Journalist

Kim Seong-hyeon
  • OpenAI Faces Rising Cost Pressures as AI Spending Surges and Rivals Gain Ground
    OpenAI Faces Rising Cost Pressures as AI Spending Surges and Rivals Gain Ground Even after big U.S. tech companies posted first-quarter earnings surprises, market attention shifted to privately held OpenAI. As a “capex shock” — capital spending rising far faster than revenue — ripples across the AI industry, some investors are even discussing bankruptcy scenarios for OpenAI. According to the IT industry on May 3, Evercore and Bank of America raised their forecast for total AI capital spending in 2027 to more than $1 trillion after big tech companies reported first-quarter results. While major U.S. tech companies posted revenue growth of 17% to 33%, their capex growth ran at more than twice the pace. In after-hours trading following earnings, shares of Meta and Microsoft fell sharply. The concern is that it remains unclear when the surge in spending will translate into profits, reinforcing fears of an AI bubble. After those earnings reports, attention focused on OpenAI. The company has signed capital spending commitments totaling $1.4 trillion for 2025 to 2032, including investments by partners such as Azure, AWS, Oracle and CoreWeave. Over the same period, its AI operating and training costs — computing spending — are estimated at $600 billion. Market analysts say OpenAI’s revenue growth is likely to lag that of big tech. Major tech companies can lean on steady income from advertising, cloud services and commerce to support AI investment, but OpenAI lacks that kind of cushion, increasing its risk. Sebastian Mallaby, an economist at the Council on Foreign Relations, wrote in a New York Times opinion piece that OpenAI’s annual losses are widening rapidly and that cumulative cash burn could reach $115 billion by 2029. If current trends continue, he warned, the company could face a liquidity crunch by mid-2027. The argument is that OpenAI’s computing infrastructure is expanding much faster than revenue, and cash could run out before that gap narrows. Similar concerns have been reported inside the company. The Wall Street Journal reported that Chief Financial Officer Sarah Friar formally raised with executives that actual revenue is not keeping pace with the scale of data center computing contracts already signed. Altman and Friar quickly issued a joint statement disputing the report, but the fallout spread beyond OpenAI. Oracle shares fell 7%, and CoreWeave and SoftBank also declined. Bloomberg Intelligence said, “OpenAI missing its revenue targets is a direct risk factor for Oracle’s financial targets.” The episode underscored how dependent the broader ecosystem has become on OpenAI, with even reports of internal friction moving publicly traded tech stocks. Shifts in competition are also adding to OpenAI’s financial pressure. Anthropic has rapidly absorbed enterprise demand since launching its coding agent “Claude Code,” and as of April 7 it had surpassed OpenAI in annualized revenue. Google’s Gemini is gaining share in the consumer market by linking to its search and Android ecosystem. As rivals push into a market OpenAI helped create, OpenAI faces a dilemma: spending more just to hold its position. Some observers say OpenAI has bought time by raising money. In March, OpenAI raised $122 billion from investors including Amazon, Nvidia and SoftBank, securing a valuation of $852 billion. That has led to assessments that near-term bankruptcy risk is limited. Still, if much of that funding has already been committed upfront through data center and computing contracts, the key question is whether revenue can catch up with the pace of expanding capital spending before the next funding round.* This article has been translated by AI. 2026-05-05 13:40:59
  • Big Tech Earnings Beat Expectations, but AI Bubble Fears Persist; OpenAI Faces Financial Scrutiny
    Big Tech Earnings Beat Expectations, but AI Bubble Fears Persist; OpenAI Faces Financial Scrutiny Global Big Tech companies posted strong first-quarter results that underscored the potential to make money from artificial intelligence, but concerns about an AI bubble have not eased. With massive capital spending far outpacing revenue growth, stock-market reactions were mixed, and financial worries surrounding OpenAI — a symbolic name in the AI industry — have intensified. According to the IT industry on May 3, four major Big Tech companies — Alphabet, Meta, Microsoft and Amazon — released their first-quarter results after the market close late last month. The numbers beat expectations. Alphabet reported revenue of $109.9 billion, up 22% from a year earlier and above the market estimate of $107.2 billion. Google Cloud revenue rose 63% to $20.0 billion. Microsoft posted revenue of $82.9 billion, up 18%, with Azure growth reaching 40%. Meta’s revenue climbed 33% to $56.3 billion. Amazon reported revenue of $181.5 billion, up 17%, while AWS revenue grew 28% to $37.6 billion, its fastest growth rate in 15 quarters. Even with the earnings surprise, some stocks fell after companies raised capital spending plans alongside their results. Meta lifted its annual capital expenditure guidance to $125 billion to $145 billion, sending