Global stock markets have fallen sharply amid concerns that a boom in valuations of artificial intelligence (AI) companies could be rapidly cooling.
Markets in the US, Asia and Europe have fallen after bank bosses warned a serious stock market correction could be ahead, after a run of record stock market highs led some companies to appear overvalued.
In the US the tech-focused Nasdaq and the S&P 500 suffered their largest one-day percentage drop in almost a month on Tuesday.
Technology shares pulled the Nasdaq lower, which resulted in it closing 2% down. Meanwhile, there were one-day falls for all of the “magnificent seven” AI-related stocks: including the chipmaker Nvidia, Amazon, Apple, Microsoft, Tesla, Alphabet (the owner of Google) and Meta (the owner of Facebook, Instagram and WhatsApp).
The S&P also closed down just over 1% as it lost ground because of tech stocks, particularly the data analytics company Palantir, which slumped by almost 8% despite having raised its revenue outlook the previous day.
Palantir has also been targeted by a well-known short-seller – who bets on falls in the value of a company.
The investor Michael Burry – who rose to prominence after predicting the 2008 financial crash and inspiring the film The Big Short – bet against Palantir and the chipmaker Nvidia, two of the biggest AI companies, sparking criticism from Palantir’s boss and a stock sell-off.
In an interview on CNBC, Alex Karp, Palantir’s chief executive, criticised Burry and other short-sellers for “trying to call the AI revolution into question”.
Asian markets followed the US falls on Wednesday, recording the sharpest slide in seven months, as the concerns about tech stocks spread, and indices in Japan and South Korea dropped more than 5% from the record highs reached the previous day. In Europe, markets in the UK, France and Germany dropped slightly on Wednesday morning.
The market falls came as the chief executives of Morgan Stanley and Goldman Sachs cautioned there could be a market correction ahead.
They added their voices to that of Jamie Dimon, the head of the US’s largest bank, JP Morgan Chase, who warned in October he was worried markets would crash in the next six months to two years.
Jim Reid, an analyst at Deutsche Bank, said there was a “growing chorusdiscussing whether we might be on the verge of an equity correction”.
skip past newsletter promotion
after newsletter promotion
Reid added: “The last 24 hours have brought a clear risk-off move, as concerns over lofty tech valuations have hit investor sentiment.”
Other analysts have raised questions about investment in AI companies, highlighting that the vast majority of investment in AI programmes has been promised to a very small group of tech companies, particularly OpenAI and Nvidia, while there has so far been little return on investment.
The price of bitcoin briefly dipped below $100,000 (£76,764) for the first time since June, as investors withdrew their money from riskier assets such as cryptocurrencies over fears about the economic outlook.
Bitcoin touched a record price of more than $126,000 in early October but went on to fall by 3.7% during the month, resulting in the worst monthly performance by the cryptocurrency in the last decade, according to figures from CoinMarketCap.
The four unamplified samples were sequenced to an average (± standard error [SE]) depth (i.e. total number of reads) of 73,614 ± 5047 reads (forward and reverse); of those, 72,243 ± 4849 reads passed quality filtering and were used for mapping. Nearly all reads were successfully mapped to the blue tit genome (approx. 99.9%), with only two reads from one unamplified sample, one read from another sample and none from the remaining two samples mapped to the parasite genome. SWGA increased the proportion of reads mapping to the parasite genome in all four samples (Fig. 1). For example, the unamplified sample with two parasite-mapping reads out of 80,839 total reads (2.5 × 10–5 or 0.0025%), increased to approximately 0.65% of reads mapping to the parasite genome with primer sets 1 and 2 on the shallow sequencing run. The shallow sequencing run of the amplified samples was sequenced to a depth (± SE) of 365,104 ± 11,496 reads (excluding negative controls which were sequenced to a depth of 13,900 ± 4484 reads).
