By Yun Park ( January 30, 2025, 5:15 PM EST) — Satellite communications company Ligado Networks LLC’s largest unsecured creditor asked a Delaware bankruptcy judge to reject the company’s proposed $115 million Chapter 11 financing package, saying Ligado’s secured lenders were seeking to help themselves to $100 million in fees as part of the deal…. Law360 is on it, so you are, too. A Law360 subscription puts you at the center of fast-moving legal issues, trends and developments so you can act with speed and confidence. Over 200 articles are published daily across more than 60 topics, industries, practice areas and jurisdictions. A Law360 subscription includes features such as Daily newsletters Expert analysis Mobile app Advanced search Judge information Real-time alerts 450K+ searchable archived articles And more! Experience Law360 today with a free 7-day trial. source

A joint venture called The Stargate Project will contribute $500 billion over four years to generative AI infrastructure in the U.S., representatives of SoftBank, OpenAI, and Oracle announced in partnership with President Donald Trump on Jan. 21. The joint venture will support infrastructure, including data centers, contributing to what OpenAI calls a “computing system.” “This infrastructure will secure American leadership in AI, create hundreds of thousands of American jobs, and generate massive economic benefit for the entire world,” OpenAI wrote in a Jan. 21 post on X (formerly Twitter). “This project will not only support the re-industrialization of the United States but also provide a strategic capability to protect the national security of America and its allies.” “It’s big money and high-quality people,” Trump said at a White House press conference, according to the Associated Press. At least one location confirmed in Texas OpenAI will initially contribute $100 billion to the project, with the other $400 billion coming at an undisclosed pace over the next four years. OpenAI said building a data center associated with the project has already begun in Abilene, Texas. “We are evaluating potential sites across the country for more campuses as we finalize definitive agreements,” OpenAI wrote. Oracle Chairman and CTO Larry Ellison said 10 data centers were already built or under construction. SEE: AI adoption tends to weaken companies’ chances of meeting sustainability goals, according to a report released in January. More must-read AI coverage Which companies are involved? The initial equity funders are: Japanese telecommunications company SoftBank. OpenAI. Oracle. MGX, a technology investment firm located in the United Arab Emirates. SoftBank will be responsible for The Stargate Project’s finances, while OpenAI will handle operations. Other initial technology partners are: “This builds on a deep collaboration between OpenAI and NVIDIA going back to 2016 and a newer partnership between OpenAI and Oracle,” OpenAI wrote. Trump’s administration will ease the way for more data centers in the United States, he said on Tuesday, according to The New York Times. That easement may include unspecified “emergency declarations” around The Stargate Project potentially generating its own electricity. Building out AI involves rethinking data and power needs “Our current infrastructure is not ready for the demands AI will require for full maturation,” said Sean Tufts, managing partner for critical infrastructure and operational technology at Optiv, in an email to TechRepublic. “This team is a perfect trifecta to embolden a new ecosystem. Bringing together the boldest AI firm, one of the largest data and cloud companies, and one of the most innovative financiers.  This is the type of public/private partnership that America’s innovators thrive on.” Tufts suggested a power generation company should join the group to address electrical needs. In addition to chips, robust data centers, and more efficient cooling, he said, power is one of the pillars required for increased support for generative AI technologies. Government increases reliance on public sector partnerships The Trump administration walked back former President Joe Biden’s initiative to provide guidelines for “safe, secure and trustworthy” generative AI, listing it among a group of what the Trump White House called “unpopular, inflationary, illegal, and radical” decisions by the previous administration. Tech companies that may be impacted by the rapidly shifting government support for AI should keep abreast of the changes, including those that may affect international deployments or partner companies outside the U.S.. States may also issue individual executive orders regarding AI and its infrastructure. State-level AI mandates, such as those in New York and Colorado, will remain in place. “Look for patterns in new or emerging state AI laws that follow existing non-AI laws,” wrote Gartner analysts Lydia Clougherty Jones, Frances Karamouzis, Svetlana Sicular, Avivah Litan in a Jan. 14 white