FlashFeed

Full text

Across 107 enterprises, AI infrastructure spending is accelerating well ahead of the ability to see or steer its economics. Most organizations run their AI on a familiar base of hyperscalers and model-provider APIs, yet the next dollar is aimed at specialized compute almost none of them use today; a majority intend to switch or add providers within the year, many within a quarter. Buying decisions turn on integration and total cost of ownership rather than headline token price — which is fortunate, because most enterprises cannot yet see their unit economics clearly: GPUs sit at half utilization or less, and fewer than half rigorously track what their compute actually costs. The result is a compute gap — heavy, fast-moving investment running ahead of the visibility needed to control it.

This wave of VentureBeat Pulse Research examines enterprise AI infrastructure and compute: where organizations are in their deployment journey, what they run AI on today, how satisfied they are, what would make them switch, where they plan to evaluate their investments, and — most revealingly — how well they can measure and control the economics of the compute underneath it all.

The central finding is a compute gap — the distance between how aggressively enterprises are investing in AI infrastructure and how little of its economics they can see. Only about one in five (21%) run AI in production at scale, yet spending intentions are outrunning that maturity: the single largest planned area enterprises plan to evaluate over the next year is AI-specialized clouds (45%), a layer almost none of these enterprises use today. Meanwhile the compute already in place runs cold — 83% report GPU utilization of 50% or less — and fewer than half (44%) can rigorously track what their AI compute costs. Enterprises are buying more infrastructure faster than they can account for what they already own.

Enterprises are not settled on their infrastructure vendors, either: A clear majority (64%) plan to switch or add an infrastructure provider within twelve months, and 38% within the next quarter — unusually high churn intent for a category this foundational. When they choose, they choose on integration with the existing stack (41%) and total cost of ownership (35%), not on headline price: cost per million tokens is the deciding factor for just 8%. And the frontier constraint that will shape the next round of decisions — the shift from GPU compute to memory bandwidth as inference scales — is barely on the radar, with roughly one in five enterprises either unaware of it or yet to address it.

Methodology

VentureBeat fielded this survey as part of its ongoing Pulse Research series, this survey focused on enterprise AI infrastructure, compute, and inference economics. Responses are filtered to organizations with more than 100 employees (n=107; the survey’s smallest size band, 1–100 employees, is excluded), drawn from a single Q2 2026 (June) 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. Several questions were multiple-select, so those shares can sum to more than 100%.

By organization size the sample concentrates in the mid-market: 101–250 employees (36%) and 251–1,000 (27%) lead, with 1,001–5,000 (22%), 5,001–10,000 (8%), and 10,001+ (7%) above them. By role it spans managers (38%), individual contributors (28%), VPs and directors (19%), and the C-suite (13%); on purchasing authority it is buyer-credible, with 45% final decision-makers and another 30% recommenders or influencers for AI solutions. Technology/Software is the largest industry at 26%, followed by Healthcare/Life Sciences (15%), Financial Services (13%), and Retail/E-commerce (12%).

At 107 respondents the sample is large enough to read directionally but should be treated as a directional signal rather than a precise measurement; it is self-selected and is not a probability sample. It also skews toward the mid-market and toward earlier-stage adopters, so it is best read as the view from organizations actively building out AI infrastructure rather than from the largest hyperscale operators.

Finding 1: Ambition outpaces production

Only one in five run AI in production at scale

We asked where organizations sit in their AI deployment journey. Most are still building toward production rather than operating at scale.

38%

are experimenting — running proofs of concept, not yet in production

37%

have some workloads in production, but not across the organization

21%

run AI in production at scale — the mature minority

4%

are not yet running AI workloads at all

The maturity curve is front-loaded. Three-quarters of enterprises (76%) are either experimenting or running only some workloads in production, and just 21% describe AI in production at scale. This matters for everything that follows: the infrastructure decisions in this report are being made largely by organizations still early in deployment, whose compute footprint — and whose costs — are about to grow. The evaluation and switching intentions in Findings 3 and 4 are the leading edge of that build-out, not the settled preferences of operators who have already found what works.

Finding 2: Enterprises run on hyperscalers and model APIs

The specialized GPU clouds barely register — today

We asked which providers and platforms enterprises currently use to run their AI. The answer is a familiar one: the incumbents.

