Apple in China occupies a position unlike the company’s presence in any other country. China is simultaneously a major consumer market, a critical manufacturing center, an extensive developer economy and the home of Apple’s most determined hardware competitors.
That combination creates an opportunity few technology companies can ignore. China has hundreds of millions of affluent mobile users, a strong appetite for premium devices and a digital economy built around smartphones.
It also creates unusual complexity. Apple must compete with Huawei, Xiaomi, Oppo and Vivo while satisfying Chinese regulations, navigating tensions between Beijing and Washington and reducing its manufacturing dependence without damaging the supplier network that helped make the iPhone possible.
The latest market figures show why Apple cannot treat China as a business it has already lost. IDC estimated that Apple captured 18.1 percent of Chinese smartphone shipments during the second quarter of 2026, placing it second behind Huawei at 22.6 percent.
Apple’s shipments increased by 24.4 percent from the previous year even as the wider Chinese smartphone market contracted. The recovery demonstrated that Chinese demand for the iPhone remains substantial when product design, storage, performance and pricing align with consumer expectations.
The same figures also show why the position can never be considered secure. Huawei remains ahead, local brands respond quickly and a few percentage points can separate several major vendors.
China is already one of Apple’s largest markets, but its scale means that even established product categories contain room for expansion.
The country has a large premium smartphone segment, extensive 5G coverage and consumers accustomed to using mobile devices for communication, payments, shopping, entertainment, transport and work.
An iPhone purchase can also lead to AirPods, Apple Watch, iPad, Mac, iCloud storage and paid applications. The commercial opportunity is therefore larger than the initial handset transaction.
Apple’s strongest advantage is its integrated ecosystem. An iPhone can share passwords, photographs, messages and files with a Mac or iPad, while AirPods and Apple Watch extend the experience without requiring separate accounts and setup systems.
That continuity becomes more valuable as households accumulate devices. Replacing one iPhone is relatively easy. Leaving an ecosystem containing several products, purchased applications, cloud storage and shared family services involves greater friction.
Apple also benefits from a premium identity that remains recognizable in China. Local manufacturers now produce sophisticated flagship phones, but the iPhone continues to operate as both a personal computer and a luxury consumer product.
The importance of that status should not be exaggerated. Chinese buyers are technically informed and increasingly willing to support domestic brands. Brand recognition can attract attention, but it does not excuse weak features or an uncompetitive product cycle.
Apple’s 2026 recovery offers a useful example of how quickly its position can improve.
China’s smartphone shipments fell 4.3 percent during the second quarter as memory and component costs pushed several manufacturers toward higher prices. Apple and Huawei held prices more stable, giving buyers a reason to purchase while competing brands became more expensive.
The iPhone 17 generation also addressed several areas where Apple’s standard models had appeared conservative beside Chinese flagships. Stronger base specifications made the regular iPhone more competitive without requiring every customer to move into the Pro tier.
This matters in a market where consumers compare display technology, camera systems, storage, battery life, charging and AI features closely. Apple cannot depend entirely on iOS loyalty when domestic companies introduce visible hardware improvements at several price levels.
The rebound does not establish a permanent trend. Smartphone demand in China remains affected by economic confidence, upgrade cycles, government incentives and the timing of major shopping festivals.
Apple’s position can also look different according to the research firm and whether a report measures shipments or consumer sales. Quarterly rankings should be treated as evidence of momentum rather than a final judgment on the market.
The important signal is that Chinese consumers have not broadly rejected the iPhone. Demand can return rapidly when the product and price create a convincing upgrade.
Huawei gives Apple a competitor with advantages unavailable to most global smartphone manufacturers.
The company combines strong national recognition with advanced cameras, communications technology, custom chips, local services and an expanding HarmonyOS ecosystem. Its return to premium smartphones removed the opening Apple had gained when U.S. restrictions disrupted Huawei’s access to important components.
Huawei can also position its products as part of China’s technological self-reliance. That message carries commercial and political weight beyond a conventional comparison between two phones.
Apple must therefore compete against product quality and national sentiment simultaneously. An iPhone can offer strong performance and still lose a purchase to a domestic flagship viewed as technically capable and strategically important.
Xiaomi, Oppo and Vivo add different pressures. They move quickly across price categories, experiment with folding designs, adopt larger batteries and provide features adapted to local applications and buying habits.
Chinese phone makers also operate inside a retail and software environment they understand directly. They can build relationships with local platforms, carriers and commerce services without translating a global strategy into Chinese requirements.
Apple follows a more centralized product model. That consistency protects the identity of the iPhone, but it can make the company slower to answer features gaining popularity inside one country.
Apple in China is not only an iPhone business. The installed device base creates demand for applications, cloud storage, entertainment, subscriptions and digital purchases.
China has an enormous developer community producing games, commerce platforms, social services, education tools and productivity software. The App Store connects those developers with high-value iPhone and iPad customers.
The relationship is commercially important but politically complicated. Apple must operate the mainland App Store according to Chinese law, including rules affecting games, generative AI, communications services and content.
Applications available elsewhere may be absent from the Chinese storefront. Developers may also need licenses or regulatory approval before distributing certain categories of software.
Apple changed commission rates and related policies for the mainland China storefront in March 2026. The company described the adjustments as part of maintaining a strong business opportunity for Chinese developers.
Regulatory and developer pressure has not disappeared. Complaints involving commissions, distribution restrictions and payment policies continue to challenge Apple’s control of the platform.
The company cannot simply copy its response from another region. European rules focus heavily on alternative marketplaces and interoperability, while China combines competition concerns with content regulation, data localization and licensing requirements.
Every concession must therefore be designed for a distinct legal environment without fragmenting iOS into an unmanageable collection of regional systems.
Apple’s relationship with China extends far beyond the devices sold there.
For more than two decades, Apple and its manufacturing partners built a dense network of factories, component suppliers, engineers, logistics operators and toolmakers across the country.
That network does more than provide low-cost labor. It can place specialists from several suppliers near an assembly site, modify production lines quickly and increase output to enormous volumes during a product launch.
China’s infrastructure also supports the less visible stages of manufacturing. Components require precision tooling, testing equipment, chemicals, packaging, subassemblies and experienced quality-control teams before final assembly begins.
