codesanitize

Theom, the corporate redefining information governance and safety for the AI period, introduced immediately it has raised $20 million in Sequence A funding to speed up its mission of securing delicate enterprise information throughout cloud, SaaS, and generative AI platforms.

The funding spherical was led by Wing VC, with participation from strategic buyers Databricks Ventures, Snowflake Ventures, and SentinelOne’s S Ventures. Theom will use the brand new capital to develop product improvement, develop its go-to-market staff, and scale deployments with Fortune 500 enterprises and fast-growing AI-native firms.

“We’re thrilled to have the help of such a powerful group of buyers who share our imaginative and prescient that information — not infrastructure — is the brand new perimeter,” stated Navindra Yadav, co-founder and CEO of Theom. “AI is remodeling how organizations use information, however it additionally introduces new dangers. Theom ensures firms can transfer quick with AI whereas staying safe and compliant.”

Constructed for a New Period of Knowledge Motion and AI

Theom is pioneering the primary AI-native Knowledge Operations Middle (DOC) — a contemporary different to outdated information governance and safety instruments. In immediately’s atmosphere, information consistently strikes between apps, clouds, LLMs, and exterior collaborators. Legacy safety instruments, constructed for static environments and community perimeters, can’t sustain.

Theom’s platform flips the mannequin by embedding governance and safety straight inside enterprise information platforms, equivalent to Snowflake, Databricks, AWS, and Azure. This enables Theom to repeatedly observe who’s accessing which information, in what context, and for what function, utilizing machine studying and behavioral analytics to flag threats earlier than they escalate.

Key use circumstances embody:

• AI-aware information entry governance, making certain delicate information is just not misused by AI fashions or unauthorized customers.

• Actual-time insider risk detection, together with consumer impersonation and over-permissioned entry.

• Automated compliance and reporting for laws like HIPAA, GDPR, NIST, and DORA.

• Safe information collaboration in clear rooms and information trade marketplaces.

By working in-place — with out transferring or proxying information — Theom avoids the egress prices and latency points that plague legacy DSPM options. It additionally integrates straight with instruments like Okta, Splunk, Microsoft 365, and Slack, enabling full-context safety throughout the enterprise information ecosystem.

Backing from Trade Leaders

Theom’s differentiated method has attracted not solely prospects equivalent to FiServ, Grammarly, JetBlue, and TradeWeb, but additionally a few of the greatest names in information and safety.

“Theom is the primary platform I’ve seen that delivers fine-grained, usage-based entry controls over enterprise information — precisely what firms want as they put together for AI,” stated Gaurav Garg, Co-Founder at Wing VC.

“Theom extends the governance capabilities of our Unity Catalog, giving our prospects real-time management over delicate information in Gen AI workflows,” added Andrew Ferguson, VP at Databricks Ventures.

“We’re seeing a dramatic enhance in information publicity dangers, particularly in finance and healthcare,” stated Harsha Kapre, Director at Snowflake Ventures. “Theom’s success with regulated industries reveals it may scale with real-world calls for.”

“Theom isn’t simply bettering information safety — it’s reimagining it,” stated Rob Salvagno, SVP at SentinelOne and Head of S Ventures. “They’re creating a brand new class on the intersection of AI, information, and cybersecurity.”

Why Now

With practically half of all enterprise information anticipated to reside within the cloud by the top of 2024, in accordance with Morgan Stanley, and generative AI instruments reshaping how information is accessed and shared, organizations urgently want a brand new class of knowledge governance platform.

Theom was based by a staff with deep roots in cloud safety and information engineering. Having beforehand created Tetration Analytics at Cisco — one of many first platforms for zero-trust analytics — the founders understood that securing information in fashionable enterprises would require a data-centric, AI-powered method.

The platform leverages superior methods together with generative adversarial networks (GANs), behavioral machine studying, and MITRE ATT&CK-based risk mapping to provide safety and compliance groups real-time perception into information utilization throughout inner and exterior customers.

Importantly, Theom can detect and cease reside information breaches that conventional instruments miss — from credential abuse to unauthorized desk cloning — with out requiring information to depart the shopper’s cloud.

What’s Subsequent

With this contemporary infusion of capital, Theom plans to develop its engineering staff, deepen integrations with ecosystem companions, and convey its real-time, AI-native information governance answer to extra enterprises globally.

