- Wartsila (MTA TH) +2.3%
- Porsche SE (PAH3 TH) +1.3%
- KPN (KPN TH) +1%
- KPN Raised to Buy at Citi; PT 5 euros
- ASM Intl (AVS TH) -2.2%
- Ryanair (RY4C TH) -2.4%
- Aixtron (AIXA TH) -2.5%
- Novo Nordisk (NOV TH) -2.5%
- Infineon (IFX TH) -2.7%
- Nokia (NOA3 TH) -2.7%
- STMicro (SGM TH) -3.1%
- Naples Outpaces Milan as Southern Italy Drives Economic Growth
- AT&S (AUS TH) -4.3%
- Frontline PLC (HF6 TH) -5.8%
- Zalando (ZAL TH) -18%
- German BaFin Opens Audit of Zalando’s 2025 Financial Statements
DAX:
- Deutsche Bank (DBK TH) -1.2%
- Deutsche Bank Investor Raised Voting Rights to 4.91% on June 22
- Bayer (BAYN TH) -1.3%
- Siemens Energy (ENR TH) -1.4%
- Infineon (IFX TH) -2.6%
- Zalando (ZAL TH) -18%
- German BaFin Opens Audit of Zalando’s 2025 Financial Statements
MDAX:
- Thyssenkrupp (TKA TH) -1.8%
- Siltronic (WAF TH) -2%
- Aixtron (AIXA TH) -2.6%
- SUSS MicroTec (SMHN TH) -2.6%
- Porsche SE (PAH3 TH) -4.9%
SDAX:
- Tonies SE (TNIE TH) -2.3%
- PVA TePla (TPE TH) -2.4%
- Morgan Stanley Raised PVA TePla Voting Rights to 6.94%
- Grenke (GLJ TH) -2.5%
- LPKF (LPK TH) -2.6%
- MLP (MLP TH) -3.3%
How Rheinmetall gambled on Germany’s doomed warship project — and lost
Arms company’s €15bn frigate blow tests faith among investors
Armin Papperger began hearing rumours on Tuesday morning that Germany was about to announce the scrapping of a troubled multibillion-euro programme to build six huge warships.
Rheinmetall’s chief executive was blindsided.
In March he had completed the €1.5bn acquisition of a naval group, a move partly driven by the expectation his company would take on the project to build six F126 frigates for the German navy.
Papperger had spent months telling investors that the new maritime division would become the lead contractor on the programme, taking over from a beleaguered Dutch group in an order that was set to be worth €15bn.
“He was astonished,” said one person who spoke to Papperger on Tuesday when the news broke.
A Rheinmetall executive described the decision, which has caused the company’s share price to plunge almost a fifth since the announcement, as a “disaster”, adding: “It was a big shock.”
The saga — as described to the FT by more than 10 figures from industry, government and politics closely involved with the project — has called into question investors’ faith in Papperger, the most high-profile figure in Germany’s defence sector.
The chief executive, who has led Rheinmetall since 2013, had disappointed shareholders with its first-quarter results, missing analysts’ forecasts as it struggled to translate booming European defence budgets into sufficient orders and profits.
Sash Tusa, aerospace and defence analyst at Agency Partners, said the company was suffering from a mismatch between what it had promised and what it could deliver.
“When your share price valuation is high, investors have a reasonable expectation of very high and consistent performance,” he said. “You don’t get cut slack any more for missing quarterly earnings quite badly or doing an acquisition and getting it wrong.”
Tusa added Rheinmetall, whose shares have fallen about 40 per cent this year, appeared to have wrongly believed that “their influence with the German government and their reputation as a big trustworthy defence company was such that they could rescue the [F126] programme and get it repriced in their favour”.
Defence minister Boris Pistorius took the decision to kill the project and instead buy eight smaller, cheaper Meko A-200 frigates made by German rival TKMS that he said would be available more quickly.
Speaking on Wednesday, he added the initial price tag of the F126 programme had risen from €10bn in 2020 when it was announced to roughly €18bn today, including €2.3bn already spent on the project. He said: “That is simply unacceptable.”
The decision, supported by the head of the German navy, was welcomed by many in parliament. Bastian Ernst, a naval expert with the ruling Christian Democrats and himself a former Rheinmetall employee, said it would “ensure that our navy receives the ships it urgently needs for defence against Russian submarines as quickly as possible”.