Apple has expanded iPhone production in India and moved parts of its Mac, iPad, Apple Watch and AirPods supply chains to countries including Vietnam. It is also increasing semiconductor and component commitments in the United States.
Those investments reduce the danger of relying on one country, but they do not recreate China’s complete manufacturing environment immediately.
A factory in another region may assemble an iPhone while still depending on Chinese machinery, materials, subcomponents or engineering support. Production diversification is therefore a gradual redistribution of individual stages rather than a clean departure.
Apple must reduce concentration risk without weakening the manufacturing system responsible for its product quality, scale and launch schedules.
The same China supply chain that provides Apple with extraordinary efficiency also exposes the company to political decisions outside its control.
Tariffs can raise the cost of devices exported from Chinese factories. Export restrictions can limit access to materials or production equipment, while sanctions may affect suppliers and technology partners.
Chinese government policies can influence the use of foreign devices inside state agencies and companies. U.S. rules can restrict the technologies available to Chinese firms that also participate in Apple’s supplier network.
Apple cannot solve these problems through product design alone. It must maintain relationships with two governments whose technology policies increasingly emphasize security, domestic production and strategic independence.
Moving more assembly to India gives Apple additional flexibility, particularly for iPhones destined for the United States. China nevertheless remains important as a production center and consumer market.
An aggressive withdrawal could damage supplier relationships and local sales. Moving too slowly could leave Apple vulnerable to another trade disruption.
The company is trying to preserve Chinese capabilities while creating enough capacity elsewhere to keep a political dispute from becoming a worldwide product shortage.
Apple’s physical stores remain an important advantage in China because they give the company direct contact with customers.
The stores allow buyers to compare models, receive technical support, attend educational sessions and enter the ecosystem without depending entirely on a carrier or reseller.
That experience is particularly valuable for premium products. A customer considering a Mac, Apple Vision Pro or high-end iPhone may need more explanation than someone replacing an inexpensive handset.
Retail locations also communicate long-term commitment. A prominent store represents more than a sales channel; it signals that Apple intends to remain part of the local technology market.
Online commerce is equally important. Chinese consumers are accustomed to large shopping festivals, aggressive promotions and rapid delivery through domestic platforms.
Apple must balance its controlled global pricing with a market where discounts can change quarterly rankings. Too little flexibility makes the iPhone appear expensive, while frequent reductions can train customers to delay purchases and weaken its premium image.
The strongest 2026 performance came partly from avoiding price increases while component costs were forcing rivals to reconsider their pricing. That advantage may disappear when competitors adjust their supply chains or launch another cycle of promotions.
China is often discussed as a sales problem for Apple, but it also functions as a demanding test of the iPhone.
Domestic flagship phones compete through larger batteries, rapid charging, advanced camera hardware, folding displays and deep integration with local digital services.
Features considered experimental elsewhere can become established expectations more quickly in China. Apple’s slower adoption cycle may look disciplined in one market and conservative in another.
The company must decide which expectations represent durable improvements and which are temporary specification contests. Copying every local feature would weaken Apple’s focused product strategy, but ignoring major changes creates openings for rivals.
This pressure can benefit Apple globally. Competition in China forces the company to defend the value of its standard iPhone, improve battery efficiency, strengthen cameras and make software features more useful outside English-speaking markets.
The opportunity is therefore not limited to selling more devices. China can reveal where Apple’s global assumptions no longer match the most competitive part of the smartphone industry.
The post Apple in China Is Its Greatest Opportunity and Hardest Test appeared first on AppleMagazine.
Todas as empresas de capital aberto (que possuem ações à venda na bolsa de valores) são obrigadas, por lei — na maioria dos países (inclusive Brasil) —, a disponibilizar publicamente suas demonstrações financeiras periodicamente. Esse costuma ser um momento bastante esperado por acionistas (analisando se a empresa na qual estão investindo está sendo bem administrada e gerando lucros), investidores (para captar novas oportunidades de alocar seus recursos), além da imprensa e de curiosos em geral. Na última quinta-feira (30/7), foi a vez da Apple de divulgar seus resultados financeiros referentes ao trimestre abril-junho de 2026.
Para alegria de todos, o CEO 1Chief executive officer, ou diretor executivo. Tim Cook se despede da sua posição com o melhor terceiro trimestre fiscal que a empresa já registrou — um recorde absoluto! Como já informado aqui no MacMagazine, no período de abril a junho (encerrado no dia 27/6), a companhia relatou uma receita de US$109,4 bilhões (alta de 16,4%) em relação ao mesmo trimestre do ano anterior, e lucro líquido de US$29,8 bilhões (um avanço de 27%). A margem bruta chegou a incríveis 50,1%, a maior já apurada em períodos semelhantes da história da empresa!
Hoje, a Apple tem orgulho de relatar o nosso melhor trimestre de abril-junho de todos os tempos, com crescimento de receita de dois dígitos em iPhone, Mac e Serviços, e em todos os segmentos geográficos.
A fala acima é de Cook. Ele deixará o cargo em 1º de setembro, quando John Ternus assumirá, e passará a presidir o conselho de administração.
Uma das grandes surpresas do relatório foi o crescimento no faturamento de Macs. Pela primeira vez, a categoria ultrapassou a casa dos US$10 bilhões em um trimestre abril-junho, e o crescimento de 29% superou o do iPhone.
A explicação está em parte na procura alucinada por Macs mini para rodar agentes de inteligência artificial e na leva de lançamentos de março: MacBook Air com M5, MacBook Pro com M5 Pro e M5 Max, e principalmente o MacBook Neo, o modelo de entrada que estreou a US$600 — preço que subiu para US$700 em junho, no reajuste geral motivado pelo custo de memória. A Apple parece ter encontrado um mercado relevante abaixo do Air, e vale acompanhar o que isso faz com a margem da linha nos próximos trimestres.
O iPad foi a única categoria no vermelho, com queda anual de 6%. Não é surpresa: a própria Apple havia avisado que a comparação seria difícil, pois no ano anterior houve o lançamento do iPad com chip A16. O iPad Air com M4, que chegou em março, não foi suficiente para compensar. Cabe ainda analisar se o bem-sucedido MacBook Neo não acabou levando um pouco do mercado dos tablets, também.
