Why Big Tech’s Cloud Shift Threatens Microsoft and Google’s Margins
Don’t count out Amazon. This was not the commerce cloud giant’s week, to be sure. When it comes to the great artificial intelligence cloud race, which now overwhelms all else in tech, Microsoft and Google appear to be ahead. Both companies posted accelerating growth in their cloud units for the June quarter, unlike Amazon. But one element is widely missing from the discussion: how a long-term shift in big tech’s business mix toward cloud services could squeeze profit margins for both Microsoft and Google while improving Amazon’s overall margin.
Why? Because for both Microsoft and Google, their cloud businesses are much less profitable than their giant software and advertising businesses, respectively. Microsoft’s productivity and business processes unit—which includes its business software operations—had an operating profit margin of 57.4% in the June quarter. Its “intelligent cloud” unit, which comprises mostly Azure, had a margin of 40.6%. Its overall margin was 45%. A similar scenario holds for Google: Its cloud margin was 20.7% in the June quarter, while its margin in Google services—mostly advertising—was 40% and its overall margin was 32%.
Microsoft and Google already enjoy faster growth in cloud than in their software and ad businesses. In the most recent quarter, Microsoft’s software business expanded 16%, while its cloud business grew 26%. (Azure grew much faster, at 39%—but the cloud business also includes old-fashioned on-premise data center products, which shrank.) Ditto for Google, whose cloud business grew 32% in the quarter, while its services segment grew 12%.
This trend is likely to be accentuated by the growing demand for AI-cloud services. At the same time, there’s a risk that new AI products will cannibalize Microsoft’s lucrative enterprise software and Google’s search advertising businesses. Of course, both companies are working hard to head off those threats. Microsoft has introduced its own AI software services, including Copilot and AI agents, while Google has enhanced its search with AI features. But they face plenty of competition, as this story and this story make clear.
Amazon is in a very different position: its bottom line will benefit from a shift to the cloud. Its Amazon Web Services cloud unit’s margin in the June quarter was 33%, far higher than its e-commerce margin of 6.6%, reflecting the razor-thin profits shopping services typically operate on. Indeed, Amazon’s margin has improved as AWS has grown: Between 2017 and 2024, AWS rose as a percentage of Amazon’s total revenue from 9.8% to 17%. In the same period, Amazon’s overall operating margin jumped from 2.3% to 10.7%. (The latter increase likely also reflects the contribution of Amazon's expanding ad operation, as advertising is typically a very high-margin business.)
The focus this week was on AWS’ relatively tepid growth rate of 17% in the second quarter, the same rate as for the previous quarter. That’s still faster than for every other part of Amazon except its advertising business, which is half the size of AWS. Moreover, there’s reason to think AWS’ growth rate will pick up. New Street Research analyst Dan Salmon made that point in a report on Thursday, noting AWS’ 25% expansion in its backlog of business—customer commitments, which offers a picture of future revenue—in the quarter.
For all three companies, their cloud profit margins aren’t likely to improve, given how much the humongous levels of capital expenditures cloud requires will raise depreciation expenses. The real issue is that the overall margins for Google and Microsoft could drop closer to their cloud margins over time. The market is well aware of that risk for Google, whose stock has underperformed that of ad rivals like Meta Platforms for a while now. But investors seem to be ignoring the risk for Microsoft, judging by that stock’s premium valuation. At the same time, they may be ignoring Amazon’s potential for growth.
Inside OpenAI’s Rocky Path to GPT-5
The troubles OpenAI has faced in developing GPT-5 point to slowing AI progress across the industry. Researchers believe advances in reinforcement learning will help to overcome that.
The Takeaway
• GPT-5 will show real improvements over its predecessors, but they won’t be comparable to leaps in performance between earlier GPT-branded models
• OpenAI confronted a series of technical problems that imperiled o3 and other models this year
• A disagreement between research chief Mark Chen and a deputy spilled into view on Slack
OpenAI made waves across the industry in December when it published the results from its tests of artificial intelligence that performs better on tasks when it gets more time and computing power to process them. The results implied ChatGPT customers were about to be blown away by what the new AI could do.
But the euphoria was short-lived.
