If you were hoping Apple’s recent price increases were a one-off, there’s another reason to watch what happens in 2027. Two of the most important components inside your iPhone, iPad, and Mac are about to cost Apple a lot more.

A report from DigiTimes, citing Chinese publication Jiemian News, claims Apple has reached an agreement with Samsung for its memory supply beginning in early 2027. Under the reported terms, Apple would pay close to $2 per gigabit for DRAM and $0.33 per gigabit for NAND flash. Put more simply, the memory that gives your devices their RAM and storage could cost Apple around 30% to 40% more than it did in the third quarter of this year.

The timing isn’t great for Apple buyers

A jump in component costs doesn’t automatically mean your next iPhone will cost 40% more. Memory is only one part of the total cost of building a device, and Apple could choose to absorb some of the increase. The concern is that Apple has already shown that it is willing to pass at least some of the effects of the memory shortage on to customers.

Prices went up across Macs and iPads earlier this year, with increases also reaching the Apple TV, HomePod lineup, and Vision Pro. Then came the iPhone 18 series, where the iPhone 18 Pro‘s US starting price climbed by $100 compared with the previous generation. Some older iPhones also became more expensive. Whether another round follows in 2027 is still unknown. Apple could have already accounted for future memory costs when setting its latest prices.

AI data centers are squeezing everyone else

The bigger problem is where all this demand is coming from. The rapid expansion of AI infrastructure has companies building huge data centers that require enormous amounts of memory. That has put pressure on the supply available for everything else, including phones and computers.

Apple has already acknowledged how unusual the situation has become. Former CEO Tim Cook previously compared the disruption in the memory market to a “100-year flood” while discussing the company’s price increases. The reported Samsung deal now suggests this isn’t a short-term problem. Apple may be securing the memory it needs for next year’s products, but securing it cheaply appears to be another matter entirely.

New York, NY, September 17th, 2026, TechnologyWire

As bipartisan legislation demands verifiable agent controls, the Advanced AI Society releases the Proof-of-Control v1.0 draft for public comment and hosts “Who’s Watching the Machines?” event with Linux Foundation Decentralized Trust

As bipartisan Congressional proposals demand continuous agent verification and third-party evaluation, and as frontier lab leaders call to pace AI capability growth, Advanced AI Society, the alliance making trust in AI possible, today announced its membership in the Linux Foundation and LF Decentralized Trust and the launch of an open verification ecosystem!

To answer growing statutory mandates and security requirements across enterprise and public systems, the alliance has released the v1.0 working draft of Proof-of-Control for public comment through October 30, 2026 at advancedaisociety.org/proof-of-control. Hosted by LF Decentralized Trust as a lab, this new standard builds directly on the production-proven Verifiable AI technologies of the alliance’s member companies. Co-designed alongside a coalition of over 80 global security leaders across enterprise sectors, it validates market demand for interoperable tools that make agent execution openly verifiable by anyone, without relying on the operator’s word.

The announcement addresses the widening Verifiability Gap as autonomous AI agents begin to act faster than manual human oversight can track. Grounded in the core principle that verification cannot belong to the verified, unaffiliated, neutral third parties, as well as agent developers and deployers, must be able to openly verify the work of AI agents.

Open verification represents the next category of open source: extending open-source principles to the agentic stack where you cannot secure what you cannot inspect, and you cannot delegate what you cannot openly verify. By housing its open standards within the neutral commons of the Linux Foundation, the alliance ensures that machine trust rests on transparent mechanisms anyone can openly verify, rather than vendor promises or self-attestation.

“Built by our members and over 80 global security leaders, Proof-of-Control brings production-proven verifiable AI technologies together into a unified open standard,” says Tricia Wang, Co-Founder and CEO of Advanced AI Society. “Safety checks and traditional audits are necessary, but when AI agents act in milliseconds, an annual or even quarterly stamp of approval isn’t enough. Trust requires continuous monitoring anyone can inspect. Launching at LF Decentralized Trust prevents a single company from owning the referee.”

“An agent’s runtime log is just a claim until anyone can verify it for themselves,” says Ken Huang, Co-Chair of Proof-of-Control. “Proof-of-Control turns open verification into standard operating practice: it requires an agent’s authority to be defined before it acts, produces evidence at runtime that it stayed inside that authority, and makes that evidence auditable across the whole agent lifecycle.”