Fig. 1
The proportion of reads that mapped to the Haemoproteus majoris parasite genome in unamplified (SWGA primer set “none”) DNA extracted from H. majoris (PARUS1 lineage)-infected blue tits (Cyanistes caeruleus) and DNA that underwent SWGA with one of three primer sets (1–3). Triangles denote individual samples, and the same samples sequenced under different conditions are connected with a solid line. Three negative controls (water) were amplified in each of the three SWGA primer set reactions (1 reaction per primer set) and are represented as filled circles connected with a dashed line. SWGA, Selective whole genome amplification
Primer sets
Primer sets 1 and 2 produced a higher proportion of reads mapping to the parasite genome than primer set 3 in the shallow sequencing run (Fig. 1). Mean read depth was positively correlated with breadth of coverage, the latter measured as the proportion of the genome with at least 1X coverage (r > 0.99 for all host and primer combinations). Mean read depth and breadth of coverage were higher for mapping to the host genome than the parasite genome in all cases (Fig. 2). For parasite mapping, primer sets 1 and 2 had greater breadth of coverage than primer set 3 (F2,16 = 15.58, p < 0.001), and this was independent of sample year (i.e. age of the sample; F2,6 = 0.47, p = 0.646; Fig. 2).
Fig. 2
Breadth of coverage measured as the proportion of the host or parasite genomes with at least 1X depth of coverage by SWGA primer set and sample year. Breadth of coverage is positively correlated with mean read depth (r > 0.99), and is higher for mapping to the host genome than to the parasite genome. SWGA primer sets 1 and 2 had higher breadth of coverage (and mean depth of coverage) than primer set 3. SWGA, Selective whole genome amplification
The negative control for primer set 1 had 906 reads that passed quality filtering, of which 189 mapped to the blue tit and none mapped to the parasite. The negative control for primer set 2 had 32,478 reads, of which 13,841 mapped to the blue tit and five mapped to the parasite genomes. Finally, the negative control of primer set 3 had 48,152 reads, of which eight mapped to the blue tit genome and none to the parasite genome. Consistent with these relatively small numbers of mapped reads, we did not find bands in the agarose gels for any of the three negative controls, suggesting that there was no contamination during the SWGA reactions.
Deep sequencing with primer set 2
We chose primer set 2 for deep sequencing as it seemed to work as well as primer set 1 and better than primer set 3 (Fig. 1). We re-sequenced the libraries of the nine samples prepared with SWGA primer set 2 on a NextSeq 2000 (56,426,167 ± 6,373,964 [SE] reads per sample) resulting in 53,238,047 ± 15,405,467 reads per sample that passed quality filtering. The mean (± SE) depth of coverage was 11.3 ± 1.85 for the host genome and 1.17 ± 0.446 for the parasite genome per sample. The mean (± SE) breadth of coverage (i.e. proportion of the genome with at least 1X depth of coverage) was 0.743 ± 0.020 for the host genome and 0.334 ± 0.075 for the parasite genome per sample (Fig. 3). The average (± SE) proportion of reads mapping to the parasite genome decreased slightly from 0.0025 ± 0.0008 in the shallow sequencing run to 0.0018 ± 0.0005 in the deep sequencing run (F1,8 = 36.79, p < 0.001).
Fig. 3
Breadth of coverage measured as the proportion of the host or parasite genomes with at least 1X depth of coverage after deep sequencing of samples amplified by SWGA. Most samples had high breadth of coverage across the host genome; variation among samples was greater for the parasite genome, as would be expected given that there is variation in parasitemia. SWGA, Selective whole genome amplification
An important aspect of any genome sequencing technique for population genomics is sequencing of the same genomic regions among samples so that genetic variants can be identified. We determined the consistency among samples by correlating the average sequencing depth of coverage over 10-kb windows between samples for the host and parasite genomes separately. The mean (± SE) correlation coefficient (r) among these 10-kb windows was 0.929 ± 0.007 for the host genome and 0.472 ± 0.028 for the parasite genome among samples. To understand the degree of overlap in parasite sequencing in a best-case scenario (i.e. where parasite genome sequencing works very well), we calculated the number of nucleotide positions with at least 1X depth of coverage in common between the two best sequenced infections. The two best sequenced parasite infections shared 10,821,340 nucleotide positions (approx. 45% of the reference genome) with at least 1X depth of coverage. Individually, the two infections had 14,090,382 (approx. 59% of the reference genome) and 14,646,363 (approx. 61% of the reference genome) nucleotide positions sequenced to 1X depth of coverage.