When OpenAI researchers turned the new AI into a chat-based version called o3 that could respond to instructions from ChatGPT customers, the performance gains the company had published largely vanished, according to two people involved in its development.
The episode was an example of the technical challenges OpenAI has faced through much of this year, threatening to slow its pace of AI advances, if not its blockbuster ChatGPT business.
But its researchers have found ways to keep AI progress going through techniques that have surged across the industry.
OpenAI is now nearing the release of GPT-5, its next flagship AI model, which improves upon existing models’ ability to complete practical computer programming and math tasks, among other things, according to people who have used it or are familiar with the company’s internal evaluations.
For instance, when the new model codes applications, it’s better at adding features that make them easier to use and more aesthetically pleasing, one of those people said.
GPT-5 is also better than its predecessors at powering AI agents that handle complex tasks with minimal human oversight, this person said. For instance, it can follow complicated instructions, such as a list of rules that determine when an automated customer support agent should grant a refund.
Previous models needed to see several examples of tricky customer cases, known as edge cases, before they could handle such refunds, this person said.
The improvements won’t be comparable to the leaps in performance of earlier GPT-branded models, such as the improvements between GPT-3 in 2020 and GPT-4 in 2023, one of the people said. And the slowing performance gains OpenAI has experienced over the past 12 months suggest it may be hard for the company to surge ahead of its biggest rivals, at least in terms of AI capabilities.
But OpenAI’s current models are generating so much commercial value from powering chatbots and other applications that any improvements, even incremental, will increase customer demand. They could also give new investors the confidence to fund the company’s plan to burn through $45 billion in the next three and a half years as it rents expensive servers for developing and running its products.
Coding Priorities
The latest gains also help explain why OpenAI executives in recent weeks told some investors they believed the company could reach “GPT-8.”
The comments are in line with CEO Sam Altman’s public comments that, using existing technological know-how, OpenAI can attain the goal of creating AI whose capabilities are near or on par with those of the smartest humans. This technology is otherwise known as artificial general intelligence.
While it’s far from being AGI, the upcoming GPT-5 model may have other attractive attributes besides better coding and reasoning. Some leaders at Microsoft, which has exclusive rights to OpenAI’s intellectual property, have told staff their tests of the model show it produces higher-quality coding and other text-based answers without consuming a lot more computing power, according to a Microsoft employee with knowledge of the situation.
That’s partly because it is capable of figuring out which tasks require relatively more or less computing resources better than prior models, this person said.
Improving AI’s ability to automate coding tasks became a priority at OpenAI after archrival Anthropic last year took the lead in developing and selling such models to software developers and coding assistants like Cursor, according to OpenAI’s internal evaluations. OpenAI staff believe automated coding isn’t just important to the company’s business—it’s also critical to automating the work of the AI researchers themselves.
Reorg Strains
Progress at OpenAI hasn’t been a straight line, as both its researchers and its managers have faced new strains this year.
Some senior researchers have resisted the idea of giving away their inventions to Microsoft, OpenAI’s largest outside shareholder, despite the software firm’s contractual rights through 2030.
The two companies have a tight financial relationship but have feuded over the terms of their deal, with each side seeking concessions from the other as OpenAI tries to restructure its for-profit arm so it can eventually go public.
Discussions between Microsoft and OpenAI have been moving in a positive direction, according to two people who have spoken to negotiators. Many bargaining points are still up in the air, though others appear to be more settled, such as the roughly 33% equity stake Microsoft is likely to get in OpenAI’s for-profit arm as part of the restructure, according to one of those people.
More recently, Meta Platforms has hired more than a dozen OpenAI researchers, some of whom had been involved in the techniques the company has been using lately to improve its technology. Meta won them over by offering compensation packages worthy of the highest-paid soccer stars.
The departures and the staff reorganizations in response to them have weighed on senior OpenAI staff. Last week, Jerry Tworek, a vice president of research at OpenAI, complained about a team change to his boss, research chief Mark Chen, on the company’s internal Slack app, which is visible to many other colleagues.
Tworek said he had to take a week off to reassess things, but he later ended up not taking time off.
Orion’s Star Falls
The company’s business progress has masked some internal concerns about its ability to keep improving its AI and stay ahead of other well-capitalized rivals such as Google, Elon Musk’s xAI and Anthropic.