its shares down more than 8% in after-hours trading. Microsoft’s shares slipped about 4% after it reported first-quarter capital spending of $34.9 billion, a 74% jump from a year earlier. Alphabet raised its annual capital spending guidance to $180 billion to $190 billion, and Amazon pledged AI infrastructure investment totaling $200 billion. Estimated AI capital spending this year by the five largest hyperscalers combined is expected to exceed $650 billion. The sharpest focus is on privately held OpenAI. Analysts have said it is losing share to Anthropic in the coding market and to Google’s Gemini in the broader consumer market. Reports of missing revenue targets and internal conflict have added to unease. In particular, OpenAI’s unusually high share of capital spending compared with rivals has emerged as a key variable as it approaches a public listing.* This article has been translated by AI. 2026-05-05 13:40:07
  • Resignation of presidential AI aide raises questions over who will lead 100 trillion won plan
    Resignation of presidential AI aide raises questions over who will lead 100 trillion won plan Ha Jung-woo, the presidential office’s senior secretary for AI Future Planning, submitted his resignation on 28, creating a vacancy at the center of the government’s AI policy push. Attention is turning to who will succeed him and steer the government’s 100 trillion won AI investment plan. Ha stepped down after formally declaring he will run in the Busan Buk-gu Gap parliamentary by-election. Having served as an architect of AI policy since the launch of the Lee Jae-myung government, his departure has drawn industry attention over whether it could affect the execution of the 100 trillion won plan. On 27, as Ha’s future became a public issue, three Naver figures who worked closely with him — Lee Dong-soo, an executive director at Naver Cloud; Kwon Se-jung, a Naver Cloud director; and Park Bae-seong, a leader — said on their social media accounts that they will leave Naver as of April 30. They said they will launch a company as co-founders in the field of “agent AI computing solutions.” Lee said he had long wanted to start a business and that the three decided to move forward together. “We wrapped things up well without any conflict with Naver,” he said, describing the move as a previously expected step. Some in politics see it differently. Because Lee’s departure became official around the same time as Ha’s resignation, speculation has grown that Lee has been tapped as the next senior AI secretary. Lee is described as having strong ties to Ha inside Naver. He holds a doctorate in electrical and computer engineering from Purdue University and previously worked at IBM’s Watson Research Center and as a principal researcher at Samsung Research. At Naver, he is credited with building the foundation of what is now Naver Cloud. He also serves as a member of the Technology Innovation and Infrastructure subcommittee of the National AI Strategy Committee, a government AI policy think tank. Some say his mix of field, research and policy experience fits the presidential AI post. Lee, however, denied any link between Ha’s resignation and the startup plan. “We exchanged messages at the level of encouraging each other in our respective positions,” he said. “Ha’s resignation and this startup are unrelated.” Other names are also being mentioned. Shin Jin-woo, a professor at KAIST’s Kim Jaechul Graduate School of AI who chairs the same subcommittee, has been cited, as has Kim Dae-hyun, head of the AI Center at Samsung Electronics’ Samsung Research and a member of the subcommittee. Moon Yong-sik, president of the National Information Society Agency, is also listed among possible candidates. An academic source said that given the Lee Jae-myung government’s pragmatic approach, someone similar in style to Ha is likely to be appointed. The source added that the shortlist is expected to be narrowed around figures from the National AI Strategy Committee. With the Lee Jae-myung government defining AI as a core national growth engine and pledging a 100 trillion won investment plan, observers say the top priority will be a successor who can ensure policy continuity. * This article has been translated by AI. 2026-04-28 17:24:18
  • Samsung SDS, LG CNS Sign Reseller Deals for OpenAI’s ChatGPT Edu in Korea
    Samsung SDS, LG CNS Sign Reseller Deals for OpenAI’s ChatGPT Edu in Korea Samsung SDS and LG CNS have each signed reseller partner agreements for OpenAI’s education-focused service, ChatGPT Edu, stepping up competition to secure an early lead in the education AI transformation market. The two companies said on the 27th that they secured the right to supply ChatGPT Edu on the same day and will ramp up sales to South Korean universities and other educational institutions. ChatGPT Edu is designed for educational institutions, supporting tasks such as generating lecture materials, organizing research and reports, and providing personalized tutoring. It applies a “non-training” policy under which user conversations and responses are not used as AI training data, strengthening data privacy. Built on the latest GPT-5 language model, it supports text generation, coding, data analysis, document summarization