48%

use Google Cloud — the most-used platform overall (Microsoft Azure 29%, AWS 22%, Oracle Cloud 22%)

41%

use Google’s Gemini models, with OpenAI close behind at 40% and Anthropic at 12%

6%

run their own on-prem or co-located GPU clusters; 4% a custom open-source self-managed stack

<2%

each use the specialized AI clouds — CoreWeave, Lambda, Crusoe, Nebius, Together, Fireworks and peers

The current stack is hyperscaler-and-API. Google Cloud leads at 48%, and the general-purpose clouds (Google, Microsoft, AWS, Oracle) together with the major model APIs (Gemini, OpenAI, Anthropic) account for essentially all current deployment. The specialized “neocloud” GPU providers that dominate AI-infrastructure headlines — CoreWeave, Lambda, Crusoe, Nebius and peers — register at or near zero among these enterprises today. Only 6% run their own on-prem GPU clusters and 4% a custom open-source stack. Enterprises are, for now, running AI on the providers they already buy from — which makes the evaluation intentions in Finding 3 all the more striking.

(A note on reading these shares. As described in the methodology section, this sample is self-selected and skews mid-market, and this question counted every provider a respondent uses — an average of 2.1 selections each — so the figures measure presence in the stack rather than spending or primary status. A sample built this way will show a different provider mix than a spend-weighted census of the broader market; Google's strength here, for example, is consistent with its long-standing position among smaller enterprises building on AI. Read these shares as a portrait of what this AI-active cohort runs today, and treat gaps between these figures and industry-wide market share estimates as a property of the sample rather than a contradiction of either.)

Finding 3: The next dollar goes to infrastructure they don’t yet run

AI-specialized clouds top the evaluations list

We asked where enterprises planned to evaluate AI infrastructure over the next 12 months. Their answers point away from the stack they run today.

45%

AI-specialized clouds (CoreWeave, Lambda, Crusoe, Nebius) — the top planned evaluation area

32%

non-NVIDIA accelerators (AWS Trainium, Google TPU, AMD Instinct, Intel Gaudi, in-house ASICs)

28%

Nvidia Blackwell (GB300) / next-generation GPUs

16%

decentralized or distributed compute networks

11%

sovereign or region-specific compute; 9% say none of the above

Here is the report’s sharpest tension. The single most-cited planned evaluation area — AI-specialized clouds, at 45% — is the very category almost none of these enterprises use today (Finding 2). Nearly a third (32%) intend to evaluate non-Nvidia accelerators, and 28% in next-generation Nvidia silicon; even decentralized compute networks (16%) and sovereign compute (11%) draw meaningful interest. Read against current usage, this is not incremental — it is the leading edge of a re-platforming. The direction-of-travel question tells the same story: every infrastructure approach is net-expanding, but specialized AI clouds carry the highest net momentum (+24), edging out even the hyperscalers (+22). Enterprises are preparing to move a meaningful share of AI compute off the general-purpose cloud.

This continues a trend we saw in our April-May survey wave. Back then, usage of the AI-specialized clouds was equally marginal — CoreWeave at 3%, Lambda at 4%, Crusoe at 2% of enterprises. When we asked enterprises what change they planned in their AI infrastructure strategy over the next twelve months, the most-cited answer was moving workloads to specialized AI clouds, at 33%. Asked in April-May which emerging compute option they were most likely to evaluate AI-specialized clouds again drew the most responses. Two waves, two differently worded questions, one consistent picture: the type of cloud enterprises are most eager to assess is the type they have barely begun to use.

Finding 4: A switching wave is building

Six in 10 plan to change providers within a year — many within a quarter

We asked whether and when enterprises plan to switch or add an infrastructure provider. Very few intend to stand still.

38%

plan to change within the next 0–3 months — tied for the most common answer

36%

have no plans to change

22%

plan to change within 3–6 months

7%

plan to change within 6–12 months

For a category as foundational as compute, this is a remarkable amount of intended movement. Only 36% have no plans to change, meaning a clear majority (64%) intend to switch or add a provider within twelve months — and 38% within the next quarter alone. Where that interest points is telling: the providers drawing the most switching consideration are again the incumbents — Microsoft Azure and Google Cloud (33% each), OpenAI (30%), and Gemini (22%) — which suggests much of the near-term movement is reshuffling among the majors and consolidating spend rather than defecting to new entrants. The neocloud interest in Finding 3 is a 12-month evaluation thesis; the switching in the next quarter is mostly incumbents trading share.

