ReportWire

Tag: Artificial Intelligence/Machine Learning

  • European Institutional Equity Trading Study: Technology | Insights | Bloomberg Professional Services

    Figure 1: Rank Technology Solutions

    Third-party tools dominate pre-trade TCA use

    Traders consider third-party tools the most essential for transaction cost analysis, with 43% saying they are very or extremely reliant on them. Broker tools are least popular, with only 24% highly reliant and more than 40% citing slight or moderate use. This low reliance might reflect concerns over conflicts of interest in measuring performance.

    Figure 2: How Reliant Are You on Help to Estimate Pre-Trade?

    How Reliant Are You on Help to Estimate Pre-Trade?

    Despite wider adoption, pre-trade TCA models still face credibility issues among traders. A large asset manager told us they have yet to see an effective model and often raise the point at industry conferences, where peers tend to agree. The lack of consensus on model quality highlights deeper skepticism. Though tools are increasingly embedded in workflows, few are viewed as robust enough to reliably predict market impact, particularly in complex or large trades.

    Though 64% of traders estimate pre-trade costs, just 52% apply that analysis at the point of execution, based on our survey. This 12% gap suggests that, for some, TCA serves more as a compliance checkbox than a genuine input into trading decisions. Among traders who forgo pre-trade TCA before executing, several cited either low confidence in the outputs or a stronger reliance on intuition. The disconnect highlights that, though adoption of pre-trade tools is increasing, belief in their practical value still lags, particularly when execution speed and minimizing information leakage are a priority.

    Post-trade TCA is close to universal across the European buyside, with 86% of traders reporting they conduct it. Adoption is strongest among medium-sized companies (92%), followed by small (85%) and large (83%). Still, a minority across all company sizes don’t engage in post-trade analysis, often citing limited internal resources or a belief that market impact is too small to measure. Overall, post-trade TCA appears embedded in most European buyside execution review processes, even if confidence in the outputs varies among traders.

    Post-trade TCA remains a largely outsourced function, with 73% of respondents saying they are either very or extremely reliant on independent third-party providers. Internal analysis sees a moderate level of reliance at 55%, reflecting its resource-intensive nature. Meanwhile, broker-provided TCA ranks lowest in trust, with only 11% of respondents indicating high reliance — a likely result of ongoing concerns over conflicts of interest. The technical and data demands of conducting meaningful post-trade TCA in-house help explain continued reliance on third-party providers.

    Figure 3: Who’s Helping You Measure Your Post-Trade TCA?

    How Often Do You Measure Post-Trade TCA?

    One trader we spoke with said of TCA analysis that the number of unknowns in execution, particularly when using algos, can distort the reliability of outputs. Though they use TCA, they admitted they don’t trust it enough to guide decision-making. In their view, execution complexity and TCA’s limited visibility make it difficult to fully assess performance, as too many factors fall outside its scope. According to the trader, assessing an algo’s effectiveness often requires dismantling it manually, something no standard TCA tool can replicate.

    Nearly half (47%) of European buyside traders report running post-trade TCA on a daily basis, with frequency varying by company size. Large ones are the most consistent, with 53% conducting daily analysis, reflecting the scale and complexity of their execution activity. Medium and small companies are notably less reliant, with daily usage 9% and 11% lower, respectively. At the other end, 18% of small companies never estimate post-trade costs — the highest non-usage rate in the sample. Monthly post-trade TCA, meanwhile, is the most evenly adopted frequency across company size, cited by 18% of traders overall.

    Figure 4: How Often Do You Measure Post-Trade TCA?

    How Often Do You Measure Post-Trade TCA?

    Arrival price leads, but VWAP gains momentum

    Arrival price, also known as implementation shortfall (IS), remains the most-used execution benchmark among European buyside traders, with 30% reporting usage, a 7 percentage-point drop from last year. The benchmark is especially favored by large firms (40%), compared with just 22% of small firms. In contrast, VWAP (volume-weighted average price) has gained momentum, now ranking as the second-most used benchmark at 27%. It’s the top choice among small (31%) and medium-sized (29%) firms, signaling a growing preference for simpler, liquidity-weighted benchmarks among smaller participants. Market-on-close (MoC) benchmarks show consistent usage across firm sizes at 21%. Notably, 10% of traders surveyed say they don’t use any benchmark for measuring execution performance.

    Figure 5: Benchmarks Used by Size

    Benchmarks Used by Size

    Traditional benchmarking on quant desks is declining, with just 17% using VWAP and 23% IS, compared with much higher usage among fundamental traders. MoC trades concentrate liquidity near market close, allowing quants to design execution around predictable volume spikes. That aligns well with algorithms targeting minimized market impact during specific trading windows.

    Figure 6: Benchmarks Used by Investing Style

    Benchmarks Used by Investing Style

    North American traders have a stronger preference for IS with 33% using it — 3 percentage points above the overall average (see Fig. 7). VWAP is less favored, with usage at just 20%, a full 7-percentage points below the global average. EU, UK, and Swiss traders show more balanced usage between IS and VWAP. MoC is most used in the EU (27%), suggesting stronger alignment with close-based execution. North American and Swiss traders (20%) have a higher preference for alternative or proprietary performance metrics to measure trade performance.

    Figure 7: Benchmarks Used by Region

    Benchmarks Used by Region

    Volatility-driven benchmark misses make work for traders

    One trader at a large asset-management firm told us that higher volatility has led to a rise in benchmark misses, making it harder to distinguish between genuine execution issues and volatility induced noise. That trader now expends more effort investigating missed benchmarks like VWAP to validate whether outcomes are justifiable, which is a time-consuming process. Instead of assuming a poor trade, they look at liquidity and market timing to assess if the “miss” aligns with expectations. The shift has added complexity to post-trade analysis, with traders expected to not only track performance, but also explain the story behind each deviation.

    A majority (57%) of survey respondents report using transaction cost analysis (TCA) to actively improve trading outcomes, suggesting it plays a critical role in measuring and improving execution quality. Meanwhile, 35% treat it as a checkbox exercise or a process driven by compliance obligations to meet best-execution requirements, instead of performance enhancement. Just 8% view TCA as a tool to support transparency, noting they primarily engage with it when clients request access to execution performance logs.

    Asked about the impact of tariff-driven volatility, 48% of traders cited higher costs, mainly from wider spreads. Another 39% said their TCA was largely unaffected, implying liquidity was sufficient to absorb shocks without notable cost swings. The remaining 13% reported a modest increase, though not enough to prompt structural changes in execution. The divergence suggests that order size and trading style shapes outcomes more than volatility itself — some desks absorbed shocks, while those executing larger orders saw a measurable drop in quality.

    One trader at a small asset management firm told us they track TCA metrics daily to use the insights as a real-time health check. On a monthly basis, they conduct a deeper dive, reviewing strategy alignment and anomalies. Annually, they compile a TCA scorecard to evaluate broker performance and reallocate flow away from those whose results show persistent underperformance. This approach creates a feedback loop that directly influences routing decisions and reinforces how TCA can serve as a bridge between execution and broker relationship management.

    Small funds concentrate trading flow more than large

    Traders at small institutions send about 38% of algorithmic trading flow to their top broker, compared with 23% at large funds and 33% at medium-sized peers. On average, traders use five brokers for 81% of such flow, though at large funds the top five account for just 75.6%. Small funds are most concentrated, routing nearly 90% (88.7%) of flow to their top five providers.

    Funds trading in the UK and Europe are executing a plurality of their shares (42.4%) via broker algorithms or through direct market access (DMA) channels to the market. They’re sending 29.8% of flow to brokers for high-touch execution, 22.6% to program trading desks and 5.2% to dark, multi-lateral trading facilities (MTFs). This year, smaller funds routed a bit more than half of their flow via low-touch channels (algo/DMA), as large funds sent 11.1% less. Funds used program-trading desks when they needed multiple orders executed simultaneously, typically tied to an index or an event.

    Figure 8: UK/Europe Buyside Equity Order Flow Allocation

    UK/Europe Buyside Equity Order Flow Allocation

    Large European buyside funds increased their use of algorithmic-trading providers by 21% year over year, averaging 10.2 providers in 2025 vs. 8.4 in 2024. The rise may reflect efforts to boost liquidity access. Midsized funds trimmed average broker use to 7 from 9.4, while smaller funds increased slightly to 6.1 from 5.1. Overall, average provider use ticked up to 7.9 from 7.8.

    Among large, medium and small European buyside institutions surveyed, 25% plan to increase use of trading-algorithm providers in 2026, while 13% expect to cut back. Growth is driven mainly by smaller firms, with 34% planning to expand next year.

    The push reflects a desire to foster broker competition through customized algorithms and algo wheels — systems that rotate orders across different brokers’ algorithms to measure performance and direct flow to those delivering the best results. Customized algorithms allow providers to distinguish themselves based on service quality, while wheels broaden order routing across brokers, improving liquidity in tougher markets.