paper. “For example, AI algorithmic discrimination laws are often duplicative of existing non-AI laws that prohibit certain discriminatory actions.” The removal of the 2023 executive order will mean “the dilution of significant federal oversight of model development,” Jones told TechRepublic in an email, “including requirements to submit per-market safety training results or notice of large-scale computer cluster acquisition, paving the way for streamlined innovation in a less federally regulated environment.” Projects like Stargate show that other AI initiatives may land in private hands rather than public ones. “Regardless of the status of the 2023 EO, strengthening sovereign AI will increase reliance on partnerships with the private sector,” said Jones. She pointed out that another major AI-related executive order, Advancing United States Leadership in Artificial Intelligence Infrastructure, was not revoked. Making the U.S. a powerhouse of AI innovation is still a goal for the new administration. “Given the substantial amount of energy infrastructure needed to train AI models, the challenge to accelerate AI development lies in the massive energy demand and consumption,” Jones said. “With the revocation of the 2023 EO and announcement of Stargate, we see the convergence of data centers and power — power being both literally in terms of energy demand and figurately in terms of the US’s competitive ambitions.” source

00:00 Welcome to CIO Leadership Live. I’m Lee Rennick, executive director of CIO communities for cio.com, and we’re here right now at the CIO 100 Symposium and Awards. And I’m with Bob McCowan, senior vice President and Chief Information Officer, Regeneron. Thanks so much for joining me here today, Bob. It’s been a really active few days here at the conference.00;00;30;07 – 00;00;55;07UnknownCould you please introduce yourself and maybe tell us a little about your current role? Yeah, absolutely. It’s great to meet you, Lee. So Bob McCowan, I’m the CIO at Regeneron Pharmaceuticals, and, I’ve been in the role six years. My role there is fairly traditional for CIO. But unlike the title, it’s, covering things from data to transformation to, technology.00;00;55;07 – 00;01;18;49UnknownSo it’s a pretty all encompassing role. I actually joined the company prior, so I was promoted to CIO from within Regeneron, which has its advantages. Right? Some disadvantages. Right. But having run the infrastructure, we had started on a whole modernization program. So going in CIO, I was really able to pick up on that and move forward.00;01;18;54 – 00;01;38;24UnknownYeah. And a lot of CIOs talk about the relationship they have the C-suite, the other C-suite and the board of directors. So certainly it seems like you had that knowledge base come in into the company that’s been put up into the role of the CIO. Yeah, absolutely. It definitely helped. And I think, Regeneron’s a founder led organization.00;01;38;24 – 00;02;00;25UnknownRight. They’re still there. Right. And a terrific job. Yeah. And it creates a culture, I think, where, there’s a little bit more risk taking than other organizations I work for because of the science based approach and the whole, approach is now you do well by doing good, right? That creates, type of excitement that I just didn’t see in some other organizations.00;02;00;30 – 00;02;23;03UnknownOh, fantastic. Well, congratulate to Regeneron for winning the CIO 100 Award. We’re having the gala tonight. It’s going to be fantastic to celebrate all of this amazing technology that people like yourselves are really embracing and pushing forward in their organizations. Could you tell me a little bit about the award winning project? It’s called Centralized Data Platform Using Data to Uplift Science, so I’d love to learn more about it.00;02;23;08 – 00;02;52;39UnknownYeah, absolutely. And, the centralized data platform is, I guess, the result of a lot of investments in the past few years. But, it’s at the simplest level, what we do is we research and find drug candidates. We, then go to what we call PMP, which is, pre manufacturing, and they learn how to scale it up, but they then have to transfer it into, tech transfer to manufacturing at large scale.00;02;52;50 – 00;03;16;36UnknownRight. And so there’s a whole process from how you take it from research to the, to what we call AIOps industrial operations. And we have 12 FDA approved products. Right. Clearly, this group know what they’re doing. Yeah. But the opportunity arose to really go back and look and look at the process and how they engaged and what this does.00;03;16;36 – 00;03;36;41UnknownIt creates a essentially a data platform that takes all the data from those processes. Provides it in a way that, each person in the step of the process can get access to it. Yeah. And in doing that, we started to see that those groups sometimes spoke a different language. They talked about the same data, but in different ways.00;03;36;42 – 