German officials insisted no promises to Rheinmetall had been broken, saying the company had never signed a contract for the frigate.
But one person closely involved in the F126 programme described that as “nonsense”. No formal promises were made, but German officials “asked Rheinmetall to engage in due diligence with the purpose of taking over the project” from Damen Naval, the Dutch shipbuilder that in 2020 won the contract to build the F126.
Only last week, the two most senior officials from the German defence ministry visited Tim Wagner, head of Rheinmetall’s naval division, and gave no indication that the plan to build six F126 frigates was about to be mothballed.
Papperger’s decision last year to acquire Naval Vessels Lürssen (NVL) — part of a family-owned group that has built yachts owned by Russian billionaires and Emirati royalty — was one of his boldest moves.
It was part of a broader push by Papperger, 63, who in 2024 was allegedly the target of a Russian assassination plot, to harness the surge in German defence spending to expand the company beyond its traditional domain of tanks, artillery and ammunition.
Papperger said the acquisition of NVL, which has four shipyards along the north German coast with 2,100 employees, would help his company become “a relevant player on land, on water, in the air and in space”.
Asked during an analyst call in September about the wisdom of taking over the difficult F126 project, he said while formal talks with government officials had yet to start, it was “relatively riskless”.
NVL already had a large share of the work on the F126 frigate. But Damen had faced technical problems that had caused rising costs and delays. German officials were discussing finding a new company to take over the role as lead contractor from Damen and NVL was the obvious choice.
Damen, which declined to comment on the decision to cancel the F126, has previously acknowledged software problems but had insisted that it met project requirements. It has maintained that it was a “sound” shipbuilder with “a long array of successful naval projects” under its belt.
In November, Berlin took the first official steps in a potential handover process, brokering a deal between Damen and NVL to begin a six-month period of due diligence, including work to see if it would be technically possible to transfer the designs.
Rheinmetall completed its acquisition of NVL in March. A month later, Papperger attended the christening of the first ever “Rheinmetall” warship in Hamburg, with guests including the deputy head of the German navy.
The same month, Rheinmetall made an offer of €12.8bn net — or €15.2bn including VAT — to take over the F126 project.
In recent weeks, officials had drawn up a contract and submitted it to the finance ministry, with the aim of sending it for approval to the Bundestag’s budget committee before the summer recess starting mid-July.
Papperger on Monday personally bought €4mn in Rheinmetall shares, according to a stock exchange filing. Then, on Tuesday night came the shock decision to scrap the frigate, sending shockwaves through German defence circles.
A person familiar with the deliberations in the defence ministry stressed there was no bad blood towards Rheinmetall.
The company was doing “excellent work” in many areas, the person said, and Berlin continued to see an important role for Rheinmetall as a “national champion”.
Pistorius, the defence minister, said he had spoken on Tuesday with TKMS chief executive Oliver Burkhard, who had indicated that he would be willing to hand some work for the Meko contract to Rheinmetall’s shipyards.
Burkhard said he was “open to discussions with our industry partners”.
But one German defence industry figure said Papperger “had not bought a shipyard so that he could become a subcontractor to TKMS”.
Tusa, the defence analyst, said the naval business only made up about 10 per cent of Rheinmetall’s 2030 revenue target of €50bn and investors should not get “carried away” by the income loss caused by the F126 cancellation.
However, he said Papperger had misunderstood the maritime business, with over-optimistic expectations about profitability as a lead contractor, and the ability to boost sales of Rheinmetall’s other products by deploying them on warships.
Rheinmetall told the FT that it remained “fully committed to the acquisition of the former NVL”, adding: “It was the right decision to expand the group’s portfolio into the naval sector and to integrate the expertise of NVL.”
But an executive from another German defence contractor said Papperger’s bet had backfired. “They bought a shipyard with the looming possibility of getting a monster contract out of it,” he said. “It was a very, very expensive gamble.”
DeepSeek plans hiring spree in escalation of China’s AI talent war
Advertised roles suggest company focused on commercialising frontier research
Chinese AI group DeepSeek is planning to double the size of many of its core teams, joining an intensifying talent war as it seeks to commercialise its frontier research.