“Trusting AI agents requires decentralized mechanisms that no single vendor can control or alter, making LF Decentralized Trust the natural home for this moment’s urgency,” says Daniela Barbossa, General Manager, Decentralized Technologies, Linux Foundation, and Executive Director, LF Decentralized Trust.

“In cryptography, the rule is: don’t trust, verify. Autonomous AI shouldn’t be an exception,” says Dr. Hart Montgomery, CTO of LF Decentralized Trust. “When non-deterministic machines make decisions in milliseconds, human inspection fails. Only automated tools can keep pace. As an LF Decentralized Trust lab, Proof-of-Control brings the evidentiary rigor established in projects such as AnonCreds and Trust Over IP to the agentic stack, giving developers the building blocks to openly verify machine authority without relying on proprietary black boxes.”

Public Launch Briefing

Advanced AI Society and LF Decentralized Trust will co-host a public launch event on Wednesday, September 23, 2026, “Who’s Watching the Machines? Launching Open Verification, the Proof-of-Control Standard for the Agentic Era,” featuring alliance leaders, builders, deployers, and policy experts. Register at advancedaisociety.org/announcements/webinar1.

Industry & Expert Support

“We’re running on a broken trust model, vendor assertions instead of evidence. That doesn’t work when patient safety and our intellectual property are on the line. We need open verification: a transparent, inspectable way to know what these agents actually did. That’s the gap Proof-of-Control closes, and it’s why I’ve joined this effort at Advanced AI Society.”
Charles Iheagwara, Global Head, AI & Cybersecurity, AstraZeneca; Distinguished Review Board of Proof-of-Control, Advanced AI Society

“As AI agents become more powerful and ubiquitous, knowing who they belong to, and what they are authorized to do, will be critical. Proof-of-Control is an important part of that.”
Bruce Schneier, Cryptographer, Technical Advisor & Distinguished Review Board on Proof-of-Control, Advanced AI Society

“Financial markets and regulatory frameworks rely upon trust that is firmly grounded in verifiable, audit-ready evidence. Bringing that same legal and structural discipline to autonomous AI agents in the open gives the agentic economy a foundation of trust that institutions can actually rely on. Proof-of-Control provides the machine-verifiable proof necessary as AI systems assume real economic authority, especially in digital network financial markets.”
J. Christopher Giancarlo, former Chairman, U.S. Commodity Futures Trading Commission (CFTC); Senior Advisor, Advanced AI Society

“As agents become more capable, more numerous, and faster, ‘human in the loop’ doesn’t cut it anymore, too generic, too vaguely specified, and vendor self-regulation is too compromised. Agents need to be logged and the logs need to be open for claims about their behavior to be validated.”
Clay Shirky, Vice Provost for AI and Technology in Education at New York University

“Midnight was built on the premise that you shouldn’t have to take anyone’s word for it. Proof-of-Control brings that same standard to AI agents, and we’re glad to support it.”
Bob Blessing-Hartley, CTO, Shielded Technologies

“At Oko Labs, we are building blockchain into the fabric of the internet, which lets verification live in the network itself. I spent years at Cisco making security provable enough to underwrite first through a cyber assurance captive and then the blockchain practice and Proof-of-Control does that exact job for agents. Prove control at the network layer, and machine risk becomes something the market can measure, price, and insure at agentic velocity.”
Anoop Nannra, co-founder and CEO, Oko Labs

“As AI agents begin making decisions that affect people’s healthcare and money, ‘trust us’ isn’t a sufficient security model. We need open, continuous ways to prove what an agent was authorized to do and what it did: which healthcare decisions were made; how data was collected and used, and what was billed, when, and why; which is why Emme is participating in Proof-of-Control.”
Erynn Petersen, CEO, Emme

“When an AI agent can move money, institutions need more than a persuasive explanation: they need defined authority and evidence they can check. Solv supports the development of Proof-of-Control as an open standard that helps separate what an agent was permitted to do from what the evidence shows it did.”
Patrick Duffy, Founder and CEO, Solv Labs