Problems had been brewing for months before the current year began. For much of the second half of 2024, OpenAI was developing a model known internally as Orion and intended to become GPT-5. According to people who worked on it, Orion was supposed to offer a big step up in performance compared to the current flagship, GPT-4o, released in May that year.
But the Orion effort failed to produce a better model, and the company instead released it as GPT-4.5 in February this year. It has since faded from relevance.
Part of the failure had to do with the limits of pre-training, the first stage of developing a model, in which it processes data from the web and other sources so it can draw connections between concepts.
Not only was OpenAI facing a dwindling supply of high-quality web data, but researchers also found the tweaks they made to the model worked when it was smaller in size but didn’t work as it grew, according to two people with knowledge of the issue.
More Nvidia Chips
As recently as June, the technical problems meant none of OpenAI’s models under development seemed good enough to be labeled GPT-5, according to a person who has worked on it.
OpenAI researchers faced other problems.
Last year, the company also developed reasoning models, which performed better if they got more computing power to process answers. The models stemmed from a breakthrough in late 2023 called Q* that had sent shockwaves among its researchers because it was able to solve math problems it hadn’t seen before. By 2024, reasoning models appeared to be helping the company overcome a slowdown in performance gains during pre-training.
Last fall, OpenAI turned the first major reasoning model into o1, a version it could sell to application developers and use to power conversations inside ChatGPT.
The launch gave OpenAI new clout within the AI field and set the stage for the development of AI agents that relied on reasoning models to handle tasks with minimal human supervision.
Before the end of 2024, OpenAI created the next reasoning model, o3, with the same underlying large language model, GPT-4o, it had used as the foundation of o1, according to a person who was involved in their development.
Despite their shared lineage, the parent model of o3—also known as a teacher model—made extraordinary gains compared to the parent model of o1 in understanding a variety of scientific and other domains, this person said.
One reason for the improvement came from OpenAI’s decision to develop the parent model of o3 with a lot more Nvidia chip servers, essentially giving it more processing power to understand difficult concepts, said two people who were involved. Another reason was that researchers gave it the ability to search the web or pull from code repositories, which also helped it improve over the parent model of o1, one of these people said.
The parent model for o3, similar to the o1 parent, also benefited from reinforcement learning, in which human experts come up with tough questions and answers in fields like biology, software engineering and medicine and ask the model to come up with thousands of its own responses to those questions.
OpenAI then trained the model on the responses that arrived at the same answers as the human experts. (The AI-generated responses are also known as synthetic data.)
Gibberish Reasoning
OpenAI generated headlines around the world and viral hype on social media when it publicly shared the results from special tests of the model’s strengths. But then reality set in.
When OpenAI converted the o3 parent model to a chat version of the model—also known as a student model—that allowed people to ask it anything, its gains degraded significantly to the point where it wasn’t performing much better than o1, the people who were involved in its development said.
The same problem occurred when OpenAI created a version of the model that companies could purchase through an application programming interface, they said.
One reason for this has to do with the unique way the model understands concepts, which can be different from how humans communicate, one of these people said. Creating a chat-based version effectively dumbs down the raw, genius-level model because it’s forced to speak in human language rather than its own, this person said. (The gibberish that reasoning models sometimes show in ChatGPT as they “think” about how to solve a problem reflects some of these communication differences.)
The Week’s 10 Biggest Funding Rounds: Ramp Ramps Up While AI And Healthcare Hold Strong
This was a big week for big checks, both confirmed and reported. Among confirmed rounds, the largest financing went to fintech provider Ramp, which landed $500 million at a $22.5 billion valuation to scale its visions around agentic AI. The next-largest financings went to MapLight Therapeutics, a developer of medicines for brain disorders, and Ambience Healthcare, a healthcare AI startup.
As for giant deals that were reported but not officially closed, Anthropric was said to be close to finalizing a round of up to $5 billion led by Iconiq Capital that would push its valuation all the way to $170 billion.
1. Ramp, $500M, fintech: New York-based Ramp, a provider of financial products and tools for businesses to automate finance tasks, raised $500 million at a $22.5 billion valuation. Iconiq Capital led the Series E financing, which brings total equity funding to date to $1.9 billion.