The company said on Friday it had launched a recruitment drive to “expand every department” and that “many teams are expected to double in size”.
The advertised roles suggest the group is broadening beyond frontier model research into product development as it prepares to take on outside investment for the first time.
DeepSeek emerged as China’s artificial intelligence darling last year after the release of its open-source R1 reasoning model, which demonstrated performance comparable to leading western systems but was trained using more efficient methods.
Competition among Chinese AI groups has intensified in recent months, with rivals including Zhipu AI and Moonshot AI releasing improved open-source models that have been rapidly adopted by developers.
DeepSeek’s consumer chatbot has ceded ground to ByteDance’s Doubao in China, as users complained of slow response times, service outages and hallucinations.
The recruitment drive reflects founder Liang Wenfeng’s ambition to build DeepSeek into a broader AI company rather than a pure research lab. Liang has also argued for recruiting young engineers rather than established researchers.
Among the openings are AI product managers, product operations specialists and data product managers with expertise in sectors including law, medicine and languages, suggesting the company intends to develop more industry-specific products.
The company is also recruiting extensively for infrastructure engineering, including specialists in AI computing clusters, distributed storage, networking and training frameworks. The move indicates continued investment in large-scale model training and inference — the process through which chatbots generate responses.
DeepSeek has previously recruited for data centre positions in Inner Mongolia, a region of northern China where a network of AI data centres is being built, taking advantage of its low electricity costs.
Data engineers feature prominently among the openings, underscoring the growing importance of high-quality training data for AI models. Chinese companies are increasingly competing on model efficiency and reasoning ability rather than simply scaling computing power.
DeepSeek has also expanded its work on AI infrastructure in partnership with Huawei, helping optimise models to run on the Chinese company’s Ascend AI chips as Beijing pushes to reduce reliance on Nvidia hardware.
This year, DeepSeek released an Ascend-optimised version of its V4 model, although adapting frontier models to domestic chips has posed significant engineering challenges.
The hiring push comes as DeepSeek pursues outside funding for the first time. People familiar with the fundraising have said it is in part motivated by China’s fierce AI talent war, in which DeepSeek’s staff have been poached by larger rivals including Xiaomi and ByteDance.
Electricity prices soar as Europe battles heatwave
Record June temperatures are fuelling demand for air conditioning and putting pressure on energy systems
Europe’s energy system has come under intense pressure with power plants forced to reduce output and wholesale electricity prices soaring as the continent battles an unprecedented June heatwave.
France, Spain and the UK have all reported record June temperatures this week, fuelling demand for air conditioning and other forms of cooling, as well as causing major transport disruption.
In France, almost 70,000 customers in Brittany were left without power after a heat-related transformer failure, while the national utility EDF also cut nuclear power generation as the heatwave increased river temperatures.
In the UK, five major gas-fired power plants were forced to reduce output as they struggled to cool when temperatures in the country reached 36C.
“What we are seeing is an aged grid and old power stations not being able to cope with fossil fuel-driven heat,” said Bruce Douglas, chief executive of the Global Renewables Alliance, of the disruption across Europe.
In Germany on Wednesday, power prices swung from €86 per megawatt-hour at midday to €566/MWh at 8pm, which is when evening demand kicks in and there is far less solar power than earlier in the day, according to data from Montel. This compared with peak prices of about €160/MWh at the same time in the evening last week.
Other countries saw similar swings, while in Britain the system operator had to spend about £1,379/MWh — about 15 times higher than typical day-ahead prices — to import power from the continent on Wednesday due to low wind speeds and high demand.
Fintan Devenney, senior energy analyst at Montel, said power demand was high across Europe, “driven in part by increased cooling load. This raises prices as increasingly expensive generators are required to run to meet rising demand.”
RTE, the French grid operator, said earlier in the week that during heatwave conditions, each additional degree of warming led to 1GW of extra consumption, driven by air conditioning.
High temperatures also lower the efficiency of solar panels and gas-fired power plants.
Charlotte Johnson, general manager of Generation Flex, the platform that controls energy assets at technology business Kraken, said the issues in the energy system this week were “a preview of the operational challenges” the continent will face as the world heats up.