“When AI agents operate inside critical infrastructure, open verification moves from a commercial preference to a national security imperative. No security team or auditor should have to take an AI agent’s word for its own actions after the fact. Proof-of-Control delivers the open, verifiable evidence required to keep critical national systems safe, transparent, and accountable.”
Noah Ringler, former AI Policy Lead, U.S. Department of Homeland Security; Senior Advisor, Advanced AI Society

“Deploying autonomous agents without open verification is like flying an airplane with no brakes and no flight recorder. When agents execute thousands of actions a second, unverified errors don’t just stay local, they compound exponentially into systemic crises. We established the Open Verification Lab at Advanced AI Society to give developers and security teams the open, tamper-evident tools, standards, and research needed to keep autonomous execution safe and within human control.”
Jim Schwoebel, Director of Research, Advanced AI Society, CEO of Quome

“To safely delegate real authority to machines, we need decentralized cryptographic infrastructure that proves an agent stayed within its boundaries without forcing individuals, governments, or enterprises to expose private, sensitive data. Unchecked machine execution presents an existential civilizational risk, but with open verification, it becomes our greatest opportunity to delegate humanity’s biggest challenges safely. Housing Proof-of-Control at LF Decentralized Trust gives society the open, tamper-evident foundation required to make trust in AI possible.”
Michael Casey, Co-founder, Advanced AI Society

“I’ve spent the past decade investing in digital asset primitives because I believed open, decentralized protocols could give us new ways to establish and exchange value. Today, I see those same foundations taking on a broader role. As autonomous agents begin to act across our businesses and daily lives, we need a shared, verifiable reality: evidence of what they control, what happened, and which actions can be attributed to them. Proof-of-Control extends the work of decentralized infrastructure into the age of machine agency.”
Bettina Warburg, Co-Founder, Advanced AI Society

“Public trust cannot be built on closed-door audits or hand-picked stamps of approval. When autonomous agents make real-world decisions that affect people’s lives, communities and public institutions shouldn’t have to take a hired auditor’s word for what happened. Open verification ensures machine accountability remains a public good rather than a walled garden.”
Sheila Warren, Co-Founder, Advanced AI Society

About Advanced AI Society

Advanced AI Society is the alliance making trust in AI possible. Advanced AI Society builds the open verification ecosystem for AI security, uniting market leaders and deployers with insurers, policymakers, and civil society to standardize, deploy, and scale verifiable AI. advancedaisociety.org/

Contact

Tricia Wang
media@advancedaisociety.org

Google and Google DeepMind researchers launched the DeepMind Institute on Wednesday to advance the conversation around artificial general intelligence (AGI). The institute lists DeepMind co-founder Shane Legg, Google executive James Manyika, and Google DeepMind chair Demis Hassabis as directors, with Legg serving as managing editor. 

The new institute aims to surface differing views between Google, Google DeepMind, and the broader global research community around AGI. “They will not always agree, and they will likely change their minds, as more data and information comes to light at the fast-moving frontier,” the announcement read.

The inaugural collection of four essays covers a range of topics: economic policies for managing potential AGI disruption, preserving human-readable model reasoning, principles for human flourishing, and a framework for evaluating frontier AI models. 

One essay, by DeepMind safety researchers Rohin Shah and Anca Dragan, argues that AI’s shrinking window of transparency — the ability to see and check a model’s step-by-step reasoning — is not inevitable. As new architectures make the most powerful models harder to monitor, the authors say developers and regulators should confront the safety trade-offs directly. That could mean limiting “opaque serial depth”— the amount of sequential computation a model can perform without producing a readable reasoning trace — or requiring developers to demonstrate that less transparent systems remain just as monitorable.

In another essay, Hassabis proposes a U.S.-led frontier AI standards body to evaluate the most advanced AI models. Under his framework, developers would initially submit models voluntarily for review up to 30 days before release. Once the evaluation system has proved effective, passing its tests could become a requirement for deploying frontier models in the United States.

The body would at first design assessments in consultation with AI companies but would eventually develop independent, undisclosed evaluations — what the essay calls “held-out” tests — to prevent labs from tailoring their models to known evaluations. Hassabis said the framework could be “ratcheted up if the seriousness of the situation demands,” potentially including a coordinated slowdown among frontier AI developers. 

