Google just did something a chip buyer rarely does: instead of simply writing a check for hardware, it negotiated the right to profit from its own supplier’s stock. In a securities filing this week, Marvell Technology disclosed that it has issued Google warrants to purchase nearly 59 million of its shares, a stake that could be worth as much as $12.2 billion if fully exercised. It’s an unusual structure for a chip supply agreement, and it says a lot about how central custom silicon has become to the AI infrastructure race.

The Mechanics Behind a $12.2 Billion Warrant

According to the filing, Google can buy up to 58,970,907 Marvell shares at a fixed price of $206.58 apiece, with the warrant remaining exercisable until August 2033. But this isn’t free money handed over on day one. Roughly 1.4 million shares vest in equal quarterly installments over the agreement’s first year, regardless of purchases. The remaining bulk of the warrant, more than 57 million shares, unlocks only as Google actually buys chips from Marvell, with one tranche vesting for every $500 million in qualifying revenue Marvell books from Google’s orders.

In other words, Google’s ownership position grows in direct proportion to how much custom silicon it purchases. If Google hits every spending threshold, the deal could translate into roughly $120 billion in cumulative chip purchases for Marvell running through fiscal 2033. That would also make Google the chipmaker’s fifth-largest shareholder, a striking outcome for what began as a components agreement.

What the Deal Actually Covers

The partnership doesn’t touch Google’s core Tensor Processing Units themselves. Instead, it covers what Marvell describes as products that “attach to the TPU ecosystem” — AI inference accelerators, storage controllers, network interface controllers, memory interface controllers, and near-memory compute hardware. These are the supporting components that let TPUs move data efficiently, store it, and communicate across a data center, which have become just as critical as the processors at the center of it all.

This expanded scope matters because it shows Google isn’t just chasing more raw computing power. It’s investing in the entire hardware stack that surrounds its custom silicon, treating data-center networking and storage as strategic assets rather than commodity purchases.

Markets React Fast

Investors didn’t waste time digesting what the deal meant. Marvell shares jumped as much as 14% intraday before settling to a close of roughly 8% higher, as the market read the agreement as a strong vote of confidence from one of the world’s largest cloud providers. Rival chipmaker Broadcom, which has been Google’s primary custom-silicon partner for years, slid more than 5% on concerns that it now faces real competition for Google’s business.

That reaction may be somewhat overblown. Morningstar analyst William Kerwin described the news as reflecting a growing opportunity at Google for new suppliers, rather than Marvell displacing Broadcom outright. Google has reportedly split its eighth-generation TPU program into separate training and inference workloads, and at the scale it’s now operating, one design partner may simply not be enough to keep up with demand.

A Pattern Reshaping the AI Chip Industry

This isn’t happening in isolation. Big Tech companies have increasingly turned to equity-linked arrangements to lock in chip supply while sharing in the upside they’re creating for suppliers. In October 2025, AMD struck a similar deal with OpenAI, agreeing to supply AI chips worth tens of billions annually while giving OpenAI the option to acquire roughly 10% of AMD. Nvidia, for its part, invested $2 billion directly into Marvell earlier this year through its NVLink Fusion partnership, and has also backstopped tens of billions in AI infrastructure spending for other partners.

For chipmakers like Marvell, these warrant-based deals offer a predictable, long-term revenue pathway and the market credibility that comes from being tied to a major hyperscaler. For buyers like Google, the structure works as a hedge: it locks in supply capacity for critical data-center hardware, all without requiring cash upfront or a fixed purchase commitment, since the shares only vest as spending actually happens.

Why This Matters for the Broader AI Buildout

The scale of this deal underscores just how much money is now flowing into the infrastructure layer beneath AI models, not just the software running on top of it. Demand for custom silicon like TPUs has surged as major tech companies search for cheaper, more efficient alternatives to Nvidia’s graphics processors, particularly for the inference workloads involved in running trained AI models day to day. The change has also caused a huge increase in capital spending by Google, which is spending tens of billions of dollars a quarter on its computing infrastructure.

What’s notable about the Marvell deal is how it blurs the line between customer and investor. Google is no longer just a buyer of chips, it is structurally aligned with the success of its supplier, and the two companies share a common incentive to continue to grow the relationship. As custom silicon becomes central to how the biggest cloud providers compete, expect more of these hybrid supply-and-equity arrangements to define the next phase of the AI hardware race, one where chip purchases and shareholder value increasingly move in lockstep.

The post Google Deepens AI Chip Bet With Marvell Deal. appeared first on POLYTIKAL.

Bengaluru: For years, India’s IT-BPM industry ran on a fairly predictable rhythm — big campus hiring drives, steady lateral movement after appraisal season, and net headcount numbers that climbed almost every year without fail. That rhythm is now breaking. A new industry study projects that IT jobs India AI is redefining the sector will see net headcount additions fall nearly 26 percent in the first half of FY27, a sign that the old hiring playbook may no longer apply.

According to the India IT-BPM Workforce Outlook Report 2027 by HAN Digital Solution, net additions across the sector are expected to slip below 70,000 in H1 FY27, down from roughly 95,000 in the same period a year earlier. It’s one of the weakest hiring stretches the industry has seen in recent memory, and the report is careful to frame it not as a passing rough patch, but as a structural shift in how technology companies think about people.

Not a Slowdown, a Reset

What makes this IT-BPM hiring 2026 story different from past downturns is the reasoning behind it. Previous slowdowns tended to track global demand cycles — a recession here, a client budget freeze there. This one is different because even as revenue stabilizes, the hiring hasn’t bounced back the way it used to. Saravanan Balasundarm, Founder and CEO, HAN Digital Solution said it was a deliberate recalibration, not a cyclical dip. Companies were reshaping their hiring models around AI-driven productivity, leaner teams and a growing preference for flexible and project-based talent over large permanent workforces.