France reached record national average temperatures this week, while the UK recorded its highest June temperature of 36.7C on Thursday. Temperatures are forecast to hit 40C in Germany, Austria and the Czech Republic over the coming days.
Scientists say climate change is making extreme weather events, including heatwaves, more frequent and intense. Europe is the fastest-warming continent, in part because of its proximity to the Arctic where the melt of snow and ice has exposed darker ground that absorbs the sun’s rays.
Across France and the UK, schools were closed and trains cancelled as the heatwave intensified, while rail operators implemented speed restrictions over concerns steel tracks would buckle.
Simon Stiell, head of the UN’s climate change arm, told the FT that the heatwave had made clear “how ill-equipped” countries and their infrastructure were for “not just what is coming but what is already here”.
Andreas Jahn, a Berlin-based power markets expert at the Regulatory Assistance Project, an energy non-governmental organisation, said the heatwave had come at a difficult time for power infrastructure, when many of Europe’s power plants were typically undergoing maintenance.
Bountiful solar supply had kept prices relatively low during the day across many European countries but prices had spiked in the evenings when solar panels stopped working and there was high demand for electricity to power air conditioning units, he added.
He said storage was the “missing piece in the puzzle”, arguing that a broader rollout of battery technology could harness solar power to keep powering air conditioning during the night.
The record-breaking heat came as an estimated 75,000 people were in the British capital for London Climate Action Week, where discussions about power — and the shift to electricity — dominated.
Canada environment minister Julie Dabrusin told the FT that the European heatwave “put in stark relief” why countries needed to decarbonise and change their energy systems.
Environment ministers also met in Luxembourg to discuss Europe’s transition away from fossil fuels and adaptation to higher temperatures. The commission is developing a resilience strategy in the autumn.
Europe’s commissioner for climate Wopke Hoekstra told journalists: “The weather outside should point us in the direction of more ambition not less.”
After Hours Summary: SYNA +10.9% to be acquired by ON -8.6%; FDXF +0.1% reports first quarter post-spin-off
After Hours Gainers:
Companies trading higher in after hours in reaction to earnings/guidance: WSE +4.2%, FDXF +0.1%
Companies trading higher in after hours in reaction to news: TII +12.1% (selected by US Army to establish public-private partnership for domestic critical minerals processing), SYNA +10.9% (ON to acquire SYNA in all-stock deal, both cos reaffirm guidance), RKLB +3.1% (selected by NASA to provide three Electron launches), IAUX +2.7% (assay results from 2025-2026 infill drill campaign), BILL +2% (names new Chief Revenue Officer), EQX +2% (signs land access agreements), NRXP +1.7% (responds to recent public statements regarding Kadima dispute), FLEX +0.9% (JetCool unit launches liquid-cooled version of Dell PowerEdge XE7745 server), MKSI +0.7% (to expand manufacturing capability), AG +0.6% (receives construction permits), ASND +0.5% (to showcase advances in treatment of achondroplasia), MSFT +0.4% (leans toward pushing its IPO into next year, according to NYT), WY +0.4% (provides update; lowers Q2 wood products guidance), KO +0.2% (leadership change for North America operating unit), AXP +0.2% (Fed stress test results; to increase dividend), FRST +0.1% (files for $150 mln mixed securities shelf offering), CTGO +0.1% (files mixed securities shelf offering), BA +0.1% (US clears potential $250 mln sales to Australia, according to Reuters)
After Hours Losers:
Companies trading lower in after hours in reaction to earnings/guidance: OMSE -6.9%, QMCO -1%
Companies trading lower in after hours in reaction to news: ON -8.6% (ON to acquire SYNA in all-stock deal, both cos reaffirm guidance), SVC -0.6% (1-for-5 reverse split), GD -0.5% (awarded a $209.3 mln modification to US Army contract)
Fed's Goolsbee (non-voter): Hard to get a throughline on inflation, how persistent or temporary it is; Reiterates inflation is going the wrong way, inflation side of the mandate is clearly the problem - CNBC interview
- Some inflation is being driven by one and done issues
- The PCE report was not all negative; Saw a little bit of improvement on services inflation
- Personally have long been uneasy with forward guidance, but I don't hate the 'dot plot'
- Welcomes Chair Warsh's taskforce examining the options on the dot plot
- Do not view balance sheet decisions as monetary policy
- World where future productivity gains from AI are factored into stock prices, and people start spending on that basis now, that can cause overheating
- Wages are not a good indicator of inflation; Inflation could go up before wages go up
- Applauds the streamlining of the Fed policy statemen
Competition intensifies for Anthropic and OpenAI ahead of IPOs
Renewed challenge from open-source models raises stakes on AI labs to make their case
In its early days, open-source software was seen by some as an existential threat. Microsoft chief executive Steve Ballmer once warned that code that is free to use and modify might become a “cancer”, making it hard for commercial software companies to make a profit from their intellectual property.
