions in space-based AI data centers.

Amazon on Friday pledged more than $1 billion over five years for US communities that host its data centers, the latest tech giant seeking to win over residents increasingly hostile to the AI-fueled construction boom. The "Built Together" program, unveiled by Amazon Web Services chief Matt Garman in a 3,000-word blog post, will fund free community college degrees and trades training as well as energy-efficiency upgrades for schools and homes. The company also announced a "Data Center Commitment," promising that its facilities will not drive up local electricity bills and that it will publish annual figures on energy and water use. Garman said AWS would stop using nondisclosure agreements with government agencies on its projects, a practice that has become a major flashpoint for the movement opposing data centers; Microsoft made a similar move in March, the first major tech firm to do so. Garman cast the buildout as a national security imperative on par with the 1950s interstate highway system, warning that with more than 100 local moratoriums under consideration nationwide, the United States "could be writing its own losing ticket" in the AI race, and he disputed as false arguments that data centers drive up power costs, deplete water supply and cause pollution from diesel fumes.

Alphabet has unveiled Argon, the flagship of its Gemini 4 generation, months later than promised after the company scrapped Gemini 3.5 Pro, which Sundar Pichai had said would arrive in June. Google says Argon matches OpenAI's Astra and Anthropic's Opus on key coding and cybersecurity tests, though its own results show it trailing on two of the four coding tests included, and the model has no public release date, going only to select cybersecurity partners under the Trump administration's voluntary pre-release access process. Google is competing on price rather than raw capability, pricing Argon at $2 per million input tokens and $10 per million output tokens with cached input tokens discounted 95%, which Jefferies analyst Brent Thill calls a particularly important competitive move at about half the cost of some rival models. The delay came amid leadership churn, with DeepMind founder and chief executive Demis Hassabis stepping aside and several Gemini leaders leaving while Anthropic and OpenAI kept releasing new top models. The infrastructure side is stronger: Google Cloud revenue grew 82% year over year to $24.8 billion in Q2, customers have lined up a $514 billion backlog, cloud now makes up a little more than 20% of Alphabet's revenue, and its cloud market share has climbed to 14% from 12% at the end of 2025, while Alphabet projects 2026 data center capital spending of $195 billion to $205 billion.

UBS said it sees the artificial intelligence infrastructure market staying hot, keeping CoreWeave in focus on Friday. In a note to clients, UBS analyst Karl Keirstead wrote that the per-hour price of hosted Nvidia GPUs is rising, that the environment is inflationary, and that revenue-per-gigawatt figures are moving higher, now a key pillar of the hyperscaler and neo-cloud bull case. Keirstead noted that CoreWeave disclosed this week it raised per-hour GPU pricing across all SKUs by 25% in July 2026 and has since raised them by another 10% in just the last two to three months. He also flagged the risk that local community and state and local government pushback could throttle the pace of AI data center additions, though he said checks suggested the industry will manage through the hurdle. Keirstead added that the availability and cost of capital could slow the AI buildout, with some projects expected to fade, but said better-positioned players including CoreWeave will not have an issue.

Nvidia shares climbed more than 1.5% in early Friday trading after Morgan Stanley restored the chipmaker to its top semiconductor position, while retaining an Overweight rating and a $300 price target. Analyst Joseph sees data-center constraints becoming an important factor for AI infrastructure spending, with limited access to electricity, land and financing potentially favoring Nvidia as customers seek more computing capacity from each unit of available power. Moore also pointed to Nvidia's customer mix, saying about half of its business comes from buyers outside the biggest cloud providers and major AI laboratories, a broader base that includes enterprise customers, system manufacturers and sovereign buyers. The analyst further expects AI agents, which can perform extended sequences of tasks, to create additional computing requirements, while Nvidia's Vera Rubin platform may provide another source of demand as customers expand their AI deployments.

CoreWeave announced availability of the NVIDIA Vera Rubin NVL72, with Cognition as its first production customer, alongside new support for the NVIDIA Vera CPU. Cognition, which uses CoreWeave for training, reinforcement learning and inference, reported that the Vera Rubin NVL72 delivered up to 4.8x higher total token throughput for SWE-2 inference workloads versus a GB200 NVL72 baseline and 3.8x higher output-token throughput for reinforcement-learning workloads. CoreWeave said a Vera rack can contain 128 CPUs and 11,264 cores, theoretically supporting more than 11,000 concurrent isolated environments, and that testing showed more than three times faster agent sandbox startup times compared with an x86 CPU. The company will offer Vera on bare metal using the same operating model and economics as the rest of its infrastructure, aiming to monetize CPU-intensive infrastructure alongside accelerator hours. CoreWeave remains heavily dependent on NVIDIA's technology roadmap and faces competition from hyperscalers and specialized GPU clouds including Microsoft Azure and Nebius Group N.V., which closed four deals in the quarter averaging more than $1 billion each and plans roughly £1.7 billion in U.K. AI compute expansion expected to deliver 65 MW when fully operational in 2027.