A margem bruta de 50,1% e o lucro por ação de US$2,02 não são resultado apenas da operação, e esse detalhe se perde facilmente na comemoração dos recordes.
À primeira vista, no press release — e nos comentários que fizemos no último podcast —, esse ponto não ficou bem claro, mas merece atenção. A Apple informou, no comunicado, que a margem incluiu cerca de dois pontos percentuais e o lucro por ação incluiu US$0,11 vindos de reembolso de tarifas, dinheiro que a empresa havia pago ao governo americano e recebeu de volta depois que a Suprema Corte derrubou, em fevereiro, as tarifas impostas sob a IEEPA (a lei de poderes econômicos internacionais de emergência). A empresa não detalhou a origem exata dos valores devolvidos, mas a decisão da Corte é o evento que abriu caminho para os reembolsos.
É dinheiro real, entra legitimamente no resultado, mas não é recorrente e não diz nada sobre a capacidade da Apple de vender aparelhos. Descontando o efeito, a fotografia muda um pouco: a margem bruta cai para aproximadamente 48%, exatamente o que o consenso projetava, e o lucro por ação vai para cerca de US$1,91, contra os US$1,89 esperados. De uma superação de 6,9% no lucro, sobra uma de pouco mais de 1%.
Ou seja: o trimestre continua bom em receita e em volume, que são genuinamente fortes. Mas o brilho extraordinário no lucro é, em boa parte, contábil, e não se repete.
A Apple abre a margem bruta em dois blocos apenas: produtos e serviços. Ela nunca divulga margem por categoria individual (não existe margem oficial de iPhone, Mac ou iPad em documento nenhum da empresa, e todo número desse tipo que circula é estimativa de analista), mas os dois blocos que ela abre já contam a história.
| Linha de negócio | Margem do FQ3 2026 | Margem do FQ3 2025 | Variação |
|---|---|---|---|
| Produtos | 40,07% | 34,52% | +5,55pp |
| Serviços | 75,62% | 75,58% | +0,04pp |
| Consolidado | 50,06% | 46,49% | +3,57pp |
A margem de hardware subiu 5,55 pontos percentuais em 12 meses. A margem de hardware não sobe 5 pontos por ganho de eficiência: é ali que o reembolso de tarifas desembarcou, porque tarifa incide sobre produto físico. A prova está na linha seguinte: serviços, que não pagam tarifa nenhuma, ficou praticamente parado. A melhora não foi operacional.
Vale registrar um desequilíbrio que explica boa parte do valuation da Apple. Serviços representam 28,1% da receita, mas respondem por 42,4% do lucro bruto do trimestre. Produtos são 71,9% da receita e apenas 57,6% do lucro. Cada dólar de serviço vale quase o dobro de um dólar de hardware, e é por isso que o mercado aceita pagar perto de 40x o lucro por uma empresa que, no fim das contas, vende telefone.
Os relatórios revelaram um problema silencioso do trimestre. Serviços cresceram 12,1%, um bom número isolado, mas bem abaixo dos 16,2% do trimestre encerrado em março e, mais grave, abaixo do que a própria Apple havia sinalizado. Na teleconferência de abril, o diretor financeiro Kevan Parekh disse que os serviços cresceriam em ritmo “similar ao reportado em março” (ou seja, algo perto de 16%, ressalvando o efeito cambial). Ficaram cerca de 4 pontos percentuais aquém.
A linha chegou a recuar 0,8% em relação ao trimestre imediatamente anterior. Foi o único item da projeção que a Apple não cumpriu, e não é um item qualquer: é o que sustenta o múltiplo da ação. Parte da imprensa de mercado atribuiu a reação negativa à combinação de serviços com a Grande China, embora as estimativas para a região divergissem bastante entre as casas: havia projeções de US$18,8 bilhões, que o resultado cumpriu, e de US$19,5 bilhões, que não.
Some-se a isso um risco que não tem data marcada. Boa parte da receita dos serviços vem do acordo pelo qual o Google paga à Apple para ser o buscador padrão do Safari — cerca de 36% da receita de anúncios de busca, algo próximo de US$20 bilhões por ano. Existe um processo em andamento que questiona a legitimidade desta taxa. A sentença de setembro de 2025 permitiu que o acordo continuasse, mas o Departamento de Justiça dos Estados Unidos recorreu em fevereiro de 2026 e o Google entrou com recurso próprio em maio. Se a decisão cair, o buraco nessa linha é de dois dígitos de bilhão!
Fora das manchetes do balanço, três indicadores das demonstrações merecem atenção, e nenhum deles apareceu no material de divulgação.
Há ainda um detalhe na linha de outras receitas, que agrupa o que a Apple ganha fora da operação: juros do caixa aplicado, juros pagos sobre dívida, variação cambial e valorização de participações minoritárias. O resultado foi de US$572 milhões positivos, contra US$171 milhões negativos há um ano — um giro de US$743 milhões. A carteira de caixa e aplicações subiu para US$146,5 bilhões, enquanto a dívida caiu para US$82,3 bilhões. Mais dinheiro rendendo, menos juros a pagar. É lucro financeiro, não operacional.
O trimestre relatado é recorde. A projeção que o acompanha, não. E como o mercado precifica futuro e não passado, foi ela que dominou a reação. A ação, que havia fechado o pregão a US$333,43, caiu cerca de 6,5% na negociação estendida ao longo da teleconferência, o que equivale a aproximadamente US$317 bilhões de valor de mercado. No dia seguinte, a queda se confirmou: -7,35%. A projeção da Apple para o trimestre de setembro, o último do ano fiscal, veio com dois números desconfortáveis.
| Indicador | FQ3 2026 (real) | FQ4 2026 (projetado) |
|---|---|---|
| Crescimento de receita | 16,4% | de 9% a 11% |
| Margem bruta | 50,1% | de 47% a 48% |
| Margem bruta sem reembolso de tarifas | 48,1% | de 47% a 48% |
| Receita implícita | US$109,4 bilhões | de US$111,7 a US$113,7 bilhões |
A desaceleração de 16,4% para algo entre 9% e 11% já seria suficiente para incomodar. O que agravou foi a justificativa: Parekh afirmou que o impacto das restrições de oferta deve “aumentar significativamente” em relação ao trimestre anterior — citando explicitamente iPhone, Mac e iPad —, e mencionou ainda ventos contrários de câmbio sobre a receita de iPhone.