It didn’t turn out that way. Open source was certainly an astounding success: Much of the infrastructure software that sits in the guts of corporate IT systems was developed and distributed using open-source methods. But this has turned out to be complementary to other types of code, leaving plenty of room for the commercial software industry.
The same “complement or existential threat” question is suddenly looming larger in AI. Until now, Wall Street has largely brushed off the threat. The shock caused by DeepSeek’s R1 model early last year — an advanced reasoning model released as open source and developed at a small fraction of the cost of US rivals — soon faded.
The six- to nine-month lead the frontier labs were able to keep with their latest models has been enough to keep customers hooked. At the same time, new agent capabilities grafted on to some models, such as Anthropic’s Claude Code, have enhanced their value for certain tasks, and usage has exploded.
Things may be changing. One sign was the release this month of the first open-source model — GML 5.2, from Chinese lab Zhipu — to reach parity with the leading frontier models on the most widely followed performance benchmarks. Technical measures like these are of limited value on their own, but equally significant has been the positive reception the model has received as a coding agent. That suggests that the software “harnesses” developed by the US labs — which guide their models through tasks like coding — may not be the sort of lasting differentiator they might have hoped.
This comes at just the moment that many customers are facing sticker shock from their soaring AI bills, adding to pressure to shift the less demanding parts of their AI workloads to cheaper models.
The pressure to segment the market around different AI requirements need not be entirely bad news for the frontier labs. Rather, it is a sign of growing market maturity. After a period of rapid price declines, the labs themselves have been getting more refined about their pricing. Anthropic, for instance, set the price of tokens generated by Fable 5 — the first broad release of its controversial Mythos model — at twice the price of its previous frontier model. If Fable 5 users can do more with fewer tokens, as it claims, then some sort of premium seems justified.
The challenge will be to convince customers when that premium is worth paying. This will become harder as a new class of intermediaries emerges to help users break up their work between different AI models. These include systems such as the Japanese model Fugu Ultra, launched this week, which claims to be able to achieve superior performance by carving up a task between different underlying models offered by other suppliers. If new intermediaries like this push frontier models into the background, the commodification risk will increase.
Open-source advocates have also been buoyed by the spat with the Trump administration that forced Anthropic to shut down access to Fable 5 this month. That has given customers an extra incentive to reduce their reliance on any one supplier, while handing a marketing advantage to open-source companies whose models can be run by customers on their own servers, beyond the reach of regulators.
Ironically, though, the Fable 5 shutdown could also be bad for the open-source world. The US labs have long argued that much of the performance achieved by open-source rivals comes from illicitly copying the latest frontier models. Even before Washington acted, Anthropic had threatened to degrade Fable 5 if it suspected this kind of misuse, provoking a backlash from users who saw this as a heavy-handed form of censorship.
All of this has intensified the pressure on Anthropic and OpenAI as they prepare for their landmark IPOs. Can they maintain broad distribution for their most advanced models while at the same time both limiting the risks of plagiarism and satisfying Washington that they can prevent misuse? And will they have enough differentiation to retain pricing power and cover their massive spending? Traditional software companies were able to make a case for their proprietary commercial code, even as open source became pervasive. Now the AI labs must do the same.
Google Revamps New AI Coding Strike Team Amid Struggle to Catch Up With Anthropic
The Takeaway
- Google revamps AI coding strike team to catch Anthropic in lucrative market.
- Revamp comes as key AI researchers Noam Shazeer and John Jumper depart.
- The changes together show key challenges facing Google in AI compute allocation and coding.