00;03;57;18UnknownRight. The way they transferred data was now legacy based, PowerPoint slides. Excel. Yeah, yeah. Or you needed to know who to talk to. And so by capturing this data in the platform, we simplified that process and made it much smoother. And, and and the biggest success, to be honest, was bringing those individuals and subject matter experts together.00;03;57;23 – 00;04;17;10UnknownAnd it was able help them. Yeah. Able to help them, empower them to put this program in place. It sounds fantastic. Now, a lot of great outcomes as a result of this. Yes. And, we’re looking at other areas we can take. Well, it sounds fantastic. And congratulations. Well, I do speak with a lot of CEOs.00;04;17;10 – 00;04;34;21UnknownI interview them and we have roundtables, and we talk a lot about cloud and edge to cloud computing and just managing the data. Right. Making sure it’s in the right place. Sometimes, you know, people are talking to me about bringing data back on prem. I spoke to one person who said to me, we just don’t do that at all because of the way our business has been structured.00;04;34;26 – 00;04;55;29UnknownBut, you know, I would love to learn some of your process. And looking at central, this centralizing data and then, you know, looking at that productivity and process, any insights you could provide to other CIOs are tech leaders listening in. So not everyone might want to hear this, but it takes a long time. Yeah. And I mentioned earlier I was helping, modernize the infrastructure.00;04;55;29 – 00;05;20;08UnknownYeah. And a big part of that was actually transition to cloud. Right. We took a native cloud approach and moved probably 60, 70% of everything we do to cloud. But we did that very thoughtfully. So we identified what made sense to stay on premise. And then in the move to cloud, we also, refactored it and redesigned it to make sure we took the benefits.00;05;20;12 – 00;05;41;24UnknownSo once you get it into the cloud, you suddenly realize that you can deal with much bigger data sets. You can connect to data, right? This idea of connected data comes into play. And so when you start building on that and in fact this is our fourth CIO 100 award in five years. Wow. And when you go back and look at them all the same is data.00;05;41;24 – 00;06;04;26UnknownData right. Right. And so the approach we took was we’ve got to have that, data platform that is

By David Minsky ( January 30, 2025, 11:37 PM EST) — A Florida district judge adopted a magistrate judge’s recommendation to dismiss a federal lawsuit brought by Jane Doe, an alleged victim who claimed OnlyFans profited from a video of her rape, but sent the case back down to state court where the remainder of her allegations will be tried…. Law360 is on it, so you are, too. A Law360 subscription puts you at the center of fast-moving legal issues, trends and developments so you can act with speed and confidence. Over 200 articles are published daily across more than 60 topics, industries, practice areas and jurisdictions. A Law360 subscription includes features such as Daily newsletters Expert analysis Mobile app Advanced search Judge information Real-time alerts 450K+ searchable archived articles And more! Experience Law360 today with a free 7-day trial. source

Security researchers from Georgia Institute of Technology and Ruhr University Bochum discovered two side-channel vulnerabilities in devices with Apple name-brand chips from 2021 or later that could expose sensitive information to attackers. Specifically, the vulnerabilities known as SLAP and FLOP skim credit card information, locations, and other personal data. Data can be gathered from sites like iCloud Calendar, Google Maps, and Proton Mail via Safari and Chrome. As of Jan. 28, Apple is aware of the vulnerabilities. “Based on our analysis, we do not believe this issue poses an immediate risk to our users,” an Apple representative told ArsTechnica. According to the researchers, Apple plans to release a patch at an undisclosed time. The researchers have not found evidence of threat actors using these vulnerabilities. Which Apple devices are affected? The following Apple devices include vulnerable chips, according to the researchers: All Mac laptops from 2022 to the present (MacBook Air, MacBook Pro). All Mac desktops from 2023 to the present (Mac Mini, iMac, Mac Studio, Mac Pro). All iPad Pro, Air, and Mini models from September 2021 to the present (Pro 6th and 7th gen., Air 6th gen., Mini 6th gen.). All iPhones from September 2021 to the present (all iPhone 13, 14, 15, and 16 models, SE 3rd gen.). What are the SLAP and FLOP vulnerabilities? Both vulnerabilities are based on speculative execution, a cyberattack technique that uses indirect cues such as power consumption, timing, and sounds to extract information that would otherwise be secret. Contemporary Apple chips inadvertently enable speculative execution attacks because they use predictors that optimize CPU usage by “speculating.” In