    Algorithm wheels — tools that automate order allocations and allow unbiased A/B testing of algorithms — are gaining traction in Europe, according to feedback in our survey. In 2025, 42% of buyside firms were using an algo wheel, up from 33% in 2024. In last year’s survey, 20% said they were considering one, but that figure has dipped to 17%. Traders cite reduced manual input and less bias in broker selection as key benefits. These features also support best execution obligations, which may be helping to drive broader use across the region.

    Figure 9: Are You Using an Algo Wheel?

    Are You Using an Algo Wheel?

    £25 trillion funds say AI won’t oust traders, analysts

    AI is gaining traction on European trading desks, but buyside traders say human insight will remain central, especially in investment research. Our survey, covering funds with £25 trillion in assets, found nearly two-thirds expect research to keep relying on judgment and in-person observations that models can’t replicate, with only 4% seeing full automation. Adoption is rising in operational efficiency, investment analysis and broker algos, yet use remains largely experimental. Execution is viewed as the most likely near-term application, though AI is broadly seen as a complement to existing processes rather than a replacement for jobs or decision making. The cautious outlook shows desks are prioritizing productivity and efficiency over structural change.

    Most buyside traders see little change ahead for trading-desk jobs, with a solid 58% expecting head count to stay the same, signaling that AI is viewed more as a tool to boost productivity than a driver of workforce shifts. Only 12% expect an increase in employment, while nearly one-third anticipate reductions, either slight or significant. Overall sentiment shows skepticism about AI’s short-term effect on staffing, with most desks not expecting meaningful near-term disruption.

    European buyside traders anticipate faster AI adoption in execution than in investment decision-making, with 26% expecting execution decisions within the next two years compared with 7% for investment decisions. Of respondents, 57% anticipate AI adoption for investment decisions in 2-3 years and a notable 33% see it taking more than five. Though there’s a higher concentration for AI in execution in years 4-5, traders appear more comfortable introducing AI in execution workflows, where performance is easier to monitor and control, than for investment decisions.

    AI’s potential to generate revenue is viewed with skepticism across European trading desks, with just 14% of respondents saying AI will increase revenue by more than 10%. Nearly two-thirds expect no impact or gains below 5%. The cautious view spans firm sizes, with 40% of small companies and 36% of large ones seeing no effect at all. This perspective highlights a broader belief that AI will serve as a complement to existing processes and a productivity enhancer rather than a direct driver of revenue growth in trading operations.

    Figure 10: AI to Improve: Revenue Generation

    AI to Improve: Revenue Generation

    Data from our study show AI adoption on trading desks remains limited and mostly in testing. Most buyside respondents report no use of AI in key areas, such as internal algos (92%), investment decision-making (81%) and broker algos (67%) (see Fig. 11). The highest AI adoption rates were in investment analysis (33%) and operational efficiency (28%), but even there, use is largely experimental. Frequent or daily AI use remains rare, suggesting that while interest in AI is growing, institutional trading workflows are still in early-stage adoption.

    Figure 11: How Much Do You Use AI on the Trading Desk?

    How Much Do You Use AI on the Trading Desk?

    AI is beginning to gain ground in investment analysis, with adoption on par with operational efficiency, as 51% of buyside firms report some level of use. Compared with other functions, the split across firm sizes is more balanced, though large firms show slightly lower uptake overall. Around a third of traders report testing AI for investment analysis, with 13% being frequent users. Daily adoption remains rare, indicating that while European buyside firms are starting to incorporate AI, the transition is still gradual and exploratory for most. Large firms lead the way in using AI for operational efficiency, with 38% actively testing, 14% using it frequently and 10% applying them daily. Small firms show the highest rate of frequent use of AI to improve operational efficiency (22%), yet uptake across medium and smaller firms remains uneven. Nearly 60% of these reported no use at all, highlighting a wide gap in adoption across firm sizes.

    Large buyside companies appear more open to exploring AI in broker algos, which execute client orders by slicing trades, routing them across venues and managing costs. About 25% are testing such tools, while another 11% report frequent or daily use. That contrasts with smaller peers, where 75% of medium firms and 74% of small ones report no use at all. Adoption overall remains limited, but early trials at larger institutions suggest groundwork is being laid for broader integration. Across all sizes, just 11% report frequent use, underscoring that AI adoption is still in its early stages. This share is likely to rise as technology matures and competition drives faster innovation.

    Figure 12: Using AI: Broker Algos

    Using AI: Broker Algos

    Larger funds use more brokers in algo-wheel rotation

    The survey shows no medium-sized institution included more than 10 brokers in its algorithmic trading wheel. Smaller firms clustered in the 0-4 and 5-9 ranges, with 40% of respondents, while 20% reported more than 10. Larger firms showed broader distribution, with 38% in the 5-9 and 10-plus buckets and 25% in 0-4. Larger institutions typically manage higher trading volumes across stocks with varying liquidity, making multiple brokers useful.

    Figure 13: Brokers on Wheel by Institution Size

    Brokers on Wheel by Institution Size

    Bloomberg

    Source link

  • Mapping AI exposure through rules and reason | Insights | Bloomberg Professional Services

    As AI dominates headlines, who’s really building it? We explore how a rules-based framework distinguishes companies materially engaged in AI development and deployment from those merely adopting the technology.

    Bloomberg

    Source link

  • Opinion | AI Is a Tool, Not a Soul

    Pope Leo XIV tries to head off claims that chatbots are sentient beings with rights.

    Kristen Ziccarelli

    Source link

  • The AI value chain: Understanding the engines of growth | Insights | Bloomberg Professional Services

    Built using a rules-based approach informed by Bloomberg Intelligence (BI) research, the index maps the AI value chain and identifies companies with significant involvement in each link of the AI infrastructure. Numerous business models fit within this theme, from cloud computing platforms to semiconductors to hardware infrastructure.  

    Only firms that have clear, material exposure to AI, as identified in BI research, are eligible for inclusion into the index. Once the eligible universe is set, the selection process groups companies into three categories: cloud, AI hardware, and AI semiconductors. From these, the 15 largest companies in each group, ranked by free-float market capitalization, are selected.

    The power of equal weighting 

    Some investors may question the need for a dedicated AI sleeve in portfolios. After all, many of the largest constituents in market cap weighted indices are active in this growing area. A simple market cap weighted approach, however, may not reflect the diversity of AI-related business models. To avoid large cap names from dominating the index, here, the selected securities are instead equally weighted. Such an approach also increases the exposure to names that are smaller in market capitalization and whose price may not yet reflect the magnitude of their involvement this early in their development cycle.  

    Additionally, taking an Equal-weighted approach may increase the representation of companies involved in AI-related M&A activity. By giving smaller and mid-sized firms greater footing, the index may highlight potential targets more fully.  

    Traditional equity indices may underrepresent how AI is reshaping market growth. While broad benchmarks capture companies that are already benefiting from digital transformation, they often reflect today’s market composition rather than tomorrow’s growth drivers. 

    Even indices that focus on technology (such as the Bloomberg 500 Tech and Bloomberg 100 Indices) do not fully capture the growth of the AI ecosystem as acceleration may be happening beneath the surface. The following chart illustrates this gap, showing that the broader indices underrepresent the scale of projected sales growth across AI-related segments over the coming decade.

    Sales CAGR: AI Infrastructure Segments

    Selected companies illustrating the approach employed by the Bloomberg AI Value Chain Index

    Western Digital: AI’s expanding library 

    AI runs on data, and every image analyzed, every prompt answered, every recommendation generated requires vast storage. Western Digital (WDC US) has developed a framework called the AI Data Cycle, mapping the six key stages of AI workflows from raw data storage to new data generation. At some stages, you need “sports cars” (SSDs) to move critical goods quickly. At others, you need “cargo ships” (HDDs) to carry massive loads efficiently. Western Digital builds both, and it designs them to work together as a complete logistics system for AI data. 

    What makes this significant is that Western Digital isn’t competing to build AI models themselves. Instead, it’s building the infrastructure backbone that is intended to support AI models with the appropriate storage options across workflow stages. The company’s strong year-to-date performance (+167.2% through September) may be attributed to some market participants increased attention to storage infrastructure’s role in AI workflows. 

    Credo: Highways for AI data 

    Technology provided by Credo (CRDO US) ensures information moves between GPUs, storage systems, and networking gear. The firm holds a leading position in the Active Electrical Cable market and has deep partnerships with hyperscalers like Amazon, Microsoft, and Meta. Credo essentially gives AI data centers the ability to add lanes to the data highways within.  

    One of the biggest challenges in scaling AI today is energy consumption. Training and running large AI models requires connecting thousands of GPUs in data centers, which in turn means moving vast amounts of data at extremely high speeds. Every time that data moves, energy is consumed. This is where Credo’s technology comes in. Their products are specifically designed to move data quickly while using less power per bit transferred. That means hyperscale data centers can train and deploy AI models without their energy bills spiraling out of control. Credo’s revenue growth and stock performance in 2025 (+117% YTD through September) may suggest a growing interest in energy-efficient solutions for AI infrastructure.

