Nimble, a New York City-based tech startup VentureBeat previously covered for its efforts to re-invent web search for enterprises by using multiple AI agents to improve accuracy and depth, is taking another step toward its vision of a world in which agents do most of the web searching instead of us typing and reviewing the results manually.

Nimble today launched Web Search Agents, a new retrieval system designed to help AI agents perform more 21% more accurate web research while using significantly fewer tokens — 51% less compared with leading AI search alternatives on comparable, according to the firm.

While Nimble did not disclose its specific benchmarking methodology or competitors evaluated, the results underscore a growing trend in enterprise AI: optimizing retrieval has become as important as improving the underlying language models themselves.

Nimble’s leadership says the product combines self-learning retrieval strategies, proprietary web indexes, and live web access to deliver domain-specific search capabilities that outperform general-purpose web search services for enterprise workloads.

“Our research team built self-learning retrieval algorithms that learn a customer’s domain,” said Nimble CEO and co-founder Uri Knorovich in an interview with VentureBeat. “They find the exact information more efficiently, reduce the amount of multi-hop reasoning required, and lower token usage while improving accuracy.”

Rather than positioning itself as another general search engine, Nimble is targeting developers building autonomous agents that require continuously updated information from the public web for research, lead generation, competitive intelligence, compliance, and other business-critical workflows.

It’s also designed to slot in seamlessly to an enterprise’s existing systems and workflows.

“You can run the agent directly through the Nimble API with zero infrastructure,” Knorovich said. “For large enterprises, we’re partnering with Microsoft, Oracle, Snowflake, and others so customers can deploy these agent systems inside their own infrastructure.”

How does it work and stack up to other, existing AI-powered search and agentic systems? Read on to find out.

Moving beyond generic AI web search into specialized search agents that fit your enterprise’s needs

Most AI applications today rely on general-purpose search application programming interfaces (APIs) for search engines and public knowledge bases that return broad collections of files, leaving the language model responsible for determining which sources are relevant.

That process often requires multiple retrieval steps, additional reasoning, and significant token expenditure before an agent produces an answer. This is obviously inefficient and raises the cost spent to run AI search looking through irrelevant sources.

Nimble argues that before long, every enterprise will need its own methods for searching, retrieving, and validating external information since each enterprise relies on its own distinct preferred sources, signals, and standards of trust.

As such, instead of applying one search strategy to every workload, Nimble’s Web Search Agents are designed to learn the characteristics of a specific domain and adapt how information is retrieved, providing agents with structured, relevant context rather than forcing them to sift through large amounts of generic search results.

“Instead of one generic retrieval model, we build specialized retrieval models for each customer’s domain, making them faster, cheaper, and more accurate,” Knorovich explained. “A single enterprise can run hundreds of different agents. Each one has its own domain expertise, guardrails, goals, and search algorithm. The optimization starts with the second search, without requiring any setup from the customer.”

Its goal is not only to reduce redundant retrieval, but also to shorten multi-step research paths and avoid repeatedly sending raw pages through a language model for parsing, resulting in the 51% reduced token figure the company cites.

The distinction is particularly relevant for long-running enterprise agents performing research over hours or days rather than answering simple consumer questions. In those scenarios, reducing unnecessary tool calls can significantly lower operating costs while improving answer consistency.

That emphasis reflects a broader shift occurring across the AI tooling ecosystem. As foundation models become increasingly capable, infrastructure vendors are competing on everything surrounding the model—including retrieval, orchestration, memory, observability, and governance.

Optimizing retrieval for production AI

The launch builds on Nimble’s broader strategy of becoming an enterprise web intelligence platform rather than simply a web scraping provider. Earlier this year, the company introduced its broader Agentic Search Platform following a $47 million Series B financing, positioning itself as infrastructure that transforms the live web into structured, machine-readable data for AI systems.

The company’s latest release extends that vision with a concept it calls “Harness as a Tool,” which powers its new domain-specialized Web Search Agents. Rather than requiring engineering teams to assemble separate search APIs, browser automation, extraction pipelines, validation logic, memory systems, and orchestration code, Nimble packages those capabilities behind a managed interface.

The harness can determine what to search, navigate pages when conventional indexes are insufficient, extract relevant information, validate the results, and return the final context in a form designed for downstream agents.

Nimble also says the system retains domain-specific memory and builds proprietary indexes that improve as customers run more searches.

“The biggest research breakthrough is adding semantic memory and a caching layer to the agent,” Knorovich told VentureBeat. “The agent learns usage patterns and domain expertise over time, so every subsequent search becomes faster and more efficient.”

As for what domains Nimble can tackle, the company says it can address virtually any knowledge work domain.

“We’ve seen customers build investment banking analysts, competitive intelligence agents for product managers, go-to-market research agents, newsroom monitoring, insurance applications, life sciences research, and supply chain optimization,” Knorovich said. “Our customers surprise us every day with new agent use cases.”

However, for enterprises concerned about data privacy and retention, Knorovich assured VentureBeat that: “Nimble is zero-data-retention by design. Customer queries are never stored in our environment, and when customers deploy semantic memory and self-learning models, that knowledge stays in their own tenant—not ours.”

Customer deployments point to operational gains

Nimble supported the announcement with early customer examples from AI-native software vendors and enterprise users.

AI-native CRM company Rox reported achieving a 20× reduction in token costs after adopting Nimble’s retrieval infrastructure while simultaneously improving the quality and completeness of information available to its AI agents.

Although the company did not disclose detailed workload measurements or a reproducible baseline, the example illustrates the operational savings retrieval optimization can provide for high-volume agent deployments.

Nimble says its infrastructure currently supports more than 90 million searches each day across Fortune 500 enterprises and AI-native companies operating mission-critical workflows where accuracy, completeness, and enterprise control are essential.

API, SDK and MCP support target AI builders

The platform is immediately available through an API, SDK, and Model Context Protocol (MCP) integration, allowing developers to connect Nimble directly into AI agents regardless of the orchestration framework they use.

Developers can use the platform for several categories of web intelligence, including:

• Low-latency live web search

• Deep multi-step web research

• Web crawling

• Structured dataset generation

• Domain-specific information retrieval

The company also provides documentation and pre-built agents for common web extraction tasks while allowing developers to build custom retrieval agents using natural-language descriptions instead of manually maintaining scraping logic.

Nimble is offering two notably different consumption models. Developers can begin with a pay-as-you-go Agent API priced from $0.025 per Web Search Agent request at the listed low-effort setting. Companies that want Nimble to configure and manage custom data delivery can instead buy annual managed plans beginning at $2,500 per month.

Where Nimble fits in the emerging agentic search stack

Nimble enters a market that has rapidly expanded beyond traditional web search into autonomous research agents capable of planning, browsing, reasoning, and synthesizing information. Products such as ChatGPT Deep Research, Google Gemini Deep Research, Alibaba’s Tongyi DeepResearch, Perplexity, and Sakana Marlin all seek to automate knowledge work that previously required hours—or, in Marlin’s case, potentially weeks—of human research.

Rather than competing head-to-head as another end-user research assistant, however, Nimble is positioning itself one layer lower in the AI stack—as the web intelligence infrastructure that powers those agents or custom enterprise applications built on leading foundation models.

That distinction reflects an increasingly important architectural shift in enterprise AI. Most “Deep Research” systems optimize the overall research workflow, generating search plans, iteratively gathering information, and producing synthesized reports.

Nimble instead argues that the retrieval layer itself has become the primary bottleneck for enterprise AI deployments. If an agent retrieves too many irrelevant pages or performs unnecessary search iterations, token consumption, latency, and operating costs all increase before the model even begins its main reasoning process.

“Customers across life sciences, insurance, healthcare, pharma, retail, and digital-native companies are all telling us the same thing: we need to feed our agents with more accurate context, and we need to reduce the amount of tokens every task consumes,” Knorovich said.

The launch blog makes that argument more concrete by describing how teams frequently rebuild the same retrieval stack themselves. A production agent may start with a search API, then accumulate browser controls, parsers, extraction components, validation steps, memory, caching, evaluations, and custom workflow logic. Nimble is positioning its harness as a managed alternative to that growing engineering burden.

In Nimble’s view, improving retrieval before reasoning begins is more valuable than simply giving a language model more documents to analyze. The company’s Web Search Agents therefore adapt retrieval strategies to a particular workload, combining proprietary indexes with real-time web retrieval and task-specific search policies rather than applying the same search algorithm across every domain.

That makes Nimble less of a direct competitor to OpenAI’s or Google’s research assistants than to developer-focused retrieval infrastructure such as Exa and Tavily. Those platforms also provide AI-native search APIs and research capabilities, but Nimble differentiates itself by emphasizing self-learning retrieval strategies, proprietary indexing, enterprise governance, managed delivery, and token efficiency for production agents.

For organizations building their own AI systems, the distinction could become increasingly important. Foundation models are becoming more capable across the industry, shifting competitive differentiation toward the infrastructure surrounding them—including retrieval, orchestration, memory, observability, and governance. Nimble’s strategy reflects that broader trend, betting that better web intelligence can deliver larger operational gains than incremental improvements in model reasoning alone.

Enterprise infrastructure versus AI research assistants

The different positioning is also reflected in pricing.While consumer-facing AI research assistants are generally sold as productivity subscriptions for individual users or teams, Nimble is pricing its managed service as enterprise infrastructure designed to power production applications. Its pay-as-you-go API, however, gives developers a lower-cost path to test the underlying agent technology before committing to a managed deployment.

The comparison reveals three increasingly distinct markets.

1. ChatGPT Deep Research, Gemini Deep Research, and Perplexity operate primarily as user-facing research assistants.

2. Exa and Tavily provide developer-facing retrieval and research APIs.

3. Nimble and Sakana Marlin occupy more enterprise-oriented territory, but at different layers: Nimble supplies retrieval infrastructure, while Marlin performs long-horizon strategic analysis.

Sakana Marlin is particularly useful as a counterpoint. It is positioned as a “Virtual CSO” rather than a search API, running autonomous research loops for as long as eight hours and producing executive-ready reports, references, and supporting materials.

Nimble, by contrast, is designed to sit beneath those kinds of systems, supplying the specialized retrieval, browsing, extraction, validation, and orchestration that enterprise agents need to gather reliable external information before reasoning begins.

The comparison therefore should not be read as a direct price-to-price evaluation. A $20/month ChatGPT Plus or $19.99/month Google AI Pro subscription buys an individual AI workspace with Deep Research capabilities.

Nimble’s $2,500/month managed plan funds concurrent production agents, managed ETL, MCP integration, web-page capacity, storage, and hands-free data delivery.

Sakana Marlin’s approximately $936/month (¥150,000/month) Pro plan pays for extended, compute-intensive strategic research workflows.

Each price reflects a fundamentally different product boundary and deployment model rather than simply a different level of AI capability.

Why retrieval is becoming the next AI battleground

As enterprise AI systems mature, the industry is increasingly recognizing that model quality alone does not determine application performance.

Large language models frequently fail not because they cannot reason, but because they lack timely, trustworthy external information. That reality has fueled rapid investment across retrieval-augmented generation, AI-native search, web intelligence platforms, knowledge graphs, browser automation, and agent infrastructure.

Nimble’s launch reflects this evolution by focusing less on building another frontier model and more on improving the quality of information flowing into existing ones.

Whether the company’s reported 21-point improvement in answer quality and 51% reduction in token usage hold up across a broad range of enterprise deployments remains to be independently validated.

The larger strategic bet is that, as frontier models become more interchangeable, companies will differentiate themselves through the data, retrieval policies, trusted-source rules, memory systems, and orchestration layers surrounding those models. Nimble is not trying to build the researcher that sits in front of the user. It is trying to become part of the infrastructure that determines what the researcher can find, how efficiently it can find it, and whether the resulting evidence is complete enough to support production decisions.

Web Search Agents are available through Nimble’s API, SDK, and MCP integrations, with a free trial available for developers evaluating the platform.

from the and-by-“efficiencies”-we-mean-“destruction” dept

Quick background: the $8 billion FCC Universal Service Fund (USF) applies a small surcharge on traditional phone lines to fund broadband expansion to unserved rural homes, schools, and libraries (of which the U.S. has a lot thanks to rampant telecom monopolization).

While the USF, like all government programs, hasn’t been absent of fraud (usually at the hands of private companies), it’s generally done a lot of good for people stuck on the wrong side of the digital divide. The kind of boring, steady infrastructure work that doesn’t get headlines, or mentioned in books about “abundance.”

Enter Trumpism, which has been taking a merciless hatchet to absolutely any effort, anywhere across government, to try and ensure that Americans — whether rural Trump supporters, struggling school kids, or inner city urban residents — have affordable access to the internet.

The USF is technically overseen by the Universal Service Administrative Company (USAC), which has been tasked since 1998 with overseeing more than $9 billion in broadband subsidies annually. Now FCC boss Brendan Carr has announced a full review of the program, something he says is necessary for the sake of efficiency:

“In taking up this effort, we will improve oversight, reduce administrative costs, and increase
accountability of USAC and its Board of Directors. Our end goal is to ensure Americans receive the best bang for their buck on universal service spending—a commonsense win for government efficiency and accountability.”

This being Trumpism and Brendan Carr, there are obviously red flags in the way they’re going about this.

One being that the Trump administration’s version of “efficiency” and “accountability” — as we saw with DOGE — generally involves the mindless dismantling of useful programs by clowns who have very little functional human empathy or understanding of the things they’re destroying.

One other issue is that Carr historically has never meaningfully opposed giant telecom monopolies like AT&T on any policy battle of note, so the idea that he’s seriously going to root out fraud and abuse of the USF program is laughable. At best it’s likely that the program is retooled to ensure big telecom monopolies get more money in exchange for significantly less subsidy oversight.

Certain Trump Republicans had been pushing for the USF to be destroyed entirely, though that gambit was surprisingly scuttled by the Supreme Court last year. So now the name of the game for that sector of Trumpism will be to destroy the program while trying to make it seem like they’re not destroying the program. Especially when it comes to helping minority communities afford internet access.

Republicans have taken a hatchet to broadband affordability programs across the board, including killing the FCC’s Affordable Connectivity Program (ACP), which provided a $30 broadband discount for low-income Americans — as well as killing a program that provided free Wi-Fi to rural school kids at no additional cost to taxpayers.

The Trump admin also illegally dismantled the Digital Equity Act, which was a bare-bones effort to stop race and class discrimination in broadband upgrades, hijacked billions in infrastructure bill broadband grants to the benefit of Elon Musk, and destroyed what’s left of U.S. federal consumer protection and corporate oversight, ensuring that U.S. ISPs face zero meaningful penalties should they rip you off.

You know, the real flyover country populism-type stuff everybody was clamoring for.

Filed Under: brendan carr, broadband, fcc, telecom, usf

The Federal Trade Commission is suing healthcare giant Hims & Hers for allegedly sharing its customers’ medical and healthcare information with advertisers and tech giants, like Meta and Snap, as well as for misleading consumers about its privacy practices.

The lawsuit is the federal consumer watchdog’s latest crackdown in recent years on healthcare companies that share sensitive information with outside companies without their customers’ knowledge. Hims & Hers, now a publicly traded company, provides prescription medication for sexual wellness, mental health conditions, weight loss, and other issues, and as such handles a large amount of sensitive patient data.

Companies typically try to learn more about their customers by installing code on their websites in order to share users’ information with advertisers, like Meta and Snap, which then use the data to provide information about who is visiting their websites and when. 

In its complaint filed in a California federal court, the FTC alleged that Hims & Hers placed pixel-sized trackers provided by Meta, Snap and other tech and advertising giants, including Microsoft, Pinterest, Reddit and X. These trackers, the FTC said, “captured and shared users’ health information,” contrary to Hims & Hers’ own privacy policy. The FCC alleges the company also used Meta’s tools to track users’ clicks and other actions that users took on its website.

The FTC also accused Hims & Hers of deceptive billing, and drawing up policies that allegedly made it difficult for customers to cancel, in violation of federal consumer protection laws.

Hims & Hers did not explicitly deny the FTC’ s claims in a statement on its website. The company claimed its privacy policy “makes clear” that users “may choose how their data is used,” and said it is “confident” in its position. Hims & Hers said it plans to defend against the FTC’s allegations.

The FTC has previously taken action against telehealth startup Cerebral, alcohol recovery provider Monument, as well as data giant GoodRx and therapy provider BetterHelp. These companies were similarly accused of sharing patients’ sensitive data through their websites to third-party tech giants and advertisers.

The use of pixel-sized trackers has previously revealed how people’s sensitive data gets shared with the companies who provide the code, highlighting how misconfigurations can result in unwanted data collection.

In 2024, TechCrunch found that the U.S. Postal Service was sharing logged-in users’ home addresses with Meta, LinkedIn and Snap by using their pixel tracking code. The USPS removed the code soon after.

For listeners with smaller rooms, loudspeaker shopping often comes down to two mildly absurd choices: buy standmounts and discover the stands consume nearly as much floor space as towers, or buy full-size floorstanders and watch your living room get annexed by MDF.

Q Acoustics has finally inserted something between those two choices.

The new Q Acoustics 3040c is a compact floorstanding loudspeaker designed to deliver more scale and bass extension than the 3030c standmount without the size and visual presence of the larger 3050c. Priced at £599, €899 or $1,099 per pair, it becomes the sixth passive loudspeaker in the 3000c range and completes a lineup that now covers everything from desktop and bedroom systems to multichannel home theaters. 

There is, however, an interesting wrinkle for American buyers: the 3040c costs only $100 less than the larger 3050c.

This is not the floorstander to buy because it saves a pile of money. It is the one to consider because it saves space.

Related Reviews:

The Missing 3000c Model

When Q Acoustics introduced the 3000c Series in 2024, the range included three bookshelf or standmount models, one large floorstander and a center-channel speaker. The original lineup incorporated technology developed for the more expensive 5000 and Concept Series, including C3 Continuous Curved Cone midrange and bass drivers, mechanically isolated tweeters, P2P cabinet bracing and HPE pressure equalization in the floorstanding model. 

What it did not include was a smaller tower.

The 3040c fills the rather obvious gap between the $699 per pair 3030c and $1,199 per pair 3050c. According to Q Acoustics, customer feedback indicated demand for a more compact and agile floorstander that could deliver greater scale than a bookshelf speaker without overwhelming tighter living spaces. 

That makes sense. A standmounted 3030c may have a smaller enclosure, but once suitable stands are added, its practical footprint is not necessarily much smaller. The 3040c places everything into one cabinet, moves the tweeter closer to seated ear height and extends the claimed low-frequency response to 43 Hz.

It also means there is no stand shopping, no filling hollow metal columns with sand and no discovering that the attractive stands cost almost as much as the loudspeakers.

Twin C3 Drivers

The two-way 3040c uses a pair of 4.75-inch C3 Continuous Curved Cone midrange and bass drivers positioned above and below a 0.9-inch tweeter.

Q Acoustics originally introduced its C3 driver architecture with the 5000 Series. The cone profile combines the bass performance of a conventional straight cone with the midrange control and dispersion characteristics of a flared design. The goal is tighter low-frequency performance, reduced distortion and smoother integration with the tweeter. 

The 3040c’s tweeter is hermetically sealed and mechanically isolated from the front baffle. Separating it from pressure changes and vibration generated by the two larger drivers is designed to reduce distortion and prevent the midrange and bass drivers from coloring the treble.

That approach originated higher up the Q Acoustics food chain in the Concept and 5000 Series before migrating into the more affordable 3000c models. The company has not merely installed another woofer in a stretched 3020c cabinet and headed to the pub.

Bracing and Pressure Control

Compact towers present their own cabinet problems. Their tall internal volume can encourage standing waves and pressure buildup, neither of which does imaging, bass definition or tonal accuracy any favors.

Q Acoustics addresses that with P2P Point to Point bracing in the areas of the enclosure most susceptible to low-frequency vibration. The 3040c also uses HPE Helmholtz Pressure Equalizer tubes to reduce internal pressure and suppress standing waves, technology previously reserved for the 3050c within the 3000c range. 

The rear-ported cabinet includes the low-profile terminal panel introduced with the 5000 Series. Its recessed binding posts help keep speaker cables from protruding excessively behind the cabinet, although “close to the wall” should not be confused with “pressed against it.” Rear ports still need room to work.

A solid aluminum stabilizer extends from the base and includes spikes that can be adjusted from above. Including the stabilizer and spikes, each 3040c measures 37.1 inches high, 11.2 inches wide and 11.1 inches deep, with a weight of 32.1 pounds. 

Four finishes are available: Satin Black, Satin White, Pin Oak and Claro Walnut.

What Might It Sound Like?

We have not heard the 3040c, so any declarations about its sonic performance will have to wait until a review pair arrives.

We do, however, have considerable experience with the technology surrounding it.

Our review of the 3020c found it more transparent, detailed and dynamically alert than the softer-sounding 3000i models it replaced. It also proved more revealing of amplifier quality and tonal balance than its affordable price might suggest. 

The 5000 Series demonstrated how much flexibility exists within the same basic driver platform. The 5040 offered a leaner, more neutral and highly detailed presentation, while the larger 5050 delivered greater bass weight, warmth and room-filling impact. 

That history gives us a useful framework, but not a verdict. The cabinet volume, crossover and twin-driver configuration will determine whether the 3040c leans toward the speed and clarity of the smaller 3000c models or adopts some of the additional tonal weight of the 3050c.

On paper, its 88 dB sensitivity, 6-ohm nominal impedance and 4-ohm minimum should make it compatible with a broad range of integrated amplifiers and A/V receivers. Q Acoustics recommends between 25 and 120 watts per channel.

A competent 50 to 100-watt amplifier should be a sensible starting point, but our experience with the 3020c and 5000 Series suggests that amplifier character will matter. Buyers using brighter or leaner electronics should audition the combination rather than assuming every box with binding posts will produce the same result.

Stereo and Home Theater

The 3040c could be particularly useful as the front left and right channels in a compact home theater, paired with the 3090c center speaker and one of the smaller 3000c models for surround duty.

Its claimed 43 Hz extension is respectable for a tower this size and may be sufficient for music in a small or medium-sized room. It does not eliminate the need for a subwoofer when reproducing deep electronic bass, pipe organ recordings or movie soundtracks with meaningful low-frequency effects.

The existing 3060S remains Q Acoustics’ slimline subwoofer option for the range, although larger rooms and more ambitious home theaters may benefit from something with greater output and deeper extension.

Specifications:

Who Is This For?

The Q Acoustics 3040c makes the most sense for listeners who want the scale and convenience of a floorstanding loudspeaker but have a smaller living room, den, apartment or dedicated listening space. It also deserves consideration from 3030c buyers who would otherwise need to purchase stands and might prefer a cleaner, more integrated installation.

Home theater users looking for compact front channels that visually and acoustically match the rest of the 3000c family are another obvious audience.

Who Should Look Elsewhere?

Listeners with larger rooms, a preference for high playback levels or a need for greater low-frequency weight should consider the 3050c, 5040 or 5050.

American buyers with enough room for the 3050c face the most difficult decision. At only $100 more, the larger tower may offer greater scale and bass capability for very little additional money. The 3040c has to win that argument through placement flexibility, proportions and room compatibility rather than price.

Q Acoustics 3000c Series Pricing

Stereo loudspeaker prices are per pair. The 3090c center channel and 3060S subwoofer are priced individually.

• Q Acoustics 3010c compact bookshelf: £249 / €379 / $399
• Q Acoustics 3020c bookshelf: £299 / €499 / $549
• Q Acoustics 3030c standmount: £399 / €629 / $699
• Q Acoustics 3040c compact floorstander: £599 / €899 / $1,099
• Q Acoustics 3050c large floorstander: £749 / €1,099 / $1,199
• Q Acoustics 3090c center channel: £229 / €379 / $399
• Q Acoustics 3060S subwoofer: £399 / €499 / $549

The Bottom Line

The Q Acoustics 3040c is not merely a discounted 3050c. It is a space-specific alternative for listeners who want more than a standmount can provide but do not need a large tower looming beside the television.

That makes it more useful than a last-minute addition designed to fill an empty position in the catalog. The 3000c Series already offers some of the strongest value in the affordable loudspeaker category, and the 3040c addresses a legitimate gap without stripping away the driver and cabinet technology that made the rest of the range successful.

The pricing is particularly attractive in the United Kingdom and Europe. In the United States, the narrow gap between the 3040c and 3050c makes room size the deciding factor.

Smaller can be smarter. It just is not automatically cheaper.

Related Reading:

Comparing Our Favorite Vacuum Cleaners

Honorable Mentions

Bissell PowerClean FurFinder for $260: This was our previous top pick for the best cordless vacuum. It got bumped though because the regular Bissell PowerClean ($200) is a touch cheaper since it doesn’t come with the FurFinder upholstery attachment, while Dyson and Ryobi’s vacuums have more powerful suction for pet hair.

Black and Decker Dustbuster Flex for $130: This is another cool handheld vacuum that’s great for cars or even indoor areas like staircases. It has a 4-foot hose, longer than most compact vacuums, and a handy little charging mount that the accessories clip into. Thank goodness for the charging mount, though, since the battery only lasts 15 minutes.

Dyson Spot+Scrub Ai Robot Vacuum for $1200: This is Dyson’s first robot vacuum that doubles as a mop, and it serves both purposes well. It uses the same app as other Dyson vacuums and can learn multiple floors of your home. The base station is large and unique, with a visible dustbin that’s certainly a callback to the clear dustbins of the brand’s cordless vacuums. Dyson says it has artificial intelligence to spot stains and scrub them away, but I couldn’t tell a difference and it didn’t spot the cherry stains I intentionally created. It’s also tall, so it was often bumping into cabinets and low furniture.

Dyson V8 Cyclone for $400: This is a new version of the long-favored Dyson V8, but I wasn’t super impressed when testing it side by side with the only slightly more expensive V10 Konical. It did a fine job and has 150 air watts of power, but I’d recommend the V10 over the V8 any day.

Dyson V10 Konical for $500: I expected to prefer the V8 Cyclone, but the V10 Konical did a better job vacuuming up debris in my tests, comes with a HEPA filter, and has a built-in light so you can better spot dust and dirt when using it on hard floors. When tested and scored against all of the new and current Dyson vacuums, the V10 Konical was in third place, ahead of both the V8 Cyclone and the V15 Detect. It’s a great buy if you want a Dyson vacuum but don’t want to manage multiple heads, and it will be compatible with Dyson’s base station that comes out later this year.

Dyson V15Detect for $846: There are a lot of Dyson vacuum models, and the older V15 Detect is a good choice if you want two vacuum heads included without spending close to a grand. There’s an option to upgrade and get a Submarine wet roller head if you want it to double as a mop. It performed solidly in testing against Dyson’s other vacuums, but wasn’t the best. I’d recommend looking for it on sale for under $700; anything higher and there’s no point in not buying the Gen5Detect or V16 Piston Animal instead.

Dyson PencilVac Fluffycones for $600: This is a new take on a stick vacuum with the motor and dustbin built into the slender handle. There’s both a Fluffycones version, which I tested, that has four fluffy cone-shaped rollers in the head, and a cheaper Fluffy version. I was impressed with the maneuverability of the PencilVac Fluffycones and often find myself grabbing it between runs of my robot vacuum to keep debris and cat litter under control. It’s also lightweight and easy to push, and it has a freestanding charging mount. But it’s a little pricey for being only for hard floors.

Ecovacs Deebot X11 Omnicyclone for $1100: If you want a newer robot vacuum, the Ecovacs Deebot X11 Omnicyclone has a unique design with no dust bag. Instead, it has a rounded canister like a Dyson or stick vacuum, circling the debris to keep it from tangling. It’s a great robot vacuum overall, especially if you don’t want to buy dust bags over and over.

Eufy X10 Pro Omni for $700: This was our previous top robot vacuum pick. It’s a well-priced option that can vacuum and mop, and has a solid 8,000 Pa of suction power. But Eufy does have a newer model that you can pick up for just a little more if you want a self-cleaning roller mop instead of roller pads, with almost double the suction power.

Roborock Saros 20 for $1,600: Roborock’s latest robot vacuum is overall pretty good, but WIRED reviewer Adrienne So noticed it left debris behind on her hard floors, specifically on the divider when it moved from her carpet onto hard flooring. This could be a fault of the super-high suction, which hits 36,000 Pa. There are a few other features she hoped would work better, too, including the AdaptiLift (which raises the vacuum up on uneven surfaces) and beta stain-detection feature. We’re in touch with Roborock about these issues.

Shark PowerDetect Cordless Vacuum for $418: This is the original PowerDetect vacuum I recommended, with 380 watts of power, three cleaning models, two accessories, and an option to upgrade to include a self-emptying vase station. It can bend to get under furniture and often outpaced Dyson vacuums in my tests, especially when cleaning up sand or cereal on carpet. I prefer the Shark PowerDetect Speed a little more since it’s cheaper and has more options for its base station, but either vacuum is a smart buy.

Tineco Pure One Station 5 for $459: If you want a cordless stick vacuum but don’t want to deal with emptying it all the time, this Tineco vacuum comes with a self-emptying docking station. You’ll eventually have to empty the station, but it’s a great bonus feature and keeps the vacuum from falling over around your home since it’s stored safely in the docking station.

Worx Cordless Cube Vacuum for $120: This adorable, cube-shaped vacuum is one of my favorite handhelds, and is my go-to for vacuuming the car since it’s so easy to hold. It can also store all of its accessories inside itself, and the hose is four feet long to reach your entire carpeted car trunk.

Power up with unlimited access to WIRED. Get best-in-class reporting and exclusive subscriber content that’s too important to ignore. Subscribe Today.

Bottom line: While Microsoft continues to ride the AI wave, Xbox continues to struggle. New CEO Asha Sharma recently began implementing plans for a radical course correction, but it is still too soon to predict their impact. The company’s fourth-quarter financial results for 2026 show a 10% fall in Xbox content and services and a 13% decline in hardware revenue, while Microsoft’s overall profits jumped by 31%.

While Xbox’s 13% YoY drop in hardware revenue is not as bad as last quarter’s 33% plummet, it still caps off a full year of uninterrupted declines. Price hikes and worrying PR surrounding the brand are likely the primary factors, and the situation is expected to worsen when Xbox Series S|X consoles become at least $100 more expensive starting in August.

Xbox content and services revenue also fell by 10% YoY this quarter, compared with a 5% decline last quarter, which Microsoft again attributed to strong performance during the same quarter last year. The company is likely still referring to titles such as Call of Duty: Black Ops 6 and Minecraft, but this past quarter saw the release of one of Microsoft’s most successful titles, Forza Horizon 6.

– ASHA (@asha_shar) July 29, 2026

During the Q4 earnings call, Microsoft CEO Satya Nadella expressed hope for Xbox to return to profitability in fiscal year 2027. Asha Sharma, who replaced Phil Spencer as Xbox CEO in February, has already begun implementing changes to that end, some of which are controversial.

To boost hardware sales, the company is partially reversing its policy of releasing in-house titles, such as Gears of War E-Day, on rival platforms. Microsoft also confirmed a renewed push for major franchises such as Fallout and The Elder Scrolls. However, rolling layoffs affecting 3,200 positions this year and the loss of five game studios have drawn sharp criticism and sown doubts about future game development at Xbox.

Meanwhile, the AI boom continues to lift Microsoft’s overall fortunes. Revenue increased 18% YoY, and profits rose 31%, both in Q4 and for FY26 overall, mostly due to cloud and AI services. The company announced that Azure revenue exceeded $100 billion for the first time this year.

Although tech giants remain bullish on AI, the boom’s longevity remains unclear. Apple recently regained its status as the world’s most valuable company and joined the $5 trillion club, largely because it has kept AI at arm’s length, causing investors to see it as an island of stability.

AMD’s Ryzen 7 7800X3D still tops plenty of gaming benchmarks against newer chips, and it has just dropped to one of its lowest prices yet.

That price is $339.99 at Amazon, down from its $449.00 list price, a saving of $109.01, or 24%, on the Ryzen 7 7800X3D, the chip our tech expert called nothing short of impressive in his review.