the case of SLAP, they predict the next memory address the CPU will retrieve data from. In FLOP, they predict the data value returned by the memory subsystem on the next access by the CPU core. SLAP enables an attacker to launch an end-to-end attack on the Safari web browser on devices with M2/A15 chips. From Safari, the attacker could access emails and see what the user has been browsing. FLOP lets threat actors break into Safari and Chrome web browsers on devices with M3/A17 chips. Once inside, they could read the device’s location history, calendar events, and stored credit card information. SEE: Chinese company DeepSeek released the most popular AI chatbot on the App Store this week, ahead of OpenAI. “There are hardware and software measures to ensure that two open webpages are isolated from each other, preventing one of them form (maliciously) reading the other’s contents,” wrote researchers Jason Kim, Jalen Chuang, Daniel Genkin, and Yuval Yarom on their Georgia Tech site about SLAP and FLOP. “SLAP and FLOP break these protections, allowing attacker pages to read sensitive login-protected data from target webpages. In our work, we show that this data ranges from location history to credit card information.” The research highlights the dangerous potential of side-channel attacks, which both SLAP and FLOP take advantage of. Side-channel attacks are difficult to detect or mitigate because they rely on properties inherent to the hardware. In March 2024, Apple silicon ran afoul of another side-channel attack called GoFetch. Must-read security coverage What can users do about the vulnerabilities? Users can’t apply mitigations to these vulnerabilities, since the vulnerabilities are rooted in the hardware. “Apple has communicated to us that they plan to address these issues in an upcoming security update, hence it is important to enable automatic updates and ensure that your devices are running the latest operating system and applications,” the researchers wrote. TechRepublic has reached out to Apple for more information. source

Large shares of adults in most of the 36 countries we surveyed say religion helps society rather than harms it. Most also say religion encourages tolerance, not intolerance. But people are slightly more divided about whether religion encourages superstitious thinking. Religion is generally seen more positively by: People in middle-income countries, compared with those in high-income countries Religiously affiliated people, compared with those who are unaffiliated People who say they pray daily, compared with those who pray less often Does religion help society? Views of religion’s impact on society are broadly positive. A 36-country median of 77% say religion mostly helps society, while a median of 19% say it mostly hurts. Views are particularly positive in parts of Asia, the Middle East and Africa. For example, at least 90% of adults surveyed in Bangladesh, Indonesia, Kenya, Malaysia, Sri Lanka, Thailand and Tunisia say religion helps society. Elsewhere, people are somewhat more divided. Around six-in-ten or more adults in some high-income countries – Chile, Greece, Hungary, Israel, Italy, Poland, Singapore, South Korea and the U.S. – see religion as a positive force. But in Australia, the Netherlands, Sweden and the United Kingdom, around half or more say religion hurts society. Does religion encourage tolerance? People also broadly view religion as encouraging tolerance rather than intolerance. Adults in middle-income countries have more positive views of religion than adults in high-income countries. Nearly all Tunisians (98%) and Indonesians (95%) say religion encourages tolerance. So do majorities across most countries surveyed in South and Southeast Asia, Latin America, and sub-Saharan Africa. High-income nations are divided on whether religion encourages tolerance or intolerance. Large shares in some high-income countries, including Singapore (79%), Hungary (72%), Italy (70%) and Israel (70%), say religion encourages tolerance. At the same time, small majorities in Sweden (62%), Germany (57%), the Netherlands (57%), the UK (57%) and Australia (56%) say religion encourages intolerance. Does religion encourage superstitious thinking? Globally, people are more divided when it comes to whether religion encourages superstition. A 36-country median of 52% say religion does not encourage superstition, while 42% say it does. While people in middle-income countries are relatively more positive about religion’s impact on superstitious thinking than those in high-income ones, the gap is less pronounced than on the other two questions about religion’s impact on society. For example, three-quarters of adults or more in middle-income countries like Bangladesh, Indonesia, Malaysia, Thailand, Tunisia and Turkey say religion does not encourage superstitious thinking. But in other middle-income countries – such as Argentina, Brazil, Colombia and the Philippines – views of religion’s