    Hon Hai: Factories of the future 

    For decades, Hon Hai Precision Industry (2317 TT), or Foxconn, was synonymous with high-volume electronics assembly, particularly smartphones. Recently, however, Foxconn redirected its enormous manufacturing muscle toward AI infrastructure. This move has paid off, with revenue from its AI and cloud/server division now surpasses that of smartphone sales. 

    Today, it co-designs servers with NVIDIA and assembles forty percent of the world’s AI servers. The company’s pivot to AI servers coincided with record revenues in 2024 and 2025, and its year-to-date stock performance (+21.6% through September) may reflect increased investor focus on AI infrastructure.  

    Measuring the growth of AI infrastructure

    AI adoption is accelerating across industries from healthcare to finance to entertainment. Each advance increases demand for chips, storage, servers, and faster data connections. The Bloomberg AI Value Chain Index is designed to capture this continuous change in demand. By focusing on the enabling infrastructure rather than any single model or application, the index seeks to capture relevant new technologies and players. 

    By equal weighting forty-five firms across cloud, semiconductors, and hardware, it offers a diversified, transparent way to track AI’s growth. In a year where AI is at the forefront of many investors’ minds, the index has delivered results that have been in line with or exceeded many of the more popular AI ETFs available on the market. If AI is the gold rush of our time, the index can be seen as a map to the companies selling the picks, shovels, and railroads.

    Bloomberg AI Value Chain Index Versus Similar Products Year-to-Date

    , and how Bloomberg Indices measure exposure to this theme? In part three of our series, we focus on AI Thematics, examining how a rules-based approach, supported by Bloomberg Intelligence, identifies companies materially engaged in AI development and deployment. 

    To learn more about Bloomberg Indices, click here. 

    Amplify has licensed the Bloomberg AI Value Chain Index for their ETF ticker AIVC. 

    Disclaimer

    The data and other information included in this publication is for illustrative purposes only, available “as is”, non-binding and constitutes the provision of factual information, rather than financial product advice. BLOOMBERG and BLOOMBERG INDICES (the “Indices”) are trademarks or service marks of Bloomberg Finance L.P. (“BFLP”). BFLP and its affiliates, including BISL, the administrator of the Indices, or their licensors own all proprietary rights in the Indices. Bloomberg L.P. (“BLP”) or one of its subsidiaries provides BFLP, BISL and its subsidiaries with global marketing and operational support and service. Certain features, functions, products and services are available only to sophisticated investors and only where permitted. Bloomberg (as defined below) does not approve or endorse these materials or guarantee the accuracy or completeness of any information herein, nor does Bloomberg make any warranty, express or implied, as to the results to be obtained therefrom, and, to the maximum extent allowed by law, Bloomberg shall not have any liability or responsibility for injury or damages arising in connection therewith. Nothing in the Services or Indices shall constitute or be construed as an offering of financial instruments by Bloomberg, or as investment advice or investment recommendations (i.e., recommendations as to whether or not to “buy”, “sell”, “hold”, or to enter or not to enter into any other transaction involving any specific interest or interests) by Bloomberg. Information available via the Index should not be considered as information sufficient upon which to base an investment decision. All information provided by the Index or in this publication is impersonal and not tailored to the needs of any person, entity or group of persons. Absence of any trademark or service mark from this list does not waive Bloomberg’s intellectual property rights in that name, mark or logo. For the purposes of this publication, Bloomberg includes BLP, BFLP, BISL and/or their affiliates.

    BISL is registered in England and Wales under registered number 08934023 and has its registered office at 3 Queen Victoria Street, London, England, EC4N 4TQ. BISL is authorized and regulated by the Financial Conduct Authority as a benchmark administrator. © 2025 Bloomberg. All rights reserved.

    Bloomberg

    Source link

  • Inside AI’s rapid expansion: What investors need to know | Insights | Bloomberg Professional Services

    Bloomberg Indices Disclaimer: The data and other information included in this publication is for illustrative purposes only, available “as is”, non-binding and constitutes the provision of factual information, rather than financial product advice. BLOOMBERG and BLOOMBERG INDICES (the “Indices”) are trademarks or service marks of Bloomberg Finance L.P. (“BFLP”). BFLP and its affiliates, including BISL, the administrator of the Indices, or their licensors own all proprietary rights in the Indices. Bloomberg L.P. (“BLP”) or one of its subsidiaries provides BFLP, BISL and its subsidiaries with global marketing and operational support and service. Certain features, functions, products and services are available only to sophisticated investors and only where permitted. Bloomberg (as defined below) does not approve or endorse these materials or guarantee the accuracy or completeness of any information herein, nor does Bloomberg make any warranty, express or implied, as to the results to be obtained therefrom, and, to the maximum extent allowed by law, Bloomberg shall not have any liability or responsibility for injury or damages arising in connection therewith. Nothing in the Services or Indices shall constitute or be construed as an offering of financial instruments by Bloomberg, or as investment advice or investment recommendations (i.e., recommendations as to whether or not to “buy”, “sell”, “hold”, or to enter or not to enter into any other transaction involving any specific interest or interests) by Bloomberg. Information available via the Index should not be considered as information sufficient upon which to base an investment decision. All information provided by the Index or in this publication is impersonal and not tailored to the needs of any person, entity or group of persons. Absence of any trademark or service mark from this list does not waive Bloomberg’s intellectual property rights in that name, mark or logo. For the purposes of this publication, Bloomberg includes BLP, BFLP, BISL and/or their affiliates.

    BISL is registered in England and Wales under registered number 08934023 and has its registered office at 3 Queen Victoria Street, London, England, EC4N 4TQ. BISL is authorized and regulated by the Financial Conduct Authority as a benchmark administrator. © 2025 Bloomberg. All rights reserved.

    Bloomberg Intelligence Disclaimer: The data included in these materials are for illustrative purposes only. The BLOOMBERG TERMINAL service and Bloomberg data products (the “Services”) are owned and distributed by Bloomberg Finance L.P. (“BFLP”) except (i) in Argentina, Australia and certain jurisdictions in the Pacific islands, Bermuda, China, India, Japan, Korea and New Zealand, where Bloomberg L.P. and its subsidiaries (“BLP”) distribute these products, and (ii) in Singapore and the jurisdictions serviced by Bloomberg’s Singapore office, where a subsidiary of BFLP distributes these products. BLP provides BFLP and its subsidiaries with global marketing and operational support and service. Certain features, functions, products and services are available only to sophisticated investors and only where permitted. BFLP, BLP and their affiliates do not guarantee the accuracy of prices or other information in the Services. Nothing in the Services shall constitute or be construed as an offering of financial instruments by BFLP, BLP or their affiliates, or as investment advice or recommendations by BFLP, BLP or their affiliates of an investment strategy or whether or not to “buy,” “sell,” or “hold” an investment. Information available via the Services should not be considered as information sufficient upon which to base an investment decision. The following are trademarks and service marks of BFLP, a Delaware limited partnership, or its subsidiaries: BLOOMBERG, BLOOMBERG ANYWHERE, BLOOMBERG MARKETS, BLOOMBERG NEWS, BLOOMBERG PROFESSIONAL, BLOOMBERG TERMINAL and BLOOMBERG.COM. Absence of any trademark or service mark from this list does not waive Bloomberg’s intellectual property rights in that name, mark or logo. All rights reserved. © 2025 Bloomberg.

    Bloomberg Intelligence is a service provided by Bloomberg Finance L.P. and its affiliates. (“Bloomberg”). Bloomberg is not an officially recognized credit rating agency in any jurisdiction, and customers should not use or rely on Bloomberg Intelligence to comply with applicable laws or regulations that prescribe the use of ratings issued by accredited or otherwise recognized credit rating agencies. Bloomberg Intelligence Credit and Company research may not be available in certain jurisdictions.

    Bloomberg Intelligence shall not constitute, nor be construed as, investment advice or investment recommendations (i.e., recommendations as to whether or not to “buy,” “sell,” “hold,” or to enter or not to enter into any other transaction involving any specific interest) or a recommendation as to an investment or other strategy. No aspect of the Bloomberg Intelligence function is based on the consideration of a customer’s individual circumstances. Bloomberg Intelligence should not be considered as information sufficient upon which to base an investment decision. Customers should determine on their own whether they agree with Bloomberg Intelligence. Bloomberg Intelligence should not be construed as tax or accounting advice or as a service designed to facilitate any Bloomberg Intelligence customer’s compliance with its tax, accounting, or other legal obligations. Bloomberg believes that the information it uses in Bloomberg Intelligence comes from reliable sources, but does not guarantee the accuracy of information contained in Bloomberg Intelligence. Employees involved in Bloomberg Intelligence may hold positions in the securities analyzed or discussed on Bloomberg Intelligence.

    Bloomberg makes no claims or representations, or provides any assurances, about the sustainability characteristics, profile or data points of any underlying issuers, products or services, and users should make their own determination on such issues.