and custom chatbot creation. The service offers the same security environment as ChatGPT Enterprise, allowing use without concerns about sensitive information leaks, and it is also positioned at a more affordable price for schools, the companies said. Arizona State University, the California State University system, Harvard University, the University of Oxford, the University of London, the Wharton School and the National University of Singapore are using ChatGPT Edu, and Estonia has introduced it across its secondary education system to expand AI use by students and teachers. Samsung SDS is conducting a proof of concept with Korea National Open University, which has about 90,000 students, and plans to discuss a formal rollout. The company said it will provide end-to-end AI transformation services through a “One-Team” setup spanning AI consulting, development, operations, cloud and security, supported by its own cloud and GPU-based AI infrastructure. LG CNS said it will run a series of introductory tours and AI education seminars for major universities in the Seoul metropolitan area. It is also reviewing plans to work with universities and OpenAI on developing AI education curricula and operating hackathons. LG CNS said its dedicated ChatGPT organization, the “OpenAI Launch Center,” will handle full-stack services from adoption consulting to technical support, with participation by AI engineers, architects and consultants alongside OpenAI specialists. Both companies pointed to progress in the corporate market. Samsung SDS said it has signed supply contracts with major companies across the public, finance, manufacturing, distribution and service sectors, including Nexen Tire. LG CNS said it began supplying ChatGPT Enterprise in February and has secured about 10 corporate customers in manufacturing, chemicals, finance and biotech. Lee Jeong-heon, vice president and head of strategic marketing at Samsung SDS, said the company will “strengthen our role as an AX partner that designs and scales corporate AI operating systems, beyond being a simple reseller.” Kim Tae-hoon, vice president and head of the AI Cloud Business Division at LG CNS, said, “As we have rapidly expanded customer use cases through our ChatGPT Enterprise business, this Edu reseller agreement expands our business into the education AX field.” * This article has been translated by AI. 2026-04-27 15:34:09
  • LG CNS Signs ChatGPT Edu Reseller Deal to Target South Korea’s Education AI Market
    LG CNS Signs ChatGPT Edu Reseller Deal to Target South Korea’s Education AI Market LG CNS has signed a reseller partnership agreement for OpenAI’s education-focused service, ChatGPT Edu, expanding beyond the ChatGPT Enterprise reseller qualification it previously secured. The company said on the 27th it will begin supplying ChatGPT Edu to educational institutions in South Korea under the agreement with OpenAI. ChatGPT Edu is designed for schools and universities, supporting campus work such as generating lecture materials, organizing research data and reports, and providing personalized tutoring. LG CNS said the service offers the same security environment as ChatGPT Enterprise, allowing use without concerns about sensitive information leaks, and is priced to reduce the burden on educational institutions. Adoption is already accelerating overseas. Arizona State University, the California State University system and Harvard University are using ChatGPT Edu, and Estonia has introduced it across its secondary education system to expand AI use by students and teachers. In South Korea, however, the market is still in the early stages. LG CNS plans to roll out an introductory tour program and AI education seminars for major universities in the Seoul metropolitan area. It is also reviewing plans to work with universities and OpenAI on developing AI education curricula and running hackathons. The company’s dedicated ChatGPT organization, the “LG CNS OpenAI Launch Center,” will provide full-stack services from adoption consulting to technical support. The group includes AI engineers, architects and consultants, along with OpenAI-specialized engineers. LG CNS began supplying ChatGPT Enterprise in February and has since secured about 10 corporate customers across manufacturing, chemicals, finance and biotech, it said. “After rapidly expanding customer use cases through the ChatGPT Enterprise business, this Edu reseller agreement allows us to extend into the education AX field,” said Kim Tae-hoon, vice president and head of LG CNS’ AI Cloud Business Division. “We will support more students and educational institutions in boosting learning and research productivity through AI.” * This article has been translated by AI. 2026-04-27 15:06:32
  • LG Uplus Turns Simple Lab Customer Ideas Into Brand Video Series