We generated separate variant files for the nine host individuals and the nine parasite infections. The raw host variant file had 16,517,700 variants (14,276,887 SNPs and 2,344,956 indels; some sites are classified by the software as both SNPs and indels so their total is greater than the number of variants); on average (± SE) there was less than one host individual missing per variant (0.912 ± 0.0004), and variant depth (123.19 ± 0.031) and quality (1012.78 ± 0.486) were high. The raw parasite variant file had 872,832 variants (787,509 SNPs and 86,449 indels); on average (± SE) about seven of nine parasite infections were missing per variant (6.737 ± 0.0015), and variant depth (14.52 ± 0.017) and quality (436.02 ± 0.645) were lower, but still relatively high. We then restricted both variant files to variants with no more than three missing individuals (host individuals or parasite infections), a minimum depth of five and a minimum quality of 30; this resulted in a filtered host variant file with 14,827,899 variants (12,806,876 SNPs and 2,118,287 indels) and a parasite variant file with 20,954 variants (17,192 SNPs and 3803 indels). For population genetics inference, one needs to know the number of sites that vary among individuals in the population (VCF, such as that used in the present study, typically calculates the number of sites that vary from the reference; however, the reference often does not come from the target population and in the case of our parasite it is not the same genetic lineage [reference is lineage WW2; samples are lineage PARUS1]). Therefore, we also counted the number of sites with more than one allele in the samples (i.e. sites that varied among individuals in the population) and found 14,512,339 such sites among the host individuals and 7068 among the parasite infections.
The 7068 variable sites for the parasites are conservative in the sense that we filtered the VCF file to include variants with six samples (infections), not all of which sequenced equally well (Fig. 3); restricting the filtering to parasite samples that sequenced well would likely reveal more variants. Therefore, we explored restricting the VCF file to the four best sequenced infections and varying the minimum read depth per variant from 1X to 5X and found additional variable sites (Table 1).
Table 1 The variant call format file was restricted to the four best sequenced infections (1HA37672, 1EP50341, 1EE56066, 1HA37673; Additional file 1: Table S1) and variants where all four infections were present with a minimum read depth of 1X, 2X, 3X, 4X or 5X, and minimum variant quality of 30
Using the larger filtered VCF file (14,512,339 and 7068 variable sites for the hosts and parasites, respectively), we found variation among the samples. A PCA of the filtered host variants captured 39.05% of the cumulative genetic variation over the first three principal components, while a PCA of the filtered parasite variants captured 49.3% of the variation over the first three principal components; both hosts and parasite genetic diversity revealed some degree of clustering over the three time periods investigated (Fig. 4).
Fig. 4
Principal components analysis of host (top two panels) and parasite (bottom two panels), with the shape and color of each individual sample corresponding to the year the sample was collected. The left side panels show the first two PCs while the right panels show the second and third PCs. All axes show the percentage of the genetic variation each PC explained in the analysis. PC, Principal component
Unmapped reads
Because the reference was a different lineage of H. majoris than the samples we sequenced (reference lineage is WW2; samples are lineage PARUS1), some of the unmapped reads may have been parasite reads that did not map because of a high degree of divergence relative to the reference (particularly intronic regions). To explore this possibility, we examined the mean GC content of reads that mapped to the bird, the parasite and remaining unmapped reads for each of the deep sequenced samples. We found unmapped reads to have much lower GC content than bird mapped reads, and similar GC to the parasite mapped reads, suggesting that they may contain unmapped parasite reads (Fig. 5). In the deep sequenced samples, the proportion of non-host reads that did not map to the parasite (i.e. plausibly including parasite reads that were too divergent to map to the parasite reference genome) was on average (± SE) 0.549 ± 0.051 (Additional file 1: Table S2), which is a substantial proportion.