A margem projetada de 47% a 48% é o número mais revelador de todo o balanço. Ela fica abaixo do resultado limpo deste trimestre sem o reembolso de tarifas. Ou seja: a Apple está dizendo que a margem vai cair de verdade, em bases comparáveis. O colchão de estoque de memória comprado a preço favorável, que segurou os custos nos últimos trimestres, está acabando, e o trimestre de setembro é justamente o que carrega o lançamento dos “iPhones 18 Pro/18 Pro Max” — e provavelmente do “iPhone Ultra”, o primeiro dobrável da empresa.
Ternus assumirá como diretor executivo em 1º de setembro, com Cook passando a presidente executivo do conselho. A julgar pelo balanço, ele recebe uma casa em ordem: receita recorde, base instalada de aparelhos ativos em máxima histórica em todas as categorias principais, patrimônio líquido de US$107,5 bilhões e — possivelmente — a companhia mais valiosa do mundo.
Recebe também três problemas bem definidos. O primeiro é o custo de memória, que já forçou aumentos de preço no MacBook Air, no Mac mini e no Mac Studio, e que agora vai comprimir margem sem o amortecedor do estoque barato.
O segundo é a desaceleração dos serviços, com os recursos judiciais contra o acordo de busca com o Google ainda pendentes.
O terceiro é entregar a Siri reformulada, anunciada na WWDC26, movida pelo Gemini do Google, prevista para o ciclo do iOS 27 e adiada por tempo indeterminado na União Europeia — decisão que a Apple atribui às exigências da Lei dos Mercados Digitais (Digital Markets Act, ou DMA).
A leitura mais honesta deste balanço talvez seja esta: Cook encerra sua gestão entregando o melhor trimestre abril-junho da história da Apple, a maior capitalização de mercado do mundo e uma operação que cresce dois dígitos em todos os continentes em que atua. O que ele não entrega pronto é a próxima curva: provar que a aposta de rodar inteligência artificial dentro do aparelho, e não na nuvem, vale o múltiplo de 40x que a ação vinha cobrando. Essa parte, agora, é com Ternus.
Metodologia
Os números do trimestre e do comparativo anual foram extraídos das demonstrações consolidadas condensadas não auditadas da Apple referentes ao terceiro trimestre fiscal de 2026 e ao terceiro trimestre fiscal de 2025, conferidas linha a linha. As margens por linha de negócio, os indicadores de balanço patrimonial e de fluxo de caixa e todos os percentuais foram calculados a partir desses documentos.
As estimativas de consenso de Wall Street citadas foram compiladas antes da divulgação junto a Yahoo Finance, TipRanks, Zacks, Visible Alpha, Hudson Labs e Goldman Sachs.
A projeção para o trimestre de setembro e as declarações do diretor financeiro vêm da teleconferência de 30 de julho de 2026, com apuração de 9to5Mac, CNBC e AppleInsider.
Os dados de mercado chinês são da IDC e da CAICT. A série de fechamentos mensais da ação tem origem em StatMuse, StockAnalysis.com, FinancialContent e Macrotrends, via S&P Global Market Intelligence.
O ajuste identificado como “sem reembolso de tarifas” não é uma linha contábil: é cálculo próprio a partir do efeito que a Apple declarou no comunicado de resultados.
Observação
Este texto não constitui recomendação de investimento.
Last week, OpenAI claimed that a group of its AI models had broken containment, successfully hacking into the systems of open source AI platform Hugging Face to cheat on a benchmark test.
In the wake of the announcement, two very distinct narratives have emerged surrounding OpenAI’s claims. Some say it was essentially a publicity stunt, with the company setting parameters for the test that pushed the models toward outrageous behavior. But others, including certain prominent researchers, warn that the hack should serve as a warning shot for an even more severe AI-enabled cybersecurity disaster that’ll inevitably take place as models become more sophisticated.
“This is the first time, to my knowledge, that an AI system has autonomously committed a crime,” said New York Times journalist Kevin Roose of the event. “If a human did to Hugging Face what OpenAI’s models did to Hugging Face, they would be charged with computer fraud, and potentially sent to prison or fined or prosecuted.”
Debate will surely continue to rage among wonks and skeptics. And new details aren’t exactly tamping out the sense of alarm: on Tuesday, OpenAI issued an update to its ongoing investigation, claiming the incident was worse than initially thought. In addition to hacking Hugging Face, the company now says, its models “used publicly exposed credentials at the account-level on other publicly available services,” totaling “four accounts on four services.”
“We’ll continue to notify service owners directly, and have not seen evidence of broader impact to these providers or other accounts on their services,” OpenAI wrote, without elaborating on which services were affected.
The news further raised alarm bells among some cybersecurity experts, highlighting ongoing concerns over the tech’s ability to evade protective measures. It’s a possibility that researchers have warned about for years, and the incident suggests that the threat is now turning from a possibility into a reality.
On the other hand, more skeptical experts have become suspicious about OpenAI’s hair-raising tale. After all, we’ve heard a strikingly similar story from its biggest competitor, Anthropic, mere months ago. Could OpenAI’s latest admission be a bid to build hype to drum up excitement and prove to investors that its latest AI models are just as much of a cybersecurity threat as Anthropic’s fabled Mythos?
Experts also point out that the Hugging Face hack could’ve easily been prevented, further adding credence to the theory that OpenAI was looking for attention from the public. As cloud security firm Edera co-founder Alex Zenla told Wired, the hack was largely a result of callousness on OpenAI’s part.
“People are YOLO-ing really hard,” he said. “It’s shocking how little people have really thought about a scenario like this.”
“I consider all AI and anything AI touches to be fully untrusted — which is fine, you just need to build against that,” Zenla added. “And this situation proves the point. The fact that OpenAI wasn’t more paranoid about this seems kind of reckless.”
“A simple analysis of the actual risk has an actual simple answer,” security and compliance consultant Davi Ottenheimer told Wired. “The OpenAI mistakes were dead simple.”
As Wired points out, simple protections like fully isolating AI services from the internet could’ve prevented the hack, which have been intimately familiar to researchers for decades now.
Considering OpenAI is nearing a $1 trillion valuation and should have all the resources in the world at its disposal, major lapses in security should have anybody start questioning the company’s narrative.