2. MapLight Therapeutics, $372.5M, biopharma and neuroscience: MapLight Therapeutics, a biopharma startup developing medicines for brain disorders, announced that it raised $372.5 million in Series D funding. Forbion and Goldman Sachs Alternatives co-led the financing for the 7-year-old, Redwood City, California-based company.
3. Ambience Healthcare, $243M, healthcare AI: Ambience Healthcare, an AI platform for healthcare systems to use in documentation, coding and clinical documentation, raised $243 million in a Series C round. Oak HC/FT and Andreessen Horowitz led the financing for the San Francisco-based company.
4. Quince, $200M, fashion: Quince, an affordable luxury brand online retailer, raised $200 million at a valuation of more than $4.5 billion, according to a report from Bloomberg. Iconiq Capital reportedly led the San Francisco-based company’s latest financing.
5. Observe, $156M, AI enterprise software: San Mateo, California-based Observe, a provider of AI-enabled observability tools for businesses, raised $156 million in a Series C funding round led by Sutter Hill Ventures. The financing brings funding to date for the 8-year-old company to more than $460 million, per Crunchbase data.
6. (tied) Motive, $150M, fleet management: San Francisco-based Motive, a provider of fleet tracking and driver safety software, raised $150 million in a new funding round led by Kleiner Perkins. The 12-year-old company is also reportedly taking steps toward an IPO.
6. (tied) Anaconda, $150M, AI software: Anaconda, a provider of AI tools for businesses using Python and open source applications, announced it raised over $150 million in a Series C funding round led by Insight Partners. The Austin, Texas-based company said it currently operates profitably with over $150 million in annual recurring revenue as of July.
8. Artbio, $132M, radiopharmaceuticals: Cambridge, Massachusetts-based Artbio, a clinical-stage radiopharmaceutical startup developing therapies (ARTs) to treat a range of cancers, raised $132 million in a Series B round that included Sofinnova Investments and B Capital as lead investors.
9. Fal, $125M, generative media: San Francisco-based Fal, a startup offering a generative image, video and audio platform for developers, raised $125 million in a Series C led by Meritech Capital Partners. The 4-year-old company said it has seen revenue increase 60x in the past 12 months.
10. Oxide Computer Co., $100M, cloud infrastructure: Oxide Computer Co., a developer of cloud infrastructure for on-premises computing, raised $100 million in a Series B round led by US Innovativr Technology. Founded in 2019, the Emeryville, California-based company has raised over $260 million to date, per Crunchbase data.
Bogner Sells Majority Stake to Katjes International to Drive Global Expansion
The Munich-based Bogner family will retain 40 percent of the shares.
Bogner, the Munich-based lifestyle and luxury fashion company, has found a new majority investor in Katjes International.
The Bogner family will sell 60 percent of its shares in Willy Bogner GmbH to Katjes International GmbH & Co. Katjes, based in Emmerich, Germany and part of the Katjes Group, which invests primarily in companies with established brands in the consumer goods sector.
The Bogner family will remain invested in Bogner for the long term with 40 percent of the company shares and will continue to be involved in the strategic direction of the company. The transaction is expected to be completed in September. It is subject to antitrust approvals. The price of the acquisition wasn’t disclosed.
Florinda Bogner, daughter of Willy Bogner Jr., said, “We are delighted about our strong new partner. In Katjes International, we gained a family-owned company that shares our values and is committed to investing in the future of Bogner alongside us. Together, we will continue writing the success story of our richly traditional brand.”
Tobias Bachmüller, managing shareholder of Katjes International, said, “With our success in the personal care sector — with Bübchen, Theramed and Shirin Beauty — we have proven that we can profitably develop brands outside our core business. The further development of brands in the consumer goods segment in Europe is our strength and is in line with our long-term strategy. With Bogner, we are expanding our brand portfolio into the luxury goods segment and further enhancing its value.”
Katjes International’s investment gives Bogner the opportunity to continue growing in the future with an experienced partner and with a strong capital structure. Both family-owned businesses, the partners intend to invest jointly in the expansion and further internationalization of the brand.