Traditionally, the second and third quarters of the fiscal year have been the industry’s busiest hiring window, powered by fresh campus onboarding and internal promotions triggering new lateral hires. This year, that seasonal surge is expected to fall well short of its usual scale — a real departure from a pattern the sector has followed for over a decade.

Where the Cuts Are Landing

The report points to specific categories of work that are shrinking fastest under this AI workforce disruption. Roles built around repetitive, lower-complexity tasks — generic Java and .NET development, digital marketing execution, campaign management, basic technical support (the L1 and L2 tiers), traditional quality assurance, and routine customer service and incident management — are the ones AI agents are absorbing first. These were historically the entry points for large volumes of fresh graduates, which partly explains why net additions are taking such a visible hit.

It isn’t just entry-level roles feeling the squeeze either. Mid-career professionals in execution-heavy delivery positions are reportedly being phased out earlier than past norms would suggest, as companies lean harder into automation for work that used to require sizeable human teams.

The numbers from India’s largest listed IT firms back up the broader trend. TCS, the country’s biggest IT employer, cut around 12,000 roles through FY26 as part of what it called a “future-ready transformation,” and has reportedly scaled back its fresh graduate hiring to around 25,000 this year compared to an average of 40,000 annually over the previous three years. Tech Mahindra and HCLTech have also reported sequential headcount declines in recent quarters, while Infosys stood out as something of an exception, adding roughly 5,000 employees — though even that pace lagged its own hiring from the year before.

Not every part of the industry is shrinking, though. The banking, financial services and insurance segment is bucking the broader trend, with hiring actually picking up there, largely driven by AI integration projects and the compliance demands that come with them.

The Reskilling Conversation Gets Louder

All of this has reignited a debate that’s been simmering in India tech employment circles for a while now — how prepared is the workforce for a job market this different from the one it trained for? The mismatch is a real structural problem: the people being displaced by automation often don’t have the specialized AI or LLM-adjacent skills the new roles demand, so companies can’t simply move people sideways into different jobs.

Some firms are responding by investing more heavily in reskilling IT sector talent rather than trimming headcount outright. Industry surveys suggest a meaningful share of companies are now prioritizing internal training programs over layoffs to bridge this gap — though that comes with its own cost, since upskilling large numbers of employees while also paying a premium for scarce AI talent puts pressure on margins.

Everest Group’s Jimit Arora put it simply: with AI and what’s being called “vibe coding,” companies can now push far more code into production with a fraction of the headcount they once needed. Teams aren’t disappearing altogether, but the traditional way of getting things done — with large delivery benches — is being rebuilt around getting more output from fewer people.

What Comes Next

The HAN Digital report doesn’t paint this as the end of hiring altogether. It expects hiring to gradually improve in select segments even as the overall pace stays muted, with a growing tilt toward contract and flexible staffing arrangements rather than full-time headcount expansion. As Balasundaram put it, the next phase of growth for the industry won’t be measured by how many people companies hire, but by the kind of capabilities those people bring — and how effectively they work alongside AI rather than in spite of it.

For India’s vast IT workforce, that’s a meaningful shift in what a career in tech is expected to look like going forward — one where staying relevant may depend less on experience alone and more on how quickly professionals can adapt to tools that are rewriting the job description underneath them.

The post India’s IT-BPM Hiring Outlook Cools As AI Reshapes Workforce Needs. appeared first on POLYTIKAL.

Maharashtra has staked its claim in India’s growing race to be the AI capital of the country. The state Cabinet, chaired by Chief Minister Devendra Fadnavis, has approved the Maharashtra AI Policy 2026, backing its ambition with real money and infrastructure. This policy is a significant step ahead for Maharashtra’s AI efforts with a dedicated ₹500-crore AI Startup Venture Fund and 12 new AI incubators across the state.

What’s Actually in the Policy

The Maharashtra AI policy isn’t a vague statement of intent — it comes with specific numbers attached. The government is establishing a dedicated ₹500-crore AI Startup Venture Fund with contributions from both the government and private sector in an effort to bolster the startup ecosystem and the policy is specifically intended to assist in creating at least one AI unicorn from the stateOf that fund, ₹250 crore is expected to come directly from state coffers

The 12 planned incubators aren’t just symbolic either. an additional 25 percent incentive on top of that for women entrepreneurs specifically each will provide startups grant support of up to ₹1 crore, with women-led ventures eligible for up to ₹1.25 crore

But the policy reaches far beyond startups to cast a wide net over the state’s broader tech and industrial ecosystem. Some 5,000 Micro, Small and Medium Enterprises will get a 20 percent subsidy on their AI implementation costs, a move aimed at pulling smaller, less digitally mature businesses into the AI economy rather than leaving adoption to large corporates alone. Six key sectors have been identified for setting up of Centers of Excellence — healthcare, agriculture, education, urban development, Marathi language technology and finance five AI Innovation Cities will be anchors for research and industrial activity around the state.

The Bigger Numbers Behind the Announcement

The AI Startup Venture Fund and incubators are just one piece of a much larger push. The cabinet approved policy of Maharashtra is expected to attract over ₹10,000 crore investment and generate 1.5 lakh jobs by 2031, making it one of the largest state-level AI initiatives in India. The policy is structured around seven pillars of infrastructure, data ecosystems, skilling, innovation, startup support, sectoral deployment and governance and is in line with the central government’s own India AI Mission framework.