impact are more evenly divided. Among the high-income countries, Hungary, Italy, Poland and Singapore stand out as the only places where majorities say religion does not encourage superstition. In the other high-income countries surveyed, much smaller shares say this. And majorities in Australia, France, Germany, the Netherlands, Spain, Sweden and the UK say religion encourages superstition. How do views of religion’s impact on society differ? Overall, people who pray at least daily are more positive about the impact of religion on society than those who pray less often. This is consistent across all three of the questions about religion and public life discussed in this chapter. Similarly, people who say they belong to a religion themselves also tend to see religion in more positive terms – again, across all three questions – than those who are religiously unaffiliated (those who identify as atheist, agnostic or “nothing in particular”). Views by religiousness People who say they pray daily are more likely than others to say religion helps society, encourages tolerance and does not encourage superstitious thinking. In many cases, these differences are sizable. For example, 85% of Australians who pray at least once a day say religion mostly helps society, compared with 37% of Australians who pray less frequently. These differences tend to be particularly large in less religious, high-income countries like the Netherlands and France. The differences are generally less pronounced or not significant in highly religious, middle-income countries like Bangladesh and Malaysia. Views by religion The religiously affiliated tend to be more positive about religion than the unaffiliated in all countries where this analysis is possible – and across all questions asked about religion’s impact on society. For example, in Peru, 82% of Christians say religion encourages tolerance, compared with 53% of Peruvians who do not have a religious affiliation. Peru exemplifies another pattern: In some middle-income countries, even the unaffiliated express mostly positive views of religion. Across the 36 countries surveyed, certain religious groups stand out. Muslims are generally the most likely to say religion helps society, encourages tolerance and does not encourage superstition. For example, nearly all Muslims in Tunisia (99%) and Indonesia (95%) say religion encourages tolerance. But it’s also the case in countries where Muslims are in the minority: Muslim Israelis are more likely than Jewish Israelis to say religion encourages tolerance and helps society. In the U.S., both Muslims (83%) and Christians (77%) overwhelmingly agree that religion encourages tolerance, compared with 52% of Jewish Americans and 32% of religiously unaffiliated Americans. In India, similar majorities of Hindus and Muslims say that religion encourages tolerance. Only in Sri Lanka are Muslims significantly less likely than other religious groups to say religion encourages tolerance, even though a large majority of Sri Lankan Muslims (75%) take this position. Particularly in the Sub-Saharan African, Latin American and Asia-Pacific countries surveyed, large majorities of Christians also generally say religion helps society and encourages tolerance. For example, 96% of South Korean Christians say religion mostly helps society. In some European countries, though, Christians are less positive about religion’s role in society – even if they remain more positive than unaffiliated people. In Sweden, only about one-third of Christians say religion encourages tolerance, and 54% of Swedish Christians say religion encourages intolerance. Buddhists across Southeast Asia overwhelmingly say religion helps society and encourages tolerance. This is especially the case in Sri Lanka, where at least nine-in-ten Buddhists agree with each statement. Smaller majorities of Buddhists

By Christopher Cole ( January 28, 2025, 8:24 PM EST) — The Federal Communications Commission’s new Republican chief said Tuesday the agency was dropping a plan launched during the Biden administration to more rigorously vet cell tower building projects for environmental impacts…. Law360 is on it, so you are, too. A Law360 subscription puts you at the center of fast-moving legal issues, trends and developments so you can act with speed and confidence. Over 200 articles are published daily across more than 60 topics, industries, practice areas and jurisdictions. A Law360 subscription includes features such as Daily newsletters Expert analysis Mobile app Advanced search Judge information Real-time alerts 450K+ searchable archived articles And more! Experience Law360 today with a free 7-day trial. source