    LG Uplus Turns Simple Lab Customer Ideas Into Brand Video Series LG Uplus has released two brand videos built around real customer ideas submitted through its customer-participation service, Simple Lab: “Your Story Arrived at U+” and the follow-up, “A Daughter’s Unfamiliar Voice.” The company rolled out the videos sequentially on April 26 through its official YouTube channel and social media platforms, including its newsroom. Simple Lab allows customers to submit inconveniences they experience while using telecom services, as well as suggestions for improvement. The feature was added to the integrated app “U+one” in November last year, and about 10,000 ideas have been received so far. The videos highlight everyday requests drawn from customer stories, including: an office worker asking for automatic summaries of call content; a child seeking to protect parents from voice-phishing scams; a parent wanting easier tools to manage a child’s first smartphone; and customers asking to understand complex rate plans at a glance. Customers can submit ideas under two tracks: a monthly theme and an open topic. LG Uplus said it reviews ideas that receive strong support first, then shares with customers whether they will be reflected in services and how work is progressing. The company said it plans to expand a people-centered brand narrative through a “Voice Series,” starting with this content. Jang Jun-young, executive director and head of LG Uplus’ marketing group, said Simple Lab is a communication channel for putting the company’s customer-focused philosophy, “Simply. U+,” into practice. He said the company will continue to deliver changes customers can feel, from telecom basics to innovative customer experiences. * This article has been translated by AI. 2026-04-26 14:48:15
  • Anthropic’s Claude Mythos preview accessed without authorization on launch day
    Anthropic’s Claude Mythos preview accessed without authorization on launch day Anthropic’s newest AI cybersecurity model, released under tight controls, was accessed without authorization the same day it was made available to a select group, according to the IT industry. Industry officials said April 26 that Anthropic has opened an investigation after detecting unauthorized access to a preview of its latest model, “Claude Mythos,” through a third-party vendor environment. The access is believed to have occurred on the 7th, when Anthropic announced “Project Glasswing” and began providing the Mythos preview to selected companies. Mythos is described as capable of finding software vulnerabilities on its own and autonomously generating attack code based on them. Anthropic has said the model can detect and exploit vulnerabilities “to a degree that could surpass top-tier human hackers.” Citing dual-use risks — it could be used for attacks as well as defense — the company chose an invitation-only, limited research release rather than a broad public launch. The suspected route was a classic insider-risk scenario. A partner-company employee used their authorized access to enter the Mythos environment and, using information about Anthropic’s file system exposed in a hack of the AI evaluation startup Mercer, inferred where the model was hosted online. The employee then shared access with a small Discord group. The group reportedly avoided cybersecurity-related prompts to evade detection, using the model only for simple tasks such as building websites. The group also claimed it accessed other unreleased models, but that claim has not been confirmed. Anthropic said there is “no evidence the intrusion spread beyond the vendor environment,” and no damage to Anthropic systems has been confirmed. Still, security experts said the incident highlights broader structural weaknesses. Gabriel Hempel, a security operations strategist at security firm Exabeam, said the preview was intentionally limited because of dual-use risks, yet it “leaked almost immediately through a partner environment.” Some analysts said Project Glasswing’s design carries an inherent contradiction. Anthropic limited the Mythos preview to companies including Amazon, Apple, Google, Microsoft, Nvidia, Cisco, CrowdStrike and JPMorgan Chase. U.S. Treasury Secretary Scott Bessent recently convened the heads of major investment banks — including Goldman Sachs, Citigroup, Bank of America and Morgan Stanley — to discuss ways to use Mythos, the report said. The strategy was framed as a nationally designed “controlled release,” but as more institutions participate, the number of people with access inevitably grows, exposing the weakest link. The incident also underscored limits of even advanced code-focused defenses: a model may detect vulnerabilities in software, but it cannot by itself prevent weaknesses in unvetted third-party tools or social-engineering attacks. The use of Mercer breach information to enable access to Mythos was cited as a real-world example of “AI supply chain chain risk,” in which one breach fuels another. The case is also raising questions about whether “only a few get access” remains a workable control strategy in the AI era. Unlike export-control models used in semiconductors and defense, the access itself can become the leak path. South Korea is also watching the debate. Ryu Je-myeong, second vice minister of science and ICT, told reporters at the World IT Show on the 22nd that the government is exploring official participation in security discussions involving Anthropic’s Glasswing and OpenAI, adding that direct information-sharing is needed “in some form.” The Ministry of Science and ICT has been holding a series of emergency issue-review meetings with domestic IT and security industry officials, including SK Telecom, KT, LG Uplus, Naver and Kakao, to assess cybersecurity readiness as AI capabilities advance. 2026-04-26 14:19:48
