For years, Microsoft was OpenAI’s best friend (and biggest backer). But expensive friendships eventually weigh on you — so the company decided to build alternatives at home. At its Build conference, it unveiled seven of its own models, the MAI family, with a not-so-subtle goal: lean less on OpenAI and trim the bill.

The headliner is MAI-Thinking-1, the company’s first reasoning model, with 35 billion active parameters. There are also models for coding, image generation, transcription and even voice. The pitch, though, is money: according to Microsoft’s own tests, these models can cost up to ten times less than the alternatives.

Why it matters

When the company runs models on its own cloud, Azure, it stops paying royalties to outsiders — and some of those savings can reach developers. For anyone building apps and services, that can mean lighter bills at month’s end.

For the rest of us, the signal is more interesting than the technical names: the AI market is moving away from a single supplier setting all the prices. And when real competition shows up, the one who usually wins is whoever’s footing the bill.

There is now a date to meet Google’s next generation of phones: the Made by Google event is set for August 12 in New York, with a twist in the format — instead of the usual midday keynote, the show moves to 6 PM local time, in prime time.

What will Google announce on August 12?

The full family: Pixel 11, Pixel 11 Pro, Pixel 11 Pro XL and Pixel 11 Pro Fold, plus the Pixel Watch 5. The official teaser shows a shiny gold metal frame and the familiar pill-shaped camera bar, with a promise to reveal “the next generation of Pixel”. Pre-orders are expected to open the same day, with shipping a week or two later — the official schedule will be on the Google Store.

How much will the Pixel 11 cost?

That is the big unknown — and the omens are not great. The specialist press anticipates price rises across the range, driven by the RAM crisis that is pushing up smartphone production costs across the industry. Nothing confirmed by Google for now; the answer arrives on August 12.

The timing tells a story too: the event lands a week earlier than last year, at a moment when the competition is wide awake — Samsung has its foldables Unpacked booked for July 22, and Apple is prepping its usual September, with a Gemini-powered Siri on the radar. August is shaping up to be an interesting month to change phones.

The Apple-OpenAI divorce has reached the courts. The Cupertino company has sued the maker of ChatGPT in federal court in Northern California, accusing it of misappropriating trade secrets to fast-track its push into consumer hardware — the bet with which OpenAI hopes to challenge the iPhone itself.

What is Apple accusing OpenAI of?

A scheme running “at every level”, in the complaint’s words: from engineers up to hardware chief Tang Tan — himself a former Apple vice president. The suit alleges that more than 400 former Apple employees now work at OpenAI, and that Tan directed candidates coming from Apple to share confidential information during job interviews. OpenAI has yet to file its defence.

What changed between the two companies?

Almost everything. In 2024 they were partners, with ChatGPT integrated into Siri and Apple Intelligence. The relationship then soured: Apple deepened its ties with Google — the new Siri, arriving this autumn, runs on Gemini models — while OpenAI assembled a hardware team stacked with Cupertino veterans. Letting users pick their iPhone assistant was a road Apple had already been paving, and OpenAI’s move into its own silicon, with the Jalapeño chip designed with Broadcom, only sharpened the rivalry. Official company news still lands in the Apple newsroom.

Whatever the outcome, the lawsuit marks an era: the two companies that shared a stage two years ago are now fighting in court over who controls the next device in your pocket.

A Portuguese graphene-based material may soon hide drones and aircraft from radar. The technology was born at GTechPlasma, a spin-off of the Institute for Plasmas and Nuclear Fusion at Instituto Superior Técnico, and will be tested by the Portuguese Air Force — a rare leap for national research in the middle of Europe’s technological rearmament race.

How does radar-absorbing graphene work?

Instead of reflecting radar waves the way metallic surfaces do, the material absorbs them, drastically reducing the electromagnetic signature of whatever it coats. GTechPlasma produces high-quality graphene at around 40 milligrams per minute and wants to turn it into ready-to-use products — paints and coatings that can be applied directly to drones, aircraft or other equipment. Roughly 260 grams have already been delivered to a Portuguese drone manufacturer for radar-absorption tests.

Who else is developing stealth technology in Portugal?

Anti-detection graphene is not a lone bet. The ASTRAL project, led by Controlar with Graphenest, will create an ultra-thin flexible film able to absorb electromagnetic radiation between 2 and 110 GHz, backed by 1.86 million euros from Portugal 2030 funds, running to May 2029 with at least one international patent planned. Between military stealth and civilian uses — electromagnetic shielding for sensitive equipment, for instance — the potential market is huge.