Google is reorganizing its recently launched strike team working on AI coding tools to try to catch up with Anthropic in the most lucrative AI applications, according to people familiar with the changes.
The goal is for the months-old strike team to change the approach to training Google’s AI models to improve their abilities in both coding and other areas such as creating presentations, the people said. That aligns with efforts by both Anthropic and OpenAI to expand AI coding tools to other business functions. In addition to expanding the scope of the strike team’s work, the changes also formalize the structure of what was originally a short-term group.
The revamp coincides with two major executive departures that have fueled concern about Google’s ability to stay at the forefront of developing and adapting competitive AI models. While Google has broadly been one of the biggest winners of the AI boom, it has struggled with how best to allocate personnel and computing resources to compete with Anthropic and OpenAI.
Last week, star researcher Noam Shazeer quit abruptly to join OpenAI, as The Information was first to report. Shazeer and Google didn’t publicly provide a reason at the time, but people familiar with the situation said this week that his departure followed a change in his access to AI servers.
Shazeer’s departure was jarring to insiders and outside observers, after Google paid $2.7 billion in a licensing deal less than two years ago to bring him back following his departure in 2021 to found his own startup. A co-author of the 2017 paper on the so-called transformer architecture that has underpinned the current generative AI boom, Shazeer had recently been working on research to find new model architectures beyond transformers.
He told colleagues before leaving that the company had merged his compute allocation with that of another team, according to the people familiar with the situation. In an internal message announcing his departure, Koray Kavukcuoglu, chief technology officer of the DeepMind unit that leads Google’s AI efforts, said that Shazeer’s departure had happened quickly, one of the people said.
Then, on Saturday, John Jumper—a DeepMind vice president who was awarded a Nobel Prize along with DeepMind CEO Demis Hassabis for their work on protein structure prediction—announced he was heading to Anthropic. Jumper had recently been moved to the strike team, according to two people familiar with his work.
The two departures shook investor faith in Google, sending the share price of parent company Alphabet down 5% on Monday, its worst day in over a year. The stock has fallen another 1% over the past two days, though it remains up nearly 30% so far in 2026.
Google since last year has been one of the biggest beneficiaries of investor enthusiasm for AI, thanks to its strength throughout the AI value chain, from homegrown chips to leading frontier models to cloud and ad businesses booming because of AI advances.
Those assets have brought their own difficulties, though. Executives have struggled to manage internal tensions over competing demands for compute from an array of outside customers and partners—including Anthropic, a major Google Cloud customer—and internal efforts such as Shazeer’s project and the coding strike team.
That coding effort has been important enough that it drew direct involvement from both Kavukcuoglu and Google co-founder Sergey Brin. Coding has been the most lucrative part of the AI services market, and Anthropic’s dominance in the area has propelled it to an annualized revenue rate of $47 billion as of last month, more than three times the level in February.
OpenAI has been aggressively building its own coding tool, Codex, and both OpenAI and Anthropic have been expanding their tools to work on other white-collar applications.
The Google changes center on how AI models are trained. The work has broadly been divided into pretraining, or building the models, and post-training to adapt the models for specific uses. Google’s code strike team is now expanding into a midtraining team in between those stages, for training models on more specialized data, according to a person with knowledge of the changes.
The midtraining team will take some responsibilities for scaling model capabilities, including coding, that previously belonged to the post-training team, which will now focus on what it’s like for people to interact with the models. The goal is to separate those two areas to speed up capabilities improvement.
“While we are moving ahead with great speed, this is also a good time to structure the organization to have clear swimlanes and responsibilities,” Kavukcuoglu told employees in the email announcing the reorganization.
Coding has emerged as a major weakness in Google’s AI portfolio. It initially focused little on the area, in the belief that having the most capable base AI model would lead to coding abilities, according to a person involved in the effort.
Some developers criticized Google’s most recent model release, Gemini 3.5 Flash, for overly sycophantic answers and a higher price compared to previous Flash models. And developers met its Antigravity coding tool with mixed reviews after a buggy initial release.
The company has yet to release its next flagship model, Gemini 3.5 Pro, which it said at its May developer conference would arrive in June. At that time, it wasn’t expected to surpass Anthropic’s most advanced model, Mythos, according to a person who has tested it.