Skilling features prominently too. The state plans to train roughly two lakh — 200,000 — young people and working professionals in AI-related skills, aiming to build a workforce that can actually staff the companies the policy hopes to attract. Alongside this, the government plans to deploy a common computing backbone offering 2,000 GPUs through a Compute-as-a-Service model , addressing one of the more persistent bottlenecks facing Indian AI startups: access to affordable, high-performance computing.

On the governance side, Maharashtra says it will become the first state in India to build a separate ethical AI framework, alongside plans for a State AI Data Exchange linked to the central government’s India AI Mission, and a “Maharashtra Centre for Advanced Artificial Intelligence Training” run as a joint platform between industry and academic institutions. The policy also mandates annual AI readiness audits across government departments, and separately proposes AI-based systems to make citizen interactions with public services smoother and less bureaucratic.

Investors and companies setting up AI operations in the state will also get a package of financial sweeteners. These include up to a 20 percent capital subsidy, a full stamp duty exemption, electricity tariff subsidies, and reimbursements for various setup costs.

Why It Matters in India’s AI Race

Maharashtra’s step comes at a critical juncture. The move comes even as India’s top IT firms have collectively lost a significant chunk of their market value, underscoring how desperately the state sees AI as the sector’s next growth engine rather than a peripheral bet The framing from Fadnavis himself was blunt — cite index he said that around 70 percent of jobs are likely to be reshaped by AI in the coming years, and argued the policy needs to be updated periodically to keep pace with the evolution of the technology.

That said, not everyone is convinced the numbers are conclusive. Analysts have noted that reskilling addresses only the supply side of India’s AI talent equation – on its own it does little to generate demand for AI-capable workers. The venture fund and incubators are aimed at solving that second half of the problem by building companies that actually employ AI talent domestically, though critics note that ₹500 crore is a relatively modest sum in a market where individual AI startups can raise more than that in a single funding round. Some observers argue that the policy’s more consequential elements may actually be its less headline-grabbing provisions — the State AI Data Exchange and the underlying compute infrastructure commitments — since India’s real structural disadvantage in AI has never been a shortage of talent, but a shortage of compute and usable data.

Maharashtra Isn’t Alone in This Race

The competition among Indian states to lead on AI investment is intensifying fast, and Maharashtra is far from the only contender. Karnataka has separately committed roughly ₹600 crore toward a “DeepTech Decade” initiative, including dedicated funds for AI and frontier-technology startups and new incubators attached to premier technical institutes. Goa, meanwhile, has released its own draft AI policy for public feedback, aiming to position itself as a hub for high-tech and responsible AI development.

For Maharashtra — long India’s leading startup hub, home to tens of thousands of DPIIT-recognized startups — the AI Policy 2026 is a bid to make sure that lead extends into the AI era rather than being ceded to newer, faster-moving state rivals. Whether ₹500 crore and 12 incubators are enough to make that happen will likely become clearer only as the policy’s early cohorts of AI startups begin to take shape over the next few years.

The post Maharashtra Puts Rs 500 Crore Startup Fund Into Artificial Intelligence, 12 New Incubators. appeared first on POLYTIKAL.

It’s been a whirlwind 24 hours for anyone following the business side of AI. Between a payments giant buying its way into the model-routing business and a chipmaker underwriting a data center the size of a small city’s power grid, the scale of money moving through AI infrastructure right now is hard to overstate. Put simply: the AI infrastructure race isn’t slowing down, it’s accelerating, and the battlegrounds keep multiplying.

Stripe Bets Big on AI Routing

Start with Stripe. The payments company has finalized an agreement to acquire OpenRouter, the startup that lets developers switch seamlessly between hundreds of different AI models, for more than $7 billion. To put that number in perspective, OpenRouter was valued at just $1.3 billion during its Series B round a mere three months ago — a markup north of five times in a single quarter.

What makes OpenRouter valuable isn’t flashy AI research; it’s plumbing. The platform routes traffic across more than 400 models for roughly 8 million developers, handling billing and directing requests to whichever model makes sense on cost, speed, or capability. Its CEO has long described the company as “the Stripe for AI,” which makes Stripe’s interest almost poetic — the payments giant is essentially absorbing a company built in its own image, extending its ambitions from moving money to moving AI traffic itself. For an industry drowning in model choices from OpenAI, Anthropic, DeepSeek, and others, having one company control both the payment rails and the routing layer is a meaningful consolidation of power.

Nvidia’s Financing Muscle Behind OpenAI

Meanwhile, Nvidia has taken its role in the AI buildout to a new level. The chipmaker has agreed to provide credit support worth up to $105 billion for a massive OpenAI data center campus being built in Pike County, Ohio. The facility, developed and operated by SoftBank-backed SB Energy, will initially support 4.25 gigawatts of computing capacity, with an option to expand toward roughly 8 gigawatts total. Nvidia is also putting $1.5 billion directly into SB Energy and has structured the arrangement so the site runs exclusively on Nvidia GPUs, CPUs, and networking gear.

Nvidia CEO Jensen Huang has pushed back on suggestions that this amounts to circular financing, insisting OpenAI will cover its lease obligations through its own revenue and investor capital. He’s framed the broader relationship as potentially representing around $600 billion in Nvidia compute spending by OpenAI through 2030 — a staggering figure that underscores just how much capital frontier AI labs now need to secure computing power, well beyond what a typical company’s balance sheet could support on its own. Huang has openly acknowledged that frontier labs are “growing faster than their balance sheets and long-term credit profiles can support,” which is precisely the gap this kind of vendor financing is designed to fill.