NETSTREIT has secured $550 million in additional term loan commitments and amended its existing credit facilities, extending its debt maturity profile and repaying a $200 million term loan due in February 2028. The financing comprises a $100 million increase to its existing 5.5-year senior unsecured term loan, a $50 million increase to its existing 7-year term loan, and a new $400 million senior unsecured 7-year delayed draw term loan. The $100 million and $50 million incremental term loans were funded at closing, while the $400 million facility was undrawn and can be drawn through September 28, 2027. The company said the transactions leave it with no material debt maturities until early 2029 and largely address its debt capital needs through 2027.

Nebius Group has acquired inference optimization startup Inferize and signed new long-term AI data center capacity deals, moves that aim to boost GPU utilization on its Token Factory platform and deepen its role in global AI infrastructure. The company's recent inclusion in the FTSE All-World Index came alongside fresh analyst coverage and large customer commitments from partners including Meta Platforms and Palantir. Nebius Group's narrative projects $29.7 billion in revenue and $2.1 billion in earnings by 2029, requiring 179.9% yearly revenue growth and about a $2.1 billion increase in earnings from $42.4 million today. Some of the most optimistic analysts were already modeling Nebius to reach about US$48.5 billion in revenue and US$1.1 billion in earnings by 2029, well above consensus. The Inferize deal directly targets Nebius's cost base and unit economics by cutting the idle GPU tax on Token Factory, though financing needs and potential dilution remain key risks.

JERA, Japan's major power producer, announced a partnership with Dell Technologies and RHAELM, a UK-based AI infrastructure developer, to develop a Hyperscale AI Data Center project in Chiba Prefecture near Tokyo, valued at approximately 15 billion dollars, with a power capacity of about 400 megawatts. It is positioned as one of the largest single-site AI infrastructure projects in Japan. Meanwhile, Apollo Global Management plans to join as RHAELM's investment and financial partner. A key point of the project is that it will be developed on land adjacent to JERA's Chiba thermal power plant, with JERA providing the site and 400 megawatts of power under a long-term agreement of 15 to 25 years, while RHAELM is responsible for development, construction, management, and financing. Dell will supply rack-scale AI infrastructure. The project is scheduled to come online in phases starting in 2028 and is expected to use the full 400 megawatts of power by 2029. The three companies also aim to create a model that expands to other JERA power plant sites during the 2030s, with the goal of developing total data center capacity at the multi-gigawatt level across Japan.

France's state-owned supercomputer giant Bull restarted its plant in Angers, western France, on the first of the month after completing an equipment upgrade. The facility has doubled its production capacity from six racks a month to twelve, responding to European demand for stronger computing infrastructure for artificial intelligence. The expansion involved an investment of 80 million euros, and the company says capacity can be raised to 24 racks a month by 2027 depending on demand. The European Union plans to invest 7 billion euros between 2021 and 2027 in building out a supercomputer network, and according to Bull, the Angers plant is the only factory in Europe dedicated to manufacturing supercomputers. Chief Executive Emmanuel Le Roux said the company booked a record number of orders this year, winning 15 of the 18 tenders held by EuroHPC, which is jointly funded by the EU and its member states.

Broadcom's Wall Street syndicate is starting to gather $60B of fresh AI chip financing to benefit Anthropic PBC and other companies, according to Bloomberg News. Banks involved in the massive debt package are poised to send out syndication letters for a $42B Class A senior-secured tranche, the report said. Blackstone is said to be leading an $18B tranche of Class B junior debt, committing $9B from various funds with plans to syndicate the rest. The financing, which has been taking shape for weeks, is being closely watched across Wall Street and Silicon Valley for reassurance that investors are still keen to back AI's buildout amid a public backlash against data-center construction. Broadcom is looking to sell more chips and other data-center equipment, challenging Nvidia, while AI companies like Anthropic need ever more computing capacity.

Palantir Technologies and Armada have entered a partnership to provide sovereign AI infrastructure, a move that could expand Palantir's addressable market across enterprise and government AI deployments. Under the deal, Palantir named Armada its inaugural Certified Modular Data Center Partner, integrating its Sovereign AI Operating System with Armada's Galleon modular data centers. The offering lets customers run and adapt open-weight AI models on infrastructure they own and control, with deployments targeted for completion in months rather than years. Financial terms, contract value, and expected revenue contribution were not disclosed.