Arndt Geiwitz, chairman of the advisory board of Bogner, said, “Bogner has successfully transformed itself in the past few years and is now well positioned as a leading player in the lifestyle and luxury sports fashion sector. With Katjes International as a strong investor, Bogner is ideally equipped for the future.”
The company’s headquarters will remain in Munich, and Bogner will continue to operate as a legally and organizationally independent company. The company is known for its upscale sportswear and luxe snow sports attire.
Bogner posted record-breaking results for the 2023-24 fiscal year, with revenue climbing 7 percent to 187.6 million euros.
Daniel Hiendlmeier, managing director and chief brand officer of Bogner noted that Katjes’ investment is a milestone for the future of both the Bogner and Fire + Ice brands. “Katjes International shares our vision of innovation and brand management and brings a deep understanding of our identity. The partnership opens up great opportunities for the brand, our employees and partners,” he said.
Frank Wiesner, managing director and chief financial officer of Bogner, added, “The transaction strengthens our capital base and creates an excellent foundation for driving our international growth and expanding global customer relations. The move confirms the appeal of the Bogner brand and the success of our strategy in recent years.”
As reported last August, Bogner said it was looking for a partner to invest in its international expansion. The company said at the time it would part ways with its chief executive officer Gerrit Schneider, who had accepted a new post, at the end of 2024. At the time, the company said it had achieved the highest sales in its history in the last fiscal year.
Willy Bogner Jr., whose father Willy Sr. started the namesake company in 1932, is its primary owner with daughter Florinda. They serve on the company’s board, but aren’t active in its day-to-day operations.
Willy Bogner Sr., an Olympic skier, started Bogner as an import business for skis, equipment and Norwegian knitwear. The company specialized in skiwear and evolved into other areas. His wife, Maria, was credited with creating the modern skiwear look, overseeing design and serving as the company’s lead model.
Following Willy Sr.’s death in 1977, Willy Jr. and his wife Sonia, a model-turned-designer, expanded into other categories, introduced the Sônia Bogner label and the snowboard-inspired Fire + Ice brand, and opened freestanding stores. The husband-and-wife team built up and modernized the family business in the ’80s, ’90s and 2000s. In 2017, Sonia Bogner died after a long illness. Willy Jr. retired from the daily business in 2019.
South China’s Foshan urged to stamp out Chikungunya spread as Guangdong cases rise
Vice-Premier Liu Guozhong calls on manufacturing hub in southern Guangdong province to strictly follow ‘port health quarantine measures’
A senior Chinese official has visited the epicentre of the country’s Chikungunya fever outbreak, as the number of cases of the mosquito-borne virus rises.
Chinese Vice-Premier Liu Guozhong visited the southern city of Foshan in Guangdong province earlier in the week, urging authorities to “strictly implement port health quarantine measures”, state news agency Xinhua reported on Friday.
Liu told city authorities to “improve prevention and control measures”, “effectively eliminate mosquitoes” and “cut off epidemic spread channels”, the report said.
Foshan, a manufacturing hub of 10 million residents, accounted for about 60 per cent of 4,824 reported cases in Guangdong as of July 26, according to the provincial centre for disease control. There have been no fatalities.
Hong Kong’s Centre for Health Protection said on Friday that “according to the information from the Foshan health authority, as of July 30 there were over 6,500 cases”.
It said 5,660 cases were in Shunde district. All cases were mild, with no severe or fatal cases.
Neighbouring Hong Kong on Saturday reported its first imported case of the disease since 2019, a boy who spent two weeks in Shunde district.
This is China’s worst outbreak of the disease in decades, and Foshan authorities issued notices earlier this week offering nucleic acid PCR tests in several neighbourhoods.
According to Chinese website Hongxing News, some residents in the Foshan township of Lecong were notified about free screening for Chikungunya fever during the week.
The report quoted community doctors as saying those who “are asymptomatic or have not received notifications” did not need to be tested for the time being.
It was not clear if the tests were compulsory.
Chikungunya fever is a viral disease transmitted by Aedes mosquitoes. It causes sudden high fever, severe joint pain, rash and fatigue. It cannot be contracted through direct contact with another person.
While rarely fatal, the illness can lead to debilitating symptoms lasting weeks. Outbreaks are often linked to tropical and subtropical regions with dense mosquito populations.