The Real Constraint Isn’t Chips Anymore

Here’s the twist that’s reshaping how everyone in this space thinks: the bottleneck increasingly isn’t GPU supply. It’s power. Microsoft has been candid about the scale of the problem, with reports describing a company sitting on AI chips it can’t fully deploy because there simply isn’t enough electricity or grid capacity available where it needs it. Grid interconnection timelines in major markets can stretch anywhere from two to seven years depending on the region, a pace wildly out of step with how fast AI hardware itself evolves.

This is a genuinely strange inversion for an industry that spent the past few years obsessed with chip shortages. Transformers, switchgear, and battery systems for power delivery are now cited as harder to secure than the semiconductors themselves. It explains why Nvidia’s Ohio commitment includes not just compute guarantees but investment in regional grid infrastructure — the company is essentially underwriting the electricity supply chain, not just the silicon, because without power, GPUs are just very expensive paperweights.

Networking Becomes the Next Frontier

If power is one binding constraint, the other emerging one is how efficiently data actually moves between GPUs once they’re powered on. Training clusters today span tens of thousands, sometimes hundreds of thousands, of GPUs, and if the network connecting them can’t keep pace, those chips sit idle waiting for data — some deployments reportedly see idle rates as high as 30 to 50 percent. Copper wiring, the traditional connective tissue of data centers, simply runs out of headroom at these speeds; it gets too hot and too power-hungry to scale further.

That’s why photonics — using light rather than electricity to move data — has become one of the most closely watched technologies in AI infrastructure. Nvidia has already poured billions into optical interconnect partnerships and is rolling out next-generation switching platforms built around co-packaged optics, promising dramatically better power efficiency for the “million-GPU” AI clusters the industry is now openly planning around. Analysts increasingly argue that the next real AI infrastructure battle will be fought over networking bandwidth as much as raw chip supply, since a cluster with brilliant chips and a mediocre network is still a slow, wasteful cluster.

What It Adds Up To

Taken together, these developments paint a picture of an industry maturing past its “just buy more GPUs” phase into something far more complex: a full-stack infrastructure war spanning payments, financing, power generation, and high-speed networking all at once. Frontier AI models don’t just need brilliant algorithms anymore — they need reliable electricity, financing structures robust enough to support hundred-billion-dollar commitments, and networking fast enough to keep expensive silicon from sitting idle.

Companies like Microsoft, Nvidia, and OpenAI are no longer just tech firms competing on product features; they’re increasingly behaving like utilities and infrastructure developers, because that’s genuinely what building frontier AI now requires. Whoever solves the physical constraints — power, cooling, and interconnects — may end up mattering just as much as whoever builds the smartest model.

The post AI Infrastructure Race Intensifies With Multi-Billion-Dollar Deals. appeared first on POLYTIKAL.

Google is gearing up for its biggest hardware moment of the year, and this time it’s bringing serious star power along for the ride. The “Made by Google” event is set for August 12 in New York City, and all signs point to the Pixel 11 lineup taking center stage, backed by a guest list that reads more like an awards show than a tech keynote.

Trevor Noah Takes the Stage

Following Jimmy Fallon’s turn as host last year, Made by Google 2026 will be led by former late-night host and comedian Trevor Noah. Google confirmed Noah’s hosting role through a teaser video, revealing an extensive list of celebrity guests set to join him as the company continues leaning into a celebrity-driven, show-like format for its annual product reveal.

The guest list itself is genuinely eclectic. Alongside Noah, attendees are expected to include NBA star Stephen Curry, podcast host Alex Cooper, Indian cricketer Shubman Gill, Japanese actress Ayami Nakajo, DJ Peggy Gou, YouTube creator Jesser, college basketball star JuJu Watkins, musician PinkPantheress, track and field athlete Chari Hawkins, former soccer manager Steven Gerrard, and actor Daniel Durant. Interestingly, most of these guests have already appeared in recent Pixel ad campaigns, so their presence at the event isn’t entirely a surprise. It’s worth noting, though, that not everyone necessarily shows up in person. Last year, Curry only appeared in a video segment rather than walking the stage, so it remains to be seen who actually turns up live in New York this time.

When and Where to Watch

The event takes place on Wednesday, August 12, starting at 3 p.m. PT / 6 p.m. EDT, and Google will stream it on its official Made by Google YouTube channel. For viewers outside the US, that translates to roughly 3:30 a.m. IST on August 13, so anyone in India hoping to catch it live is in for a late night. The stream will also be available to watch on the Google Store website. The choice of a prime-time New York venue continues Google’s shift away from traditional daytime keynotes toward more theatrical, evening presentations.

What’s Actually Launching

While the event has been dressed up with celebrity glitz, the hardware is still very much the point. Google is widely expected to announce the Pixel 11, Pixel 11 Pro, and Pixel 11 Pro XL alongside the new Pixel Watch 5. The Pixel 11 Pro Fold, meanwhile, is expected to arrive separately in October rather than launching alongside the rest of the main lineup.

Google has already dropped a few teasers ahead of the big reveal. In its promotional material, the company briefly showed off the Pixel 11 Pro featuring something called “HiLight active” technology, a feature name Google hasn’t explained yet, though the phrasing suggests it’s tied to display or camera technology rather than software. The teaser also offered the first official glimpse of the Pixel Watch 5, following weeks of leaks about the wearable’s upgraded specs and design.

On the software side, expect plenty of attention on where Google’s AI ambitions are headed next. The event is also expected to touch on Gemini and the broader software ecosystem tied to the new hardware. One feature that’s been rumored for months and finally looks set to arrive is worth flagging too: Night Sight Video is expected to debut with the Pixel 11 series, a long-awaited addition to the phone’s camera capabilities.