Yet there could be truth to both versions of the story. It’s entirely possible that there is legitimate cause for concern as AI models become more sophisticated at identifying cybersecurity vulnerabilities. We’ve already seen frontier models flagging thousands of software bugs, underlining their growing competence.
But OpenAI is also heavily invested in showcasing its models’ capabilities to the world as it tries to keep up with Anthropic. That leaves the possibility that the company could’ve coordinated with Hugging Face to orchestrate the hack — or at least given its AI models a strong push in the direction of controversy.
More on the hack: OpenAI Says a Group of Its Models Broke Out of Secure Containment and Hacked Another AI Company
The post OpenAI’s Escaped Models Were Allegedly Rampaging More Extensively Than Previously Reported appeared first on Futurism.
The US job market is in a weird place right now.
Life is clearly rough for the working class, stuck navigating Kafkaeqsue hiring infrastructure while billionaires devour an ever-larger slice of the world’s wealth. But frustratingly, the data being used to steer the economy tells a rosier story than the one workers actually live. The latest unemployment numbers — a flawed metric, admittedly — showed a slight improvement from May to June, at least on paper. And even as the economy appears to cool its jets, Americans keep spending like there’s no tomorrow, even in the face of aggressive inflation.
With so much data — anecdotal and empirical — swirling around, it’s hard to find the pulse of the US labor force. Luckily, one enterprising AI startup found a perfect barometer. There’s no conflicting data or tedious calculations involved — just a little ink, and the vibrating sting of a tattoo machine.
On Thursday, the Y Combinator-backed AI startup LemonLime hosted a bizarre recruitment event in which attendees were offered an “instant interview” in exchange for getting the company’s logo tattooed on their body.
It sounds like a joke, but LemonLime co-founder Jordan Zietz bragged on a LinkedIn post that an astonishing seven people took them up on the offer.
According to his post, the goal was to “meet exceptional people,” and sort out “which of them are just as crazy as we are.”
“That’s why we brought an actual tattoo artist to our party and offered an instant interview to anyone who got a LemonLime tattoo,” Zietz continued. “If you’re reaching out to us to apply for a job, you might want to take notes.”
While coercive body art has become increasingly common in the 21st century, there’s usually at least some form of financial compensation involved. The idea that it takes getting a permanent tattoo just to get one foot in the door for an interview is especially galling, and reveals a lot about the sorry state of the US job market.
One of the people who received a free tattoo, Dimple Amitha Garuadapuri, clarified on LinkedIn that while she didn’t get the company’s logo inked on her skin, “almost everyone else got a LemonLime tattoo.”
Regardless, news of the stunt spread quickly on social media, prompting a lightning-fast apology from the founder, who admitted on X-formerly-Twitter that he “messed up.”
“Even though participants chose their own tattoos and no one was asked to get our logo, I should have understood the pressure and power dynamic created by connecting tattoos to hiring,” Zietz said. “We’ve taken the posts down and have spoken to everyone who participated. We’re also ready to cover associated removal costs with anyone facing any delayed regret.”
Just for good measure, the founder stressed that “interviews were available to everyone regardless,” though it remains to be seen if the seven share that view.
Did you attend the YC afterparty with LemonLime? If so, we’d love to hear from you. Reach out at tips@futurism.com. We can keep you anonymous.
More on tech startups: Startup Sued After Delivery Bot Clobbers 73-Year Old Woman
The post We Are Mildly Horrified by This AI Startup That Coerced People Into Getting Tattoos of Its Logo in Exchange for a Job Interview appeared first on Futurism.
For a time, it appeared as if AI would find broad acceptance among the American masses.
The peanut gallery on social media happily embraced early image generation models as researchers appeared to make significant breakthroughs with the aid of their newfound tools. On paper, surveys found public attitudes toward the tech continued to improve into 2025 — before sharply reversing.
In 2026, America’s honeymoon phase with AI seems to be quickly wearing off. According to a joint poll conducted by Bentley University and Gallup, the percentage of US adults who think AI does “more harm than good” is rising, from 31 percent in 2025 to 39 percent this year, while a piddling 9 percent still think the tech does “more good than harm.”
As outlooks sour, the Bentley-Gallup poll found that Americans’ familiarity with AI continues to grow. In 2026, 70 percent of those polled said they’re “somewhat or extremely knowledgeable” about AI, up 6 percent compared to 2024. The implication probably gives OpenAI Sam Altman a migraine: the more people learn about AI, the less they seem to want it in their lives.
Meanwhile, Americans are increasingly wary about data centers, with a recent Politico survey finding that 41 percent of US adults would oppose a data center being built within three miles of their home, up drastically from 28 percent in January. At the same time, the volume of support for data centers plummeted, down 13 percent over the past six months alone.
Speaking to Bloomberg about the rising aversion to AI, professor of ethics at Carnegie Mellon University’s Tepper School of Business Derek Leben said the attitude is increasingly taking root along social lines. Not only do AI haters dislike the tech in their own personal sphere, but they’re increasingly scrutinizing their friends, family, and coworkers about their attitudes.
“The most obvious social effect is suspicion,” Leben told Bloomberg. “There is also a moral and political judgment that is coming along with AI usage, especially in creative industries where the stigma is the strongest.”
Carla Milan, a 37-year-old public health professional in New York City, told the publication she’s frustrated that family and friends deploy AI reflexively, without considering its social or environmental costs.
“I hear people champion AI and I’m like, ‘sure, why don’t we all participate in our collective downfall?'” Milan told Bloomberg. “Why don’t we all play a part in diminishing our humanity and degrading the environment?”
More on AI: Meta Working on Crushingly Sad AI App That Tells Bedtime Stories to Children, for “Tired” Parents
The post The More People Learn About AI, the More They Want It Out of Their Lives appeared first on Futurism.
Apple Notes in iOS 27 receives a focused collection of upgrades designed to make long documents easier to navigate and information simpler to move between apps.
Section links allow users to jump directly to headings inside a note. Divider lines create clearer visual boundaries, while expanded Markdown support improves workflows involving websites, writing tools and plain-text editors.
Siri AI introduces the most ambitious addition by creating notes, adding information to existing documents and transferring answers from a Siri conversation without requiring manual copying.