It is another sign that Portuguese science is building things that fly: the same week brought confirmation that LusoSpace’s Lusíada constellation now has eight satellites in orbit, sketching an aerospace and defence ecosystem gaining altitude. From the plasma lab to the sky — without showing up on radar.

A 37-year-old woman goes missing in downtown Coimbra. Days later, her family receives a photo of her, apparently tied up, along with a ransom demand. Except the kidnapping never happened: the image had been manipulated with artificial intelligence, and the woman — from Miranda do Corvo — had left of her own free will.

The case, revealed this week by Portugal’s Polícia Judiciária, began on 1 July when relatives reported her disappearance. While the family lived through days of anguish and the story gained media attention, fraudsters exploited precisely that exposure to stage an extortion attempt, lending the scheme credibility with the AI-generated photo. The woman was found safe on 8 July, in an operation coordinated with the PSP and GNR police forces, and investigators concluded no crime was involved in the disappearance itself.

How does this kind of AI scam work?

The recipe is simple, which is exactly what makes it dangerous: scammers take publicly available photos — from social media or the news coverage itself — and use image-manipulation tools to fabricate “proof” of an abduction, betting on a family’s panic to force a quick payment. It is the same pattern as the cloned-voice scams that have been circulating abroad, now arriving on Portuguese doorsteps.

The PJ is still working to identify the perpetrators, who face fraud charges, and updates go through the official channels of the Polícia Judiciária. The backdrop is familiar: the European Union will soon require AI-generated content to identify itself as such, but no watermark stops someone who has already decided to use the tool for crime.

The practical rule is old-fashioned and still works: faced with a ransom demand or a distressing “proof”, stop, verify through another channel and call the authorities — before transferring anything at all.

Three democracies on three continents have decided to team up in the technology race. India, Australia and Canada this week launched ACITI — the Australia-Canada-India Technology and Innovation Partnership — a trilateral framework covering artificial intelligence, green energy technologies, critical minerals and resilient supply chains.

What is the ACITI partnership?

A framework to align investment, research and rules across the three countries in sectors where none wants to depend on a single supplier — read: China. Australia and Canada hold the critical minerals the energy transition and chipmaking need; India brings scale, engineers and a tech ecosystem in full boil. Together they are trying to close the loop between raw material, manufacturing and talent.

Why does it matter for Portugal and Europe?

Because it redraws the map of technology alliances beyond the US-China axis — leaving Europe, running its own race, with one more bloc to compete and negotiate with. Not coincidentally, Portugal has just joined the European Tech Champions fund, while Brussels tightens regulation, as seen in its standoff with Meta. Anyone sitting on lithium knows what critical minerals are worth — just ask the Alentejo coast.

The partnership’s official outlines were published by the governments involved, including the Australian government. In the global technology chess game, a three-headed player has just sat down at the board.

The platforms and apps entering Portuguese classrooms will now need a green light before they reach students. Decree-Law 138/2026, published this Friday in the official journal, makes prior certification mandatory for information systems intended for school and pedagogical use — across public, private and cooperative education below university level.

What changes with Decree-Law 138/2026?

Until now, adopting school software — from learning management platforms to pedagogical apps — lived in loosely regulated territory, with each school or cluster deciding for itself. Under the new decree, systems must be certified before they can be used with students, creating a common national filter for what enters schools’ digital ecosystem, from state schools to private and cooperative colleges.

Why does prior certification matter?

Because schools handle the most sensitive data there is: information about minors. Regulatory pressure on digital platforms keeps mounting across Europe — the same week Brussels accused Meta of breaching EU digital rules with «addictive» design, Portugal has decided to run the same magnifying glass over the software its students use daily. Privacy, security and pedagogical fitness stop being taken on faith.

The full text is available in the Diário da República, and the detailed regulation — who certifies, on what criteria, on what timeline — is the next chapter to watch via the Directorate-General for Education. For edtech companies selling into Portuguese schools, recess is over.

Infinite scroll could cost Meta dearly. The European Commission concluded this Friday, on a preliminary basis, that Instagram and Facebook violate the Digital Services Act (DSA) by being engineered to hook users — including minors and vulnerable adults — into compulsive use whose risks to physical and mental health the company never properly assessed.

What does Brussels consider ‘addictive design’?

The list is familiar to anyone with a phone: infinite scrolling, videos that autoplay, constant push notifications and hyper-personalised recommendation systems, all tuned to maximise screen time. For the Commission, the problem is not each feature in isolation but the whole — an architecture that pushes towards compulsive use without Meta having assessed and mitigated those risks, as the DSA requires of very large platforms.