Why This Launch Matters

This year’s event isn’t happening in a vacuum. Smartphone makers across the board have been racing to bake more on-device AI into their flagship phones, both to justify premium price tags and to differentiate themselves in a market where hardware specs alone rarely move the needle anymore. Google’s own Gemini-powered features have become a central selling point for the Pixel line, and this launch is expected to push that integration even further into the phone’s camera, display, and everyday software experience.

The celebrity-heavy approach is also a calculated bet. Noah’s hosting role follows a similar approach Google took last year, when former “Tonight Show” host Jimmy Fallon led the Pixel 10 launch event. By leaning into a star-studded, entertainment-first format rather than a straightforward product briefing, Google appears to be chasing the kind of cultural buzz that Apple and Samsung launches have traditionally captured through sheer scale and anticipation, betting that a splashier show will translate into wider reach beyond the usual tech-press audience.

What Comes Next

Viewers can expect Google to formally confirm pricing, detailed specifications, and release timing for the Pixel 11, Pixel 11 Pro, Pixel 11 Pro XL, and Pixel Watch 5 once Noah takes the stage in New York. Until then, the run-up to August 12 will likely keep generating fresh leaks and teaser drops, par for the course with any major smartphone launch these days. But with the venue booked, the host confirmed, and a guest list stacked with names from sports, music, and film, Google seems intent on making sure the Pixel 11’s arrival gets talked about well beyond the usual circle of tech enthusiasts.

The post Google’s Pixel 11 Launch Set for August 12 in New York, Trevor Noah to Host. appeared first on POLYTIKAL.

For years, Europe watched from the sidelines as driverless cars became a familiar sight on the streets of American and Chinese cities. That gap is finally starting to close. This week’s confirmation that Uber and UK autonomous-driving firm Wayve have secured a robotaxi Europe licence from London’s transport authority marks one of the clearest signs yet that the continent’s long-promised self-driving era is moving from pilot programmes into something closer to real commercial deployment.

What Actually Got Approved

Transport for London granted private hire vehicle licences to a fleet of Wayve’s autonomous Ford Mustang Mach-E vehicles, allowing them to legally carry paying passengers in the city. The approval came after months of technical review, with the vehicles inspected against TfL’s safety and operational standards before being cleared. It’s worth being precise about what this licence does and doesn’t allow: the cars will still carry a TfL-licensed driver on board to supervise the trip and step in if needed, so this isn’t yet a fully driverless free-for-all. Wayve can run up to 15 vehicles under this trial for one year.

Still, in the context of autonomous vehicles regulation 2026, this is a meaningful unlock. It puts Uber and Wayve ahead of rivals Waymo and the Lyft-Baidu partnership, both of which have also announced plans to bring robotaxis to London this year. The approval builds on a broader partnership Uber and Wayve struck with Stellantis at London’s MOVE 2026 conference, combining Wayve’s driving software, Stellantis vehicle platforms, and Uber’s ride-hailing network — a template that could be replicated across other markets once the London trial proves itself out.

Why London, and Why Now

London has effectively become the proving ground for self-driving cars Europe-wide, largely because of the UK’s Automated Vehicles Act, passed in 2024, which laid out a legal framework for driverless commercial services, liability, and insurance ahead of most of the continental EU. That head start is why so many of the biggest names in autonomous driving — Uber, Wayve, Waymo, Lyft, and Baidu among them — have all been racing to get a foothold in the city this year.

But London isn’t the only place where the regulatory dam has been breaking. Croatia beat everyone to the punch earlier this year when Zagreb-based startup Verne, backed by Rimac Group founder Mate Rimac, launched what it called Europe’s first commercial robotaxi service, using vehicles powered by Chinese autonomous driving firm Pony.ai and backed by a strategic investment from Uber. Switzerland followed with Baidu’s Apollo Go securing a Level 4 operating permit for a service called AmiGo, run jointly with Swiss Post’s PostBus arm across roughly 80 square kilometres of the country’s eastern cantons. Taken together, these approvals suggest 2026 will be remembered as the year Europe’s self-driving ambitions finally caught up with its regulatory paperwork, even if full driverless operation in most cities remains a step or two away.

The European Commission has also been under pressure to move faster. Brussels had capped the number of autonomous vehicles manufacturers could deploy across the EU at 1,500, a limit officials have signalled they intend to lift this year, though similar promises around automated valet parking slipped by nearly a year before finally being delivered in March 2026. The real bottleneck, analysts note, sits less with Brussels and more with individual member states, many of which still require a human driver behind the wheel under national traffic law.

The Bigger Picture: Robotics, Investment, and Security

This regulatory breakthrough isn’t happening in isolation. It’s arriving alongside a much broader wave of robotics infrastructure investment that’s reshaping logistics, manufacturing, and warehousing well beyond passenger transport. Consulting firms and corporate venture arms have been pouring money into physical AI this year, betting that the same sensor, edge-computing, and machine-learning advances powering robotaxis will also transform how goods move through warehouses and factories. Industry trackers estimate the broader robotics sector raised tens of billions of dollars in the past year alone, with major technology companies racing to build the software and hardware layers — from foundation models to edge-compute chips — that autonomous systems of all kinds will run on.

That expansion is bringing new risks along with it. As robotics systems become more cloud-connected and AI-driven, industry bodies have flagged a growing wave of security threats aimed at robot controllers and the cloud platforms coordinating their movements. That’s part of why the growth of AI agents cybersecurity work has become such a live theme this year — the same autonomy that lets a robotaxi navigate city streets or a warehouse robot reroute around an obstacle also creates new digital attack surfaces that need defending, especially as fleets scale into the thousands of vehicles some operators are now planning for.