Image Playground also becomes more capable inside Notes, gaining additional visual styles and photorealistic image generation through Apple’s new cloud-based model.
Not every feature requires Apple Intelligence. Divider lines, section links, Markdown tools and the updated interface work on every iPhone compatible with iOS 27. Siri AI and advanced Image Playground capabilities need newer hardware and may have additional language or regional restrictions.
Section links solve one of the largest problems affecting long notes: reaching information located several screens below the beginning.
A user can now select text, choose Add Link and connect it to any heading or subheading inside the current note. Tapping that linked text immediately moves the reader to the selected section.
The feature can turn the opening of a document into a manual table of contents. A project note might contain links to its schedule, budget, participants, research and outstanding tasks without requiring separate documents.
Select text > Add Link > Link to Section > Choose a heading
Notes displays the headings and subheadings already present in the document. Regular body paragraphs do not appear as destinations, so the note must use Apple’s title or heading formats before links can be created.
Select the section title > Formatting > Heading or Subheading
The visible link text does not need to match the destination heading. A phrase such as “View the current budget” can link to a section titled “Financial Information.”
Section links differ from the existing ability to connect separate notes. Apple Notes already allows users to create internal links leading from one document to another. The new control moves within the same document instead.
That distinction is useful for meeting records, research files, travel plans and study material that become too long for ordinary scrolling but do not need to be divided into several notes.
Links should continue working when the document is shared through iCloud because they point to structural elements inside the note rather than fixed screen positions.
Renaming or deleting a destination heading may affect its link. Users maintaining complex documents should test their navigation after reorganizing major sections.
Apple Notes in iOS 27 adds horizontal divider lines as a native formatting element.
The line creates a visible separation across the document without requiring users to type repeated hyphens, underscores or other characters.
Place the cursor > Open the edit menu > Insert Divider Line
The divider is inserted at the cursor position and moves with the surrounding content as the note changes.
A horizontal line can separate meeting notes by date, distinguish reference material from personal comments or divide a recipe into ingredients, preparation and serving information.
It also provides structure without making every section a heading. Headings define the hierarchy used by section links, while dividers create a visual break that does not become another navigation destination.
Typing a row of hyphens previously produced an inconsistent substitute. The characters could wrap on a narrow screen, change length after editing or become difficult to select. The native element maintains its appearance as the document moves among iPhone, iPad and Mac.
Divider lines should be used selectively. Adding one between every short paragraph can make the note look more fragmented rather than organized.
The strongest layout combines headings for major subjects, subheadings for smaller divisions and divider lines where one type of content clearly ends before another begins.
Apple Notes gained the ability to import and export Markdown files with iOS 26. The iOS 27 upgrade removes additional steps from copying content between Notes and Markdown-based tools.
Markdown uses plain-text characters to represent formatting. A hash symbol can identify a heading, pairs of asterisks can create bold text and square brackets can define a link.
The format is widely used by writers, developers, documentation systems and publishing platforms because it remains readable without requiring a proprietary editor.
When Markdown-formatted text is pasted into Apple Notes in iOS 27, the app automatically converts the syntax into rich text. Headings become formatted headings, links become clickable and emphasis appears as bold or italic text.
A pasted line such as Project Schedule appears as a formatted heading rather than displaying the two hash symbols.
Moving content in the opposite direction is also easier. Apple has added a Copy as Markdown command that converts selected rich text into Markdown syntax before placing it on the clipboard.
Select the text > Open the edit menu > Copy as Markdown
The result can be pasted into a content-management system, code editor, Markdown document or another writing application while preserving supported formatting.
Copy as Markdown does not transform Apple Notes into a complete Markdown editor. Users continue writing through the familiar rich-text interface rather than viewing raw syntax throughout the document.
Complex attachments may not translate perfectly. Scanned documents, drawings, audio recordings, embedded files and some tables do not have universal Markdown equivalents.
Basic elements such as headings, links, emphasis and lists provide the most dependable transfer. Users preparing an important publication should review the pasted result before assuming every element survived the conversion.
Siri AI gives Apple Notes in iOS 27 a more direct connection with information generated or located through the assistant.
A user can ask Siri a question, refine the answer through conversation and save the result into a new note. The transfer happens without selecting the response, copying it and leaving Siri to open Notes manually.
“Create a note called Kyoto Trip and add this itinerary.”
“Add these restaurant suggestions to my Kyoto Trip note.”
The integration extends beyond saving complete answers. Siri AI can reorganize poorly formatted information, convert text into a clearer structure and place the revised version inside Notes.
Someone collecting an unorganized group of tasks could ask Siri to separate them by priority before saving the result. A block of research could be converted into headings and shorter paragraphs.
Siri’s personal-context capabilities can help locate an existing note when the precise title is forgotten. The assistant is designed to search information across apps and understand requests using details connected to the user.
These actions remain subject to permissions. Siri should not silently place information into a document without recognizing the requested destination and action.
Users should review AI-generated material before treating it as authoritative. Reformatting supplied text carries less factual risk than asking Siri to generate research, dates, quotations or instructions from general knowledge.
The main Apple Notes iOS 27 update runs on iPhone 11 and later, as well as the second-generation iPhone SE and newer models.
Siri AI has a narrower hardware requirement. Apple lists iPhone 15 Pro, iPhone 15 Pro Max and every iPhone 16 or later model as compatible with the new Apple Intelligence architecture.
An iPhone 15, iPhone 15 Plus, iPhone 14 or iPhone 11 can receive section links, dividers, Markdown support and the redesigned Notes interface. Those devices cannot use the Siri AI note-writing actions.
This separation means two iPhones running the same operating system can display different Notes capabilities.
Siri AI is scheduled to reach users as a beta later in 2026, beginning in English. Apple plans to expand language availability, but individual capabilities may not arrive simultaneously in every supported language.
The assistant will not initially be available on iPhone or iPad in the European Union. Apple has also said its new Apple Intelligence features will remain unavailable in China while the company works through regulatory requirements.
Ordinary Siri controls may continue to perform basic actions on unsupported devices, such as creating a simple dictated note. They should not be confused with the new ability to hold a conversation, reorganize information and add the result to a chosen document.
Image Playground remains integrated with Apple Notes in iOS 27 and receives a substantially more capable generation model.