Fig. 5
A histogram of the percentage GC content of reads that mapped to the host or the parasite, or were unmapped is presented for each sample. Parasite reads had much lower GC content than host reads and unmapped reads were more similar in GC content to the parasite reads than to the host reads
Sequencing success and parasitemia
We quantified parasitemia (number of infected red blood cells after examining approximately 10,000 red blood cells) for five infected birds (Additional file 1: Table S1). Despite a low sample size, parasitemia was strongly positively correlated with both the proportion of the reference genome over which the infection was sequenced to at least 1X coverage (r = 0.99, t = 12.04, df = 3, p = 0.001) and mean depth of coverage (r = 0.96, t = 6.12, df = 3, p = 0.009).
Wegovy-maker Novo Nordisk cuts outlook again as obesity drug sales slow Reuters
Ozempic maker Novo Nordisk lowers growth outlook for its weight loss drugs as pricing pressures mount CNBC
Should You Buy Novo Nordisk Stock Before the Huge Investor Update? Nasdaq
Novo Nordisk CEO faces baptism of fire amid board shakeup, Pfizer fight Reuters
Novo Nordisk’s sales increased by 12% in Danish kroner and by 15% at CER in the first nine months of 2025; R&D pipeline progress continues Yahoo Finance
The service behind this page isn’t responding to Azure Front Door.
Gateway Timeout
Azure Front Door cannot connect to the origin server at this time. The origin might be overloaded, misconfigured or under maintenance now. Please contact the site owner for assistance.
Kuwait International Airport is currently undergoing extensive expansion. With the construction of the new Terminal 2, the site will become one of the most modern aviation hubs in the Gulf region. Once completed, the airport is expected to handle an initial 25 million passengers per year – scalable up to 50 million – as part of the national development strategy “Kuwait Vision 2035”.
The contract was awarded by the Directorate General for Civil Aviation (DGCA) in cooperation with the Ministry of Public Works. Limak, an internationally active construction company based in Turkey, is acting as the project’s main developer.
The gensets are engineered to operate reliably even under extreme climatic conditions and are designed for ambient temperatures of up to 55 degrees Celsius. Delivery is scheduled for early 2026, followed by commissioning, including testing and handover.
“We are proud that our products ensure the stable operation of critical infrastructure at the airport—even in this region with its extreme environmental conditions,” said Salim El Banna, Country Sales Manager UAE, Bahrain, Iraq & Kuwait for the Power Systems division of Rolls-Royce.
Rolls-Royce secures critical infrastructure worldwide with more than 85,000 mtu emergency power systems, including airports, data centers, hospitals, industrial plants and energy suppliers. The systems are based on diesel and gas gensets as well as dynamic Uninterruptible Power Supply (UPS) systems, ensuring uninterrupted power even under extreme conditions. At major international airports – including Frankfurt, Dubai, Madrid, Prague, Palma and Hurghada – mtu gensets and combined heat and power systems have been in reliable operation for many years, ensuring uninterrupted operation of terminals, baggage systems and control centers.
Imagery is available for download from: Media Center (mtu-solutions.com)
Danish green hydrogen company Everfuel has inaugurated HySynergy, its 20 MW green hydrogen production plant in Fredericia, Denmark, and sent its first supply to Germany.
It is understood that this is the first time that RFNBO-certified green hydrogen produced in Denmark is being exported abroad.