( Method note: Respondents who selected both "no plans to change" and a specific switching window are counted as switchers, on the logic that naming a timeframe is the more specific answer; three respondents were reclassified under this rule. )

Finding 5: Nobody buys on token price

Integration and total cost of ownership decide — not sticker price

We asked what matters most when enterprises select an AI infrastructure provider. Headline price finished last.

41%

integration with the existing cloud and data stack — the top factor

35%

total cost of ownership (TCO)

24%

performance — latency and throughput

19%

each cite security/compliance, autoscaling for spiky workloads, and GPU access/availability

8%

cost per 1M tokens — the least-cited factor

Enterprises do not buy AI infrastructure on pricing, which is the place vendors compete on hardest. Integration with the existing stack (41%) and total cost of ownership (35%) dominate, while the headline metric — cost per million tokens — is the deciding factor for just 8%, dead last. The pattern is coherent: buyers are optimizing for how a provider fits and what it truly costs to operate, not for the advertised unit rate. It also foreshadows Finding 7 — enterprises say TCO matters most, yet most cannot yet measure it rigorously. The stated priority and the measured capability are out of step.

Finding 6: Expensive GPUs, idle most of the time

83% report GPU utilization of 50% or less

We asked what share of their GPU capacity enterprises actually utilize. The answer is a well-known but rarely quantified inefficiency.

37%

run at 26–50% utilization

34%

run at 10–25% utilization

15%

run under 10% utilization

12%

run over 50% — the efficient minority

8%

don’t measure utilization at all; a further 7% consume via API and run no GPUs of their own

Disclosure: Band percentages count every selection against all 107 qualified respondents; 14 respondents selected more than one band, so bands overlap. At the respondent level, 83 of the 100 GPU-operating enterprises reported utilization at or below 50%

The compute already in place runs cold. Adding the bands at or below half capacity, 83% of enterprises that operate GPUs report utilization of 50% or less, and nearly half (49%) run at 25% or below. Only 12% clear the 50% mark, and a further 8% do not measure utilization at all. Idle accelerators are expensive accelerators, and this is the clearest single measure of the compute gap: enterprises are planning to buy more GPUs and specialized compute (Finding 3) while the capacity they already own sits substantially unused. The efficiency headroom in the current fleet is large — and largely unmeasured.

Finding 7: Spending fast, measuring slowly

Fewer than half rigorously track what their compute costs

We asked whether enterprises can quantify the cost and return of their AI infrastructure spend, and how satisfied they are with what they run. Confidence in the ledger lags the spending.

44%

track compute cost and ROI rigorously

39%

track it only partially

20%

can’t quantify it yet

6%

say it isn’t a priority

Measurement trails money. Fewer than half of enterprises (44%) rigorously track the cost and return of their AI compute; the majority track only partially (39%), cannot quantify it yet (20%), or have not prioritized it (6%). That gap is consequential given Finding 5, where total cost of ownership was the second-ranked buying criterion — enterprises are choosing providers on an economic basis they mostly cannot yet measure. Satisfaction with current infrastructure is moderately positive but not enthusiastic: on a five-point scale, overall satisfaction averages 4.0, with ease of implementation (3.8) and value for money (3.9) trailing slightly — the softness landing, tellingly, on cost. Enterprises are spending quickly and accounting slowly.

Finding 8: The next bottleneck few are watching

As inference shifts from compute to memory, the field scatters

Finally, we asked how enterprises would address the emerging constraint in large-scale inference — the shift from GPU compute to memory, specifically KV-cache capacity. The responses reveal a frontier that is not yet a priority.