The essays arrive as the industry’s safety debate shifts from broad statements of concern toward concrete proposals for disclosure, outside scrutiny, and, if safeguards fall behind, coordinated slowdowns. That shift accelerated this week as industry leaders endorsed elements of Anthropic CEO Dario Amodei’s call to “pace” frontier AI development.

The hackers have reportedly given Revolut 24 hours to pay the ransom, or else they will sell the stolen data to other criminals.

A group of hackers claiming to be responsible for a recent Revolut data breach has reportedly threatened to sell the confidential records of hundreds of customers to other criminal groups unless the fintech company pays them a ransom of $3m in cryptocurrency Monero. 

According to the Financial Times, the group – which called itself ‘iamnotavillain’ – published the demand on its website yesterday (16 September) alongside a digital countdown clock, giving the British online bank 24 hours to comply. 

Last Saturday (12 September), Revolut admitted that it disclosed sensitive customer data to an unauthorised fraudulent party after receiving requests for information from an email with a government agency domain. The leaked data included information such as full names, dates of birth, occupations, addresses, contact details, passports, drivers’ licences, financial statements and facial verification images.

It is believed around 680 people, including 12 in Ireland, have had their personal information compromised. 

The Financial Times contacted the hacking group through Telegram and were then sent a 60-second screen-capture video showing an unseen user moving through the cache of purported Revolut documents.

Reportedly the online message told Revolut to pay the $3m, “otherwise all the data will be sold and the blood will be on your hands”. The group also told the Financial Times that the demand notification was the first communication attempt and that, as of yet, there had been no negotiations with Revolut.

In an official statement to SiliconRepublic.com, a Revolut spokesperson said: “Revolut has not received any direct contact or demand from the individuals or group making these claims.”

The reported ransom comes after Revolut recently announced that it had applied for a Swiss banking licence as part of its continued global expansion campaign. The company plans to invest more than 150m Swiss francs into Switzerland over the next five years, as a means of creating jobs and supporting product development. 

In August, the company received a full banking licence in France, giving the fintech giant a second European Union hub. The company claims to have more than 75m customers worldwide.

Updated, 5.00pm, 17 September 2026: This article was amended to clarify accurate phrasing in relation to the data breach.

Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.

Elon Musk filed to have Apple dismissed from the lawsuit he brought against it and OpenAI, and OpenAI demanded answers. The judge has denied OpenAI’s request to reveal settlement terms.

When Elon Musk’s SpaceXAI sought to dismiss Apple from the ongoing lawsuit brought against it and OpenAI, it wasn’t clear exactly what had transpired. In OpenAI’s efforts to find out, it has been revealed that there was some kind of settlement reached between the companies.

OpenAI asked that the settlement terms be revealed, but after Judge Pittman reviewed the settlement, he denied the request. The decision was first shared by 9to5Mac, which said that if the settlement terms were relevant to the ongoing case with OpenAI, they could be shared, but they apparently are not.

Here is what he said in the filing:

The Court hesitates to compel the production of confidential settlement agreements entered by parties. Confidentiality in settlement serves the “[s]trong federal policy” of “induc[ing] the parties to settle a case.”

In other words, if Apple and SpaceXAI want to settle out of court, it is not only their business to do so, but to do so confidentially. To make such agreements public could jeopardize such agreements from ever taking place.

An odd lawsuit either way

The origins of xAI’s, now SpaceXAI’s, lawsuit against Apple can be tied to a tirade by owner Elon Musk. He made claims that the App Store was favoring OpenAI’s ChatGPT over Grok and Apple was biased against X.

He claimed that Apple was manipulating App Store charts to keep Grok and X from ranking higher, keeping ChatGPT at the top. The problem is, all of these claims were easily proven incorrect.

By the time the lawsuit came about, the accusations were shrunk to much more manageable proportions. Musk focused on Grok and accused Apple and OpenAI of anticompetitive collusion.

Whatever agreement or settlement Apple and Musk came to, it is unlikely that OpenAI can escape as easily. Elon Musk is known to have a very strong rivalry with his former business partner Sam Altman, the current CEO of OpenAI.

It will be interesting to see exactly what the lawsuit reveals, especially as Grok and X continue to face problems surrounding non-consensual deepfake pornography and CSAM distribution on the platforms. However, it is no longer Apple’s problem as it steps away to focus on its own lawsuit with OpenAI.