In the high-stakes world of B2B sales, no path to a closed deal is perfectly linear. According to Forrester’s Demand, ABM, And Customer Marketing Survey, 2024, 56% of opportunities handed off to sales fail to close successfully. The preponderant focus that still exists on identifying new opportunities in demand and account-based marketing (ABM) means that revisiting stalled or detoured deals has significant untapped potential. Adaptive programs provide a structured approach to reviving these opportunities, ensuring that they’re not prematurely discarded and that marketing and sales investments yield maximum returns. Why Opportunities Stall, And Why They Shouldn’t Be Abandoned Deals can detour for various reasons, including shifting buyer priorities, budget constraints, or internal changes within the buying organization. When this happens, these opportunities are often deprioritized and fall out of the sales pipeline altogether. But this doesn’t mean they lack potential. Frontline marketing teams can effectively develop strategies to reengage buyers using insights gained during the buying process. Capturing and sharing first-party data — such as updated time frames, missing features, or evolving business needs — ensures that these stalled opportunities remain within the broader B2B Revenue Waterfall™, ready for reactivation when the time is right. Designing Adaptive Programs To Restart Deals That Detour To revitalize stalled deals, organizations must align marketing, sales, and customer-facing teams around shared insights and adopt an adaptable and automated approach to buyer engagement. Adaptive programs leverage standardized reason codes to diagnose why opportunities have stalled and determine the appropriate actions for reengagement. These codes clarify the reasons, context, and next steps for detoured opportunities, ensuring targeted treatments and messaging. Businesses can effectively address setbacks and move opportunities back into the sales pipeline with greater precision and impact by systematically reactivating and recycling deals through tailored demand programs. Accelerating Deals With Program Plays Forrester’s framework emphasizes using specific acceleration program plays to tackle the core reasons for detours. Each play is designed to address a particular buyer concern: Budget. Demonstrate the value of the solution to obtain the necessary budget. Authority. Identify and engage unresponsive or new buying group members and ensure that they align with other buying group members. Need. Showcase how the solution addresses critical business challenges and demonstrate the value of the solution’s functionality. Urgency. To reinforce the importance of timely action, prioritize and create a sense of urgency around the buyer’s challenge ahead of other challenges the business faces. Closing The Gap In B2B Revenue By adopting adaptive programs, businesses can reduce inefficiencies, maximize their marketing ROI, and increase the likelihood of closed deals. Detoured opportunities shouldn’t represent failure but rather an opportunity to refine your approach, leveraging data and insights to meet buyers where they are in their journey. Read the Forrester report, Harnessing Adaptive Programs To Revitalize Deals That Detour, to explore these strategies in greater depth and learn actionable steps for implementation. This comprehensive report provides insights into how adaptive programs can transform stalled deals into successful outcomes. Forrester clients can also schedule a guidance session or inquiry to discuss this important topic for all B2B marketers. source

Over the weekend, Chinese AI company DeepSeek released an AI chat app including a “reasoning” AI model comparable to OpenAI’s o1, causing a stir among American AI companies as DeepSeek rose to the top of Apple’s App Store. DeepSeek is a Hangzhou, China-based company providing generative AI models and AI integration. Its first products to make waves in the American market are the GPT-4-like DeepSeek-V3 and R1, an advanced “reasoning model.” Like ChatGPT, DeepSeek-V3 and R1 quickly answer natural-language prompts. NVIDIA and Microsoft stock fell on Monday after the buzzy debut. Overall, the stock market reflected a sudden dip in confidence in U.S. AI makers. DeepSeek’s success sparked conversation about whether U.S. restrictions on Chinese access to AI chips limited or encouraged competition. For tech professionals, DeepSeek offers another option for writing code or improving efficiency around day-to-day tasks. Along with DeepSeek’s R1 model being able to explain its reasoning, it is based on an open-source family of models that can be accessed on GitHub. More must-read AI coverage What is remarkable about DeepSeek? Like OpenAI’s o1 (formerly known as Strawberry), the reasoning model slows down its prediction capabilities to “reason through” its work, which helps it provide more accurate answers. In particular, reasoning models have scored well on benchmarks for math and coding. DeepSeek said DeepSeek-V3 scored higher than GPT-4o on the MMLU and HumanEval tests, two of a battery of evaluations comparing the AI responses. DeepSeek said one of its models cost $5.6 million to train, a fraction of the money often spent on similar projects