31%

would rely on Dell (PowerScale / Project Lightning) — the leading single answer

16%

would rely on Nvidia (Dynamo / ICMSP)

18%

are not aware of this as a constraint (9%) or haven’t addressed inference-memory limits yet (8%)

10%

Hammerspace (Tier Zero); 9% DDN (Infinia); the rest split across open-source KV-cache tooling, model-level efficiency, VAST Data, and WEKA

The memory frontier is real but barely governed. Asked which approach they would rely on as the binding constraint in inference shifts from compute to memory bandwidth, enterprises scatter: Dell leads at 31%, Nvidia follows at 16%, and the rest fragments across storage vendors, open-source tooling, and model-level efficiency techniques. Most telling is that roughly one in five (18%) either do not recognize the constraint or have not begun to address it. For a shift that will reshape inference cost and architecture, this is an early and unsettled market — and, consistent with the measurement gap in Finding 7, one where many enterprises simply do not yet have a view. It is the next chapter of the compute gap, arriving before most have closed the current one.

The bottom line: A compute gap that faster spending will widen, not close

Organizations with more than 100 employees are investing in AI infrastructure faster than they can measure it. Most are still early in deployment, yet their spending intentions point past their current stack — toward specialized clouds and alternative accelerators almost none of them run today — and a clear majority intend to change providers within the year. They buy on integration and total cost of ownership rather than headline price, which is rational; the difficulty is that most cannot yet see those economics clearly.

The visibility gap is concrete. The GPUs enterprises already own run at half utilization or less for the overwhelming majority, and fewer than half can rigorously track what their compute costs or returns. Satisfaction is decent but unenthusiastic, softest on value for money — the dimension hardest to judge without measurement. And the next constraint, the shift from compute to memory in large-scale inference, is arriving while most enterprises are still unaware of it. At 107 respondents in a single Q2 wave this is a directional read, skewed toward the mid-market and earlier-stage adopters — but the direction is consistent: the appetite to spend is running well ahead of the instrumentation to spend well. The compute gap is not a capacity problem that more hardware will solve on its own; it is, first, a problem of seeing what the hardware already costs. The open question for later waves is whether enterprises build that visibility before the re-platforming arrives — or buy the next layer of infrastructure as blind to its economics as the last.

Based on survey responses from 107 qualified enterprise respondents (100+ employees), drawn from a single Q2 2026 (June) wave. Because this is one wave rather than a pooled multi-month sample, the results read cross-sectionally rather than as a month-over-month trend, and at 107 respondents this is a directional signal rather than a precise measurement — the sample is self-selected, skews mid-market, and leans toward earlier-stage adopters rather than the largest hyperscale operators. Respondents include managers, individual contributors, VPs/directors, and the C-suite, with buyer-credible purchasing authority, across Technology/Software, Healthcare/Life Sciences, Financial Services, Retail/E-commerce, and other industries.

Full text

UnitedHealth Group smashed through investors’ earnings expectations in the second quarter, with executives confidently saying this new level of profit is durable for the foreseeable future.

The numbers lifted the company’s stock about 3% as it attempts to return to financial dominance under CEO Stephen Hemsley . But a string of comments made during the company’s earnings call on Thursday signals a more ominous path ahead for America’s businesses and workers.

More than 150 million Americans get their health coverage from a job. A new series from STAT chronicles how the costs of that insurance are stretching employers like never before. An increasing number of America’s small businesses are abandoning traditional health benefits completely, forcing workers to find alternative options.

Continue to STAT+ to read the full story…

Full text

Few movie stars are eyeing as big of a year in 2026 as Anya Taylor-Joy , who has already starred in the first $1 billion film of the year, The Super Mario Galaxy Movie . She will also be seen later this year reprising her role as Alia Atreides in Dune: Part Three , the third and final movie in the epic trilogy from director Denis Villeneuve . Taylor-Joy featured in only one scene during Dune: Part Two , but the massive time jump between the second and third films allows her to star in a much larger capacity in the trilogy-capper. Anya Taylor-Joy is also teaming up with Timothy Olyphant for the new Apple TV crime thriller, Lucky , which just debuted its first two episodes this week. New installments will drop every Wednesday before the Season 1 finale airs on August 19.

Full text

Maintaining a live-service game is a tough line of work. You've got to keep the updates coming, constantly roll out hotfixes, and you're constantly relying on your community to make the release viable. Warframe 's been going steady for 13 years now, an impressive feat, and Rebecca Ford, the game's creative director, takes nothing for granted.

"You can be as wise as those that came before you, you can learn from those who have not continued on, and the truth is always somewhere in the middle of it," she explains to GamesRadar+ during an interview at TennoCon 2026. "Just try and make a good game before they kick you out."