What Comes Next

For now, London’s supervised trial is a cautious first step rather than a full driverless rollout, and European regulators across the EU and UK alike remain focused on proving safety cases before loosening restrictions further. But the direction of travel is now unmistakable. With Croatia, Switzerland, and the UK all granting some form of operating approval within months of each other, and Germany, Spain, and Luxembourg reportedly not far behind, Europe’s robotaxi race has genuinely begun. Whether ordinary commuters end up hailing a fully driverless ride from an app within the next year or two will depend on how quickly this current wave of supervised trials builds the public trust and safety track record that regulators — and riders — are still waiting to see.

The post Robotaxis Clear Major European Regulatory Milestone. appeared first on POLYTIKAL.

For about three weeks this summer, two of Anthropic’s most advanced AI models simply weren’t available. Now they are again. The company has confirmed that access to Claude Fable 5 and Claude Mythos 5 has been fully restored, closing out an unusual episode that briefly pulled frontier AI development into the world of federal export-control policy.

How It Started

The story begins on June 9, 2026, when Anthropic released both models. They share the same underlying architecture, but the two were positioned very differently. Fable 5 came with strong general-purpose safeguards, meant for everyday users across Claude’s various platforms. Mythos 5, on the other hand, had fewer of those guardrails and was kept tightly restricted, made available only to a small group of trusted partners through a program called Project Glasswing, largely for defensive cybersecurity work.

That careful rollout hit a snag just three days later. On June 12, the US Department of Commerce stepped in with an export control directive, and it wasn’t a narrow one. Anthropic was told to cut off access for any foreign national — whether they were based in the US or abroad, and even if they were Anthropic’s own employees. Because the order took effect immediately and the company had no reliable way to verify user nationality on the spot, it made the call to suspend access to both models entirely, for everyone, rather than risk non-compliance.

Reports suggest the trigger for this was a security concern. Researchers reportedly found a way to get around Fable 5’s safeguards by prompting it in a way that led the model to surface software vulnerabilities, including code that demonstrated how certain exploits could work. That kind of finding, understandably, raised flags at the national security level.

The Standoff and the Reversal

For nearly three weeks, the situation stayed frozen. Fable 5 and Mythos 5 sat inaccessible to the public, and Anthropic worked behind the scenes with regulators to sort things out. There was a partial development on June 26, when the company received approval to extend Mythos 5 access to a limited set of US organizations involved in operating and defending critical infrastructure — a signal that talks were progressing, even if the broader restriction hadn’t lifted yet.

The real turning point came on June 30, when the Department of Commerce notified Anthropic that the export controls had been lifted entirely. Commerce Secretary Howard Lutnick’s office was reportedly involved in finalizing that decision. Anthropic wasted no time sharing the news, posting a public update thanking users for their patience throughout the shutdown.

Access to Fable 5 was restored the very next day, July 1, rolling out globally across the Claude Platform, Claude.ai, Claude Code, and Claude Cowork. As something of a goodwill gesture for the disruption, Anthropic made Fable 5 available for up to half of weekly usage limits through July 7 for Pro, Max, Team, and select Enterprise plan users, before shifting to standard usage-credit access. The company also said it would work to restore availability through cloud partners like AWS, Google Cloud, and Microsoft Foundry as quickly as it could. Mythos 5, true to its more limited design, remained available mainly to Project Glasswing partners, though reports indicate the government has been looking at expanding that circle to more domestic and international organizations over time.

Why This Matters Beyond Anthropic

This whole episode says less about any single company and more about where AI policy is heading. As frontier models get more capable, national governments are increasingly treating them the way they’d treat other sensitive technologies — subject to export rules, national security review, and rapid intervention when something looks risky. For a company like Anthropic, that means the line between “shipping a great product” and “navigating federal policy in real time” is getting thinner.

There’s also an international dimension worth noting. During the roughly three-week blackout, competitors elsewhere kept moving. Other major AI labs continued to release and refine their own frontier models during that stretch, and some industry observers pointed out that a shutdown like this, even a temporary one, can hand a bit of breathing room to international rivals working on similar technology. Some allied governments also voiced concern about how closely their own access to advanced AI tools is tied to decisions made by US regulators — a reminder that AI policy in one country can ripple outward fast.

For everyday users, the practical takeaway is simpler: Claude Mythos 5 and Claude Fable 5 access restored means things are largely back to normal, at least for now. But the bigger story — the growing overlap between frontier AI capabilities and export-control policy — isn’t going away. If anything, this is likely to be one of the earlier examples of a pattern regulators around the world will keep testing as AI systems become more powerful and, in the eyes of some governments, more strategically sensitive.

The post Anthropic Restores Access to Claude Fable and Mythos Models. appeared first on POLYTIKAL.

Google is giving people a limited window to try out its newest AI video tools without paying a cent. The Google Gemini video trial lets users create up to ten AI-generated videos through the app’s “Create video” tool, but the clock is ticking — the offer runs only until August 4 at 11:59 p.m. Pacific Time, and it’s specifically aimed at people who don’t already have a Google AI subscription.

The move is part of a broader push to get Gemini’s newest video capabilities in front of as many people as possible, and it comes at a moment when the AI video generation free trial format has become something of a go-to tactic for tech companies trying to build habit and buzz around tools that are otherwise locked behind paywalls.

What the trial actually offers

To access it, users open the tools menu inside the Gemini app and select “Create video.” From there, the feature is powered by what Google calls its Omni model capabilities — a system built to handle video creation, editing, and remixing all within the same conversational interface, rather than requiring a separate editing timeline or specialized software. Instead of dragging clips around or fiddling with keyframes, users can describe changes in plain language — swap a background, change the lighting, adjust how a character moves — and the model updates the footage accordingly.

