FlashFeed

Full text

Across 101 enterprises, agent orchestration is consolidating onto model-provider platforms — Anthropic’s Claude leads by a wide margin — chosen for the gravity of the underlying model and judged on reliable multi-step execution. But the ambition runs well ahead of the reality: most deployed “agents” are still chatbot wrappers, the control plane enterprises expect is deliberately hybrid to avoid lock-in, and real-time fiscal control over token burn remains the exception.

This wave of VentureBeat Pulse Research examines enterprise agent orchestration: which platforms enterprises run on, what drives the choice, what they optimize for, how they expect agent control to be structured, and — most revealingly — how orchestrated their deployed “agents” actually are and how tightly they control the cost of running them.

The central finding is a gap between orchestration ambition and orchestration reality. Enterprises are consolidating fast onto the major model platforms: Anthropic’s Claude is the primary platform for 40%, more than double any rival, followed by Microsoft (18%) and OpenAI (13%). The choice is driven by “model gravity” — native alignment with a state-of-the-art base model (21%) — and success is judged by reliable, multi-step execution (task completion reliability 32%, multi-step workflow management 28%). Yet asked to assess their portfolios honestly, 71% say a quarter or fewer of their deployed “agents” are true multi-step orchestrated workflows rather than single-prompt chatbot wrappers, and only 10% have crossed the halfway mark. The orchestration layer is being built well ahead of the orchestrated portfolio it is meant to run.

That gap shapes the architecture enterprises are putting in place. By the end of 2026 a clear majority (51%) expect a hybrid control plane — provider-native plus external orchestration — and only 6% expect to hand control to a provider-managed service, because vendor lock-in (35%) is the risk they fear most if control lives inside a model provider. Investment follows the build-out: agent workflow tooling leads the spend (34%), with security and permissions enforcement (25%) behind. And fiscal control lags throughout — more than a quarter (27%) have no real-time way to stop a runaway agent before the bill arrives.

Methodology

VentureBeat fielded this survey as part of its ongoing Pulse Research series, this instrument focused on enterprise agent orchestration. Responses are filtered to organizations with 100 or more employees (n=101), drawn from a single June 2026 wave; because this is one wave rather than a pooled multi-month sample, the report reads cross-sectionally and does not infer month-over-month trends.

By organization size the sample is spread evenly across the enterprise bands: 100–499 employees, 2,500–9,999, and 50,000+ (21% each), with 10,000–49,999 and 500–2,499 (19% each). By role it is senior and buyer-credible: product and program managers (15%), CIO/CTO/CISO (13%), consultants and advisors (13%), and a spread of data, AI, and engineering directors and VPs, with an “Other” function at 18%. On purchasing, 81% are recommenders, influencers, or final decision-makers for AI solutions (66% recommender/influencer, 15% final decision-maker). Technology/Software is the largest industry at 44%, followed by Financial Services (17%) and Healthcare/Life Sciences (8%).

At 101 respondents the sample is robust enough to read directionally with reasonable confidence, though it remains self-selected and is not a probability sample.

Finding 1: Orchestration runs on model-provider platforms

Anthropic’s Claude leads; open frameworks are marginal

We asked which agent orchestration platform enterprises primarily use today. The answer concentrates on the major model providers — and on one in particular.

A note on reading these shares. As described in the methodology section, the respondents are self-selected, and this question asked them for a single primary platform — so the figures measure which platform leads each enterprise's deployment, within a self-selected audience of AI-active technical decision-makers. A sample built this way can diverge substantially from spend-weighted market measures, and each VB Pulse survey draws its own sample with its own company-size mix, so vendor figures should not be compared across our surveys either. Read these shares as a portrait of where this cohort has placed its primary orchestration bet today, rather than as market share.

The model platforms dominate. Anthropic, Microsoft, OpenAI, Google, and Amazon together account for roughly 80% of deployments (81 of 101), while the open frameworks (LangChain/LangGraph) and custom in-house builds that anchor engineering discussion sit in single digits. Anthropic’s lead — 40%, more than double the next platform — mirrors the “model gravity” selection logic in Finding 2: enterprises are choosing the orchestration layer that comes with the model they want to build on. As with the security vendors in the prior agent-security wave, the tools that define the category in technical circles are not yet where enterprise deployment concentrates. A small 3% are not orchestrating at all.

Respondents rate the platforms they run at 3.94 out of 5 overall (109 answered), with “value for money” specifically at 3.94 and “ease of implementation” the weakest score, at 3.85 — placing orchestration near the bottom of our five-tracker satisfaction range, ahead of only evaluation tooling. A rating just under 4 out of 5, from users of whom 96% plan to change their orchestration approach within the year, reads as provisional acceptance: the platforms work well enough to run today, and not well enough to stop the search for something better. The ratings sit alongside near-universal intent to change; this is a layer enterprises tolerate more than they love.

Finding 2: Model gravity drives platform selection

The base model, not the tooling, decides the platform

We asked what most influenced the orchestration platform choice. The single largest factor is the pull of the underlying model — though flexibility and ease of development follow close behind.

Model gravity leading is the selection-side explanation for Anthropic’s platform lead: enterprises pick the orchestration environment closest to the frontier model they have standardized on. But the next tier complicates the picture — flexibility across models and tools (17%) and ease of development (17%) say enterprises also want to avoid being trapped by that choice, foreshadowing the lock-in fear in Finding 6. Security and permissions (14%) and total cost of ownership (11%) round out a pragmatic buying logic. Performance (latency/memory) sits last at 4%, a reminder that at this stage of adoption the binding constraints are model fit and optionality, not raw speed.

Finding 3: The job is reliable multi-step execution

Enterprises just orchestration by whether it completes the work

We asked what enterprises optimize for — their primary success metric for orchestration. Reliability and multi-step workflow management dominate; developer- and user-facing metrics trail.

Task completion reliability (32%) and multi-step workflow management (28%) together account for 59% of responses (60 of 101): orchestration succeeds, in the enterprise view, when it reliably carries a task through multiple steps to completion. Developer productivity (17%) matters but is secondary — the inverse of its prominence in framework discussion — and end-user experience (9%) is a minor concern, consistent with orchestration being an internal execution problem rather than a UX one. This reliability-first standard is exactly what makes the Chatbot Trap finding so pointed: enterprises define success as dependable multi-step execution, yet most of their deployed “agents” do not yet do multi-step work at all.

The trap is not evenly distributed. Splitting the sample by organization size, 77% of smaller enterprises say a quarter or fewer of their agents do true multi-step work, against 62% of larger ones. Larger enterprises are meaningfully further into genuine multi-step deployment; the chatbot trap is, directionally, a mid-market condition.

Finding 4: Consolidate, productionize, and build in-house

Three strategic moves are nearly tied for the year ahead

We asked what major change enterprises anticipate in their orchestration strategy over the next 12 months. Three moves cluster at the top, almost evenly split.

The top three — building in-house control (25%), standardizing on one framework (24%), and moving agents from sandbox to production (23%) — are statistically indistinguishable and tell a single story: enterprises are moving from experimentation to operational consolidation. They want fewer frameworks, more production exposure, and more ownership of the control layer; only 4% expect no change. The appetite for custom in-house control planes is notable alongside the platform concentration in Finding 1 — enterprises are standardizing on model-provider platforms while simultaneously planning to wrap them in control logic they own, the hybrid posture that Finding 6 makes explicit.

Finding 5: Investment flows to workflow tooling

Tooling and permissions lead the spend; monitoring trails

We asked which orchestration-related investment will grow most next year. Agent workflow tooling leads, with security and permissions enforcement behind.

Workflow tooling leading (34%) is the budget-side expression of the reliability-and-multi-step priority in Finding 3: the money is going to the machinery that strings steps together dependably. Security and permissions enforcement (25%) and scaling infrastructure (20%) follow — the investments required to take agents from sandbox into production, the strategic move in Finding 4. Monitoring and debugging draws a smaller 11%, with another 11% reporting flat budgets. The weight on tooling, permissions, and scaling over pure observability signals that enterprises are spending to build and harden orchestration, not merely to watch it run.

Finding 6: The control plane will be hybrid — and lock-in is why

Enterprises expect to split control between providers and their own layer

We asked where enterprises expect the primary control plane for agents to live by the end of 2026, and what worries them most if that control sits inside a model-provider platform. A clear majority expect a hybrid model — and vendor lock-in is the reason.

Hybrid control is the dominant expectation by a wide margin (51%), and only 6% expect to hand control to a provider-managed service outright. Read together, the hybrid, custom, and externally-abstracted options — every architecture that keeps control at least partly outside the provider — sum to 88% (89 of 101). The reason surfaces directly when we asked about the risk of provider-resident control: vendor lock-in leads at 35% (35 of 101), ahead of security and permissioning limitations (28%) and inflexibility across models and tools (21%). The pattern echoes the prior wave’s “don’t trust the model to police itself” posture — here, enterprises will build on a provider’s platform but decline to be governed entirely by it. The hybrid control plane is the architectural hedge against the lock-in they most fear.

The June figure asserting a preference for a hybrid control plane marks movement from earlier. In the April–May survey (n=145), only 34% expected a hybrid control plane, and a greater number (12%) expected to hand control fully to a provider-managed service. These two snapshots don’t yet measure a confirmed longitudinal trend — but the direction of the conversation is unambiguous: toward keeping control.

Lock-in is also a new arrival as a top concern. In the April–May wave, the leading concern was security and permissioning limitations (32%), with lock-in second at 24%; by June the two had traded places. The worry about provider platforms appears to be maturing from whether they can be secured to whether they can be replaced.

Finding 7: The chatbot trap — most “agents” aren’t agents yet

Enterprises admit most deployments are still chatbot wrappers

We asked enterprises to assess their portfolios honestly: what share of their deployed “agents” are true multi-step orchestrated workflows versus simple single-prompt chatbot wrappers. The answer is the defining finding of this wave.

This is the gap at the center of the report. Combining the bottom two bands, 71% of enterprises (72 of 101) say a quarter or fewer of their deployed “agents” are genuinely orchestrated — and just 10% (10 of 101) have crossed the halfway mark. The ambition documented in the earlier findings — model-provider platforms, reliability-first success metrics, production rollouts, a deliberate control architecture — runs well ahead of the deployed reality, which remains overwhelmingly single-prompt assistants dressed as agents. This is less a contradiction than a roadmap: the platforms, budgets, and strategies are being put in place precisely because the orchestrated portfolio is still so thin. The open question for later waves is how fast the reality closes on the ambition.

Finding 8: Fiscal control is still reactive

Only a minority can stop a runaway agent before the bill arrives

Finally, we asked how enterprises enforce fiscal control over agent token consumption — the risk that an autonomous loop exhausts a budget before anyone intervenes. Most rely on native caps or after-the-fact monitoring; real-time programmatic control is the exception.

More than a quarter of enterprises (27%) admit they have no real-time, programmatic way to stop an agent before a budget-breaking bill arrives — they learn of it from the logs afterward. Another 32% lean entirely on the native caps and throttles built into their primary platform, a control only as good as the provider’s tooling and one that ties back to the lock-in concern of Finding 6. The enterprises building custom gateways (23%) or exploiting cross-model routing to arbitrage cost (19%) are the ones treating token burn as an engineering problem to be controlled deterministically. As with orchestration maturity, fiscal control is an area where the operational reality lags the ambition: agents are moving toward production faster than the cost-control plane around them is being built.

It’s worth noting, a split appears according to company size: roughly one in three enterprises under 2,500 employees (34%) exercises only reactive control of agent spend, against 20% of larger enterprises — directional figures, but consistent with the chatbot-trap split. The mid-market is running the least mature agents on the least instrumented budgets.

The bottom line: The layer is real; most of the agents aren't yet

Organizations with 100 or more employees describe an orchestration strategy that is consolidating quickly and maturing slowly. They are standardizing on model-provider platforms — Anthropic’s Claude leads at 40% — chosen for the gravity of the underlying model, and they judge success by reliable multi-step execution. Investment is flowing to workflow tooling and permissions, the strategy is to consolidate frameworks and push agents into production, and the control plane they expect is deliberately hybrid, because vendor lock-in is the risk they fear most.

But the honest self-assessment punctures the ambition. Seventy-one percent say a quarter or fewer of their deployed “agents” are truly orchestrated, only 10% are past the halfway mark, and more than a quarter cannot stop a runaway agent in real time. The orchestration layer — the platforms, the budgets, the control architecture — is being built ahead of the orchestrated portfolio it is meant to run. At 101 respondents in a single June wave this reads as a clear directional signal rather than a precise measurement: enterprises have decided how they want to orchestrate agents well before most of their agents are doing anything an orchestration layer is for. The question for subsequent waves is whether the deployed reality closes the gap on the ambition — or whether the chatbot trap proves stickier than the roadmap assumes.

Based on survey responses from 101 qualified enterprise respondents (100+ employees), drawn from a single June 2026 wave. Because this is one wave rather than a pooled multi-month sample, results read directionally rather than as a confirmed trend. Respondents include product and program managers, CIOs, CTOs and CISOs, consultants and advisors, and directors and VPs of data, AI, and engineering, across Technology/Software, Financial Services, Healthcare, and other sectors.

Full text

Eligible Honda models now replace Google Assistant with Gemini.

The AI understands natural conversation instead of fixed voice commands.

Gemini Live enables hands-free chats for trips, news, and brainstorming.

Modern technology offers the opportunity for cars to improve after purchase. AI promises more efficient engagement with that same tech. Now, Honda and Google are combining the two. Owners across the USA are about to get a software update that kills Google Assistant and instead introduces Google Gemini. The update rolls out over the air to models equipped with Google built-in.

var adpushup = window.adpushup = window.adpushup || {que:[]};

adpushup.que.push(function() {

if (adpushup.config.platform !== "DESKTOP"){

adpushup.triggerAd("0f7e3106-c4d6-4db4-8135-c508879a76f8");

} else {

adpushup.triggerAd("82503191-e1d1-435a-874f-9c78a2a54a2f");

}

});

The big benefit for owners is that Gemini doesn’t require them to use specific commands. Google engineered Gemini to understand natural language, maintain context between requests, and handle more complex conversations. That means instead of asking for “the nearest restaurant,” drivers can request something like a highly rated outdoor lunch spot along their route without having to break the request into multiple commands.

Read: Google Maps Is Getting A Mind Of Its Own Thanks To Gemini AI

Honda says Gemini is available to customers signed into their Google account in vehicles equipped with Google built-in. Eligible models include the 2023-and-newer Accord Touring Hybrid, 2025-and-newer Civic Sport Touring Hybrid, Civic Si, Civic Hatchback Sport Touring Hybrid, all 2024-and-newer Prologues, plus the 2026 CR-V Sport Touring Hybrid, Passport, Pilot, and upcoming Prelude.

var adpushup = window.adpushup = window.adpushup || {que:[]};

adpushup.que.push(function() {

if (adpushup.config.platform === "DESKTOP"){

adpushup.triggerAd("ea4d4b5a-f8e5-4d5d-9d9a-53a3b956b6cd");

} else{

adpushup.triggerAd("44c0a4dc-e25b-4ba5-ac64-2c510461b92a");

}

});

The update also introduces Gemini Live, allowing drivers to have free-flowing conversations by saying, “Hey Google, let’s talk.” Honda suggests it can help plan vacations, brainstorm ideas, summarize the day’s news, or answer questions while keeping drivers’ hands on the wheel. With that in mind, Honda is framing this as a convenience upgrade.

The Privacy Catch

Of course, it’s also another step in the industry’s race to make in-car voice assistants genuinely useful. As with any AI assistant, the added convenience comes with privacy tradeoffs .

Gemini can access location, conversation history, and connected apps when users opt in, although Google says drivers can manage or disable much of that data collection through the vehicle’s privacy settings. That doesn’t necessarily mean Google or Honda are using all of that information, but it illustrates how much data these systems are capable of handling.

var adpushup = window.adpushup = window.adpushup || {que:[]};

adpushup.que.push(function() {

if (adpushup.config.platform !== "DESKTOP"){

adpushup.triggerAd("bb7964e9-07de-4b06-a83e-ead35079d53c");

} else {

adpushup.triggerAd("9b1169d9-7a89-4971-a77f-1397f7588751");

}

});

Which Hondas Get It

Model Year Model Trim

2023 – Current Accord Touring Hybrid

2025 – Current Civic Sedan Sport Touring Hybrid

2025 – Current Civic Sedan Si

2025 – Current Civic Hatchback Sport Touring Hybrid

2024 – Current Prologue All trims

2026 – Current Pilot All trims

2026 – Current Passport All trims

2026 – Current CR-V Sport Touring Hybrid

2026 – Current Prelude All

SWIPE

var adpushup = window.adpushup = window.adpushup || {que:[]};

adpushup.que.push(function() {

if (adpushup.config.platform !== "DESKTOP"){

adpushup.triggerAd("bb7964e9-07de-4b06-a83e-ead35079d53c");

} else {

adpushup.triggerAd("9b1169d9-7a89-4971-a77f-1397f7588751");

}

});

Full text

On July 7, Doom: The Dark Ages received a substantial expansion called Revelations , which players and critics alike agree is terrific. The day before that release, Microsoft announced layoffs and cut almost half of id Software’s employees . The developer provided an update a few days later, assuring fans it still has “the crew we need to build the games and tech we’re known for.” But the layoffs cast a pall over what should have been a happy occasion. For Doom diehards hoping for brighter things ahead, Deepward is definitely a game worth watching. It’s a fast-paced FPS with roguelite mechanics, one where each wound and every death matters because a fatal encounter is, well… fatal.

Full text

4 Min Read

NASA Uses Subscale Aircraft to Accelerate Flight Innovation

An atmospheric probe model attached upside down to a quad rotor remotely piloted aircraft ascends with the Moon visible on Oct. 22, 2024. The quad rotor aircraft released the probe above Rogers Dry Lake, a flight area adjacent NASA’s Armstrong Flight Research Center in Edwards, California. The probe was designed and built at the center.

Credits:

NASA/Steve Freeman

Testing new aerospace concepts in flight remains one of NASA’s most effective ways to advance knowledge and reduce risk.

The Dale Reed Subscale Flight Research Laboratory at NASA’s Armstrong Flight Research Center in Edwards, California, supports this mission by using small, remotely piloted and autonomous aircraft as cost‑effective platforms to mature innovative ideas, accelerate learning, and enable smoother transitions to full‑scale flight.

When experiments require a flight platform, several NASA remotely piloted aircraft are available: the Alta‑X quadrotor; the Dryden Remotely Operated Integrated Drone (DROID) with its 10‑foot wingspan; and the Multi‑Use Cub, a 14‑foot‑span fixed‑wing aircraft with an expandable payload capacity for flight experiments. For electric vertical takeoff and landing testing, the HQ‑90 quadrotor provides an additional option.

Once aircraft and experiments are cleared for operations, laboratory pilots support the mission, including ground operations and flight activities.

Justin Link, left, holds the subscale aircraft in place, while Justin Hall manages engine speed during preliminary engine tests on Friday, Sept. 12, 2025, at NASA’s Armstong Flight Research Center in Edwards, California. Link is a pilot for small uncrewed aircraft systems at the center’s Dale Reed Subscale Flight Research Laboratory and Hall is the chief pilot.

NASA/Christopher LC Clark

Flight expertise

Each staff member serves as an experienced and certified subscale aircraft pilot and is prepared to fly unique one-of-a-kind or modified commercial aircraft wherever the mission requires.

NASA’s FireSense project conducted flights in the Geneva State Forest, located about 100 miles south of Montgomery, Alabama. NASA Armstrong flight research staff integrated the instrument onto an Alta-X drone and tested the system before deployment. Two team members then transported the drone and sensor to the forest, prepared the vehicle for flight, and operated it during the mission. The NASA sensor was flown on the drone to demonstrate how remotely piloted aircraft can gather localized weather data that influences smoke movement and fire behavior. This information may help operational agencies improve wildfire decision-making and better allocate firefighters and resources.

Other missions occur closer to NASA Armstrong, such as the Enhancing Parachutes by Instrumenting the Canopy ( EPIC ) project. EPIC involved air‑launching a capsule containing a parachute and flexible sensor from the Alta‑X. Laboratory staff piloted the flights, supported flight operations, and worked with the EPIC team to design and integrate the parachute‑drop mechanism and safety system into the aircraft.

These tests demonstrated that a flexible sensor could help researchers study supersonic parachutes. Continuation of this work can help fill gaps in computer models, making supersonic parachutes safer and more reliable for delivering science instruments and payloads to Mars.

Justin Link, left, pilot for small uncrewed aircraft systems, and Justin Hall, chief pilot for small uncrewed aircraft systems, install weather instruments on NASA’s Alta X drone at the agency’s Armstrong Flight Research Center in Edwards, California. Members of the center’s Dale Reed Subscale Flight Research Laboratory used the Alta X to support the agency’s FireSense project in March 2025 for a prescribed burn in Geneva State Forest, which is about 100 miles south of Montgomery, Alabama.

NASA/Steve Freeman

Advancing challenging research

The Dale Reed Subscale Flight Research Laboratory uses rapid design and testing capabilities to help small aircraft fly big ideas. These concepts could lead to future breakthroughs that support NASA’s missions across aeronautics, science, and exploration.

For decades, NASA and its partners have advanced Automatic Collision Avoidance Technology . The research demonstrated an autopilot could detect and recover from an imminent ground collision – a capability now helping save lives in high‑performance U.S. military jets. NASA Armstrong had key roles in that work and developed a simplified version, the Automatic Ground Collision Avoidance System , which was installed on the DROID for testing.

The system demonstrated on the DROID — developed to assist general aviation pilots as well as remotely piloted and autonomous aircraft — performed well and led to further research toward a version that provides alerts and steering cues. The NASA Armstrong Technology Transfer Office is working to license the technology for U.S. businesses to develop the system as a commercial product.

The Prandtl‑D (Preliminary Research Aerodynamic Design to Lower Drag) flying‑wing glider was also designed, fabricated, and flown at NASA Armstrong. Researchers found that its twisted wing design could reduce drag and generate thrust at the wingtips, advancing concepts that may support greater fuel economy for future aircraft. The original Prandtl‑D is now part of the Smithsonian National Air and Space Museum collection in Washington, and the Prandtl-D3 is at the California Science Center in Los Angeles. Researchers continue developing the next generation of the design in the laboratory.

A wide range of capabilities in the laboratory help transform promising concepts into flight-ready test structures. These include rapid prototyping using traditional and advanced 3D manufacturing techniques, as well as composite and conventional fabrication processes. The team of engineers and technicians also provides custom component design and specialized fabrication to meet unique research needs.

The laboratory supports electrical and mechanical design, hardware and software integration, and the safety and flight-readiness processes required for successful missions. Additional technical facilities, such as the Experimental Fabrication Branch and the Environmental Laboratory at NASA Armstrong, further enhance these capabilities. Together, they support development, testing, and validation activities that advance NASA’s aeronautics and exploration goals.

Deborah Jackson, Al Bowers and Abbigail Waddell successfully launch the subscale Prandtl-D 3C glider.

NASA

Share

Details

Last Updated

Jul 15, 2026

Editor

Dede Dinius

Contact

Jay Levine

jay.levine-1@nasa.gov

Location

Armstrong Flight Research Center

Related Terms

Armstrong Flight Research Center

Aeronautics

Flight Innovation

Explore More

3 min read

NASA Study Points to Smoother Air Taxi Rides

Article

2 days ago

3 min read

A Day of Flight Testing at NASA Armstrong

Article

2 weeks ago

6 min read

NASA’s Newest Wind Tunnel Builds on Legacy of Innovation

Article

2 weeks ago

Keep Exploring

Discover More Topics From NASA

Armstrong Flight Research Center

Aircraft Flown at Armstrong

Armstrong Research & Engineering

Aeronautics

Full text

Lisbon's hotel scene has as much character and charm as the city itself — and its newest arrival comes courtesy of World of Hyatt's Andaz brand. Just opened in March, Andaz Lisbon made not one but two of TPG's lists of most exciting hotels opening in 2026.

The Andaz brand's Portugal debut promised a prime location in the bustling Baixa district, views of historic Lisbon landmarks and the brand's signature cultural connections — and it didn't disappoint. From the gorgeous design and top-notch service to the carefully curated menus of Portuguese dishes and cultural activities offered, this points hotel feels anything but cookie-cutter. Read on to learn what the buzz is all about.

Related: The Andaz Turks and Caicos Grace Bay is now open for bookings

First impressions

RACHEL CRAFT/THE POINTS GUY

From the outside, Andaz Lisbon — housed in a former bank building — fits right in with the Baixa neighborhood's elegant 18th-century architecture. Inside, the brand-new hotel gives a nod to its history. The foyer displays an Our Lady of Mercy tile panel from 1767, unearthed during the renovations, and a desk from one of the bank's offices displays books on art and design in the lobby.

RACHEL CRAFT/THE POINTS GUY The corridors and common spaces feel alive with color, and artworks and curios await around every turn. In keeping with the Andaz brand's cultural ethos, the pieces on display were all created by artists with some connection to Portugal. Other decor elements were inspired by destinations around the globe: rich, saturated prints bring to mind African textiles, for example, and abstract sculptures are reminiscent of the German Bauhaus style.

RACHEL CRAFT/THE POINTS GUY The result is a reflection of Lisbon: A diverse, cosmopolitan city that has plenty of style but doesn't take itself overly seriously. A place where high-end fashion designers coexist with tiny, centuries-old shoe stores (more on those later), where tuktuks rattle down winding cobblestoned streets and e-scooters zip past castles and cathedrals.

The rooms

Behind its compact facade, Andaz Lisbon houses 170 rooms and suites arranged over six floors. Options range from basic one-king rooms to the palatial 1,065-square-foot presidential suite.

A basic king room at Andaz Lisbon. RACHEL CRAFT/THE POINTS GUY I stayed in a deluxe one-king room, with a little extra space in the seating area and double windows overlooking Rua do Comercio. Pops of color and mixed textures felt aligned with the hotel's art-forward design.

RACHEL CRAFT/THE POINTS GUY My snack station was well equipped with a coffeemaker, tea kettle and minibar. (The snacks and drinks are complimentary, but only on the first night.) There were also two fun stainless steel water bottles — free souvenirs offered in every room — which the hotel encouraged guests to refill at the water station in the lobby.

The bold, red-tiled bathroom managed to fit a roomy shower, toilet stall and vanity into a small footprint. Even better, I found plush bathrobes and slippers waiting for me.

RACHEL CRAFT/THE POINTS GUY It's worth noting that, like some other upscale hotels, Andaz Lisbon has a fragrance throughout the property, which may be a concern for migraine-prone travelers like me. I didn't have trouble because the smell didn't extend to my room or the dining areas, but extremely sensitive travelers should be mindful.

Dining at Andaz Lisbon

One notable takeaway about dining here: Andaz Lisbon offers an extensive breakfast spread, similar to what you'll find at many European hotels, for an extra 29 euros ($33) per person per day. There's a cold and hot buffet, and you can order more sophisticated cooked dishes from the menu at no added cost. This is a tasty and filling way to kick off a day of exploring; I often skipped lunch because I was still full from breakfast.

RACHEL CRAFT/THE POINTS GUY Luzzi Restaurant & Terrace

Breakfast is served inside Andaz Lisbon's rooftop restaurant, Luzzi, which for the rest of the day serves traditional Portuguese fare with a modern twist. The space is infused with colors, patterns and textures dappled with sunlight from the windows overlooking the city. The result manages to be both bright and cozy, yet energetic and relaxed at the same time.

Luzzi seems to attract not only guests but local Lisboetas, which is a good sign; even on a Wednesday night, I saw packs of locals dressed up for a night on the town.

RACHEL CRAFT/THE POINTS GUY Everything I tried — feijoada Lusitana, a Portuguese bean stew (20 euros or $23), tofu dumplings (10 euros or $11) and sweet potato fries (8 euros or $9) — was superb. I also sampled a vegan version of the Portuguese dessert quindim (10 euros or $11), and the passion fruit sorbet blew my mind.

RACHEL CRAFT/THE POINTS GUY On the terrace next door, visitors can enjoy the same dishes and drinks in a more casual setting, with views of Rua Augusta Arch from balconies tucked into the tiled walls and live music every night from Wednesday to Saturday.

RACHEL CRAFT/THE POINTS GUY Andaz Lounge

On the ground floor, Andaz Lounge — inspired by the traditional Portuguese quiosque — serves refined drinks and elevated street food.

RACHEL CRAFT/THE POINTS GUY I'm not exactly an oenophile, but I couldn't leave Lisbon without sampling some Portuguese wines — especially when all the lounge's by-the-glass offerings are made by female wine producers nearby. The bartender was happy to help me choose a wine, as well as to find vegan-friendly menu items. I went with the Caesar salad (12 euros or about $14) with a mushroom dressing and tofu, and let me tell you: I've had a lot of salads in my life, and this easily made the top five.

RACHEL CRAFT/THE POINTS GUY Whether you're fueling up at the breakfast buffet, having a lively dinner with friends at Luzzi or curling up with a glass of wine and a good book at the Andaz Lounge, you can expect warm service that's just polished enough to make you feel a little bit like a VIP — without having to worry about a dress code.

Amenities and Signature Experiences

Beyond the on-site restaurants, here are the other amenities you'll find at Andaz Lisbon:

A 24/7 fitness center

A spa coming later this year (until then, in-room spa treatments are available, and the hotel partners with nearby spa facilities that the concierge can arrange for guests)

Pet friendly accommodations (for dogs up to 66 pounds)

Andaz Signature Experiences available for booking

The hotel partners with a local tour company to offer various half- and full-day Andaz Signature Experiences. These culturally inspired activities include an artist-led street art tour, a coastal horseback ride and a catch-and-cook boat ride with a local fisherman.

I tried a shortened version of the "Gems in the Making" experience (900 euros or $1,028 for the full version), which involved a tour of the city's oldest artisanal shops on foot or by tuktuk. With the help of my guide, I explored the world's oldest operating bookshop, a fifth-generation hat maker, a coffee shop once frequented by the 19th century poet Fernando Pessao and more — followed by a hands-on filigree workshop with a family of jewelers that has been in business since 1924.

RACHEL CRAFT/THE POINTS GUY I was impressed not only by the tour's quality but also by how distinct to a "local's" version of Lisbon that it was. It introduced me to a lesser-known facet of the city that I otherwise would never have discovered.

Location and logistics

The Rua Augusta Arch near Andaz Lisbon. RACHEL CRAFT/THE POINTS GUY The Andaz is located in the tourist-friendly Baixa area, or downtown, half a block from the main drag of Rua Augusta. It's steps away from the Tagus River, and it puts guests within easy walking distance of many of the city's highlights — from the historic Alfama neighborhood to the buzzy nightlife of Pink Street.

The hotel is about a 20-minute drive from Lisbon Airport (LIS). A rideshare will set you back around 10 to 20 euros ($12 to $24), or you can take public transportation for around 2 euros, though it will take roughly twice as long. The nearest bus stop and metro station, Praca do Comercio and Terreiro do Paco, are a couple of minutes away on foot.

Accessibility

RACHEL CRAFT/THE POINTS GUY Andaz Lisbon is not the most accessible hotel I've visited, but it does have some accessibility features.

Most common areas are accessible by elevator, and there is a specially adapted room for guests using wheelchairs, complete with widened doors and a roll-in shower. There is only one of this room type, so be sure to call ahead and confirm it for your dates.

I didn't see any Braille on signs, and the hotel doesn't advertise features like visual doorbells. However, it's worth calling to see if it can accommodate your needs before ruling it out.

How to book a stay at Andaz Lisbon

RACHEL CRAFT/THE POINTS GUY Standard rooms start from 266 euros ($303) per night, while my one-king deluxe room starts at 304 euros ($346) per night, not including breakfast. Award rates for this category-6 property range from 20,000 to 40,000 World of Hyatt points per night, per the program's updated award chart .

Andaz Lisbon is also a member of Chase Travel's The Edit , so you can access elitelike perks while stacking Chase Ultimate Rewards and World of Hyatt earnings on your stay. Plus, it's a Points Boost property, so you can get solid value when redeeming your Chase points through Chase Travel. Just be sure to compare prices on the portal to Hyatt's website before deciding where to book .

CHASE TRAVEL If you book directly, you'll want to use one of the best credit cards for Hyatt stays , and ideally one that offers World of Hyatt elite status for value-added benefits. These include:

World of Hyatt Credit Card (see rates and fees ): Earn 4 points per dollar spent at Hyatt hotels and resorts, automatic Discoverist status and five elite night credits per year toward a higher tier.

World of Hyatt Business Credit Card (see rates and fees ): Earn 4 points per dollar spent at Hyatt hotels and resorts, automatic Discoverist status and five tier-qualifying night credits for every $10,000 you spend each year toward a higher tier.

Bottom line

With a central location, top-notch service and style in spades, Andaz Lisbon makes a luxurious home base for a trip to Portugal's capital.

But this jewel of a hotel is more than just a place to rest your head. In fact, the hotel's excellent fare and cultural experiences were among the most memorable aspects of my entire trip.

Editorial disclaimer: Opinions expressed here are the author’s alone, not those of any bank, credit card issuer, airline or hotel chain, and have not been reviewed, approved or otherwise endorsed by any of these entities.

Full text

I had thought science denial would hit a bottom — that once people began to experience the direct consequences of being wrong, of holding fast to beliefs that were demonstrably at odds with well-established science, they would drop those beliefs. It turns out I was wrong.

As a research scientist who’s also dedicated many years to science communication, coming to grips with this truth has been hard. It leaves us all facing difficult questions about how science and society can — or cannot — co-evolve when so many are opting out of science. How we address these questions may determine the fate of democratic societies in a world where excellence in science translates directly into both soft and hard power.

Science denial has little to do with science

My first experience with this kind of science opt-out (I’ll also call it “denial”) came in the late 2000s when I was writing a science blog for a major U.S. news outlet. Anytime I’d write a piece about climate change, the comments section would quickly fill up with the most aggressively ignorant opinions I’d ever encountered. People would claim the science was a hoax. They would claim scientists were only in it for the money. They would say things that even a freshman taking a planetary science class would know were glaringly wrong.

Being committed to the ideals of science communication, I did my best to engage with the commenters. I would explain what scientists knew about radiation transfer through an atmosphere (the basis of the greenhouse effect). I would lay out how hard-won evidence — from ice cores to tree rings — allowed us to unravel millennia of Earth’s response to changing atmospheric and ocean chemistry. Finally, I would detail how consensus works in science and how communities of researchers build a public understanding of their field.

None of it mattered.

One notable example of my encounter with science opt-out came when I tried engaging with a reader on a personal level. In response to a climate post, he had sent me an email telling me that I should be arrested for my climate lies. I replied and asked him about himself. What did he do for a living? He said he owned a dry-cleaning business. I asked some questions about the business. After a bit, I made a purposefully ridiculous claim about dry-cleaning chemicals. When he responded that I was clearly ignorant of the work he did every day, I said, “Yes, and that’s what your claims sound like to a climate scientist.”

The point was lost. Nothing would shake him from his beliefs. From this, and many other encounters, I learned that science denial is not about science — that reader and others I’ve engaged with expressed no curiosity about how planetary climates actually worked. Instead, it is about a kind of tribalism: If you identify as a member of certain groups — usually ones that have a particular political orientation — then you oppose certain kinds of science.

This opposition comes regardless of whether they know any of the deep (and often difficult) details required to make sweeping claims about that science. So, what might have been an interesting back-and-forth with a reader about what determines a planet’s temperature becomes a tiring slog through their fixations on Al Gore or long-refuted claims that 1970s climate science predicted “global cooling.”

Seeing was not believing

Despite these experiences, I still believed that once we all saw firsthand the reality that a field of science was predicting, denial would have to fall. Then came COVID-19. I was already watching the new resistance to vaccines rise with its all-too-predictable outcomes, like the 2017 measles outbreak at Disneyland . But with COVID-19, the willingness to ignore what was right before people’s eyes rose to a new level.

Before I go further, though, I want to be clear that the CDC and others in public health did make mistakes during this difficult period. It was, after all, a “novel” coronavirus and the first pandemic of the modern age. Also, I’m all in for discussions about whether and how some measures taken by public health officials went too far. If we don’t have these discussions, then how else can we learn what to do in the next pandemic? What we can’t argue over, however, is the reality of COVID-19 or the ability of the vaccines that scientists developed to save lives.

With COVID-19, just like with climate, lines were drawn, and tribalism kicked in. The famous cases of people dying of COVID-19 and still proudly proclaiming their resistance to a vaccine that might have saved them told the tale in no uncertain terms. They’d opted all the way out.

Now we live in a time when science denial inhabits the highest levels of power and receives broad acceptance. At the institutional level, climate science in the U.S. (along with many other types of science) is being decimated. The U.S. government lost 10,000 STEM PhDs in 2025 (compared to about 4,500 in 2024 ) in the wake of the present administration’s aggressive funding cuts to its own world-renowned scientific institutions. At the popular level, social media is awash with influencers routinely hosting guests claiming the Moon landings never happened.

As part of my work in science communication, I’ve appeared on podcasts associated with the “manosphere,” including The Joe Rogan Experience . Because of these appearances, I often have conversations with young men who have a clear interest in science, but who’ve also been led into strange vortices of science denial. “So, you really think we went to the Moon?” they ask. Or, “Scientists are always hiding the truth about stuff, right?”

I was not routinely fielding questions like these in 2010. There has been a shift in the cultural current. We now live in a kind of mirror-verse where opinions, if repeated often enough, can replace the decades of work needed to establish a scientific truth. Building off a narrative that falsely associates expertise with oppressive authority, influencers seeking clicks have driven viewers like the people I encounter into a haze of impossible claims presented as revealed truth.

So, where does it end? Surely there must be consequences for so many people opting out of science in a world that is, in fact, dominated by science.

The price of denial

Yes, there are consequences, but they aren’t occurring on an individual scale, as I expected them to a decade ago. Instead, they are playing out for society as a whole.

If you were interested in the cutting edge of science in 1600, you would have traveled to Italy. That is where wealthy government patrons were supporting Galileo and other scientists’ efforts to push the frontiers of telescope design, mechanics, and other fields. But if you were interested in the cutting edge of science in 1700, you did not go to Italy anymore. The baton of scientific excellence had already been passed to England and France. Likewise, if you wanted to be where the most important scientific discoveries were being made in 1900, you went to Germany, but by 1960, you would be on a plane to the U.S. instead.

The reasons for this ever-shifting localization of the best science have everything to do with a nation’s broad commitment to science. History shows that once a nation begins to slip in those commitments, the frontier of scientific progress moves elsewhere. Just as important, history shows that the nations making this commitment are the ones that prosper from the latest scientific breakthroughs.

It’s already becoming clear how the infection of science denial — of opting out of a commitment to science — is affecting the U.S. Because the current administration has embraced climate denial, it also remains in denial about the world’s ongoing energy transition away from fossil fuels. In 2025, global electric vehicle sales increased by 20% , but in the US, the growth rate was less than 10%. Today, the largest producer of new EVs is China.

Science and science leadership don’t stand still. When a society becomes infected with science denial, when its populace thinks it’s OK to opt out of treating science as a source of public truths, it elects leaders who lead with that vision of the world. But the bill always comes due. When those leaders slash science budgets for political reasons and create chaos in their own world-renowned scientific institutions, it becomes impossible to plan on the decades-long timescales required to ensure the nation is still a scientific leader well into the future.

Worse, the rest of the world is always watching. The best and brightest students from across the globe — the ones who used to come to your universities to learn and then stick around to start businesses — go where they will be supported and not harassed.

In the long run, the price you pay for science denial is the denial of your own scientific capacities.

This article Who pays the cost when Americans opt out of science? is featured on Big Think .

Full text

The AI buildout is well and truly underway with OpenAI CEO Sam Altman leading the charge, having made various deals with companies including Oracle and Nvidia to guarantee the infrastructure needed to train future AI models is installed. But as this ensues, the spotlight has been thrown on how much energy these models will need.

Eating machines

Altman was speaking during an AI summit in India earlier this year when making the remarks to The Indian Express .

Quote of the day

This article is part of TechRadar Pro's QOTD project to provide an insight into the minds of the brightest and most recognized figures in the technology industry today and in years gone by. Read the full series here .

The OpenAI chief was being questioned about the substantial amount of energy that AI has already been consuming – and will be projected to continue consuming – for both training and inference. Bringing these models online, after all, and keeping them running require a huge amount of resources, not only in terms of the energy to power the data centers, but the water for cooling, and the components and resources in building the hardware.

Altman's defence hinged on the idea that people, too, require plenty of energy in order to reach utility — while AI can be trained much quicker. This view, in essence, frames machine intelligence as a like-for-like competitor with human intelligence.

Energy efficiency

Altman argued that humans are deeply inefficient, and that you should compare the total energy spent to create a human expert versus a machine expert. There's also an argument that energy efficiency of AI could improve over time.

But detractors were scathing in their criticism of this entire point of view – saying this framing doesn't take into account the fact that the human brain operates on roughly 20 watts of power . This isn't to mention the ethically gray and dehumanizing nature of the remarks.