  • WMO Warns Strong El Nino Possible by Mid-2026, Raising Flood Concerns for South Korea
    WMO Warns Strong El Nino Possible by Mid-2026, Raising Flood Concerns for South Korea The World Meteorological Organization has formally announced that a strong El Nino could develop by mid-2026. South Korean climate experts urged early preparations for possible heavy rain in the country’s south and for water-resource management, with some warning the event could intensify into a “super El Nino.” According to the Science Media Center Korea on Saturday, the WMO said Thursday afternoon (Korea time) that sea-surface temperatures are rising rapidly in the equatorial Pacific, making it likely El Nino will return around May to July 2026. The WMO said the episode is likely to develop into a strong El Nino. El Nino refers to sea-surface temperatures in the central and eastern equatorial Pacific running at least 0.5 degrees Celsius above average. Through ocean-atmosphere interactions, it can disrupt global climate patterns and trigger a mix of extremes depending on the region, including drought, heavy rain, heat waves and shifts in hurricane activity. In its outlook for May to July, the WMO projected above-average surface temperatures in nearly all regions, with especially pronounced warmth expected in the southern United States, Central America, the Caribbean, Europe and North Africa. Rainfall signals vary widely: precipitation is expected to increase in southern South America, the southern United States, the Horn of Africa and parts of Central Asia, while drought is forecast for Australia, Indonesia and parts of South Asia. During the Northern Hemisphere summer, hurricane activity tends to increase in the central and eastern Pacific, while Atlantic hurricane formation is typically suppressed. The WMO cautioned that uncertainty remains in intensity forecasts because of the “spring predictability barrier,” adding that forecast accuracy improves after April. In South Korea, scientists are debating whether the event could grow to a record scale. Kuk Jong-seong, a professor in the Department of Earth and Environmental Sciences at Seoul National University, said the odds of a strong El Nino are “very high,” citing unusually warm tropical oceans, heat buildup in the western Pacific warm pool and the potential for ocean-atmosphere coupling. He said it would be difficult to rule out a super El Nino if westerly wind bursts occur repeatedly in the western and central Pacific in April and May, pushing the Nino 3.4 index to 2.5 degrees Celsius or higher. During the 2015 super El Nino, widespread abnormal weather was observed, including global heat, drought and heavy rain. Yeasook Oh, a professor in the Department of Climate and Energy Systems Engineering at Ewha Womans University, said El Nino events that develop in summer do not affect South Korea in a consistent way and can be highly variable. He said variability in rainfall and temperature is likely to be greater than normal. Huh Chang-hoi, also a professor at Ewha Womans University, countered—citing his own research—that South Korea’s climate is not strongly influenced by El Nino or La Nina. Experts also urged caution in timing any assessment of impacts. Koo Ja-ho, a professor of atmospheric sciences at Yonsei University, and Park Jae-heung, a research professor at Seoul National University, noted that El Nino typically develops in summer and fall and peaks in winter. They said regional impacts such as temperature and precipitation should be evaluated based on the winter of December 2026 to February 2027, not May to July 2026. Climate specialists warned that any El Nino developing on top of already elevated global temperatures could have broader effects than in the past. * This article has been translated by AI. 2026-04-26 11:09:16
  • Google Unveils 8th-Gen TPU Chips, Challenging Nvidia as AI Accelerator Market Shifts
    Google Unveils 8th-Gen TPU Chips, Challenging Nvidia as AI Accelerator Market Shifts Google has unveiled what it called its highest-performing AI-only chips yet at its annual tech conference, taking direct aim at Nvidia’s dominance. Analysts say that as the AI accelerator market diversifies beyond Nvidia, Samsung Electronics and SK hynix stand to gain because they supply the key memory used in those chips. According to the IT industry on the 23rd, Google Cloud on April 22 (local time) introduced two eighth-generation Tensor Processing Units at “Google Cloud Next 2026” in Las Vegas: the training-focused TPU 8t and the inference-focused TPU 8i. Both are scheduled for official release within the year. TPUs are AI-specific application chips (ASICs) that Google co-designed with Broadcom. Unlike general-purpose graphics processing units, they are optimized for AI workloads and are widely viewed as more power-efficient. The key change in the eighth-generation TPU is a design that separates training and inference for the first time. It is the first major architectural shift since Google first launched the TPU in 2015. The TPU 8t uses a “superpod” design that can link up to 9,600 chips in a single