    Bloomberg

    Source link

  • Can the AI workhorses carry the world’s markets – yet again? | Insights | Bloomberg Professional Services

    The data included in these materials are for illustrative purposes only. The BLOOMBERG TERMINAL service and Bloomberg data products (the “Services”) are owned and distributed by Bloomberg Finance L.P. (“BFLP”) except (i) in Argentina, Australia and certain jurisdictions in the Pacific Islands, Bermuda, China, India, Japan, Korea and New Zealand, where Bloomberg L.P. and its subsidiaries (“BLP”) distribute these products, and (ii) in Singapore and the jurisdictions serviced by Bloomberg’s Singapore office, where a subsidiary of BFLP distributes these products. BLP provides BFLP and its subsidiaries with global marketing and operational support and service. Certain features, functions, products and services are available only to sophisticated investors and only where permitted. BFLP, BLP and their affiliates do not guarantee the accuracy of prices or other information in the Services. Nothing in the Services shall constitute or be construed as an offering of financial instruments by BFLP, BLP or their affiliates, or as investment advice or recommendations by BFLP, BLP or their affiliates of an investment strategy or whether or not to “buy”, “sell” or “hold” an investment. Information available via the Services should not be considered as information sufficient upon which to base an investment decision. The following are trademarks and service marks of BFLP, a Delaware limited partnership, or its subsidiaries: BLOOMBERG, BLOOMBERG ANYWHERE, BLOOMBERG MARKETS, BLOOMBERG NEWS, BLOOMBERG PROFESSIONAL, BLOOMBERG TERMINAL and BLOOMBERG.COM. Absence of any trademark or service mark from this list does not waive Bloomberg’s intellectual property rights in that name, mark or logo.

    All rights reserved. © 2025 Bloomberg.

    Bloomberg

    Source link

  • How is investment research transformed by AI? | Insights | Bloomberg Professional Services

    How are investment firms rethinking data architecture to support advanced analytics and responsibly apply AI at scale? 

    At Bloomberg’s 2025 Enterprise Data & Tech Summit in New York, Angana Jacob, Global Head of Research Data at Bloomberg, spoke with Carson Boneck, Chief Data Officer at Balyasny Asset Management, C.J. Jaskoll, Chief Technology Officer at Russell Investments, and Nan Xiao, Chief Technology and Data Officer at Greenland Capital Management, about how investment firms are building AI-ready infrastructure, harmonizing data systems, and looking at ROI of their AI investments. 

    Featured insights from the discussion panel: 

    On AI transforming investment firms

    Carson Boneck: What AI allows us to do is extract certain heuristics that can then be emulated and replicated. In a platform like ours, we have so many great investors. I think a lot of firms, including ours, in many places, are at the task level, but we could really help them improve these tasks. We can write research reports by using the deep research agent and compiling all the various information… My hypothesis is that firms like ours are going to get completely transformed [by AI] and the firms that win are those that are able to do the best job at using AI to extract the heuristics of the very best investors.* 

    On measuring AI investment ROI  

    C.J. Jaskoll: Gen AI is new, but change management is not. For all change at every buy-side firm I’ve ever been at, there are really two major ways that I think about ROI. Number one: operational efficiency. That means avoiding a trade error, reducing costs, and reducing risk. All of those things can be tangible and quantified. If you are able to quantify that, then you can pay for the platform you’re building or that you’ve bought. 

    The second one, which I think our industry has and almost no other industry has, is generating alpha. I was talking to a friend last week who asked the same question, and he said, “How do I bring this data into my firm?” I said, “One trade at your firm will pay for the entire platform and three developers.” But is ROI actually needed right now? I think we’re still in the honeymoon period. I don’t think many companies are actually thinking about ROI. 

    On scaling research with agentic AI 

    Nan Xiao: We’ve deployed [solutions that are in] production already, and a lot of things in experiments. We have research agents who do multi-steps research and come up with conviction, and go from there to generate sizing, suggestions, trading ideas, and essentially simulate outcomes. Those goals become proposals to PMs. A possible use case will also be… you can ask [AI] to think [about] things differently. Instead of having one research analyst, it’s like [having] five research analysts thinking from different angles. As a result, giving you five different proposals that you can…pick which you would like to go for or combine. 

    *Quotations have been edited for brevity and clarity.

    Bloomberg

    Source link

  • October Global Regulatory Brief: Digital finance | Insights | Bloomberg Professional Services

    MAS launches PathFin.ai knowledge hub to boost industry literacy and innovation in AI

    Summary

    In a speech by Minister Chee Hong Tat, the Monetary Authority of Singapore (MAS) announced the launch of the PathFin.ai knowledge hub, a strategic initiative to boost AI adoption, literacy, and innovation across Singapore’s financial sector. This launch is part of a two-pronged strategy—upgrading the AI ecosystem and upskilling the workforce—designed to ensure the sector remains competitive amidst global shifts in technology, trade, and climate change. The government confirmed its commitment to ensuring AI augments workers, with the goal of uplifting AI literacy for all employees while providing clarity on supervisory expectations for responsible AI use.

    In more detail

    Strategic vision and context

    • Growth Driver: The financial sector is critical to Singapore’s economy, growing by 6.8% last year, and is essential for creating high-value jobs for locals. AI is identified as a major force capable of adding significant value to the global sector.
    • Two Key Pillars: To maintain competitiveness, the sector will focus on: 
    1. Continuously upgrading the AI ecosystem to promote knowledge exchange and innovation;
    2. Uplifting and upskilling the workforce to be AI-ready.

    Pillar 1: Upgrading the AI ccosystem

    • PathFin.ai Knowledge Hub: Launched under the existing PathFin.ai program (which involves over 80 FIs), the hub is a new resource for peer learning. It features an initial set of successful AI use cases and learnings curated by industry participants in key areas like sales, risk management, and tech. The goal is to reduce the time and effort required for individual FIs to implement AI solutions by learning from shared experiences.
    • Enhancing Supervisory Clarity: MAS plans to boost industry confidence in innovation by clarifying its expectations on AI risk management.
      • Building on the FEAT principles (Fairness, Ethics, Accountability, Transparency), MAS will consult the industry later this year on new supervisory guidelines on AI risk management.
      • Concurrently, MAS is developing the Project MindForge AI risk management handbook for publication later this year, which will provide practitioners’ perspectives to guide responsible AI deployment.

    Pillar 2: Preparing the AI-ready workforce

    • Jobs Transformation Map (JTM): MAS and IBF, in partnership with WSG, developed a JTM to study how Generative AI will reshape jobs and skills. Pilot programs with 10 FIs are testing and refining the approach.
    • Universal AI Literacy: The core principle is to “leave no one behind” by lifting foundational AI skills for all workers (e.g., prompt design, AI governance). The three local banks have committed to training all 35,000 of their Singapore employees in the next 1–2 years using IBF-accredited programs.
    • Augmentation over Replacement: The strategy is to augment employees with role-specific AI tools, streamlining routine tasks to enable them to take on higher-value and more complex work, thereby improving productivity and career progression (e.g., Manulife underwriters, Bank of Singapore RMs).
    • Talent Pipeline: IBF is working with Institutes of Higher Learning (IHLs) and FIs (like UBS and UOB) to establish internships and traineeships for young talent to gain practical exposure to AI use cases in finance early in their careers.

    Next steps

    • MAS Consults on AI Risk Guidelines: MAS will formally consult the industry later this year on new supervisory guidelines for AI risk management.
    • Publish Practitioner Handbook: The Project MindForge AI risk management handbook will be published later this year to assist FIs with responsible AI implementation.
    • Expand PathFin.ai Hub Content: MAS and industry partners will progressively enhance the PathFin.ai knowledge hub with more peer-validated use cases, resources, and solutions.
    • Complete Mass AI Literacy Training: The three local banks are expected to complete the training of their 35,000 employees in foundational AI literacy within the next 1 to 2 years.
    • Union and FI Collaboration: FIs and unions are encouraged to utilize platforms like the NTUC Company Training Committee (CTC) grant and roll out IBF-accredited courses to accelerate the upskilling of the financial sector workforce.

    The Australian government proposes legislation for crypto platforms

    Treasury presented proposals for new rules affecting digital asset platforms (DAPs) and tokenised custody platforms (TCPs) in Australia. The focus of the legislation is on businesses that hold assets on behalf of clients, rather than on the digital assets themselves. It is part of the Government’s commitment in the 2024-2025 budget to modernise Australia’s digital asset regulation.

    Background

    The draft legislation seeks to capture DAPs and TCPs by introducing each as new financial products. Where digital assets already fall within existing financial product definitions, the proposed laws will largely apply to activities involving those assets in the same way they do now. However, the proposals introduce targeted elements of risk mitigation, regulatory clarity, and “right-sized” obligations – in a way that facilitates innovations without sacrificing consumer protections.

    Anyone providing specified services in relation to DAPs or TCPs will be treated as providing a financial service. Providers of financial services will need to hold an Australian Financial Services Licence (AFSL), the same licence required for other financial service providers. Using the existing AFSL framework avoids the need for a new licensing regime. It also reduces complexity and gives industry and consumers the benefit of familiar rules and protections.