That's a slightly dispiriting outlook, but it's pragmatic, since we just saw the sudden finish of Destiny 2 . Bungie 's sci-fi FPS came to a slightly unceremonious end in June, with a sudden final update after it was reported Sony took a considerable loss on acquiring the studio, leading to a large amount of layoffs .

Warframe developer Digital Extremes knows full well the environment it's in. "It shows that even if you care so much, the business side of this industry always gets the last remark," Ford says of what happened to Bungie. "Those are the types of stories and experiences that, when you're in a position where you have your own game, your own IP, and you work as hard as you do on it… That's not the first time it's happened, and it'll happen again, where the business aspect of the video game economy makes the decision for you."

She mentions that "the idea that we aren't in charge of our own goodbye is something I wake up thinking about every single day," demonstrating the kind of pressure she and her colleagues work under. Right now, Warframe seems to be going strong. TennoCon gets bigger each and every year, and being both free and well-balanced as a game makes Warframe appealing and easy to settle into.

There have been layoffs in the studio before, though, and Ford notes that despite this, the internal structure contains many longtime employees – including herself, as she's risen from community management to her current role – giving them a distinct perspective.

"Anyone here in a leadership position didn't come into the company in that position, which I think gives us a little bit of a unique way of working as a team," she explains. "None of us signed up for this, but we do it because it must be done, and we trust each other, which is one of the most fundamental parts of our operations, and every day, every week, we can improve how we work as a business with an office, or how we work remote as a studio."

Warframe lead Rebecca Ford says Destiny 2's death is "horrible news" for everyone: "Not the first time it's happened, and it'll happen again"

Full text

A Florida man has been arrested by federal agents amid an FBI investigation into a Steam malware scheme said to have plundered crypto wallets in the past two years.

South Florida outlet Local 10 News reports that 21-year-old Zyaire Dontaevious Zamarion Wilkins was arrested on Tuesday, July 14 and charged with participating in the scheme, whose operators are collectively said to have stolen "at least $220,000" in cryptocurrencies from roughly 80 crypto wallets. On current charges, Wilkins faces up to 10 years in prison under the Computer Fraud and Abuse Act .

Steam and owner Valve are not specified in the federal complaint tied to Wilkins' arrest, but the FBI does specify Steam in a March request for victim information regarding malware threat actors believed to have been active "between the timeframe of May 2024 and January 2026," and the details of these incidents are closely aligned.

The FBI also listed several games identified in the malware scheme: "BlockBlasters, Chemia, Dashverse/DashFPS, Lampy, Lunara, PirateFi, and Tokenova." All of these games are no longer on Steam and, in SteamDB archives, have been flagged as "suspicious, as it may be malicious or impersonating another product."

The removal of BlockBlasters made headlines last year . Reported crypto figures indicate this one game was responsible for a majority of the malware distribution.

Agents say Wilkins and his collaborators promoted "eight games embedded with malware" online and ultimately reached "approximately 8,000 individual customers," apparently including several specific people targeted for their "large cryptocurrency holdings." Wilkins allegedly paid $10,000 for a Trojan program capable of the remote access necessary for this type of hack.

Local 10 News reports that Wilkins was tracked down via Bitrefill purchases of over 150 digital gift cards primarily for Uber Eats. The complaint adds that Wilkins' crypto history showed he had personally sent or received $382,000.

Malicious games and mods are not unheard of on Steam, and games are a relatively common vehicle for this type of malware. Earlier this year, a developer called out a worm-like mod which was evidently spreading through the UGC Steam Workshop and replacing other files "with itself."

Hacker behind GTA 6 leak that cost Rockstar $5 million reportedly awaiting retrial after being released from secure hospital .

Full text

An Early Bronze Age metalworker and shaman discovered over 200 years ago in a lavish burial near Stonehenge and long assumed to be male was actually female, a new genetic analysis reveals.

The results of the ancient DNA analysis of the " Upton Lovell Shaman ," carried out by researchers at the Francis Crick Institute in London, break the previously held stereotype of Early Bronze Age women, according to a statement from the Wiltshire Museum, where the remains and grave goods are housed.

"It completely tears up previous assumptions," David Dawson , director of the Wiltshire Museum, told The Guardian . "We're so used to the assumption [that] men do everything, men are the leaders, men are the metalworkers. Here we have smoking gun evidence of a female metalworker. And metalworking was the space science of its day."

The nearly 4,000-year-old burial was unearthed in 1801 near the village of Upton Lovell, about 10 miles (16 kilometers) west of Stonehenge . The human remains were surrounded by an unusually rich tool kit containing stone axes, metalworking implements with traces of gold on them, a touchstone for testing metal purity by comparing streaks left by different metals, and pierced animal bones that were likely once sewn onto a garment as decorations, hinting at a ceremonial cloak.

The mix of high-status metalworking tools and objects thought to have ritual significance led archaeologists to interpret the individual as a spiritual specialist, earning the remains the nickname, the "Upton Lovell Shaman."

William Cunnington, the English archaeologist who excavated the burial mound, known as a barrow, noted at the time that, "from the largeness of the bones," the burial "appeared to be a stout man," according to the statement. For the next two centuries, the individual's assumed sex was male, with its museum display depicting a bearded male figure.

Among the grave goods were four fossil sponges hollowed out into cups, hinting that their owner was once a crafter. (Image credit: Wiltshire Museum) The DNA analysis was originally intended to trace the individual's ancestry, but the results showed sex chromosomes of XX, instead of XY, catching the researchers off guard. To be certain, the team tested DNA from a tooth and a toe — and got the same answer each time, with no evidence that the grave held more than one person, according to the statement.

Other clues in the skeleton revealed more details about the individual. She stood around 5 feet, 4 inches (165 centimeters), which was unusually tall for a Bronze Age woman, and died at about age 45. She was robustly built, with arthritis in her right wrist but not her left ‪—‬ a pattern that fits years of the repetitive work with metalworking tools, the statement reported.

A 2022 study found that the individual was likely a skilled goldsmith who fashioned gold ornaments. This ability may have seemed magical during the Bronze Age, Susan Greaney , an archaeologist at the University of Exeter who wasn't involved in the study, told Live Science at the time .

RELATED STORIES

Bronze Age 'treasure' was crafted with extraterrestrial metal

How do archaeologists figure out the sex of a skeleton?

'A weird result from an already weird hominin': Archaeologists discover all Homo naledi skeletons found in South African cave are female

"The ability to transform other objects by the delicate and skilled process of covering them with gold sheet may have been seen as a magical or ritual process, a secret method known only to a few people," Greaney said in a 2022 email . "This research shows how metalworking was closely related to magical, ritual and religious beliefs."

This is not the first time an ancient elite individual has been mistakenly identified as male. For instance, an elite person from Sweden's Viking Age who was buried with weapons and strategy games was thought to be male but was later verified to be female , and a high-ranking individual from Copper Age Spain was thought to be male until a DNA analysis showed they were female.

"We now have a whole new understanding of this burial, rewriting their story, breaking stereotypes, and putting women front and centre in our understanding of early Bronze Age society," Lisa Brown , curator of the Wiltshire Museum, said in the statement.

The findings will be unveiled Thursday (July 16) in a new exhibition on ancient DNA, " We Go Way Back ," opening at the Francis Crick Institute.

Full text

Ukrainian President Volodymyr Zelenskyy said Thursday that the head of the country’s security services, Yevhenii Khmara, will be appointed as acting defense minister.

Khmara was appointed the acting head of the Security Service of Ukraine at the beginning of this year and, under Ukrainian law, cannot simultaneously serve as defense minister. He is expected to retire from military service in the coming days, allowing Zelenskyy to request that parliament appoint him as defense minister formally.

“I have tasked Yevhenii Khmara with serving as acting minister, continuing to reform the defense sector, and ensuring that Ukraine achieves all of the results we have discussed,” Zelenskyy wrote on Facebook .

He added: “Khmara has gained extensive and, in many respects, unprecedented experience with technological combat operations. This is exactly where our defense efforts should be focused during this war.”

Zelenskyy dismissed former Defense Minister Mykhailo Fedorov on Wednesday as part of a cabinet reshuffle. The 35-year-old had held the post for only six months. While reforming the military’s technology, Fedorov also clashed with top generals, who said they found him difficult to work with.