The Gemini Omni model itself represents a fairly significant technical leap for Google. Rather than functioning as a standalone text-to-video generator, it’s designed to combine several of DeepMind’s underlying systems: the reasoning capabilities that power Gemini’s chat responses, the video rendering technology behind Google’s Veo models, and a world-simulation layer aimed at keeping physics, lighting, and object behavior consistent across a scene. In practice, that’s meant to translate into videos that hold together more coherently from one edit to the next, with characters retaining their appearance and voice across multiple cuts within the same conversation.

Every video generated through the tool also carries Google’s SynthID watermark, an invisible marker meant to help identify AI-generated content even after the footage has been edited, resized, or re-uploaded elsewhere — part of Google’s broader effort to keep some transparency around AI-made media as these tools become more widely available.

Why now, and why free

The timing of this promotion isn’t an accident. Generative video has quickly become one of the most competitive corners of the AI industry, with a growing list of platforms — including Runway, Kling, and various other consumer-facing tools — all racing to convince users that their model produces the most convincing, most editable, or most useful output. Against that backdrop, offering a taste of Gemini Omni’s capabilities for free is a fairly straightforward play: lower the barrier to entry, let people experience the conversational editing style firsthand, and hope that experience turns into a paid subscription once the trial credits run out.

It’s also worth noting this isn’t unprecedented for Google. Earlier this year, the company ran a similar promotion offering free access to its Veo 3 video generator, letting users create a handful of short clips without a paid plan. This latest trial builds on that same playbook, but with a more capable and more conversational underlying model.

What the trial isn’t

Google has been fairly clear that this is meant as a taste test rather than a long-term free tier. Ten videos is enough to get a feel for how the tool handles prompts, edits, and remixes, but it’s not designed to replace an actual content production workflow. Once the trial period or credit allotment runs out, continued access is expected to require one of Google’s paid AI plans. It’s also strictly a consumer-facing promotion tied to the Gemini app itself — developers looking to build with the underlying model through Google’s API are on a separate, paid path entirely.

The bigger picture

Promotions like this tend to say as much about the state of the market as they do about any single product. With generative AI tools 2026 having become a crowded, fast-moving space, companies are leaning harder on free trials and limited-time offers to pull users away from competitors and into their own ecosystems. For Google specifically, folding this capability into Gemini — an app that already has an enormous existing user base — gives the company a low-cost way to put its video tools in front of people who might otherwise never have tried them.

Whether that translates into long-term paid adoption remains to be seen, but for the next few days at least, anyone curious about what Google AI features look like on the video side has a fairly easy, no-cost way to find out.

The post Google Gemini Offers Free AI Video Generation Trial. appeared first on POLYTIKAL.

The AI benchmark race just got a fresh jolt of drama. Google’s Gemini 2.5 Pro, paired with its new “Deep Think” reasoning mode, has posted standout scores across a batch of advanced science and knowledge tests, putting it ahead of rival models from OpenAI and Anthropic on several key measures. For an industry that’s been trading incremental wins back and forth for months, this felt like a bigger jump than usual.

What Deep Think Actually Changes

Deep Think isn’t just a rebrand of the same old Gemini. Instead of firing off an answer after a single pass, the model explores multiple reasoning paths internally before settling on a response — essentially weighing several possible solutions in parallel rather than committing to the first one that seems plausible. That extra deliberation comes at a price, though: responses can be noticeably longer, sometimes stretching to multiple seconds or even minutes on the hardest prompts. Google gives developers a “thinking budget” dial to manage that trade-off, letting them choose speed for everyday tasks or depth when accuracy matters more than latency.

The payoff shows up clearly on the numbers. Gemini 2.5 Pro with Deep Think reportedly reached 89.8% on MMLU-Pro and 82.4% on GPQA Diamond — a notoriously tough PhD-level science benchmark — putting it ahead of comparable scores from OpenAI’s GPT-5.5 and Anthropic’s Fable 5 on those specific tests. It’s worth noting that “topping a benchmark” rarely means dominating across the board; different labs tend to lead in different categories depending on the task, whether that’s coding, math, video understanding or general knowledge.

The Math Olympiad Flex

Perhaps the most eye-catching claim to come out of this launch involves mathematics. An advanced research version of Gemini 2.5 reportedly achieved gold-medal-level performance on 2025 International Mathematical Olympiad problems — a tier of mathematical reasoning that’s historically separated elite human competitors from everyone else, since IMO problems demand genuine creative insight rather than brute-force calculation. That kind of result was, until fairly recently, considered a multi-year target for AI labs, not something arriving this soon.

The consumer-facing version of the model, the one actually available inside the Gemini app, performs at a bronze-medal standard on similar problems — still an impressive result, just running with far less compute than the research variant. That gap illustrates a pattern that’s becoming familiar across the frontier AI models race: labs often hold back a more resource-intensive version for research purposes, while shipping a faster, leaner variant for real-world, real-time use.

Why the AI Benchmark Race Keeps Heating Up

Google’s Gemini 2.5 Pro launch lands squarely in the middle of an increasingly crowded and competitive stretch for frontier labs. OpenAI, Anthropic and Google have all been pushing out reasoning-focused updates in relatively quick succession, each claiming leadership on one benchmark or another. For enterprise customers trying to pick a model for coding, research or customer-facing tools, that constant leapfrogging can make decisions genuinely difficult — the “best” model this month isn’t guaranteed to hold that title by the next one.