    Last week, ASIC had proposed extending class relief for intermediaries engaging in the secondary distribution of a second stablecoin issued by an Australian financial services (licenced) issuer. Earlier this month, ASIC granted class relief for intermediaries engaging in the secondary distribution of a stablecoin issued by an AFS licensed issuer. ASIC advised at the time that as and when more issuers of eligible stablecoins obtain an AFS licence, it will consider extending the same relief to intermediaries distributing those stablecoins. ASIC is working closely with Treasury as it looks to implement the Government’s digital assets reforms.

    In developing its legislative proposals, Treasury considered the recommendations proposed by the FSB and IOSCO, and have been guided by them in developing the current reforms. These recommendations aim to ensure a level-playing field between traditional and emerging financial intermediaries.

    Next steps

    Treasury’s consultation closes on 24 October 2025.

    South African regulators issue consultation on upcoming cybersecurity & incident reporting standards

    Context

    The Financial Sector Conduct Authority (FSCA) and the Prudential Authority (PA) have issued two key Joint Standards that reshape regulatory expectations for financial institutions:

    • Joint Standard 1 of 2023 – IT Governance and Risk Management (effective 15 November 2024)
    • Joint Standard 2 of 2024 – Cybersecurity and Cyber Resilience Requirements (effective 1 June 2025)

    Together, these standards define notification obligations for material IT and cyber incidents across regulated financial institutions.

    Current development, consultation on notification framework

    In September 2025, the Authorities released Joint Communication 3 of 2025 for industry consultation, including:

    • Annexure A: Draft Determination outlining the formal notification process.
    • Annexure B: Draft template for reporting material IT and cyber incidents.
    • Annexure C: Comment template for feedback (submissions due 5 October 2025).

    Implications

    Regulators are tightening expectations around timely and standardized incident reporting.

    Bloomberg

    Source link

  • Global insight: AI’s three revolutions for macro forecasting | Insights | Bloomberg Professional Services

    Filling data gaps

    Economists have long mined text for research and forecasting. Early applications, from economic uncertainty (Baker et al. 2016) to central bank sentiment (Bulir et al. 2012), mainly relied on keyword counting.

    Today, natural language processing and large language models unlock far richer signals. Sentiment scores of commentary in purchasing manager indexes can improve GDP nowcasts in combination with traditional data (de Bondt and Sun 2025). Topic modeling of news helps forecast aggregate stock market returns (Bybee et al. 2023). And language models turn headlines into forward-looking indicators of central bank policy — as with Bloomberg Economics’ indexes for the Federal Reserve, European Central Bank and Bank of England. These are available to Bloomberg Terminal subscribers via BECO MODELS CBSPEAK .

    Satellite imagery is another important data source. Computer vision models can leverage interpretable features, such as roads, parking lots and crop fields, to provide information that official statistics either miss or capture too slowly. For example, we’ve built a monthly global GDP tracker using night-light data. Other researchers have found that retail parking occupancy measured from space can offer investors an edge (Katona et al. 2018), while snapshots of port-occupancy can help nowcast GDP and trade (Spelta et al. 2025).

    Supercharged workflows

    AI is changing how economists work as much as what they analyze.

    • Automation of routine tasks: Data cleaning, classification and feature engineering are increasingly handled by AI, freeing researchers to focus on interpretation and strategy.
    • Faster iteration: LLMs can increasingly act as research assistants — summarizing literature, coding econometric routines or stress-testing assumptions. Incorporating generative AI can cut project timelines from weeks to days (Korinek 2025).
    • Collaborative tools: AI-driven platforms integrate data, models and visualization, creating more transparent and reproducible research pipelines.

    At Bloomberg Economics, we’re working faster across a broader suite of models, and spending more time contextualizing the results in our research.

    Beyond linear models

    AI is improving the predictive power of econometric models, enabling policymakers to be more proactive in monitoring financial risks. From classical machine learning to transformer-based foundation models, new methods capture signals — including nonlinear relationships — that can provide early warning systems for stress events.

    • Classical ML methods like Lasso, Ridge, and Elastic Net handle large datasets effectively, while tree-based ensembles such as Random Forest and Gradient Boosting capture nonlinearities that emerge during periods of high macroeconomic uncertainty (Aldasoro et al. 2025). Newer variants, such as the macro random forest, combine ML flexibility with the structure of economic models, outperforming traditional econometric techniques even when even when the time series are short (Chinn et al. 2023).
    • Deep learning models (neural networks) extend these gains to complex, high-dimensional tasks, such as predicting FX dislocations (Aquilina et al. 2025). These models excel at identifying nonlinear patterns and rare events, though they require substantial tuning (Athey and Imbens 2019).
    • Text-based large language models and NLP systems extract sentiment and information from unstructured sources like news, policy statements and corporate filings. This can boost forecast accuracy when combined with numerical data (de Bondt and Sun 2025). Similarly, time-series foundation models — such as TimeGPT (Nixtla), Moirai (Salesforce), and TimesFM (Google) — bring transformer architectures to economic forecasting. While they don’t yet outperform econometric mainstays like Bayesian VARs, combining their predictive flexibility with structural model discipline yields accuracy gains across variables and horizons (Carriero et al. 2025).
    • Although some studies have explored using LLMs directly as forecasters, their lack of a true “point-in-time” notion makes them unreliable for out-of-sample evaluation (Lopez-Lira et al. 2025). The frontier instead lies in hybrid approaches — blending the interpretability of structural economics with AI’s adaptive, data-driven strengths.

    Central banks and international institutions are increasingly deploying these same AI models to anticipate financial stress. The ECB’s Cassandra system, for example, fine-tunes language models to analyze financial news sentiment and then applies boosting and neural network methods to flag early warning signals for banks (Petropoulos et al. 2025).

    Similarly, BIS research (Aquilina et al. 2025) combines recurrent neural networks with LLMs to forecast and interpret episodes of FX market stress. The BIS system combines quantitative assessments with qualitative interpretation of financial news to predict stress events up to two months ahead.

    Bloomberg

    Source link

  • Multi-theme index outpaces Mag 7 amid AI growth | Insights | Bloomberg Professional Services

    Meanwhile, the China Tech 8, a company basket created by Bloomberg Intelligence and including companies such as Alibaba Group Holding, Baidu, and Pinduoduo Inc, gained 42%, but rising trade tensions have since slowed the momentum of that rally.

    Notably, as the chart below shows, leadership dynamics have shifted: Intel, Alibaba and Baidu now occupy the chart’s upper-right corner, signaling rising strength. Meanwhile, Meta, Nvidia and Microsoft have slowed. The scatter chart uses proprietary indicators of relative performance of securities versus the benchmark. The X-axis shows the RS-ratio, measuring strength. The Y-axis shows RS-momentum, measuring direction and pace of the RS-ratio line.

    Among firms holding the most cash, Alibaba led with a 99% jump and Tencent followed with 52%. Amazon and Apple moved the other way, posting declines. The rise of AI enablers like Samsung, Engie, Broadcom and Siemens Energy added to the index’s gains.

    Bloomberg Intelligence notes that index members hold roughly $410 billion in cash, giving them the capacity to accelerate growth across themes including AI, disruptive energy and robotics. Since 2018, they’ve already funneled about $850 billion into deals across public and private markets, underscoring their willingness to reinvest in innovation.

    Tracking

    To see rotation of leadership momentum in Multi-Thematic Index, run Relative Rotation Graph using RRG function on the Bloomberg Terminal.

    To analyze returns, cash holdings, forecast growth and valuations run BMULTIT Index WATC for a view of multi-thematic index fundamentals.

    For the latest Bloomberg Intelligence thematic research, run BI THEM on the Bloomberg Terminal.

    To analyze the characteristics of BMULTIT Index’s outperformance, run BMULTIT Index PORT WS /I

    Figure 3 - Multi-theme index outpaces Mag 7 amid AI growth

    For more information on this or other functionality, click here to request a demo with a Bloomberg sales representative. Existing clients can press on their Bloomberg keyboard.

    Bloomberg

    Source link

  • Bloomberg AI in Finance Summit Highlights | Insights | Bloomberg Professional Services

    The data included in these materials are for illustrative purposes only. The BLOOMBERG TERMINAL service and Bloomberg data products (the “Services”) are owned and distributed by Bloomberg Finance L.P. (“BFLP”) except (i) in Argentina, Australia and certain jurisdictions in the Pacific Islands, Bermuda, China, India, Japan, Korea and New Zealand, where Bloomberg L.P. and its subsidiaries (“BLP”) distribute these products, and (ii) in Singapore and the jurisdictions serviced by Bloomberg’s Singapore office, where a subsidiary of BFLP distributes these products. BLP provides BFLP and its subsidiaries with global marketing and operational support and service. Certain features, functions, products and services are available only to sophisticated investors and only where permitted. BFLP, BLP and their affiliates do not guarantee the accuracy of prices or other information in the Services. Nothing in the Services shall constitute or be construed as an offering of financial instruments by BFLP, BLP or their affiliates, or as investment advice or recommendations by BFLP, BLP or their affiliates of an investment strategy or whether or not to “buy”, “sell” or “hold” an investment. Information available via the Services should not be considered as information sufficient upon which to base an investment decision. The following are trademarks and service marks of BFLP, a Delaware limited partnership, or its subsidiaries: BLOOMBERG, BLOOMBERG ANYWHERE, BLOOMBERG MARKETS, BLOOMBERG NEWS, BLOOMBERG PROFESSIONAL, BLOOMBERG TERMINAL and BLOOMBERG.COM. Absence of any trademark or service mark from this list does not waive Bloomberg’s intellectual property rights in that name, mark or logo. All rights reserved. © 2025 Bloomberg.