in Silicon Valley. DeepSeek-V3 and R1 can be accessed through the App Store or on a browser. Visitors to the DeepSeek site can select the R1 model for slower answers to more complex questions. When selected, the R1 model creates lengthy answers that explain in a conversational style how it arrived at its conclusions. As of Monday morning, the DeepSeek chat site warned service may be disrupted, though the chatbot was functioning normally. DeepSeek also offers an APII, which operates through the OpenAI SDK or software compatible with the OpenAI SDK. SEE: OpenAI announced Operator, an AI agent that can take multi–step actions in a web browser, such as choosing flights. What does DeepSeek’s V3 and R1 launch mean for the AI industry? “We can fully expect an ecosystem of applications will be built on R1 as well as several global cloud providers offering its models as a consumable API,” said Gartner Distinguished VP Analyst Arun Chandrasekaran in an email to TechRepublic. “Deepseek’s future success is predicated on its ability to continuously innovate (rather than being a one-off success), build a developer ecosystem on its products and overcome cultural barriers, given its country of origin.” Chandrasekaran said DeepSeek’s low cost, efficiency, benchmark results, and open weights make it remarkable. DeepSeek-V3 was trained on 2,048 NVIDIA H800 GPUs. U.S. manufacturers are not, under export rules established by the Biden administration, permitted to sell high-performance AI training chips to companies based in China. “The potential power and low-cost development of DeepSeek is calling into question the hundreds of billions of dollars committed in the U.S,” said Ivan Feinseth, a market analyst at Tigress Financial, according to a note to clients acquired by ABC News. DeepSeek further differentiates itself by being an open source, research-driven project, while OpenAI increasingly focuses on commercial efforts. “Deepseek R1 is one of the most amazing and impressive breakthroughs I’ve ever seen — and as open source, a profound gift to the world.,” Silicon Valley insider and venture capitalist Marc Andreessen posted on X on Friday. Gartner said the global AI semiconductor industry will reach $114,048 in 2025. Gartner predicted the power required for data centers to run newly-added AI servers will reach 500 terawatt-hours by 2027. DeepSeek introduces multimodal models On Monday, DeepSeek followed up its success with another surprise: the Janus-Pro family of multimodal models, which can analyze and generate images. OpenAI alleges DeepSeek ‘distilled’ existing models On Jan. 29, Microsoft announced an investigation into whether DeepSeek might have piggybacked on OpenAI’s AI models, as reported by Bloomberg. Microsoft security researchers found large amounts of data passing through the OpenAI API through developer accounts in late 2024. OpenAI said it has “evidence” related to distillation, a technique of training smaller models using larger ones. Distillation violates OpenAI’s terms of service. OpenAI has not detailed the nature of the alleged evidence. Security concerns raised about DeepSeek’s models Since DeepSeek’s debut rocked the AI world, several security concerns about its models have swirled in the industry. Some concerns – input data feeding the model, copyright concerns, and possible disinformation or misinformation – apply to generative AI broadly; others caution U.S. users from potentially giving information to or opening a backdoor for a Chinese company. “The technology sector needs frameworks that ensure all AI systems protect user privacy and intellectual property rights according to international standards, while recognizing the different data access and governance requirements that exist across jurisdictions,” said Cliff Steinhauer, director of information security and engagement at U.S. nonprofit The National Cybersecurity Alliance, in an email to TechRepublic. “The path forward requires balancing innovation with robust data protection and security measures, while acknowledging the varying regulatory landscapes in which AI systems operate.” DeepSeek research temporarily exposed in a public database On Jan. 29, research firm Wiz Research revealed that they found a publicly accessible database of information exposed by DeepSeek, including chat history. The database has since been secured. Wiz Research found chat history, backend data, log streams, API Secrets, and operational details within the DeepSeek environment through ClickHouse, the open-source database management system. “This exposure underscores the fact that the immediate security risks for AI applications stem from the infrastructure and tools supporting them,” Wiz Research cloud security researcher Gal Nagli wrote in a blog post. “While much of the attention around AI security is focused on futuristic threats, the real dangers often come from basic risks—like accidental external exposure of databases.” Alibaba Cloud debuts new model