Singapore state investor Temasek said it is not concerned about the massive investments by major IT companies driving the expansion of artificial intelligence. Jane Atherton, head of its North America business, spoke on the first at the Reuters Investment USA conference in Boston. Amid a global bond selloff and expectations of tighter monetary policy, some have questioned whether Big Tech's huge spending is sustainable in an era of high interest rates, but Atherton expressed the view that companies engaged in infrastructure investment can earn returns commensurate with the risk as AI spreads, stressing that "what is supporting all of this AI infrastructure buildout is a group of companies with some of the strongest balance sheets in the world. We maintain a very positive view on AI." Analysts forecast that Big Tech's capital expenditure will exceed 1 trillion dollars by 2027, driven mainly by funding needs for expensive semiconductors and data centers to meet enormous computing demand. Temasek, which holds stakes in Anthropic and OpenAI, is one of the world's largest state-owned investment institutions with total assets of about 400 billion dollars, and its holdings in the Americas account for 26 percent of its overall portfolio, with BlackRock, Mastercard and Nvidia among its major investments.

Jacobs Solutions has been selected by NVIDIA to deploy its Data Center Digital Twin platform at a large U.S. research and development facility, sending the company's shares up 3.3% in the afternoon session. The work is a three-year SaaS agreement built on NVIDIA Omniverse libraries, under which Jacobs will use the digital twin to optimize energy demand, monitor operations, and balance power loads across the AI facility. The stock closed the day at $139.76, up 3.3% from the previous close. Jacobs Solutions is up 3.3% since the beginning of the year, but at $139.77 per share it is still trading 15% below its 52-week high of $164.44 from October 2025.

The Commerce Department's September 30 release of the final reading for second-quarter real GDP showed growth of 2.2% from the prior quarter, revised up 0.7 percentage points from the preliminary estimate. Driven by a rush to build AI data centers, capital investment rose 9.0% and led overall growth, while personal consumption, which accounts for 70% of GDP, remained solid with 3.8% growth. Meanwhile, the August PCE price index rose 3.4% year over year, below the expected 3.7% increase, and market expectations for an October rate hike receded somewhat. However, changes to the calculation method for some components also pushed the index down, and inflation remains well above the Fed's 2% target; Minneapolis Fed President Kashkari noted it is still too high. Daiwa Institute of Research economist Tsubasa Fujiwara analyzed that if upcoming indicators show strong upward price pressure, it will become difficult for the Fed to hold off at its October meeting.

Nvidia's fair value estimate in the latest analyst model rose to US$327.70 from US$302.83, an increase of about 8%, as research commentary weighed strong AI demand against financing and execution risk. The model now assumes revenue growth of 47.28%, up from 41.23%, while net profit margin was trimmed to 51.62% from 54.98%, future P/E was cut to 21.25x from 25.21x, and the discount rate moved to 11.28% from 10.99%. On the bullish side, Barclays cited a chart implying hyperscaler revenue materially above its base case, Raymond James and BMO Capital lifted their price targets to US$352 and US$340 respectively, and Rosenblatt and Raymond James called the planned Hugging Face acquisition financially modest but important for Nvidia's AI software role. BofA and Goldman flagged rising vendor financing and the new US$500b financing platform as introducing off balance sheet risk and leverage, while Morgan Stanley said further upside may depend on clearer evidence on long term market share, gross margin durability and the eventual impact of Rubin. In company news, Nvidia approved a record US$150b expansion of its share repurchase program, lifting remaining buyback authorization to US$235b through fiscal 2028 and total authorization to about US$391.3b, reported fiscal Q2 revenue of US$96.2b and net income of US$59.7b with around 92% of sales from the Data Center segment, and issued a first full year revenue outlook pointing to approximately 70% growth for fiscal 2028. Nvidia also outlined more than US$500b of compute financing agreements and plans with partners including Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to mobilize over US$500b of third party capital for AI infrastructure over time, agreed to acquire open source AI platform Hugging Face for about US$12.9b, and is in talks to invest up to US$10b in Anthropic while pursuing further investments in AI firms including Perplexity and Decart AI.

A Chinese state-backed financing entity funded the purchase of restricted Nvidia processors, according to a Bloomberg News report citing Chinese regulatory filings. The financing, provided by Semi-Tech Leasing Group Co., supported the purchase of more than 700 servers by AI infrastructure developer Glory View Technology Co., and filings with the People's Bank of China cited by Bloomberg show at least one transaction explicitly funded 32 Asustek Computer Inc. servers powered by Nvidia's high-performance B300 Blackwell chips. Nvidia told Bloomberg it is investigating the matter with its equipment manufacturing partners, while Asus reiterated its commitment to international export control regulations. Semi-Tech Leasing, originally established with backing from China's national semiconductor fund and now predominantly controlled by municipal government entities in Shenzhen and Beijing, has deployed more than 11 billion yuan, or $1.6 billion, into compute infrastructure. Under sale-and-leaseback arrangements initiated in mid-2025, Glory View secured over 3 billion yuan, or $450 million, to expand data center capacity, with most of the procured hardware slated for a major China Mobile Ltd. computing hub in Ningxia. After initial scrutiny, Semi-Tech Leasing resubmitted altered filings to Chinese credit registries that removed specific hardware descriptions, supplier details, and server model designations.