    Bloomberg

    Source link

  • AI a game changer for power demand | Insights | Bloomberg Professional Services

    Renewable energy sources such as wind and solar are emissions-free, like nuclear, but intermittent, which suggests they must be paired with storage or backup capacity. Coal and natural gas-fired power plants have higher capacity factors yet generate greenhouse-gas emissions, making them an unlikely choice for hyperscalers that prioritize environmental stewardship.

    Capacity factor, a measure of efficiency, is the ratio of power produced by a source over the total amount it could have produced if running at full strength over a certain period. Nuclear has the highest capacity factor by far of any power source at 90%, followed by coal and combined-cycle gas (60%), wind (35%) and solar (25%). Yet nuclear also costs the most, at $12,500 a kilowatt, well ahead of coal ($5,000), combined-cycle gas ($2,500) and wind and solar ($1,500 each).

    New generation capacity of 131-310 GW may be needed to supply the 345-815 terawatt-hours of power required to support US growth in AI computing by 2030. This would equate to an increase of 11-26% from 2023. It assumes a capacity factor of only 30%, in line with wind and solar, since any new baseload capacity may have to be outfitted with costly and unproven carbon-capture technology, according to Environmental Protection Agency regulations, and new nuclear plants may take much longer to build.

    Our analysis assumes all AI power needs are met with new generation, rather than increasing output from existing plants. It also excludes the cost of backup, data-center or transmission infrastructure, as well as any fuel efficiencies obtained via advances in chip technology.

    Contract premiums of $15-$25 Per MWh

    Data centers are willing to pay a premium for that power, leaving generators such as Constellation Energy and Vistra poised for major Ebitda gains from nuclear power deals set at above-market prices. Constellation’s January VPPA with the US General Services Administration – $840 million for 1 million MWh annually for 10 years – implies a contract price of mid-$80/MWh.

    This equates to $15-$25 above market, assuming wholesale power in the low-to-mid $50s and capacity in the low-to-mid $10s. As Figure 8 shows, if each company contracts half of its nuclear capacity at the midpoint $20/MWh premium, Constellation could get a $1.8 billion annual Ebitda boost and Vistra $500 million.

    These premiums may prove sustainable, underpinned by data-center demand and limited new nuclear supply. Constellation is by far the largest US merchant nuclear owner, with capacity of more than 22,000 megawatts, followed by Vistra (6,500), Public Service Enterprise Group (3,800), NextEra (2,300), Talen (2,200) and Dominion (2,000).

    Recent nuclear data-center deals have shifted toward front-of-the-meter (FTM) virtual power-purchase agreements (VPPAs) that draw from the grid, avoiding regulatory hurdles tied to off-grid behind-the-meter (BTM) setups that supply power directly. Constellation’s VPPAs with Microsoft (835 MW from the Crane plant restarting in 2027) and Meta (Clinton, after Illinois subsidies expire in 2027) illustrate this trend.

    Yet BTM PPAs may offer key advantages to data centers by providing greater load control and operational flexibility. Since the generation is located on-site, electricity can flow directly to the data center, bypassing the network. New nuclear capacity developed under this model could alleviate grid congestion and help prevent cost-shifting to non-data center users. Also, avoiding the interconnection queue may reduce capital costs and deployment timelines compared with front-of-the-meter projects.

    PSEG estimates incremental transmission costs can reach $7 a megawatt-hour for FTM solutions, costs that behind-the-meter deployments would largely avoid. Yet BTM connections require FERC approval.

    PPA opportunities could emerge for Constellation’s other Illinois-subsidized plants – Quad Cities, Byron, Dresden and Braidwood – totaling 8 GW. In New Jersey, Salem and Hope Creek lost their subsidies in May and may be candidates, with PSEG owning 2.5 GW and Constellation 1 GW. Talen Energy and Amazon Web Services recently expanded their 960-MW Susquehanna agreement in Pennsylvania into a 1.9 GW VPPA, including a 300 MW BTM deal.

    This analysis comes from Bloomberg Intelligence’s “Nuclear Power 2026 Outlook”. Terminal subscribers can find the full version of this analysis on BI . 

    Bloomberg

    Source link

  • An AI Wake-Up Call From Walmart’s CEO

    This is an edition of the WSJ Careers & Leadership newsletter, a weekly digest to help you get ahead and stay informed about careers, business, management and leadership. If you’re not subscribed, sign up here.

    In the Workplace

    Walmart’s CEO issued an AI wake-up call, saying the technology will wipe out some jobs and reshape the company’s workforce. Doug McMillon’s remarks—which echo those made by leaders at Ford, JPMorgan Chase and Amazon—reflect a rapid shift in how executives discuss the potential human cost of AI.

    Copyright ©2025 Dow Jones & Company, Inc. All Rights Reserved. 87990cbe856818d5eddac44c7b1cdeb8

    Source link

  • AI data center workload pivot favors databases over applications | Insights | Bloomberg Professional Services

    Slowdown in application workloads ahead

    As AI agents automate more steps in everyday workflows, less of that work needs to run inside large application suites. That points to slower growth in data-center demand for enterprise resource planning, customer relationship management, human capital management and supply-chain management software. Reasoning-model agents and deep research tools can now autonomously browse the web, pull sources and run analyses on their own — tasks that previously lived in those apps’ user interfaces.

    Engineering software — computer-aided design and computer-aided manufacturing — may skirt these headwinds, as simulation and synthetic-data creation keep workloads anchored in specialized tools.

    Coding agents supercharge testing workloads

    AI coding agents — assistants inside developer tools that suggest, write and fix code — should give a big boost to application development and testing workloads. Agents from Cursor, Anthropic’s Claude Code, GitHub Copilot, OpenAI’s Codex and Gemini Code Assist handle tasks like debugging and appending to existing code. Companies report 30-40% productivity gains on new code written with these agents, which should channel more development and testing to AI data centers. Prompt-based code generation is quickly becoming one of the most-used generative-AI features in existing business applications.

    Workload by Accelerator Type

    Content delivery, cybersecurity also benefit

    As autonomous AI agents plug into business workflows, more mission-critical tasks will run in AI data centers. The rise of reasoning models like OpenAI’s o3 shifts the focus to ensuring that infrastructure is fast, efficient and reliable from simply having a model. That’s a tailwind for content delivery networks (CDNs) from companies like Cloudflare and cybersecurity providers such as Zscaler. Most companies seek to integrate internal knowledge databases and documentation with LLMs while relying on CDN and cybersecurity vendors to manage token consumption for LLM fine-tuning and inferencing.

    Content Delivery Vertical Growth

    Bloomberg

    Source link

  • Inside AI’s rapid expansion: What every investor should know | Insights | Bloomberg Professional Services

    The data and other information included in this publication is for illustrative purposes only, available “as is”, non-binding and constitutes the provision of factual information, rather than financial product advice. BLOOMBERG and BLOOMBERG INDICES (the “Indices”) are trademarks or service marks of Bloomberg Finance L.P. (“BFLP”). BFLP and its affiliates, including BISL, the administrator of the Indices, or their licensors own all proprietary rights in the Indices. Bloomberg L.P. (“BLP”) or one of its subsidiaries provides BFLP, BISL and its subsidiaries with global marketing and operational support and service. Certain features, functions, products and services are available only to sophisticated investors and only where permitted. Bloomberg (as defined below) does not approve or endorse these materials or guarantee the accuracy or completeness of any information herein, nor does Bloomberg make any warranty, express or implied, as to the results to be obtained therefrom, and, to the maximum extent allowed by law, Bloomberg shall not have any liability or responsibility for injury or damages arising in connection therewith. Nothing in the Services or Indices shall constitute or be construed as an offering of financial instruments by Bloomberg, or as investment advice or investment recommendations (i.e., recommendations as to whether or not to “buy”, “sell”, “hold”, or to enter or not to enter into any other transaction involving any specific interest or interests) by Bloomberg. Information available via the Index should not be considered as information sufficient upon which to base an investment decision. All information provided by the Index or in this publication is impersonal and not tailored to the needs of any person, entity or group of persons. Absence of any trademark or service mark from this list does not waive Bloomberg’s intellectual property rights in that name, mark or logo. For the purposes of this publication, Bloomberg includes BLP, BFLP, BISL and/or their affiliates.