There are vaccines for preventing Chikungunya virus infection, but none are licensed for use in China.
The US Centres for Disease Control and Prevention on Friday issued a level 2 travel notice – the second level in a four-tier system – for visitors to China to take extra precautions because of the rising number of Chikungunya cases.
China also issued new treatment guidelines for the disease earlier this week, calling on medical staff to avoid indiscriminate use of antimicrobial medications.
The World Health Organization recommends using antipyretics and analgesics such as paracetamol to treat fever and joint pain.
Fed Officials Take Cool Jobs Report in Stride
New York Fed’s John Williams says labor market remains solid, even after unusually large downward revisions to job gains for May and June
July’s jobs report rattled financial markets and meaningfully raised the prospect that the Federal Reserve will cut interest rates at its next meeting in September.
But two Fed officials on Friday maintained that the U.S. labor market is still well balanced, describing the latest softness as part of a gradual cooling rather than a worrying deterioration.
In an interview Friday, New York Fed President John Williams said he would go into the September meeting with “very much an open mind” about cutting interest rates. But he also offered a more measured assessment that tempered the sense of urgency driving expectations for more aggressive easing.
“What we’re seeing I would describe over the past year as a gentle gradual cooling of the labor market, but still leaving it in a still solid place,” said Williams, a top ally of Fed Chair Jerome Powell.
Fed officials were split over their decision to hold rates steady at their meeting on Wednesday, with two officials voting in favor of a cut. Nine others supported the decision to hold steady.
Even among those who favored taking no action in July, a divide is emerging over whether to worry more about the labor market or the prospect for firmer inflation in the months ahead, creating a challenge for Powell in forging a consensus going forward.
While the unemployment rate rose only modestly in July to 4.2%, reversing a decline to 4.1% in June, soggier payroll figures give Powell room to build agreement for a rate decrease because they undercut a narrative that the labor market has strong momentum. Officials will see one more month of job data before their next meeting.
Unusually large downward revisions to job gains in May and June had been “really the news of this report,” said Williams. “This is important information…to understand the direction of what we’re seeing in supply and demand for labor” and the labor market’s cooler momentum.
A wider range of indicators, including job openings and initial claims for jobless benefits, paint a picture of an economy that remains in the same slow-to-hire, slow-to-fire equilibrium that’s been a feature for the past year, Williams said. “The labor market still is solid in my view,” he said.
Cleveland Fed President Beth Hammack echoed that assessment. In an interview Friday with Bloomberg Television, she said the labor market “looks like a healthy labor market that’s still well in balance, but with some disappointing signs that we should watch very carefully.” She added that weaker job growth makes sense given reduced immigration.
White House economic adviser Stephen Miran said Friday’s jobs report was “a disappointment” but attributed much of the weakness to technical and temporary factors.
In an interview, he said about 60% of the downward revisions to May and June job gains were due to seasonal adjustment issues. Policy uncertainty around tariffs had temporarily weighed on hiring, but its resolution with recently announced trade deals should allow job growth to rebound in coming months, he said.
Williams said the concentration of job gains in certain sectors like healthcare, education and government partly reflects a catch-up dynamic from the pandemic era. During the hot labor market of 2021-22, employers who could afford to pay higher salaries competed aggressively for workers, while others—particularly government agencies, healthcare systems and schools—“just didn’t have the ability to compete as well” and ran understaffed.
As the labor market has cooled, those employers have been able to fill positions they couldn’t during the hiring frenzy, he said.
Williams answered questions about a possible September cut cautiously, declining to defend market expectations that at one point on Friday placed an 80% probability on a September rate cut.
“Market participants are confronted with the same challenging questions that we are as policymakers,” Williams said. Markets are “responding to signals in ways that directionally…I think are understandable.”
“The question ahead of us is not whether we need” to maintain a modestly restrictive interest-rate policy, but rather about fine-tuning the level of restriction, he said, as it moves gradually to a more neutral stance. “Is it time to turn that dial down a little bit?” he said.
Fed governors Michelle Bowman and Christopher Waller, who dissented in favor of a rate cut at Wednesday’s meeting, issued statements before Friday’s report warning of weaker labor market conditions if the central bank didn’t resume the rate cuts it began last year.