“We’re constructing the blueprint for a way enterprises will govern and safe their information within the age of AI,” stated Yadav. “This funding accelerates our potential to assist prospects innovate confidently with AI — with out compromising on safety or compliance.”

A Waymo robotaxi working in LA. | Credit score: Waymo

Waymo is sending a small fleet of robotaxis to map town of Boston. The automobiles gained’t be driving autonomously. As a substitute, the autos could have educated, human autonomous specialists behind the wheel always.

The aim of the journey is to discover Boston’s distinctive – and chaotic – driving atmosphere and perceive the variations in comparison with areas the place Waymo routinely operates. The Mountain View, Calif.-based firm stated its robotaxis will probably be exploring Boston in Could and June. The corporate will start handbook driving by probably the most complicated elements of town, together with town middle and freeways. The robotaxis may also be in Cambridge.

Having lived close to Boston my complete life, I can let you know driving right here is difficult. The streets are a large number of slender, winding roads that don’t observe any sample. It doesn’t provide a grid system like Phoenix and San Francisco, two places the place Waymo affords a robotaxi service for folks. Intersections can come at bizarre angles. You’re continuously coping with heavy foot site visitors, cyclists squeezing by tight spots, and drivers who don’t precisely observe the principles.

And we haven’t even talked about the climate but. Boston’s climate, from heavy snow and icy roads within the winter to dense fog and rain all year long, could make driving situations particularly difficult for autonomous autos.

Boston is a part of Waymo’s 10-city highway journey throughout the USA in 2025. Different stops on the tour embrace Dallas, Las Vegas, and Nashville. Waymo explored New York in late 2021 and has but to launch a robotaxi service there. Now we have requested Waymo to make clear its future plans for Boston, however I wouldn’t get your hopes up for it launching a service in Boston any time quickly. We are going to replace this text after we hear again from Waymo on that.

By including Boston to the checklist of cities explored, Waymo continues its dedication to creating know-how that adapts to a variety of city settings. The insights gained from this highway journey will assist form the way forward for autonomous driving in complicated and diverse environments. The corporate has beforehand defined how these highway journeys advance its autonomous driving know-how, saying “expertise is one of the best trainer.”

“To convey the advantages of our know-how to extra folks in additional locations and ship on our mission to enhance highway security, we take the Waymo Driver on highway journeys to areas the place the driving tradition and situations differ from the areas we frequently function,” the corporate wrote in 2023.
Waymo is now offering greater than 250,000 paid journeys every week throughout its robotaxi fleets in Phoenix, San Francisco, Los Angeles, and Austin. Waymo can also be planning to launch robotaxi providers in Miami, Atlanta, and Washington D.C.
Waymo can also be now testing its robotaxis on public roads in Tokyo. This marks the corporate’s first deployment outdoors the U.S. In the course of the testing, educated drivers will manually drive the robotaxis throughout seven Tokyo wards, together with Minato, Shinjuku, Shibuya, Chiyoda, Chūō, Shinagawa, and Kōtō. Waymo stated this handbook operation will present its engineers with expertise to check, refine, and adapt its autonomous driving know-how to native roads.
Nihon Kotsu, Tokyo’s largest taxi firm, will oversee the administration and servicing of the autos. Waymo educated the groups from taxicab app supplier GO and Nihon Kotsu on fleet-management greatest practices and established first-responder coaching protocols.
Waymo was named the 2025 Robotic of the 12 months by The Robotic Report. Waymo and all of the RBR50 Robotics Innovation Awards winners had been honored on the Robotics Summit & Expo in Boston two weeks in the past.
Waymo additionally simply partnered with Toyota. The businesses are collaborating on a brand new autonomous automobile platform to convey autonomous driving capabilities to the following technology of personally-owned autos.

North American robotic orders peaked in 2022, when the corporate skilled a number of document quarters in a row.

Robotic orders in North America remained basically flat within the first quarter of 2025, in response to the Affiliation for Advancing Automation (A3). The group discovered that firms bought 9,064 items valued at $580.7 million in Q1 of 2025.

In comparison with Q1 2024, this represents a 0.4% improve in items ordered and a 15% rise so as worth, signaling continued demand and elevated funding in higher-value automation methods.

2024 was a slower 12 months for the North American robotics {industry}, with a powerful ending. General, North American firms ordered 31,311 robots valued at $1.963 billion, representing slight will increase of 0.5% in items and 0.1% in income over 2023.

Whereas the automotive {industry} has traditionally been the most important adopter of robotics, meals and client items emerged as 2024’s fastest-growing sector, with robotic orders surging 65%. This pattern, nevertheless, appears to be altering, with automotive gross sales driving development initially of 2025.

“The primary quarter knowledge highlights a continued resilience in automation funding, significantly within the automotive sector, at the same time as producers navigate a posh macroeconomic atmosphere,” mentioned Alex Shikany, government vp at A3. “On the similar time, some sectors are taking a extra cautious method as broader financial uncertainty persists.”

Demand from automotive authentic tools producers (OEMs) was the first development driver, with 3,668 items ordered valued at $263 million—a 42% improve in items and a 78% improve in income over Q1 2024.

The income development displays not solely greater quantity, but additionally a shift within the varieties of functions being automated this quarter, with a higher share of orders tied to higher-value methods, A3 mentioned. In distinction, orders from automotive element suppliers declined 29% in items and 12% in worth year-over-year, totaling 1,407 items and $88 million.

Most non-automotive sectors skilled contraction relative to Q1 2024, aside from plastics and rubber, which noticed 12% development in items ordered and a 33% improve so as worth. Sectors similar to meals and client items, metals, and semiconductors skilled year-over-year declines.

Now accepting session submissions!

A3 shares its first collaborative robotic numbers

A3 has begun reporting collaborative robotic (cobot) statistics as a part of its official quarterly knowledge, making it the primary industry-wide dataset of its type in North America.

“We’re proud to launch this new degree of reporting,” mentioned Shikany. “Cobots are one of many fastest-growing areas of robotics adoption, and offering clear, dependable knowledge on the place they’re getting used will assist producers, integrators, and suppliers make extra knowledgeable selections.”

In Q1 2025, North American firms ordered 1,052 collaborative robots valued at $39.2 million. Cobots accounted for 11.6% of all robots ordered and 6.8% of whole income this quarter. A3 discovered that cobot demand was strongest in industries prioritizing flexibility and secure human-machine collaboration:

• Life Sciences/Pharma/Biomed: 127 items ($7.1M)
• Meals & Shopper Items: 114 items ($4.5M)
• All Different Industries: 419 items ($13.5M)

In every of those segments, collaborative robots made up over 20% of whole demand. Common Robots, a number one cobot supplier, launched its UR15 at present. The UR15 has a most TCP velocity of 5 m/s to scale back cycle instances, improve productiveness, and cut back prices throughout functions and industries.

Our our bodies are continually breaking down. Over time, their built-in restore mechanisms additionally fail. Knee cartilage grinds away. Hip joints not help weight. Remedies for breast most cancers and different well being points require elimination by surgical procedure. As a result of the physique can’t regenerate these tissues, reconstruction utilizing biomaterials is usually the one method.

Historically, this sort of restoration includes one-size-fits-all breast implants or hip joints. Extra not too long ago, 3D bioprinted tissues have begun to be tailor-made for sufferers. However these synthetic tissues are printed outdoors of the physique, and so they nonetheless require further surgical procedures to implant, including to the possibilities of scarring, irritation, or an infection, and growing therapeutic time.

This month, a staff from the California Institute of Know-how unveiled a system to 3D print tissues contained in the physique—no surgical procedure wanted. Dubbed deep tissue in vivo sound printing (DISP), the system makes use of an injectable bioink that’s liquid at physique temperature however solidifies into constructions when blasted with ultrasound. A monitoring molecule, additionally delicate to ultrasound, tracks tissue printing in actual time. Extra bioink is safely damaged down by the physique.

In exams, the staff 3D printed tissues inside a rabbit’s abdomen and mouse’s bladder. In addition they added conductive nanoparticles to make delicate biosensors and depots of medicine—anticancer medicine or antibacterial drugs—that launched their payloads when hit with ultrasound.

“This work has actually expanded the scope of ultrasound-based printing and proven its translational capability,” Yu Shrike Zhang at Harvard Medical College, who was not concerned within the analysis, advised IEEE Spectrum. “It’s fairly thrilling.”

From Mild to Sound

Due to its versatility, 3D printing has captured the creativeness of bioengineers. The know-how can be utilized to make synthetic organic tissues, organs, or medical gadgets.

Bioprinters often deposit one layer at a time. Every layer is solidified utilizing gentle, then the subsequent layer is laid on high of it. This layer-by-layer course of takes time. Extra not too long ago, an upgraded methodology known as volumetric printing solidifies 3D constructions with a single blast of rigorously tailor-made gentle. The method is quicker but additionally constrained by how deeply gentle can pierce tissues.

Infrared gentle, for instance, can form implants beneath a skinny layer of pores and skin and muscle, Xiao Kuang on the College of Wisconsin–Madison, who was not concerned within the research, wrote in Science. However gentle dims and scatters the deeper it goes contained in the physique. This limits the “direct printing of implants beneath millimeter-thick tissues”—or barely under the pores and skin.

Ultrasound, greatest recognized for its use monitoring pregnancies, has a bonus right here. It may attain deep into organs—almost eight inches—with out damaging them. Scientists have been exploring targeted ultrasound, which blasts a sure frequency of soundwaves in direction of a tissue, as a technique to monitor mind and muscle exercise.

Ultrasound may also set off chemical reactions. In 2023, Zhang and colleagues engineered a molecular concoction dubbed “sono-ink” that solidifies when blasted with a selected frequency of sound waves. The staff 3D printed a number of shapes inside remoted pork stomach, liver, and kidneys and patched tissue injury in a goat coronary heart.

However the ink was delicate to emphasize and different disruptions within the physique, leading to slower printing and poor decision. Sound waves additionally generate warmth, which hardened a few of the sono-ink earlier than it had fashioned the supposed constructions. Additional, different molecules within the ink and native warmth spikes throughout tissues elevated biocompatibility threat.

“Ultrasound 3D printing inside a physique is more difficult than it appears,” wrote Kuang.

A New Ink

The brand new system depends on upgraded sono-ink.

The ink has a number of elements mixed right into a single concoction. First up are chains of molecules that usually float about freely however grasp one another when given a molecular cue. These are accompanied by fatty bubbles crammed with binder molecules—the molecular cue—that launch their payloads when uncovered to ultrasound. A remaining encapsulated element consists of a number of chemical compounds that scatter sound waves and light-weight up when sure soundwaves hit. These assist the staff visualize the ink’s location and decide if it’s fashioned the specified construction.

The brand new setup “prevented untimely chemical reactions at physique temperature and supplied higher management of the printing course of,” wrote Kuang.

The ink is injected to the goal web site or shuttled there utilizing catheters. To check the method, the staff 3D printed shapes—stars, pinwheels, teardrops, and lattices—in quite a lot of tissues, comparable to thick pork chops and rooster.

In comparison with earlier light-based strategies that solely reached fatty tissues, the brand new know-how made it into muscle tissues and extra exactly activated the bioink to type shapes. The system works at roughly 40 millimeters a second, the common pace of an inkjet printer.

A Sound Remedy

In mice with bladder most cancers, the staff 3D printed a form of depot that slowly launched anticancer medicine to stave off most cancers progress. In the present day’s bladder most cancers medicine are sometimes flushed away inside hours. A bioink patch would focus the medicine on the tumor for longer.

In one other take a look at, the researchers printed synthetic tissue inches under the pores and skin in rabbits’ leg muscle tissues and bellies, showcasing the know-how’s means to restructure tissues in bigger animals.

The bioink may also be custom-made. It may embrace carbon nanotubes, nanowires, and different biocompatible constructions for bioengineering. In a single take a look at, the staff loaded the ink with conductive nanoparticles and printed digital sensors to measure the exercise of residing tissues.

“This performance may very well be helpful for exams that monitor physiological alerts,” wrote Kuang, comparable to people who monitor coronary heart and muscle well being.

The ink is shelf-stable for a minimum of 450 days and doesn’t appear to set off immune responses. The physique eliminates extra ink via regular metabolism or it may be dissolved with a therapy usually used to counter heavy-metal poisoning.