In case you’re into that kind of thing.

FiiO has officially launched the FT15, its new $499 open-back planar magnetic headphone and successor to the FT5. The headline changes include a substantially thinner diaphragm, larger driver, lower weight and enough manufacturing terminology to make a semiconductor engineer feel strangely at home in a headphone article.

The interesting part for us is that we have already heard them.

eCoustics contributor James Fiorucci spent time with the FT15 at CanJam London 2026, when FiiO was still calling it a forthcoming model and final specifications had not been released. Now we know exactly what FiiO was hiding underneath those walnut ear cups.

Related Reading:

What Changed From the FiiO FT5?

The FT15 uses a 101mm planar magnetic driver, up from 90mm in the FT5, with an effective radiating area of 6,443.5mm². FiiO says that represents a 35% increase in active diaphragm area and is intended to improve bass authority and soundstage scale. 

The bigger engineering story is the diaphragm itself. FiiO has reduced its thickness from 6 microns in the FT5 to just 800 nanometers, an 86.7% reduction, before applying a full-surface diamond coating designed to increase rigidity and suppress high-frequency breakup. FiiO claims extension to 40kHz, which should keep both the Hi-Res Audio certification people and neighborhood bats satisfied. 

A patented dual-sided magnet array works with a 28nm electroplated voice coil, while 95dB/mW sensitivity and 24-ohm impedance are intended to make the FT15 manageable with portable players and competent dongle DACs rather than requiring a power amplifier normally reserved for starting farm equipment. 

FiiO has also trimmed the finished weight to 425 grams and uses a carbon-fiber headband with North American FAS-grade black walnut around the ear cups. 

We Heard the FT15 at CanJam London

Specifications are useful, but we had the advantage of hearing the FT15 before FiiO finished filling in the spreadsheet.

James found the prototype largely balanced, open and easy to listen to, with treble that remained airy without becoming immediately aggressive. His reservation concerned macrodynamics: the FT15 did not deliver the same physical punch or dramatic contrast as some of the stronger planar designs he heard at the show. 

That was a show-floor audition rather than a controlled review, so consider it an early impression rather than the final word. It does, however, make FiiO’s claims of a warmer, more relaxed presentation somewhat more interesting than the usual launch-day adjectives written by people who have not actually put the headphones on their heads.

FiiO FT15 Specifications

• Type: Open-back planar magnetic
• Driver: 101mm planar
• Diaphragm: 800nm with full-surface diamond coating
• Effective radiating area: 6,443.5mm²
• Impedance: 24 ohms
• Sensitivity: 95dB/mW
• Claimed high-frequency extension: 40kHz
• Weight: 425g
• Headband: Carbon fiber
• Ear cups: North American FAS-grade black walnut
• Cable: 6N monocrystalline copper, cryogenically treated
• Connection: 4.4mm balanced
• Adapters: 6.35mm and 4-pin XLR
• Earpads: Suede and protein leather
• Case: Yaoli linen storage case
• Price: $499
• Availability: Shipping internationally in stages through FiiO and authorized dealers

The Competition

The Audeze MM-100 ($399) features 90mm planar drivers with a more reference-oriented presentation aimed partly at studio users. It is also built like something designed to survive an argument with a road case. 

Then there is the HiFiMAN Edition XV ($399), one of our Editors’ Choice models, which offers a warmer and smoother take on the traditional HiFiMAN sound with a 452g chassis and unusually easy-going treble for the brand. 

That leaves the FT15 entering an extremely competitive part of the planar market at $499. FiiO is betting that larger drivers, lower weight, easier drivability and a less aggressive tonal balance will justify the premium.

The Bottom Line

The FiiO FT15 looks like considerably more than an FT5 with nicer wood attached. FiiO has reworked the diaphragm, enlarged the driver and reduced the weight while keeping the headphone friendly to portable amplification.

Our brief listen in London suggested FiiO has also resisted the temptation to turn all of that engineering into an overly bright detail microscope. Whether the final production version improves on the restrained macrodynamic punch we noticed at the show will require a proper review.

At $499, it certainly has competition. It also no longer has the luxury of being judged as FiiO’s first serious planar experiment.

Where to Buy

Related Reading: