News
In the summer of 1995, the future of computing briefly seemed to belong to Netscape.
Netscape went public that August, barely sixteen months after it had been founded. Its stock doubled on the first day. The company had no empire of hardware, no installed operating system, no grip on the office desktop. What it had was a window into a new world.
Open Navigator, type an address, and the Internet appeared with a little throb of electricity. The browser did not feel like an application. It felt like a passage out of Microsoft's world.
Microsoft noticed.
The battle that followed was called the browser wars, a phrase that makes it sound tidier than it was. Really, it was a fight over who had the right to stand between the user and the next era of computing.
Netscape believed the browser would make the operating system less important. Microsoft believed that anything capable of making Windows less important needed to become part of Windows, preferably yesterday.
By the end of the decade, the company that had introduced so many people to the Web was no longer the Web's gatekeeper. It had been out-distributed, out-bundled, and finally absorbed into a stranger corporate afterlife.
A familiar temptationThere is a familiar temptation now to ask which artificial intelligence company is "the Netscape of AI." The answer usually offered is OpenAI, and the comparison is not wrong. ChatGPT did for artificial intelligence what Navigator did for the Web. It turned a technical architecture into a public experience. It gave the future a text box.
But the more important question is this: Who is today's Microsoft? Who understands that the winner is often the company that owns the default?
Every platform shift begins with a miracle and ends with a map of choke points. The miracle is what users remember. The choke points are where the money goes.
The early Web was sold as an escape from gatekeepers. It created new ones. Search. Browsers. Marketplaces. Mobile operating systems. Cloud platforms. Social media. The AI era is being sold with the same democratic glow.
And yet the deeper stack is already hardening. It is made of chips, power contracts, data centers, model weights, enterprise identity, workflow data, cloud credits, procurement channels, and the small tyrannies of default settings. The romance is in the chatbot. The control is somewhere colder, louder, and much more expensive.
That was true in the nineties, too. The Web looked like a page. The winners understood it was a stack.
The danger to OpenAIOpenAI is the obvious Netscape figure because it supplied the first mass-market revelation. Before ChatGPT, artificial intelligence was a research field, a back-office tool, or a phrase executives used when they meant analytics with a larger budget. After ChatGPT, it was something anyone could talk to.
People underrate the power of the first interface that makes a new technology feel inevitable. Netscape did not invent the Internet. It made the Internet feel reachable. OpenAI did not invent the transformer. It made the transformer feel conversational.
But Netscape's story is not a founder myth. It is a warning label.
Netscape had the user's excitement but not enough control over distribution. Microsoft had the operating system. It could place Internet Explorer where users already lived. It could make the browser free. It could turn a product category into a feature.
The lesson was simple: if your rival owns the layer beneath you, your brilliance may become their menu option.
OpenAI is better protected than Netscape was, but not safely protected. It has a huge brand, astonishing usage, and deep ties to Microsoft. It also has the curse of being expensive in a way software companies used to avoid. Each improvement requires compute, chips, talent, energy, and capital. The old software dream was scaling with almost no marginal cost. AI can't do that.
That is why OpenAI's partnership with Microsoft is both strength and vulnerability. Microsoft gives it cloud computing infrastructure, enterprise access, and capital. Microsoft also sits close enough to learn from it, package it, hedge around it, and sell AI into the places where people already do work.
If Netscape's problem was that Microsoft stood underneath it, OpenAI's problem is subtler: Microsoft stands underneath it, beside it, and increasingly in front of the customer.
Netscape had the dazzling demo. Yahoo had the traffic. AOL had the subscribers. None of that was enough. The durable winners were the companies that turned their software layers into a control point and a tollbooth.
Nvidia may be doing exactly that.
The rise of NvidiaIn the nineties, Intel was the metronome inside the personal-computer boom. Microsoft owned the software platform. Intel owned the pace of the machine.
Nvidia occupies a similar place in AI, but the analogy understates the ambition. Nvidia is not merely selling chips into a boom. It is selling the industrial base of the boom: GPUs, networking, software libraries, developer habits, and a vision of the data center as an AI factory.
Every major AI player is both Nvidia customer and Nvidia escape artist. Google has TPUs. Amazon has Trainium. Microsoft is developing its own silicon. Everyone wants an alternative. The problem is that wanting one does not create an ecosystem.
Nvidia's position today may be the purest example of moving up the stack from below. A chip company becomes a systems company. A systems company becomes a software company. A software company becomes a developer environment. A developer environment becomes a tax on ambition.
For now, Nvidia is the toll collector.
The AI market will not resolve into one winner. Platform shifts rarely do. The Web did not produce one winner. It produced layers of power. Microsoft kept the desktop. Google won search. Amazon won commerce and cloud. Apple won mobile hardware and the app economy. Meta won social attention.
AI will do the same.
Microsoft may become the default enterprise AI company, not because every Copilot is brilliant, but because Microsoft sits where work already happens. Nvidia may remain the dominant compute toll collector. Amazon will likely win much of the infrastructure layer. Google must reinvent search while defending it. Meta will use AI to extend attention. Apple may yet turn personal AI into a device-native experience.
The likely losers are the companies attached to the wrong layer.
The changing landscape of AIIn the nineties, AOL looked invincible because it owned access. Broadband made that access less special. Yahoo looked inevitable because it owned attention. Search made that attention less decisive. Netscape looked revolutionary because it owned the browser. Microsoft made the browser a dependency of the operating system.
In AI, the same demotions will happen. Some model companies will become features. Some application companies will become demos. Some incumbents will decorate old products with AI and call it transformation, which is the corporate version of putting a spoiler on a minivan.
The mistake, in every technological boom, is to confuse the moment of wonder with the arrangement of power that follows it.
The wonder is sincere. The arrangement is not.
The early Web made people feel as if they had slipped the old gatekeepers. Then came the search box, the app store, the marketplace, the cloud account, the login, the subscription, the default. Each solved a real inconvenience. Each left behind a narrower path.
AI will likely travel the same road, only faster and with a larger electricity bill. It will begin as a conversation and mature into an administrative system for human intention: what we ask, what we buy, what we write, whom we trust, which choices are shown, and which never quite appear.
The future does not usually arrive wearing chains. It arrives offering to save time.
We've tested, reviewed, and ranked the best business cloud storage services.
This article was produced as part of TechRadar Pro Perspectives, our channel to feature the best and brightest minds in the technology industry today.
The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/pro/perspectives-how-to-submit
The enterprise technology industry has a peculiar relationship with accountability. When it comes to cloud uptime, latency, and data security, we expect contractual guarantees, SLAs, and clearly defined remedies. But when it comes to AI-generated outputs, the actual content these systems produce, we've quietly accepted a different standard: best effort.
I've spent years in commercial and operational roles at companies like Gap, Amazon, and Door Dash / Wolt. In every one of those environments, product visuals weren't a marketing nice-to-have. They were infrastructure. A wrong color on a listing didn't just look bad; it drove returns. A missing ingredient on a food image wasn't an aesthetic issue; it was a trust issue that compounded at scale and was dangerous to our customers.
So when AI-generated images started entering enterprise workflows in earnest, I watched with real interest. The efficiency gains were compelling: the ability to generate, retouch, and adapt product visuals at a speed and scale that traditional studio workflows simply cannot match.
But something fundamental was missing from the enterprise conversation: accountability for outputs.
The gap between impressive and dependableThere's a difference between AI that produces impressive results in demos and AI tools you can stake commercial operations on. For enterprise buyers, that gap matters enormously.
Consider what happens when am AI-generated product image fails at volume. A wrong product color in a hero image doesn't trigger one return; it triggers thousands. A distorted shape on a fashion listing doesn't affect one conversion; it affects an entire category. The commercial exposure from visual inaccuracy compounds at scale in a way that individual errors simply don't.
Yet for most of the AI visual tools currently available to enterprise buyers, the contractual position on this exposure is essentially zero. You buy credits, you run images, and what comes out is what you get. If the output doesn't match the brief, you absorb the cost: in regeneration time, in quality control overhead, and ultimately in the downstream commercial impact of content that doesn't perform.
This isn't an indictment of the technology. AI-generated images have genuinely transformed what's operationally possible for enterprise visual production. But the commercial model hasn't kept up with the commercial reality.
Why ownership changes everythingThe reason most AI visual vendors can't offer meaningful output guarantees isn't reluctance; it's architecture. If you're building on third-party foundation models, you have no ability to evaluate, course-correct, or stand behind the quality of what those models produce at the output level. The accountability stops at the API.
The vendors who can make guarantees are the ones who own the full stack: the generation models, the evaluation models, and the remediation process. This is the structural distinction that makes contractual guarantees viable, not as a commercial gesture, but as something that can actually be operationalized.
When a proprietary fidelity evaluation model is running on every output before delivery, you have a mechanism for identifying failures before the client does. When you own the rater, the fixer, and the generation pipeline, you have the ability to correct those failures.
When you've run a feasibility check on a customer's actual catalogue before any commercial commitment, you know what the pass rate will look like in production.
That's the architecture that makes a guarantee meaningful: not a promise, but an auditable process with contractual teeth.
What contractual accountability looks like in practiceThe mechanics matter here, because "guarantee" can mean many things. In practice, an enterprise visual guarantee should do three things: define pass/fail criteria upfront based on the customer's actual brief; evaluate every output against those criteria before delivery; and trigger a clear remedy, regeneration or credit refund, when failures occur.
Critically, the criteria need to be specific. Product fidelity failures, an altered color, a missing ingredient, a distorted product shape, are measurable and contractually defensible. Subjective aesthetic preferences, a lighting angle, a background tone, are not. The boundary between these two things is where a real guarantee lives, and where vague commitments fall apart.
For enterprise buyers, this specificity is valuable in itself. It forces the conversation about what "quality" actually means for a given catalogue before procurement, rather than after. That clarity typically improves outcomes on both sides.
The accountability moment for enterprise AIWe're at a point in the enterprise AI cycle where the conversation needs to shift from what these systems can do to what vendors are willing to stand behind. Capability is no longer the differentiator; the market is full of capable tools. Dependability is.
For enterprise procurement teams, this means starting to ask harder questions. Not just "what's your accuracy rate?" but "what happens when it's wrong, and what are the contractual terms?" Not just "can you handle our volume?" but "what remedies apply when you don't meet the standard we've agreed?"
For the vendor community, it means recognising that the era of best-effort AI in enterprise contexts is ending. Buyers who are running tens of thousands of product images through AI pipelines need the same accountability from those systems that they expect from any other mission-critical infrastructure.
The goal in commerce was never the most beautiful image. It was always an image that sells, reliably, accurately, at scale. Enterprise AI that can guarantee that outcome is the next competitive frontier. The vendors willing to back their outputs contractually are the ones that will earn a place in enterprise infrastructure for the long term.
We've featured the best AI website builder.
This article was produced as part of TechRadar Pro Perspectives, our channel to feature the best and brightest minds in the technology industry today.
The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/pro/perspectives-how-to-submit
- HP India has been fined 1.42 billion rupees ($14.7 million) for two separate cases
- Self-reporting ultimately landed it with lighter fines
- Both cases relate to government tender manipulation between 2017 and 2020
India's Competition Commission (CCI) has accused HP India and some related resellers of coordinating bids for Indian government contracts on the Government e-Marketplace.
According to the CCI, the company and certain partners manipulated government tenders by predetermined or communicated bid prices, submitting deliberately uncompetitive bids to create the appearance of competition and controlling discounts.
The regulator revealed two separate cases for investigation – one relating to PCs, and the other relating to printing consumables like ink and toner.
HP accused of manufacturing bids to win Indian government contractsIn the printing case, the CCI uncovered emails, witness statements, WhatsApp group conversations and even a 2019 video from a reseller meeting. Discussions around which companies would submit supporting bids, prices and discounts, and which reseller should win particular contracts were found.
The CCI declared that a total of 16 Tier-2 resellers had violated its Competition Act through bid rigging, with HP India fined 119.8 million rupees and its resellers fined a combined total of 23 million rupees.
A separate case revealed similar conduct covering laptops, desktops, workstations, POS systems, peripherals and more. Similarly, five additional resellers were highlighted on top of HP India's core businesses, bringing this case's fines total to 1.3 billion rupees and 12.2 million rupees respectively.
The CCI's final orders bring HP India's total fines to around 1.42 billion rupees, or $14.7 million, excluding the fines imposed on its partners.
But the fine could have been a lot worse had HP India not come forward and admitted to its wrongdoing between 2017 and 2020, having submitted a lesser-penalty application to buy itself a discount on the fines.
TechRadar Pro has asked HP for a comment, but we did not receive an immediate response.
The UK's payment landscape is undergoing a rapid transformation. While regulatory initiatives, such as the mandatory authorized push payment (APP) reimbursement scheme, provide safeguards, sophisticated cyber-attacks and elaborate scams relentlessly evolve.
To reduce fraud, strong data-sharing frameworks and collaboration across the financial services industry are essential. Collaboration between big tech, telecoms, banks and the public sector can help combat fraud through a joined-up approach to data-sharing aimed at driving out scammers and identifying potentially fraudulent transactions.
Current landscapeThe scale of UK fraud is stark, with losses reaching £1.28 billion in 2025, a 4% increase year-on-year. Fraud is now operating on an industrial scale, with criminals using increasingly advanced tools and techniques to target victims. Fraud increasingly funds serious and organized crime in the UK and globally, reinforcing its status as a national security threat.
Authorized Push Payment (APP) fraud is a growing area of concern – this type of fraud continues to rise, with 248,070 cases recorded in 2025 (up 7%), showing that fraudsters are consistently adapting, pivoting to exploit new vulnerabilities, even as defenses strengthen. New scams have focused on investment, purchase-related, advance fee, invoice, and mandate scams as well as romance and impersonation scams.
Total losses from APP fraud rose sharply to £576.4 million (up 19%). This reflects a clear shift in criminal behavior, from exploiting systems to manipulating people through increasingly sophisticated social engineering.
Purchase scams made up 71% of all APP cases, demonstrating the scale and diversity of modern fraud tactics. This impact extends beyond financial loss, affecting individual livelihoods, disrupting businesses, and undermining national economic confidence.
Crucially, most APP fraud now originates outside the banking system. 66% of cases begin online (accounting for 32% of losses) and 17% via telecommunications networks, highlighting the growing role of digital platforms and telecoms in enabling fraud. Criminals are no longer primarily hacking systems; they are manipulating people, using sophisticated social engineering to bypass even the strongest technical controls.
Key considerations for the payments industryTackling these challenges requires a multi-faceted approach, combining robust technology, seamless collaboration to enable effective and compliant data-sharing, effective regulation, and public awareness. Key considerations are:
What does government strategy mean in practice?Initiatives, such as the Government Fraud Strategy, provide an important framework for government, law enforcement, and the private sector. Infrastructure and data-sharing initiatives need to be effective, compliant, and aligned with national priorities to disrupt and prevent fraud. This ensures the fight against fraud remains a national priority that continuously adapts.
New data-sharing initiatives with Faster Payment System participants can play a key role here. Pay.UK’s work with participants on Enhanced Data Exchange (EDEx) will facilitate secure, timely data-sharing to empower financial institutions to detect and prevent fraudulent payments before they happen.
While still in development, its principles will complement the FCA’s APP fraud guidance and the Home Office’s Data Strategy ambitions: to enable secure, proportionate information exchange that helps prevent fraud before funds leave the system.
How can the industry continue to build cross-sector collaboration?Cross-sector intelligence sharing and advanced data analytics are increasingly vital. Real-time, secure data exchange illuminates patterns, identifies emerging threats, and enables proactive intervention before attacks occur. When combined with strong governance and clear accountability, this kind of collaboration shifts fraud defense from isolated warning signs to a coordinated, system‑wide response.
Confirmation of Payee significantly reduces misdirected payments and various APP fraud types such as impersonation and invoice scams. It's a vital, preventative layer of security before funds are transferred. It has implications beyond its intended purpose and has strongly influenced the development of Verification of Payee in Europe.
There is a growing call for greater enforceable responsibilities for technology platforms and telecommunications providers, not only to prevent fraud at source, but also to contribute financially and operationally to combating it.
How can the industry empower and educate end users?Beyond technology and industry collaboration, fraud prevention has a vital human dimension. Educating and empowering end users remain central – recognizing the warning signs of a scam is still one of the strongest protections available. But education alone is not enough. Consumers also need better information at the moment a decision is made.
It’s encouraging to see that, as a payments community, we are already building richer data-sharing across the ecosystem to provide clearer, more relevant context when prompting customers to pause before making a payment. Banks are moving beyond generic warnings, providing genuinely useful guidance and strengthening the point of payment as a powerful, collective line of defense.
The APP reimbursement scheme is a significant consumer protection funded by UK banks. Data from the PSR shows that £215 million was reimbursed to victims of APP fraud in 2025 alone. Across the first 15 months of the scheme (October 2024 to December 2025), 89% of the money lost to APP scams has been successfully claimed back from a payment firm and returned to victims.
While not a direct comparison, this is a significant uptick from the 65% reimbursement rate reported by UK Finance for personal accounts in 2024. Providing a safety net of up to £85,000 for victims, this scheme offers a clear recovery mechanism and brings more consistency for customers than the previous voluntary Contingent Reimbursement Model (CRM) Code.
Further, the scheme continues to evolve in line with the shifting payment landscape. The PSR has appointed Frontier Economics to carry out an independent evaluation and review of the APP fraud policies, the results of which are due to be published in the second half of 2026.
These findings, which look at the current effectiveness of the policies, fraud performance reporting and the reimbursement requirement, will influence the future of APP fraud prevention strategies and regulatory requirements, ensuring the creation of safe and trusted payment infrastructure
The battle against payment fraud is ongoing, demanding constant vigilance and strategic adaptation. While the digital age has transformed how we transact, it has also presented fraudsters with new avenues for exploitation. Yet, as outlined, it is a battle we are actively and collectively winning.
By embracing a multi-faceted approach, combining robust technological defenses, seamless industry collaboration, effective regulatory frameworks, and comprehensive public awareness, we are building a formidable shield against these threats.
We've featured the best billing and invoicing software.
This article was produced as part of TechRadar Pro Perspectives, our channel to feature the best and brightest minds in the technology industry today.
The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/pro/perspectives-how-to-submit
In May, WordPress shipped the most consequential release in its history. Version 7.0 brought AI into the core of the CMS platform for the first time, and the people who built it made a choice that's easy to miss in all the noise about the feature itself. They left it switched off.
The infrastructure is in the codebase, but nothing reaches an AI service until the site owner connects a provider and turns it on. Upgrade a site and walk away, and it behaves exactly as it did the day before.
The on-switch was handed to the person who owns the site, not flipped on their behalf.
It's worth sitting with how deliberate that was. The team had just shipped the most powerful capability the platform has ever carried, and the posture they chose for it was opt-in, plugin-based, with nothing injected into anyone's site automatically.
In an industry that loves a sensible default, that restraint was itself a statement: this decision is yours to make.
Then the more interesting thing happened.
A reasonable instinct, taken one step too farWithin days of the release, SiteGround, one of the most established managed WordPress hosting companies in the business, did close to the opposite. It pre-installed and activated its own AI product across its customer base, configured it as the default connector, and bundled in a generous allowance of free usage to get people going.
The active-install count crossed a million almost immediately. Plenty of site owners logged in to find capable new software already running on sites they had never touched to put it there.
I want to be fair to SiteGround here, because fairness is where the useful lesson lives. This is a serious operator with a long, well-earned reputation, and the product it built is genuinely good, a real piece of engineering rather than a thin upsell. The reasoning behind the rollout isn't hard to reconstruct either.
The "correct" path to native AI is fiddly, and most people would stall somewhere in the middle and never finish it. Pre-installing the whole thing, free usage attached, removes that friction in a single stroke. From an operator's chair, that's a tempting piece of customer service, and I've sat in that chair for the better part of two decades. I understand the pull of it completely.
So this isn't a story about a company behaving badly. It's a story about a reasonable instinct (reduce friction, help the customer get to the good part faster) carried one step past the line. And the reaction told us exactly where that line is.
The objection wasn't AI. It was consent.
The pushbackThe pushback was quick and pointed, and the striking thing about it was its subject. Almost none of it was about whether AI belongs in WordPress, or whether the tool was any good. Many of the people objecting use AI every day. What they objected to was finding it already switched on.
That distinction matters more than it first appears, because it separates two things the industry tends to blur: the quality of a change, and the consent to it. A genuinely good feature, installed without asking, still lands as something done to you rather than for you.
The standard defense (it's optional, you can remove it whenever you like) is all true, and none of it is the same as agreement. "We switched it on and you can switch it off" quietly moves the work of noticing, understanding, and undoing onto the customer, for a change they never approved. "Here's one-click setup if you'd like it" delivers the identical convenience and leaves the decision where it belongs.
This isn't a new tension. Webhosting companies have always made changes customers never see, and most of the time they're glad we do. But AI is going to surface this question over and over, because it's the most consequential thing most of us will ever be tempted to switch on by default. Getting the principle right now, while the stakes are still mostly reputational, is a lot cheaper than getting it wrong later.
The line worth holdingThe honest objection to all of this is that hosts intervene on customer sites all the time, and nobody asks permission for that. True, and the distinction is the whole point.
When a host patches a vulnerability, blocks a malicious request, or disables a plugin that's being actively exploited, it's protecting the customer's site and the wider platform from harm. Customers extend us that trust precisely because it's defensive, narrow, and in their interest. Installing a new product is a different category of act.
It isn't protecting anything; it's changing what the website is. The trouble starts when the second borrows the permission we were granted for the first, when goodwill extended for security work quietly gets spent on shipping features. That's the line. Maintain the platform freely; change the product only with a yes.
Holding it doesn't mean making customers do more work. New capabilities can arrive off by default and one click away for anyone who wants them. Multi-site managers can get a single place to see and control what's running, rather than a hunt site by site.
Anything a host pushes can be pulled back as easily as it went out. And changes can be announced in plain language before they happen, including how to say no, because the absence of a clear, opt-out-inclusive heads-up is usually what turns an ordinary product decision into a breach of trust.
Parts of the ecosystem are already moving this way. None of it is anti-AI. If anything, it's what lets hosts lean into AI confidently, because customers can trust that nothing shows up uninvited.
Whose site is it, anyway?As AI moves from novelty to default across the web, every host will face its own version of this question. Here's a genuinely useful new capability. Do we switch it on for everyone, or do we let people choose? The convenient answer and the right answer won't always be the same one, and the gap between them is where reputations are quietly made or lost.
It helps to remember who actually lives with the answer. When a host changes something on a site, the host moves on to the next ticket. The owner is the one who stays: the one whose visitor hits a page that behaves differently than it did the day before, whose inbox fills up when something looks off, whose name is on the business the site exists to represent. We get to make the change. They have to live with it. That asymmetry, more than anything written into the terms of service, is the real reason asking first isn't a nicety. It's an acknowledgement that the consequences were never ours to carry in the first place.
The site owners who pushed back this spring weren't standing against progress. Most of them, I'd wager, will happily adopt the very tooling they objected to, the moment they get to be the ones who switch it on. They were defending something simple that's easy to lose sight of when the technology is moving this fast: it's their site. Not the site we host for them. Theirs. A host's authority runs right up to the edge of the customer's ownership and stops there, and the best operators I've worked alongside never needed reminding of it. They saw their role as stewardship rather than possession.
That trust is the real product. Not the servers, not the dashboard, not even the support, though every bit of it matters. What a customer is buying is the confidence that nothing happens to their site that they didn't choose, and that when we do step in uninvited, it's to protect what's theirs and never to quietly redraw it. Trust like that takes years to earn and an afternoon to spend. Asking first is simply how you keep from spending it.
Get the boundary right, and AI in hosting becomes exactly what it should be: useful, and genuinely welcome. Get it wrong, and even the best feature in the world arrives as something taken rather than offered. The difference was never the technology. It was only ever whether anyone thought to ask.
We tested, reviewed, and rated the best website builders.
This article was produced as part of TechRadar Pro Perspectives, our channel to feature the best and brightest minds in the technology industry today.
The views expressed here are those of the author and are not necessarily those of TechRadarPro or Future plc. If you are interested in contributing find out more here: https://www.techradar.com/pro/perspectives-how-to-submit