What could Meta be forced to change?

The changes on the table are concrete: switching off features like autoplay and infinite scroll by default, implementing effective screen-time breaks, and making the recommendation system less driven by pure engagement. The finding is preliminary — Meta can now present its defence before a final decision — but if confirmed, the fine can reach 6% of the company’s annual global turnover. At Meta’s revenues, that bill would make history.

The case slots into Brussels’ wider offensive against big tech, which we have been following since June’s record fines in the digital sovereignty standoff. Still, this Friday’s message is new: for the first time, the target is not what the platforms show, but how they hold us inside. If the feed never ends, says Brussels, that is not chance — it is engineering. And engineering is now regulated.

Nearly 28 years of internet history, and Google had never been asked so many questions at once. The company has confirmed that Search hit its all-time record for queries per second — and the exact moment it happened tells you everything about this World Cup.

What sent Google searches through the roof?

Football, obviously. The historic spike came immediately after Argentina’s wild comeback against Egypt, settled by Enzo Fernández deep in stoppage time. Millions of people reached for their phones at the same instant to confirm what they had just watched — and that second became the busiest in the company’s history, according to Google search executives Nick Fox and Robby Stein.

What was the most searched query?

“Argentina vs Egypt”, unsurprisingly, followed closely by searches about the next match for Messi’s side and questions like “how many World Cup goals does Messi have”. Google didn’t publish hard numbers or methodology — it announced the record and kept the accounting to itself — but the message it wanted out there is clear: in an era when chatbots are eyeing the throne, the old search box is still the planet’s reflex when something extraordinary happens. The official highlights roll out on Google’s blog.

Not bad for a tournament that hasn’t even reached the semi-finals. If a last-16 game broke the counter, wait for 19 July.

Leiria doesn’t usually feature on venture capital maps — which is exactly why this story matters. Portugal Ventures, the venture arm of the Portuguese state development bank (Banco Português de Fomento), has invested more than €2.7 million in three tech companies from the region, under its Leiria Crescimento and Leiria Recuperar calls.

Which companies received the investment?

The biggest slice went to Brainr, which takes €1.5 million, followed by Sound Particles with €1 million — both under the Leiria Crescimento call. Nutrivalley received €250,000 under Leiria Recuperar. Sound Particles is the best-known of the trio: its 3D audio software is used in Hollywood productions, and the company has become one of Portugal’s best examples of export-grade technology built outside the two big urban hubs. The terms of the regional calls are detailed on Portugal Ventures’ website.

Why does investment outside Lisbon and Porto matter?

Because talent doesn’t pick an address, but capital usually does. Portugal’s ecosystem remains concentrated in Lisbon — visibility and investors — and Porto — engineering and product. When public growth capital reaches Leiria, Braga or Coimbra, the multiplier is bigger: it keeps engineers who would otherwise migrate, and proves to other founders that you don’t need to change postcode to scale. Not coincidentally, Porto just climbed the global startup rankings — the decentralisation of the ecosystem is, slowly, happening.

For the three companies, the challenge now is turning the cheque into traction: more hires, more foreign revenue and, ideally, a next round with private investors at the table. Leiria says thank you — and so does the ecosystem.

Porto climbed more than 15 places in the world ranking of emerging ecosystems, according to the Global Startup Ecosystem Report (GSER) 2026, and is singled out as one of Europe’s fastest-growing startup ecosystems. For a city seen only a few years ago mainly as a destination for tourism and wine, it is a jump that confirms a change of status.

What is driving Porto?

Four areas explain much of the leap: software, health, fintech and artificial intelligence. It is the mix of engineering teams cheaper than Lisbon’s, active incubators and a still-competitive cost of living that has pulled founders and talent to the banks of the Douro.

How does it compare with Lisbon and the rest of the country?

Portugal as a whole sits near 30th in the global index — a respectable tier for a country its size. Lisbon remains the shop window for investors and international visibility, but Porto has gained ground as a product and engineering base, while Braga stands out for incubator support and lower costs.

We also covered fast-tracked visas for AI talent and the outlook for Portugal’s economy. Official ecosystem data is on the AICEP Portugal Global portal.

Portuguese science’s biggest annual gathering is back: Encontro Ciência e Inovação 2026 takes place on 15 and 16 July in Lisbon, bringing together researchers, universities, companies and policymakers to take stock of what Portugal is researching — and what it still isn’t.

What is the Encontro Ciência e Inovação?