system. Google said each pod delivers 121 exaflops of performance and trains models three times faster than the previous generation, with double the power efficiency. Through its Pathways platform, Google said more than 1 million TPUs can be pooled to operate like a single cluster, cutting development time for very large AI models from months to weeks. The TPU 8i targets demand tied to agentic AI. Google said it delivers 9.8 times the performance of the seventh generation and includes 384MB of on-chip SRAM — more than triple the prior generation — along with 288GB of high-bandwidth memory (HBM). Market share remains heavily tilted toward Nvidia. Google’s TPUs account for about 5% of the AI accelerator market, while Nvidia’s GPUs hold about 92%, according to the report. Nvidia’s next-generation “Vera Rubin” platform has already entered mass production and is set for commercial release in the second half of this year. Still, conditions are shifting in Google’s favor as a prolonged shortage of Nvidia GPUs pushes customers to seek more cost-effective alternatives. Amazon Web Services is also pursuing a strategy to reduce reliance on Nvidia by promoting its own AI accelerators, including Trainium for training and Inferentia for inference. Hedge fund Citadel has built quantitative research software using Google TPUs, and 17 U.S. national laboratories under the Department of Energy are operating TPU-based “AI co-scientist” software, the report said. The trend is drawing attention to South Korean chipmakers. Each TPU carries six to eight HBM stacks, and supply of HBM for Google’s TPUs has effectively consolidated around Samsung Electronics and SK hynix. In Nvidia’s GPU supply chain, Samsung, SK hynix and Micron compete, but Micron has effectively fallen out of the TPU camp due to limited production capacity, the report said. The industry expects the eighth-generation TPU to use HBM4, a sixth-generation HBM standard, and forecasts Samsung’s HBM shipments to Google will more than double from this year. SK hynix, described as Google’s preferred supplier, is currently supplying HBM3E and is also reported to be the exclusive supplier of 12-high HBM3E for the power-efficiency-improved TPU 7e. * This article has been translated by AI. 2026-04-23 14:22:22
  • Amazon’s New Anthropic Commitment Sharpens Split Market View Ahead of IPOs
    Amazon’s New Anthropic Commitment Sharpens Split Market View Ahead of IPOs Market sentiment is sharply diverging over two AI giants expected to pursue IPOs, a split underscored by Amazon’s latest move involving Anthropic. Amazon has agreed to a conditional commitment of up to $25 billion in additional investment in Anthropic, according to the IT industry on Monday. The amount is more than triple Amazon’s existing $8 billion investment, but it is not a finalized disbursement. Investor demand for Anthropic has already become visible. In February, the company ran a $30 billion funding round that drew subscriptions more than six times its initial target, leaving demand outstripping supply. Some venture capital firms have also been reported to have approached the company with interest based on an $80 billion valuation. Driving the rush is competition to secure stakes ahead of a potential IPO. Anthropic is widely seen as likely to list on U.S. markets in the fourth quarter of this year. Goldman Sachs, JPMorgan and Morgan Stanley have been mentioned as potential underwriters, and Chris Liddell, described as a former Microsoft chief financial officer, has joined the board. The fundraising target is more than $60 billion; if completed, it would be the second-largest IPO on record after SpaceX, the report said. That prospect is pushing institutions to seek pre-IPO shares. OpenAI is also aiming for a fourth-quarter listing, but its internal situation differs. The report said there are disagreements between the CEO and CFO over timing, and some observers believe the IPO could slip to 2027. The market’s temperature toward the two companies is diverging most clearly on valuation. Industry sources said OpenAI would need a post-IPO valuation of at least $1.2 trillion to justify investment, while Anthropic’s current valuation is $380 billion. That implies OpenAI’s threshold for investors to recoup their money is more than three times Anthropic’s valuation. Signals from the private market have been more direct. Bloomberg reported on March 1 that about six large institutions, including hedge funds and venture capital firms, approached the over-the-counter market seeking to sell $600 million worth of OpenAI shares. Some of that stock was said to be difficult to sell. By contrast, demand for Anthropic shares has been described as close to unlimited. Amazon’s positioning adds another layer. Amazon, Anthropic’s largest shareholder since it began investing in 2023, also struck a deal in February involving a $50 billion investment in OpenAI and a $10 billion AWS cloud contract. The latest Anthropic commitment is being read not as a bet on a single model, but as an extension of a cloud strategy to capture AWS infrastructure demand regardless of which company wins. Anthropic, through the deal, has agreed to spend more than $100 billion on AWS over the next 10 years. 2026-04-21 14:45:59