Anthropic has committed $518 billion in capital expenditures over the next 10 years for infrastructure, data centers, GPUs and compute to build, train and inference its models, a figure examined in a new In The Loop segment. The $518 billion commitment is a decade-long total, and the segment notes that Google, Amazon and Meta combined have spent over $250 billion this year alone on data centers, GPUs and compute, with Microsoft and Google each reportedly ramping toward $300 billion next year. The segment frames Anthropic as a small private startup running the same AI capex race as those larger public companies, but without their hundreds of billions in cash balances, meaning the money it has raised and will raise in its IPO will go toward that compute. It cites a $2.2 trillion IPO valuation for Anthropic and argues the spending could be justified if the company captures a large share of AI's spread across multiple industries, while cautioning that the case for buying the IPO rests on whether investors believe that outcome.

Oracle shares reversed an early gain and turned lower Thursday after the Financial Times reported that Tencent Holdings had locked in a five-year lease spanning several Oracle data centers across Southeast Asia. Under the reported terms, Tencent would gain access to about 100,000 cutting-edge AI chips it otherwise can't obtain inside China, with the arrangement said to be worth roughly $7 billion and close to a third of that due upfront. Neither company responded when asked to confirm the details. Oracle stock had jumped nearly 2% before Thursday's opening bell on the report, but the shares gave up that advance amid overall market weakness and were down 0.82% at $136.18 at the time of publication. The deal reflects a broader scramble among Chinese tech firms for overseas computing capacity as domestic chip supply stays tight and U.S. export restrictions persist.

Vertiv Holdings Co. issued third-quarter 2026 guidance calling for net sales of $3.65 billion to $3.85 billion and adjusted earnings of $1.77 to $1.83 per share, with adjusted operating profit projected between $898 million and $938 million and an adjusted operating margin of 24% to 25%. The guidance follows a second quarter of 2026 in which net sales rose 24% year over year, with the Americas and APAC each growing 29%, and adjusted operating margin reached 22.6%, up 410 basis points year over year and above prior guidance. Management raised its full-year 2026 adjusted operating margin guidance to 23.3%-24.3%, citing organic growth, operating leverage and productivity, and said it expects pricing to exceed inflation including current tariffs and countermeasures. Acquisitions contributed 5% of second-quarter 2026 revenues, and Vertiv recently agreed to acquire King Environmental Services, a Europe-based provider of fluid management, commissioning and load-testing services for high-density liquid-cooled data centers, in a transaction expected to close in the fourth quarter of 2026. The Zacks Consensus Estimate for third-quarter 2026 revenues is $3.77 billion, implying growth of 40.79% year over year, while the consensus earnings estimate stands at $1.83 per share, up 0.75% over the past 30 days and implying a year-over-year increase of 47.58%.

The data centre board is preparing to classify data centre investment by electricity usage into 5 tiers, with the official numerical thresholds to be announced next week, along with a central public data dashboard. Digital Economy and Society Minister Chaichanok Chidchob said the meeting of the Data Centre Business Policy Committee, chaired by Deputy Prime Minister and Finance Minister Ekniti Nitithanprapas, is in the process of drawing up detailed criteria and regulatory policy for large-scale data centre investment, including a clear data centre identity system, or DC ID, for every facility. Thailand currently has 26 data centres in operation, 20 under construction, and several more applying for investment promotion from the BOI, though fewer than 100, all of which have signalled intent to reserve more than 20,000 megawatts of electricity in advance, reflecting leapfrog growth. Electricity rates will differ clearly, with operators that choose to buy clean power directly, or Direct PPA Green Energy, receiving a cheaper rate than buying from the regular grid, in order to encourage a shift toward clean energy. At the same time, the board has added a greenhouse gas reduction condition, or Zero, into its core policy, and is pushing for the installation of energy storage systems and setting safety criteria covering electricity usage, water usage, and structural safety, while giving data centres already in operation time to upgrade to the new standards. The first subcommittee will gather and integrate data from the other subcommittees on legal matters, energy criteria, and urban planning criteria, to submit to the main committee and bring to the Cabinet meeting for completion within 1 month, before it is announced as national policy.