    BISL is registered in England and Wales under registered number 08934023 and has its registered office at 3 Queen Victoria Street, London, England, EC4N 4TQ. BISL is authorized and regulated by the Financial Conduct Authority as a benchmark administrator.

    © 2025 Bloomberg. All rights reserved.

    Bloomberg

    Source link

  • Etsy drifts further away from its roots with first Super Bowl ad

    Etsy drifts further away from its roots with first Super Bowl ad


    Etsy Inc., once known as a quirky marketplace for handmade, artisanal and vintage items, seems to be moving further away from its origins amid a much tougher e-commerce landscape and the impact of AI.

    Etsy
    ETSY,
    +4.83%
    will be marketing to a whole new audience on Sunday, when its first Super Bowl commercial will run. The 30-second ad is quirky; it depicts a generic 19th-century American leader who’s flummoxed over how to reciprocate France’s gift of the Statue of Liberty. With the help of an anachronistic smartphone, he and his team search on Etsy using its new Gift Mode option, and find its “Cheese Lover” category after determining that the French love cheese. Voilà — they decide to send the French some cheese.

    The commercial is part of Etsy’s push of a new user interface featuring Gift Mode, which lets shoppers search for gifts for a specific type of person or occasion — combining generative AI and human curation to give gift buyers some unusual options.

    But are these moves desperate and costly efforts to try to reach potential new buyers, coming on the heels of Etsy’s plans to lay off 11% of its staff?Or could running a TV ad at the most expensive time of the year actually lead to more sales on the once-fast growing marketplace?

    Etsy believes these moves will help the company grow again, and its research shows the average American spends $1,600 a year on gifts. “There is no single market leader and Etsy sees a real opportunity to become the destination for gifting,” Etsy’s Chief Executive Josh Silverman said in a recent blog post.

    Etsy is clearly under pressure after seeing its gross merchandise sales more than double in 2020 during the pandemic, when it became a go-to place to buy handmade masks and all kinds of items for the home, from vintage pieces to antiques to castoffs. From personal experience as an Etsy seller, I saw sales at my own small vintage-clothing shop more than double in 2020 and then fall back in 2021, while still remaining higher than in 2019. In the last two years, sales have slowed, and some other sellers have witnessed similar patterns, based on their comments in seller forums.

    The number of sellers and buyers on the platform has increased on the same level as gross merchandise sales. But e-commerce competition has also gotten more fierce.

    “Our main concern with Etsy is growing competition in the space from new players like Temu,” said Bernstein Research analyst Nikhil Devnani, in an email. Temu and fellow Chinese online retailer Shein have raised a lot of investor jitters, as Etsy’s gross merchandise sales have slipped over the last year and are forecast to fall again in its upcoming fourth-quarter earnings report later this month.

    Devnani said a Super Bowl ad could potentially help the marketplace gain visibility, something it has always lacked.

    “One dynamic they’ve talked about a lot is that brand awareness/recollection is still low, and this keeps frequency low,” he said, noting that Etsy buyers shop on the site about three times per year, on average. “They want to be more top-of-mind … Super Bowl ads are notoriously expensive of course, but can be impactful/get noticed.”

    The company’s big focus on Gift Mode, however, could be a risky strategy. How many times a year do consumers look for gifts? And in a note Devnani wrote in October, before the company’s Gift Mode launch, he said that one of the concerns investors have is that Etsy is too niche. “’How often does someone need something special?’ is the rhetoric we hear most often,” he said. Etsy, then, is counting on buyers returning for other items for themselves.

    Etsy CEO Silverman believes buyers will come back again and again to purchase gifts. Naved Khan, a B. Riley Securities analyst, said in a recent note to clients that he believes Gift Mode plays to Etsy’s core strengths, offering “unique goods at reasonable prices” versus the mass-produced products sold on Shein, Temu, Amazon.com Inc.
    AMZN,
    +2.71%    ,
    and other sites.

    Consumer spending has changed, though. At an investor conference in December, Silverman said that consumers are spending on dining out and traveling, instead of buying things.

    But while investors still view Etsy as a niche e-commerce site, some buyers and sellers see it overrun with repetitive, non-relevant ads. Complaints about a decline in search capabilities, reliance on email and chat for support, and constant tech changes are common on seller forums and Facebook groups. AI-generated art offered by newer sellers as a side hustle has also become a thought-provoking, debated issue. And there are complaints about mass-produced items making their way on the site.

    Etsy said that in addition to its human and automated efforts, it also relies on community flags to help take down infringing products that are not allowed on its marketplace, and that community members should contact the company when if they see mass-produced items for sale on the site.

    It also continues to work on search. On its last earnings call, Silverman said the company was moving beyond relevance to the next frontier of search, one “focused on better identifying the quality of each Etsy listing utilizing humans and [machine-learning] technology, so that from a highly relevant result set we bring the very best of Etsy to the top — personalized to what we understand of your tastes and preferences.”

    The pressure could build on the company if its latest moves don’t generate growth. Etsy recently gave a seat on its board to a partner at activist investor Elliott Management, which bought a “sizable” stake in the company in the last few months. Marc Steinberg, who is responsible for public and private investments at Elliott, has also has been on the board at Pinterest
    PINS,
    -9.45%
    since December 2022.

    Elliott Management did not respond to questions. But in a statement last week, Steinberg said he was joining the board because he “believe[s] there is an opportunity for significant value creation.” Some sellers fear that the pressure from investors and Wall Street will lead to Etsy allowing mass-produced products onto the site. In its fall update, Etsy said the number of listings it removed for violating its handmade policy jumped 112% and that it was further accelerating such actions.

    Etsy’s stock before the news of Elliott’s stake was down about 18% this year. Its shares are now off about 3.65% this year, after recently having their best day in seven years on the news that Steinberg joined the board.

    Etsy is a unique marketplace that for many years had a much better reputation than some of its rivals, like eBay
    EBAY,
    +0.98%    .
    But since going public and answering to Wall Street, the need to provide growth and profits for investors has become much more of a driver. The Super Bowl ad and Gift Mode may bring a broader awareness to Etsy, but will it be the right kind of awareness? Sellers like me hope these new efforts will stave off the continuing fight with the likes of Temu and other vendors of mass-produced products, and help Etsy retain the remaining unique aspects of its marketplace.



    Source link

  • Hewlett Packard Enterprises to buy Juniper Networks in $14 billion deal

    Hewlett Packard Enterprises to buy Juniper Networks in $14 billion deal

    In an effort to keep up in the accelerating AI arms race, cloud-services provider Hewlett Packard Enterprise Co. on Tuesday agreed to buy Juniper Networks, Inc. in a deal worth around $14 billion.

    Under the terms of the deal, Hewlett Packard Enterprises
    HPE,
    -8.92%
    will acquire Juniper
    JNPR,
    +21.81%
    — which makes communications-networking products and also has an AI segment called Mist AI — for $40 a share. The companies expect the deal to close late this year or in early 2025.

    “The acquisition is expected to double HPE’s networking business, creating a new networking leader with a comprehensive portfolio that presents customers and partners with a compelling new choice to drive business value,” the companies said in a release.

    After the deal is completed, Juniper Chief Executive Rami Rahim will lead the combined HPE networking business, and report to HPE CEO Antonio Neri.

    “This transaction will strengthen HPE’s position at the nexus of accelerating macro-AI trends, expand our total addressable market, and drive further innovation for customers as we help bridge the AI-native and cloud-native worlds, while also generating significant value for shareholders,” Neri said in a statement.

    HPE said the addition of Juniper will boost margins and result in up to $450 million in annual cost savings within three years of the deal’s completion, as well as accelerate growth. HPE’s networking segment was the company’s top source of quarterly earnings before taxes, $401 million, on $1.4 billion in revenue.

    HPE’s deeper plunge into networking closes a chapter of sorts. Then-Hewlett-Packard Co. acquired Aruba Networks for about $3 billion in March 2015, months before Silicon Valley’s original garage startup split in half, resulting in the formation of HPE, which sells servers and other equipment for data centers, and HP Inc.
    HPQ,
    -2.71%    ,
    which makes PCs and printers.

    The Wall Street Journal reported the possibility of a deal on Monday, sending shares of Juniper higher.

    Shares of Juniper
    JNPR,
    +21.81%
    rose 0.5% after hours, after jumping 21.8% during regular trading hours. Hewlett Packard
    HPE,
    -8.92%
    shares were down 0.4% after hours, after falling 8.9% during the day.

    As of Tuesday’s close, Juniper had a market cap of $9.64 billion, while HPE’s was $23.04 billion.

    The companies hope the deal can provide a much-needed jolt after a series of lackluster quarterly earnings. Juniper shares have gained 15.7% over the past 12 months, while HPE shares are down 5.4% over that span. The S&P 500
    SPX,
    in comparison, is up about 21.4% over the past year.

    For decades, Juniper has lagged rival Cisco Systems Inc.
    CSCO,
    -1.09%
    in the networking-equipment market. In its most recent quarter, Juniper reported net income of $76 million on revenue of $1.4 billion, down 1% from the same quarter a year earlier.