Jacob Krogsgaard, CEO and Founder of Everfuel, said: “HySynergy is a symbol of Denmark leading the way in Europe’s green transition. We have now shown that green hydrogen can be produced at industrial scale, delivered to industry – and exported. For years, Denmark has talked about the potential of green hydrogen. Now we are showing that it is possible in practice, paving the way for the establishment of a real value chain and, in the coming years, a multiplication of production capacities as the hydrogen pipeline to Germany becomes available.”
New Falcon for XIoT innovations bring deeper industrial and OT telemetry into the Falcon platform, eliminating blind spots and complexity across environments
AUSTIN, Texas and Fal.Con Europe 2025, Barcelona – November 5, 2025 – CrowdStrike (NASDAQ: CRWD) today announced new Falcon® for XIoT innovations, delivering zero-touch asset discovery, real-time segmentation visibility, and unified insight across OT and XIoT environments. These innovations extend the protection of the CrowdStrike Falcon® platform, eliminating blind spots and complexity while consolidating security across IT, cloud, and OT/XIoT environments.
The Expanding Operational Attack Surface
As industrial systems become increasingly connected, security teams face growing blind spots across segmented networks, unmanaged devices, and legacy infrastructure. Many organizations struggle to see what’s connected, how assets communicate, or whether segmentation policies are working as intended – creating opportunities for adversaries to exploit these gaps and move laterally between IT and OT environments.
Zero-Touch Discovery and Unified Visibility for OT Security
Falcon for XIoT delivers continuous operational insight without the hardware dependencies, intrusive scans, or manual effort common in legacy OT tools. With Falcon’s lightweight architecture, defenders gain safe, scalable visibility into operational environments that doesn’t disrupt critical systems. By bringing in additional OT and XIoT telemetry, the Falcon platform gains richer context for faster, smarter security decisions across operational environments.
New Falcon for XIoT Capabilities
CrowdStrike’s latest innovations enhance discovery, monitoring, and asset interaction visibility, providing deeper insights and control over industrial system risks:
Zero-Touch XIoT Discovery: Automatically identifies and inventories industrial assets across segmented networks without dedicated sensors, manual configuration, or intrusive scans – delivering instant visibility without disrupting operations.
Segmentation Visibility: Provides real-time context into device-to-device communication and segmentation policy enforcement to detect violations and reduce lateral movement risk across IT and OT networks.
Dynamic User Experience: Unifies industrial asset and vulnerability data in a single, customizable interface within the Falcon platform, giving teams the context to explore, assess, and act faster.
“Customers are demanding a single platform to understand risk, unify protection, and eliminate complexity across every attack surface,” said Elia Zaitsev, chief technology officer at CrowdStrike. “With these innovations, customers can replace the fragmented tools they’ve been forced to rely on for too long, accelerating consolidation on Falcon.”
To learn more about Falcon for XIoT innovations, read our blog and visit here.
About CrowdStrike
CrowdStrike (NASDAQ: CRWD), a global cybersecurity leader, has redefined modern security with the world’s most advanced cloud-native platform for protecting critical areas of enterprise risk – endpoints and cloud workloads, identity and data.
Powered by the CrowdStrike Security Cloud and world-class AI, the CrowdStrike Falcon® platform leverages real-time indicators of attack, threat intelligence, evolving adversary tradecraft and enriched telemetry from across the enterprise to deliver hyper-accurate detections, automated protection and remediation, elite threat hunting and prioritized observability of vulnerabilities.
Purpose-built in the cloud with a single lightweight-agent architecture, the Falcon platform delivers rapid and scalable deployment, superior protection and performance, reduced complexity and immediate time-to-value.
CrowdStrike: We stop breaches.
Learn more: https://www.crowdstrike.com/
Follow us: Blog | X | LinkedIn | Instagram
Start a free trial today: https://www.crowdstrike.com/trial
This press release includes descriptions of products, features, or functionality which may not currently be generally available. Any such references are provided for informational purposes only. The development, release, and timing of all features or functionality remain at our sole discretion and may change without notice. These statements are subject to risks, uncertainties, and assumptions that may cause actual results to differ materially from those expressed or implied. Customers should make purchasing decisions based only on services and features that are currently generally available. For more information on our existing offerings please talk to your CrowdStrike representative.