Williams said it was natural to have divided views, given elevated uncertainty over a range of changes on trade, immigration and fiscal policy.
Williams’s measured approach reflects his view that the Fed still has work to do on inflation. He noted that a widely watched gauge that excludes volatile food and energy prices is “still meaningfully above” the Fed’s 2% target.
While he said businesses’ and consumers’ expectations of inflation had been holding at low and stable levels, he noted the Fed still had an important job to do in making sure that remained true. “It needs to be reinforced or complemented with policies that keep the economy in good balance,” he said.
Williams expects economic growth to slow to around 1% this year but said he anticipates growth to rebound in 2026 as policy uncertainty and other headwinds fade and as tailwinds develop, including America’s dominant position on new artificial-intelligence technology.
“I’m not…particularly worried right now about the economy contracting or really weakening a lot,” Williams said. “It has slowed, but I expect it to just be at this pace of growth for a couple quarters and then come back.”
Sizewell C costs could hit £100bn including financing, modelling shows
Total for nuclear power station project set to be billions of pounds higher than official government estimates
The true cost of the Sizewell C power station in Suffolk could be tens of billions of pounds higher than official government estimates once financing costs are factored in, according to official modelling seen by the Financial Times.
The UK government last week said the mostly debt-funded project would cost an estimated £38bn in real 2024 prices to build.
Under the financial structure of the deal, investors will be rewarded if the project is built for less than £40bn, and not obliged to put in further funds if costs rise above £47.7bn — which is considered unlikely.
But financial modelling — prepared as part of the wider fundraising process and seen by the Financial Times — gives a range of roughly £80bn-£100bn in nominal terms over the period of construction for the two scenarios, once debt interest and payments to shareholders are factored in.
That would imply costs of roughly £65bn-£80bn in real 2024 terms, although the exact costs will depend on inflation rates and spending rates across the lifetime of the project.
The modelling also states the revenue raised from consumer bills to help fund the project is expected to be about £35bn-£50bn in nominal terms, depending on the overall cost of the project.
This will be in the form of a surcharge on energy bills during construction, before the plant starts generating electricity. The government estimates households will pay about £1 each per month, on average, during this period.
The government’s energy department last week initially declined to share details of Sizewell C’s financing costs, saying it had not published them given how variable the costs could be.
When the FT later asked about the unpublished modelling figures, a spokesperson said the figures did not reflect consumers’ exposure to the costs, and the project would be cheaper than its sister project Hinkley Point C.
The government was the project’s biggest single investor and would recycle profits back to consumers, they added.
Sizewell C, which is due to come online in the mid- to late-2030s, is the second large-scale nuclear power project currently being built in Britain, part of an attempt by the government to revive the sector in order to bolster supplies of low-carbon electricity.
Its costs are under scrutiny given UK taxpayers’ and consumers’ hefty exposure to the project, and large cost overruns and delays at Hinkley Point C, which France’s state-owned utility EDF is building in Somerset.
Sizewell C is being built under the “regulated asset base” financing model, which is designed to encourage investment and keep financing costs down by reducing the risk to investors, as they start being paid during construction rather than having to wait until the project is built.
Private investors Centrica, La Caisse and International Public Partnerships are committing combined equity of £3.25bn, while banks backed by France’s export credit agency have committed £5bn in debt.
Centrica, which is taking a 15 per cent stake for £1.3bn, said it was expecting returns of more than 12 per cent.
But most of the funding is coming from public sources: the UK government is supplying £36.55bn in debt to be raised on the gilt market by the National Wealth Fund, as well as £3.8bn in equity, the largest portion. State-owned EDF is supplying £1.05bn and the Nuclear Liabilities Fund £400mn.
The energy department said: “Sizewell C will deliver cheaper clean electricity for generations for at least six decades, and analysis shows the project could save £2bn a year across the future low-carbon electricity system once operational.”
It added: “The £38bn cost of constructing Sizewell C will be spread between consumers, taxpayers and private investors, and represents a saving of around 20 per cent compared with Hinkley Point C, demonstrating the value of building a virtual replica project.”