Google’s pitch here goes beyond a single benchmark screenshot, too. Alongside Deep Think, Gemini 2.5 Pro has been positioned around a large context window that lets it process long documents, codebases or research papers in a single prompt, plus native multimodal support across text, images, audio and video. That combination — a big context window, competitive reasoning scores and reasonably efficient pricing — has made Gemini 2.5 Pro an appealing option for teams handling long-document analysis, even in cases where a rival model might edge it out on pure coding benchmarks.

What This Means Going Forward

For everyday users, benchmark charts can feel abstract, but the practical effect is real: reasoning models like this one are increasingly capable of tackling multi-step problems — scientific analysis, complex math, dense technical questions — that would have tripped up AI systems just a year or two ago. Deep Think mode, and the “thinking budget” concept behind it, also point to where the industry seems to be heading: giving users more direct control over how much computation (and how much waiting) they’re willing to trade for a more carefully reasoned answer.

Whether Gemini 2.5 Pro holds its benchmark lead for long is anyone’s guess — this is an industry where records tend to get broken within weeks, not years. But for now, Google has planted a flag squarely in the middle of the reasoning AI conversation, and rivals at OpenAI and Anthropic will almost certainly be racing to answer back. If the past year is any indication, the next headline in this saga probably isn’t far off.

The post Google’s Gemini 2.5 Pro Tops Reasoning Benchmarks. appeared first on POLYTIKAL.

Big Tech just handed Wall Street another reason to squint at its spreadsheets. Microsoft and Meta Platforms reported second-quarter 2026 earnings on Wednesday, and the results landed with a very different reception for each company — proof that the old rulebook for judging AI stocks is falling apart in real time.

For years, the deal between Silicon Valley and investors was simple: spend whatever it takes on artificial intelligence, and as long as revenue kept climbing, the market would look the other way. That arrangement is now cracking. Alphabet kicked off the unease the week before when it raised its 2026 capital expenditure ceiling and revealed negative free cash flow for the first time since going public, even though cloud revenue jumped sharply. The stock still had its worst day in over a year. That reaction set the tone for what came next.

Microsoft Delivers, Meta Disappoints

Microsoft, at least for now, gave the market what it wanted. The Intelligent Cloud division posted strong growth, with Azure crossing the $100 billion mark and expanding by roughly 43% year-over-year. Shares moved higher in after-hours trading as investors credited the company for showing that its enormous AI infrastructure bill is translating into tangible cloud demand. Analysts on the earnings call largely came away reassured, and some described the quarter as close to flawless.

Meta told a rockier story. Revenue climbed nicely, but earnings per share missed expectations and net profit slipped compared to the same period last year. The company also nudged up the lower end of its capital spending guidance, reinforcing fears that its AI budget is only going in one direction. On the earnings call, analysts pushed back harder than usual. Bernstein’s Mark Shmulik questioned whether AI was actually changing consumer behavior in any meaningful way, while Goldman Sachs’ Eric Sheridan pressed for a clearer timeline on when Meta’s AI bets would start paying off. The tone wasn’t hostile, but it carried a noticeable edge of impatience.

The Real Story Is Capital Expenditure

What’s striking is how little the headline numbers seem to matter anymore. Investors used to size up AI stocks on three things: revenue, earnings, and whether both beat expectations. That framework is quickly becoming outdated for the biggest tech companies. Now the number everyone watches is capital expenditure — the billions being poured into data centers, chips, and computing power to stay competitive in the AI race.

The scale involved is almost hard to process. Microsoft alone spent close to $97 billion over the trailing twelve months, and analysts expect that figure to climb further in the year ahead. Combined, Alphabet, Microsoft, Amazon, and Meta are projected to funnel roughly $724 billion into capital spending in 2026, with that number expected to approach $950 billion the following year. Some longer-range estimates put combined hyperscaler spending near $1.3 trillion annually within five years.

That kind of spending has a side effect nobody likes discussing out loud: it eats free cash flow. Alphabet already crossed into negative territory. The question hanging over every earnings call this season is whether Microsoft, Meta, and Amazon are headed the same direction, and whether shareholders will tolerate it if they are.

Why Wall Street Is Losing Patience

The broader market backdrop isn’t helping. The so-called Magnificent Seven stocks fell sharply in the week leading into this earnings run, and the tech-heavy names that once dominated the S&P 500 are increasingly ceding ground to companies further down the AI supply chain — chipmakers and infrastructure providers who benefit no matter who wins the platform war. Meanwhile, broader economic data has stayed resilient enough that the Federal Reserve is expected to hold rates steady for now, even as markets price in the possibility of hikes later in the year.

Against that backdrop, every dollar of AI capital expenditure gets more scrutiny than it did twelve months ago. Investors aren’t necessarily doubting that artificial intelligence will eventually generate real returns. What’s changed is their patience for waiting to see it. A year ago, heavy spending was treated as evidence of ambition. Today, it’s treated as a risk that needs justifying on every call.

What Comes Next

Apple and Amazon report later this week, and their results will add more data points to a debate that isn’t going away anytime soon. If either company shows the same pattern — strong growth paired with spending that outpaces it — the pressure on the entire sector could intensify further.

For now, Microsoft has bought itself some goodwill by pointing to concrete cloud growth as the payoff for its AI bets. Meta has more convincing to do. Zuckerberg and his team will likely spend the next few quarters trying to show skeptical analysts that the enormous sums going into AI infrastructure are building toward something investors can actually see in the numbers, rather than just a bigger bill.

Until that happens, expect capital expenditure to stay the headline number every Big Tech earnings season, no matter how good the rest of the report card looks.

The post Microsoft and Meta Post Mixed Big Tech Earnings as Wall Street Questions Runaway AI Spending. appeared first on POLYTIKAL.