The upgraded system can produce higher-quality images in a wider range of visual styles, including photorealistic results. Earlier Image Playground versions concentrated on illustration, animation and sketch-like appearances.
Users can describe an image and insert the generated result into the current note. A travel plan might receive a destination image, while a creative project could use generated concept art beside written ideas.
Photorealistic generation runs through Private Cloud Compute rather than entirely on the iPhone. Apple says requests processed through this architecture are not stored or made accessible to the company.
Generated images automatically include a hidden SynthID watermark identifying them as AI-created. The watermark does not appear as a visible label across the picture, but it provides a technical signal that compatible systems can detect.
Image generation may have daily usage limits because it depends on powerful server models. Apple says most iCloud+ plans will provide increased access, although the exact allowance can depend on the subscription and service conditions available at release.
The complete Image Playground experience may not appear in every early iOS 27 beta. Apple can activate or adjust cloud-dependent features separately while testing the operating system.
Apple Notes adopts the wider iOS 27 design changes through refreshed buttons, menus and Liquid Glass interface elements.
The Notes app icon has been updated while preserving the familiar yellow-and-white paper design. The change keeps the app recognizable without leaving it visually disconnected from the new system icons.
Shared material receives clearer identification. The shared folder displayed near the top of the folder list now uses a blue icon, separating it more clearly from private iCloud and local folders.
Individual shared notes and folders also receive revised symbols. These indicators make it easier to understand whether another person can view or edit material before sensitive information is added.
The distinction matters because a collaborative note can look almost identical to a private document once it is open. A clearer icon in the list provides another warning before the user begins editing.
Liquid Glass changes the appearance of controls and menus but does not alter the underlying note format. Existing notes, scans, checklists, attachments and locked documents remain available after the update.
People using accessibility settings such as increased contrast, reduced transparency or reduced motion may see a modified version of the visual design.
The iOS 27 changes build on a much larger set of capabilities introduced during earlier releases.
Apple Notes can already scan documents, attach PDFs, record audio, generate transcripts, solve mathematical expressions, create checklists and collaborate through iCloud.
Users can link one note to another, place notes inside folders, apply tags and create Smart Folders that collect documents according to selected rules.
Locked notes can use a device passcode or separate password, while Face ID provides faster access on supported iPhones.
The app can also identify handwriting, search text inside scanned documents and recognize information contained in images.
None of those existing tools has been replaced by Siri AI or Markdown. The new version concentrates on moving through large notes and transferring information with less manual work.
Section links are particularly effective when combined with audio transcripts or lengthy scanned research. A user can create headings above important material and place navigation links at the beginning of the document.
Markdown support addresses another gap by making Notes more practical as the first stage of a publishing workflow rather than a closed destination for text.
Apple Notes in iOS 27 is available for testing through prerelease software, with the complete operating-system release scheduled for the fall.
Beta software can contain synchronization failures, interface problems and changes that Apple has not finished documenting. Notes is especially sensitive because its content may synchronize immediately across every device using the same Apple Account.
Before installing a beta on a primary iPhone, users should confirm that important notes have finished syncing and that another copy of critical information exists outside the beta device.
A Mac, iPad or iCloud.com can be used to verify whether recent documents have reached the server. Exporting especially important notes as PDFs or Markdown files provides an additional copy that does not depend on the Notes database.
Collaborative notes require more caution. People using older operating systems may not see every new formatting element exactly as intended, particularly while iOS 27 remains in testing.
The public release should resolve many compatibility issues, but Siri AI will follow its own beta schedule and regional limitations. Installing iOS 27 alone will not guarantee that every intelligent Notes action appears immediately.
Apple can still modify menu locations, visual elements and cloud-based features before the final version arrives. The safest time to reorganize a large Notes library around section links will be after every frequently used Apple device can install its corresponding fall update.
The post Apple Notes iOS 27 Upgrade Adds Smarter Organization appeared first on AppleMagazine.
Whether driven by a desire for transparency or to keep OpenAI from hogging the spotlight when it comes to advertising advanced AI models, Anthropic revealed that Claude also hacked into three production systems belonging to unsuspecting targets during cybersecurity capabilities testing. Two of the affected companies didn't know they had been hacked, while a third one is unreachable.
The alleged incidents reportedly happened during the previous quarter and involved several versions of Claude: Opus 4.7, Mythos 5, and "an internal research test model." Similar to what happened when OpenAI Sol hacked into Hugging Face, Anthropic was running Claude through cybersecurity capture-the-flag scenarios where the bot was told to find a piece of information somewhere in its network. Anthropic says there were 141,006 test runs, and the three incidents occurred over six problematic runs. As expected, the tests ran with most AI safeguards disabled.
The problem is, the network was not isolated, a newbie mistake that some might even find suspicious. This was due to an alleged "miscommunication" between Anthropic’s virtual test lab firm Irregular, which gave the bots full access to the internet. For safety's sake, the Claudes were apparently told they were in an isolated environment, which didn't help much.
The first time, Claude Opus 4.7 had to find a piece of data belonging to an imaginary company. As it turns out, there was actually a real company whose website domain matched the fake one. With internet access, it gained application and infrastructure credentials and grabbed "several hundred rows" of data from a production database.
For two out of four runs where hacking took place, Claude didn't see any problem and "rationalized that the real company must be part of the exercise," but soldiered on. The other two times, it just kept going. Anthropic notes this was the only case of the three where the bot continued to operate after the realization the company actually existed for the aforementioned reason. However, the firm's interpretation is open to discussion.
The second, and potentially more interesting, was a bona fide supply-chain attack, this time perpetrated by Mythos. The fake data available to the bot included an installation script for a hypothetical developer that included installing Python packages. Much like a human would, Claude noticed the obvious attack vector and, once again, due to having internet access, inferred that publishing a booby-trapped package in the PyPI repository would do the trick. Unfortunately, it was the real one that the whole world uses.
In speedrun form for brevity: Claude needed a PyPI account, meaning it needed an email and a phone number. It tried buying one but found itself short of both pockets and funds; but it found another free email provider without that requirement, registered the PyPI account, and uploaded a malware package to PyPI. The package was seemingly only up for one hour but was downloaded and run on 15 systems.