IRH Global Trading signs a 20-year Heads of Agreement (HoA) for the purchase and sale of LNG from Delfin’s U.S. export facility.
Delfin to supply LNG on a free-on-board (FOB) basis to Vitol, which will then deliver to IRH Global Trading under a long-term sale and purchase arrangement.
The agreement marks the largest transaction to date for IRH’s energy trading desk.
Strengthens ePointZero’s integrated energy strategy and supports downstream operations across its ecosystem
International Resources Holding RSC Ltd. (IRH), a leading mine-to-market platform and subsidiary of ePointZero, has signed a 20-year Heads of Agreement (HoA) with Delfin LNG LLC (Delfin) and Vitol Inc. (Vitol) for the purchase and sale of 1.0 million tonnes per annum (MTPA) of liquefied natural gas (LNG) from Delfin’s export facility in the United States.
Under the agreement, Delfin LNG, a leading U.S.-based export infrastructure company, will supply LNG on a free-on-board (FOB) basis to Vitol, one of the world’s largest independent energy traders, which will act as the offtaker and deliver the volumes to IRH Global Trading (IRHGT), IRH’s global trading arm, for a period of 20 years. Definitive agreements are expected to be concluded in the coming weeks.
This long-term partnership represents a major step in IRH’s energy expansion strategy under ePointZero, a diversified investment platform focused on sustainable energy, technology, and resources. By leveraging synergies across the ePointZero ecosystem, IRH aims to build a fully integrated energy and commodities platform that enhances the resilience of supply chains and supports the global energy transition.
Ali Rashed AlRashdi, CEO ofIRH, said: “This transaction is a major milestone in the development of IRHGT’s global LNG portfolio. We are pleased to collaborate with Delfin and Vitol to help bring the project to Final Investment Decision soon. As part of our vision to build an integrated global trading platform headquartered in Abu Dhabi, IRHGT is actively expanding its presence across physical and financial markets in natural gas, power, crude oil, refined products, metals, and equities. This agreement sets us on the path to be a reliable supplier of LNG to our valued clients worldwide.”
Mohamed Hesham, CEO of ePointZero, commented: “This agreement represents another step forward in building an integrated, future-ready energy platform. By securing reliable resources through long-term partnerships, we secure the downstream operations of ePointZero Group and ensure energy resilience for our ecosystem. This collaboration reinforces our commitment to delivering diversified, efficient, and reliable energy solutions that power economies and communities across regions.”
Dudley Poston, CEO of Delfin, added: “It’s an honor to have been selected by IRHGT and Vitol as a long-term liquefied natural gas supplier, and we look forward to working together as we make Final Investment Decision on the first FLNG Vessel in the coming weeks. We are pleased to continue our very strong relationship with Vitol and add another world-class trading organization such as IRHGT to our growing list of strategic partners.”
Pablo Galante Escobar, Global Head of LNG and European Gas & Power at Vitol, highlighted: “We are excited to conclude this agreement with IRHGT and Delfin. Vitol continues to strengthen its position to safely and reliably deliver cost effective, flexible LNG solutions to our customers around the world and we look forward to continuing to expand on our relationship with both IRHGT and Delfin. Vitol’s offtake commitments and investment grade rating will help Delfin on its path to financial close.”
The agreement also follows other strategic initiatives, including IRH’s acquisition of a 56% stake in Alphamin Resources for AED 1.35 billion, one of the world’s largest and highest-grade tin producers; a Memorandum of Understanding signed in May with ePointZero to decarbonise global mining operations through a joint task force; and a September agreement with the Egyptian General Petroleum Corporation (EGPC) and the Egyptian Mineral Resources & Mining Industries Authority. Together, these initiatives highlight IRH’s growing role in advancing sustainable resource development, expanding its trading capabilities, and strengthening global energy connectivity.