    Source link

  • Synopsys and Ansys in talks to merge: report

    Synopsys and Ansys in talks to merge: report

    Shares of Ansys Inc. soared 18% in trading Friday on reports the company is in discussions to be acquired by Synopsys Inc. in a deal that would create a design-software behemoth.

    The potential deal would kick off 2024 with a mega-merger, even as the Federal Trade Commission attempts to crack down on such transactions. Talks remain fluid and a third party might still emerge as a possible suitor of Ansys, according to a Wall Street Journal report, which cited people familiar with the situation.

    Ansys
    ANSS,
    +18.08%    ,
    which has a market value of nearly $26.3 billion, makes software that helps predict how products in aerospace, healthcare and automotive applications will work in the real world. A deal could be struck early in 2024, according to people familiar with the matter. Ansys reported revenue of $2.1 billion in 2022.

    Synopsys
    SNPS,
    -6.34%    ,
    with a market value of $85.1 billion, makes software that engineers use to design and test silicon chips used in smartphones, self-driving cars and other forms of artificial intelligence. Its stock has climbed 65% this year as investors have hopped on the AI bandwagon boom. Shares of Synopsys dipped 6% in late trading Friday.

    Synopsys’s customers include Nvidia Corp.
    NVDA,
    -0.33%    ,
    Intel Corp.
    INTC,
    +1.95%
    and Advanced Micro Devices Inc.
    AMD,
    -0.22%    .

    Representatives from Synopsys and Ansys were not immediately available for comment.

    Should the companies strike a merger, it would offer a fresh test for the FTC and its chair, Lina Khan, who have opposed large tech mergers and acquisitions. The agency unsuccessfully sued Facebook parent Meta Platforms Inc.
    META,
    -0.20%
    in its pursuit of VR developer Within, as well as Microsoft Corp.’s
    MSFT,
    +0.28%
    $69 billion purchase of Activision Blizzard Inc.

    Source link

  • AMD wins high praise for AI advancements as its stock soars 6%

    AMD wins high praise for AI advancements as its stock soars 6%

    While Advanced Micro Devices Inc. shares didn’t enjoy a Wednesday bump during the company’s artificial-intelligence event, they were rallying sharply Thursday as analysts reflected on the chip maker’s presentation.

    Chief Executive Lisa Su and her team “put together one of the most impressive new product event/launches by our reckoning in the last decade, perhaps ever,” Rosenblatt Securities analyst Hans Mosesmann wrote in a note to clients.

    The launch of AMD’s
    AMD,
    +7.09%
    MI300X AI/graphics-processing-unit accelerator “was not just a speeds and feeds geek fest (it was that for sure, with AMD claiming superiority in AI inferencing), but an industry movement coalescing around the concept of ‘open’ sourced technologies are preferred (demanded really), to address the insanely fast/accelerating life-changing thing that AI has become,” Mosesmann continued.

    Opinion: AMD’s new products represent the first real threat to Nvidia’s AI dominance

    He was also impressed by the company’s talk of its software platform ROCm, which he thinks is catching up to Nvidia Corp.’s
    NVDA,
    +1.54%
    CUDA.

    “Of course, Nvidia is not going away, and we are quite sure will remain the dominant AI player for years to come but AMD we feel made the case yesterday that they will be an important AI innovator on a secular basis,” Mosesmann noted, as he kept his outperform rating and $200 target price on the stock.

    AMD shares were up 6% in Thursday morning trading.

    Baird’s Tristan Gerra was also impressed.

    “Rapidly unfolding hyperscaler engagements, highly competitive AI architecture specs, along with accelerated new product roadmap, bode well for share gains and continued acceleration in AI-related revenue for AMD beyond 2024, while faster-than-expected rate of adoption so far could potentially drive upside in the AI revenue outlook for 2024, in our view,” he wrote.

    Read: Nvidia and Microsoft CEOs say industrial companies will benefit most from AI. Here are stocks to put on your watch list.

    Gerra also sees the potential for “high-volume deployments,” thanks to the “significant software milestones” AMD is showing. He rates the stock at outperform with a $125 target price.

    TD Cowen’s Matthew Ramsay said that AMD’s event reinforced his belief that the company “is well positioned to meaningfully participate” in the large total addressable market for AI accelerators.

    The company called out Microsoft Corp.
    MSFT,
    -0.01%    ,
    Meta Platforms Inc.
    META,
    +2.41%
    and Oracle Corp.
    ORCL,
    -0.08%
    as customers, announcements that were “strong” but not “surprising,” in Ramsay’s view.

    “We remain encouraged that AMD is making an impressive case (and is getting customer support) to provide adaptive computing solutions for both training and inference in increasingly large [generative-AI] infrastructure builds,” he wrote. “We believe this signifies a strong AI strategy of delivering a broad portfolio of [central processing unit], GPU, and [field-programmable gate array] assets, with open software that enables easily deployed AI workloads while leveraging the company’s existing partnerships to accelerate its AI ramps at-scale.”

    Ramsay has an outperform rating and $130 target price on AMD shares.

    Source link

  • Why Sam Altman is a no-brainer for Time’s ‘Person of the Year’

    Why Sam Altman is a no-brainer for Time’s ‘Person of the Year’

    Nothing has changed our lives more this year than the advances made in artificial intelligence — and they have the potential to alter our lives in even more dramatic ways down the road.

    So it’s a no-brainer that Sam Altman, co-founder and recently returned chief executive of the once-little-known OpenAI, should be named “Person of the Year” by Time Magazine when the selection is announced Wednesday.

    Altman has already cracked Time’s shortlist, joining candidates from varied backgrounds, including world leaders like Xi Jinping and entertainment phenomenon Taylor Swift. The selection ultimately comes down to an “individual or group who most shaped the previous 12 months, for better or for worse.”

    But Time has often given “agents of change” its yearly honor — just look at 2021 winner Elon Musk — and Altman certainly fits that bill.

    No other innovation in the past year has had an impact in such disparate realms. OpenAI publicly launched its ChatGPT chatbot late last year, and as the technology grew viral in 2023, it upended the stock market, Silicon Valley and companies that wouldn’t normally be classified as technology businesses. The ensuing product development and surge in generative AI investment revitalized a tech industry that had sunk into the doldrums amid a pandemic hangover.

    Admittedly, it will take time for companies to realize the true financial benefits of AI: Nvidia Corp.
    NVDA,
    -2.68%
    is among the few to generate serious money from the frenzy so far. But market researcher IDC predicted that global spending on AI, including software, hardware and services for AI-centric systems will reach $154 billion this year, up 27% from a year ago. That total could zoom above $300 billion by 2026.

    Also read: One year after its launch, ChatGPT has succeeded in igniting a new era in tech

    And AI isn’t only impacting the corporate world. The technology is already affecting our daily lives, and it will have even deeper effects going forward. Chatbots are getting smarter on websites, facilitating better customer service. They’re starting to alter the workplace as well, spitting out mostly coherent marketing copy, research and even, gasp, news articles — albeit with plenty of errors.

    At first, ChatGPT seemed like a fun way to kill time or get homework help, but the chatbot and its ilk will seriously alter the working world, helping to eliminate perhaps millions of jobs. Morgan Stanley recently predicted that more than 40% of occupations will be affected by generative AI in the next three years.

    Altman himself has been the face of OpenAI in the past year. He’s talked up the technology, but he also appeared at congressional hearings in May to discuss potential regulation of AI, testifying that “if this technology goes wrong, it can go quite wrong.” His recent firing and quick rehiring by OpenAI and its small, nonprofit board late last month fueled a veritable media storm before the Thanksgiving holiday in the U.S.

    Time chooses its persons of the year for their impact, not because they’re saints. And Altman’s own story is not without controversy. The recent brouhaha over his leadership of OpenAI is believed to have been caused by a deep schism over the ethics of AI development. The board seemingly wanted more guardrails and precautions, and feared that rushed development could irrevocably doom mankind.

    Read in the Wall Street Journal: How effective altruism split Silicon Valley and fueled the blowup at OpenAI

    Altman, who also wooed Microsoft Corp.
    MSFT,
    -1.43%
    to become an investor in OpenAI, emerged the victor in the upheaval with his own company’s altruistic board. Had Altman truly been fired from OpenAI, Microsoft was planning to hire him, and nearly every employee at OpenAI was ready to quit and follow him there. While OpenAI faces plenty of competition, including from Alphabet Inc.’s
    GOOG,
    -2.02%

    GOOGL,
    -1.96%
    Google, Altman should continue to be the face of AI development, for good and for bad, even as he has advocated industry regulation.

    The debut and influence of ChatGPT and follow-on AI products are having the biggest impact on tech development since the invention of the iPhone. Altman is at the center of it and leading the charge. Whether he can keep the lid on Pandora’s Box or not depends on many factors, but he and the company he leads are clearly driving a new tech movement that affects us all, whether we like it or not.

    Source link