In a twist of irony, one such system belonged to a security vendor that scans PyPI for malware, and lo and behold, promptly failed to find Claude's booby-trap and ran it. Once Claude presumably had remote code execution privileges, it used the credentials it found for further infiltration. The amusing bit is that Claude had no idea this company existed and didn't target it; the downloads just happened because the package was up and live for a short while.
According to Anthropic, the bot did detect it was acting on the real internet and even said that publishing a package like this was "NOT okay." However, it talked itself into believing it was in a test environment as it didn't recognize the real SSL certificates for the connections. It even believed the 2026 calendar date on the systems "proved" the environment was staged. It even recognized the systems that installed the malware as part of the experiment.
As for the third incident, Anthropic isn't saying much, other than Claude scanned 9,000 real live potential alternative targets once it noticed the intended one wasn't reachable. One of them reportedly had a live page with debugging information and was vulnerable to plain ol' SQL injection. Interestingly, this time around, once Claude noticed that the servers it was accessing resided on a cloud environment and not on the local network, it stopped the attack.
For its part, Anthropic recognizes that despite Claude following the instructions for the objectives, the fact that it stopped by itself once it found the target was real in only one of the cases is food for some thought. The firm says it's talking to METR for a third-party review, and it needs to "better co-design evaluation environments."
Oddly enough, Anthropic believes Claude probably wouldn't have gone online "if the prompt had clearly explained which systems were in and out of scope for the evaluation," while also stating the incidents were "closer to a harness and operational failure than a model alignment failure."
An AI-assisted counterexample disproves the Jacobian conjecture above two dimensions while leaving its original two-dimensional form open. A formula short enough to fit inside a single post on X has sent a surge of excitement through the mathematical community, just as millions of people were settling down after the FIFA World Cup Final. Levent Alpöge, [...]If you’re among the 2.6 million active users of the popular telehealth apps Hims and Hers to obtain medication for erectile dysfunction, hair loss, and a cavalcade of other intimate issues, there’s probably a very short list of people you want to know about it — and Mark Zuckerberg probably isn’t one of them.
On Wednesday, the US Federal Trade Commission filed a major lawsuit against Hims and Hers, over allegations the firm was sharing user’s health information with third party companies, Gizmodo reported. Chief among those third-parties apps was Zuckerberg’s Meta, as well as Snap, two social media companies with horrible track records on protecting sensitive information.
Though Hims and Hers advertises its app-based health services as “100 percent online, private, and secure,” the FTC suit alleges “those assurances were false or misleading.”
“Instead of honoring its privacy promises,” the filing says, “Hims opted to grow the company and increase its revenue streams… with advertising platforms such as Meta and Snap.”
The suit doesn’t blame Meta directly, but argues that Hims and Hers used two “automated tracking technologies offered by Meta” to surveil users’ actions on the telehealth app. “These two tools automatically tracked and disclosed certain Events [user actions] to Meta,” the lawsuit reads.
Hims sells all kinds of men’s health meds, but most people know it for one thing: erectile dysfunction pills. While any leak of patient data would be an outrage, the thought that Hims could be sharing this particular kind of medical information is especially horrifying.
In a statement, Hims and Hers argued that the FTC’s case “is not enforcement grounded in consumer protection,” but rather “an effort to generate headlines at our expense.”
The lawsuit comes on the back of a three-year investigation by the FTC, and is currently in the very early stages — meaning it could be years before a final verdict is reached.
More on Meta: Meta’s Attempted Crackdown on AI Glasses Perverts Is Already Failing
The post FTC Concerned About Meta Surveillance of Users of Erectile Dysfunction Apps appeared first on Futurism.
Suddenly, things ain’t looking great for the AI industry. Or at least for the people pouring loads of money into it.
Situational Awareness, a much-hyped AI hedge-fund whose 24-year-old founder Leopold Aschenbrenner preached endless AI gains, basically imploded over this month amid a mass tech sell-off.
The Wall Street Journal reported that it’s now down 67 percent in July, forcing Aschenbrenner to desperately seek clients to buy its shares on the cheap to cover its losses.
“We let you down this month,” the fund wrote in a letter to clients, per the WSJ.
Aschenbrenner, a former OpenAI researcher who had zero investment experience before launching the fund, has been hailed as the “Nostradamus of AI.” He apparently didn’t predict his rapid reversal of fortunes, though.
The Financial Times reported that his fund returned 439 percent for the year through June, and was up 1,551 percent since its founding.
Then came July. According to CNBC, the fund assets reached a peak of $45 billion, before beginning a precipitous decline this month. Pretty much everything went against its way. Its commitments to AI infrastructure tanked when the likes of the semiconductor manufacturer SK Hynix dropped off a cliff. And it shorted software companies — which investors once feared would be outmoded by AI labs — like Adobe, which instead continued to steadily climb.
Other companies it invested in, including Sandisk, Micron, and CoreWave, are all down more than 35 percent this month, CNBC noted.
As the dominoes fell, Aschenbrenner sent a letter to investors to quell their fears, highlighting how well it had performed until this point. Boldly, he also said that the mass tech sell off was a “particularly good time to add funds,” the FT reported.
But when the assets kept tanking — and more money didn’t come in — Aschenbrenner started frantically calling investors trying to get them to sell assets or pour in fresh capital, according to the reporting. When it got even worse, his clients couldn’t reach him.
“Leopold just stopped taking calls,” one investor told the FT.
With nowhere to turn to, Situational Awareness sold the majority of its public assets to the investment firm Citadel, though it hung onto its stake in Anthropic. The FT reported that Citadel’s vulturous purchase of Situational Awareness’s shares “marks one of the largest and most sudden stock transactions in Wall Street history.”
Situational Awareness, despite commanding billions in assets, operated with a skeleton staff of only four investment professionals and eight employees overall, the FT found in a regulatory filing. The fund used borrowed money to buy shares, which helped boost its returns, but made its losses far worse when those assets declined.
AI stocks will likely weather this storm. The real warning sign is that investors are so blinded by AI that they’re entrusting their billions to someone who had no idea what they were doing.
More on AI: Microsoft CEO Warns That Companies Embracing AI Could Drive Themselves Out of Business
The post A Prominent AI Investor Is Now Crumbling, in What Could Be a Sign of Things to Come appeared first on Futurism.
Login