Luxury’s problem? The price isn’t right
High-end brands have ignored the aspirational shopper. No wonder they’re in trouble
Do you remember your first designer purchase? Mine was in Fenwick, the department store formerly on Bond Street and even then a somewhat fusty retail experience. I bought a grown-up work bag from Sportmax, in dark-brown suede, with big silver details and a secret pocket to put my keys. It cost £250, which for my salary, then in the teens of thousands, was far more than I could afford.
I think about this purchase when I read of the luxury industry’s current woes. Results this week have seen Gucci sales plunging another 25 per cent at Kering, with the parent group suffering a loss of 15 per cent in sales. The picture is not much brighter at LVMH, the group which owns 75 brands including Celine, Louis Vuitton and Dior. LVMH, which has seen its share price slump 19 per cent in the past year, is said to be considering offloading some of its lower performing houses, with Marc Jacobs rumoured to be readying for sale. “We will not keep brands if we believe they are not a good add-on, or we are not the right operator to operate them,” said Cécile Cabanis, LVMH’s finance chief, on an earnings call.
What makes us buy a luxury item? I’ve spent nearly two decades considering the moment that triggers an individual sale. No one needs a luxury: the purchase is instead propelled by aspiration, status, self-affirmation and desire. Then, there’s the slightly transgressive thrill of spending lots of money on something you simply want. The sweaty, credit-card-at-the-ready denial of what others might interpret as a reasonable spend. Some will never feel the impulse. Others know it well. They, like me, are probably familiar with the splurge index of things we are prepared to sacrifice — food, socialising, taxis — in order to justify a spend.
For many years there was a consensus among retailers that the gateway price to snagging a new client was around £250. Having made that psychological leap on that first purchase, so went the thinking, the customer would then return.
That was certainly my experience. The handbag cemented a relationship with luxury that has (ahem) flourished to this day. And while the entry figure of £250 has no doubt increased since the late 1990s, it has by no means kept pace with the price inflation we’ve seen across the industry. A straw poll of my colleagues’ first designer purchases reveals a fascinating inventory — Prada shoes, Chanel sunglasses, Moschino key ring, an Hermès scarf, a vintage Luella “Gisele” bag, so much Marc by Marc Jacobs — that spans several decades but still crucially falls into a price range of between £200 to £350.
Try taking £350 up Bond Street these days and see what you can buy: a pair of shoes is generally about £800. Handbags are at least £2k. Even I am shocked to find a plain white cotton vest top, from a highly covetable Italian label, priced at £870.
After the pandemic, the industry redoubled its efforts on courting the tiny percentage of high-spending clients who made up the majority of sales. And fair enough — if you can sell it. But in driving prices so far beyond anything that might be considered accessible, the aspirational shopper got left behind.
It makes me especially sad that a brand such as Marc Jacobs, once the quintessential entry-level brand for young fashionistas, is considered to have failed. A strategic restructuring by LVMH in 2015 subsumed the cheaper offering into the more expensive fashion line and in doing so it became too esoteric and high-end. Likewise, all the kids who once bought its tote bags, T-shirts and shoes found their portal into fashion closed.
Much has been written about the slowdown in the luxury industry, and how the real success stories are those brands that have cleaved to more conservative aesthetics instead of chasing febrile trends. This week, Hermès reported 9 per cent growth in its second quarter, a rare star in a beleaguered field. But although most analysts obsess about demand for its £12k Birkin handbag, the house is still buoyed by lowly folk like myself who buy its perfumes, notebooks and fashion jewellery.
And not all of luxury is suffering: in the wake of the pricing vacuum a number of smaller independent brands have emerged. Studio Nicholson, Tove, Me+Em (all founded by women, incidentally) are boasting healthy growth. The secret to their success has been the dazzling revelation that most people don’t have infinite bank accounts. Toteme, the Swedish brand co-founded by husband and wife Elin Kling and Karl Lindman in 2014, for example, is now turning revenues “just shy” of £200mn. Their strategy from the outset has been to keep its pricing aspirational.
For the past few years, the aspirational shopper was seen as a dirty word. The big houses didn’t seem to care about them — they only wanted the crème. I have since bought many other handbags, but that first Sportmax purchase remains dear. At the time it was a huge investment, but it was also an investment in my self-worth. It made me feel confident and stylish: it was something I had earned. There’s no luxury in going shopping and simply feeling poor.