It is the national scientific community’s yearly meeting point, organised with the backing of the government and the country’s science-funding agencies. Over two days, it packs in plenary sessions, project showcases, science-policy debates and networking moments worth their weight in gold to anyone hunting for partners or funding. From state laboratories to deep-tech startups, everyone who makes research happen in Portugal passes through.

This year’s edition has plenty to chew on: artificial intelligence seeping into every field of knowledge, European investment in clean technologies accelerating — as this week’s mega-round for European nuclear fusion shows — and the perennial debate over scientific careers and keeping talent in Portugal.

How can you attend Ciência e Inovação 2026?

The full programme, speakers and registration are available on the event’s official website. For those working in research, it is the year’s best chance to show their work; for the curious public, it is a rare window onto what science made in Portugal is preparing for the next decade — all in one place.

The race for the energy of the Sun has a new European heavyweight. Germany’s Proxima Fusion has closed a 411-million-euro funding round — about 468 million dollars — with two eyebrow-raising names on board: Google, making its first bet on a European fusion company, and German utility RWE. The Munich firm is now valued at 2.4 billion euros, making it the best-funded fusion company in Europe.

What is a stellarator and why does it matter?

Proxima is betting on stellarator technology — a twisted magnetic chamber that confines plasma at temperatures hotter than the Sun’s core, an alternative to the tokamaks that dominate the field. The promised advantage is stability: a stellarator can, in theory, run continuously, without the abrupt disruptions that plague its rivals. The company aims to have a demonstrator running in the early 2030s and to build its first plant with RWE at Gundremmingen, Bavaria — on the site of a former nuclear fission plant, symbolism included.

The round was led by XTX Ventures and East X Ventures, with the money earmarked for scaling up production of high-temperature superconducting cables and magnets, the heart of the machine.

Where does Portugal fit in this race?

On the bench, for now — but not out of the game: Portuguese space and energy science has been growing, as the Lusíada constellation of Portuguese satellites shows, and a Europe with cheap clean energy in 2035 changes the maths for every economy, ours included. If fusion delivers half of what it promises, this round will be remembered as money well spent. Official details are on the Proxima Fusion website.

Portugal has just sent three poets into space. Florbela Espanca, Miguel Torga and Cesário Verde are the names of the newest satellites in the Lusíada constellation, launched this week aboard a SpaceX Falcon 9 on the Transporter-17 mission from Vandenberg, California. With this batch, Portuguese company LusoSpace now has eight satellites in orbit.

What is the Lusíada constellation for?

Think of it as a Waze of the oceans: a communications and data-sharing network for the maritime sector, fitted with AIS and VDES technology to track ships, improve surveillance and safety at sea, and support search-and-rescue operations. The three poets join Camões, Fernando Pessoa, Agustina Bessa-Luís, José Saramago and PoSat-2 — with four more satellites still to come before the twelve-strong constellation is complete.

How much does the Lusíada constellation cost?

The project represents a 15-million-euro investment, 10 million of it financed by Portugal’s Recovery and Resilience Plan, as laid out in the official PRR project note. For a Lisbon-founded company, it’s one of those cases where European money literally leaves the atmosphere.

Portugal’s space ambitions keep climbing — and in a summer when the world is meeting in Geneva to work out how to govern AI, there’s something comforting about national technology being busy watching the sea, under the names of people who wrote about it.

Cybercrime has just crossed a line experts have dreaded for years: a ransomware attack executed from start to finish by an artificial-intelligence agent, with no skilled human operator at the wheel. It is called JadePuffer, and it was documented by Sysdig’s threat research team.

What is JadePuffer?

It is the first known “agentic” ransomware operation: an AI agent exploited a vulnerability in internet-facing software, harvested credentials, moved laterally through the network, established persistence and ended up encrypting more than 1,300 service configuration items in the victim’s database, deleting the originals. Researchers found the malicious payloads peppered with natural-language commentary explaining each step — including picking the “most profitable” target — a telltale sign of code generated by language models rather than human operators.

Why does this change cybercrime?

Because it knocks down the expertise barrier. Until now, a full attack chain demanded skill at every stage; here, one agent strings it all together on its own and even adapts to setbacks — in one sequence it went from a failed login to a working fix in 31 seconds. One industry report estimates autonomous agents already account for one in eight AI-related security incidents. The calendar has a sense of irony: the disclosure lands the same week the UN gathered the world to discuss AI governance.

The full technical analysis is published on the Sysdig blog. For businesses everywhere, the message is blunt: patching internet-facing software is no longer a job for tomorrow.

AI is automating work — all of it, apparently.