There are nonetheless some kinks to iron out. Goal tissues differ in depth, form, and measurement, which may alter how the ultrasound bounces and cures the bioink. Printing on organs that transfer—our lungs, coronary heart, and abdomen—is much more complicated.

On this respect, AI might assist, wrote Kuang. Algorithms may decipher the hyperlinks between sound waves, temperature, supplies, and bodily interactions to higher simulate and information printing. An AI-based controller that mixes real-time monitoring of the method may quickly alter to modifications within the physique’s standing so the bioink solidifies as deliberate.

Though there’s a methods to go earlier than medical use, the staff says their latest work exhibits the know-how is flexible sufficient to be extensively helpful.

Like many, I really like good recommendation. However typically, I need assistance to get one thing executed.

The following rev of AI — agentic AI — will transfer us from recommendation to getting stuff executed. It’ll allow companies that harness it to take a transformative leap ahead.

However leap to what? And remodel how?

Agentic AI can scale back the price of buyer help by 25-50% whereas dramatically enhancing high quality and buyer satisfaction as a result of it goes past easy activity execution. It will probably additionally autonomously resolve complicated workflows and buyer interactions. When utilized to buyer help, for instance, brokers don’t simply reply to queries however comprehensively resolve inquiries from begin to end, decreasing human intervention and rising effectivity.

As with all new applied sciences, adopting agentic AI presents challenges. An organization should have its workflows well-documented and deeply understood and possess a strong data base on which the agentic AI can draw. And simply as with generative AI, knowledge privateness and safety issues require corporations to grasp the big language fashions (LLMs) they faucet into and the way data is saved and handed by them.

Nevertheless, the best adoption technique for clever automation can guarantee success. To reap essentially the most advantages, corporations might want to do three issues:

• Begin in the best place
• Stability agentic AI with human experience
• Faucet right into a community of agentic experience

Whereas it’s nonetheless early days, right here’s what we’re studying as we work with purchasers in numerous industries to combine agentic AI into their workflows and operations.

Don’t begin small — begin good

Maybe counterintuitively, the very best place to begin is together with your highest-volume use circumstances. Isn’t that dangerous? Not if executed correctly. The truth is, though beginning with low-volume use circumstances would possibly seem to cut back danger, it truly will increase the danger of not seeing ample impression to justify the funding.

Beginning with high-volume use circumstances affords the best potential return on funding (ROI), enabling an organization to shortly notice important impression, maximize effectivity positive aspects, and show the clear worth of utilizing AI brokers.

How do you mitigate the danger of beginning too huge? By initially implementing the brokers with simply 1% of the largest use case volumes. This method means that you can establish and repair potential points whereas getting ready for broader automation.

For a retail firm, this would possibly imply automating “the place’s my order?” or return-processing workflows. Along with monitoring shipments throughout the corporate’s achievement community, an AI agent might confirm a buyer’s identification, test real-time standing and replace the client — even supply choices if the order has been unexpectedly delayed.

For returns, an agent might test the corporate’s return insurance policies, collect buyer details about the return, counsel subsequent steps, and full applicable related duties, like printing a return label, scheduling a pickup, issuing a refund, and so forth. The return agent might additionally look ahead to patterns of abuse and, if warranted, alter its choices and subsequent steps accordingly.

After an organization deploys an AI agent on a pattern portion of a high-volume workflow, it should monitor workflow exercise to establish the place it’d want changes. When the agent features easily, the corporate can develop its use in pre-defined quantities till it will definitely handles the whole workflow quantity.

In fact, not all duties and workflows lend themselves to whole automation with agentic AI. in truth, preserving human specialists linked to the general workings of AI brokers will yield the very best outcomes.

Stability AI with human experience

As an organization examines its workflows and processes for automation candidates, it’ll discover situations finest suited to human oversight or direct motion. Agentic AI is an unimaginable, extremely succesful innovation, but it surely has limitations.

Three particularly:

AI brokers, just like the LLMs that help them, don’t at present possess normal intelligence. They operate finest in slender, well-defined areas. So, whereas people would possibly discover ways to carry out a selected activity and summary from that data ideas they then apply to totally different, unrelated duties, AI at present can’t.

Then, there are workflows with extraordinarily complicated choice matrices that demand important expertise and experience-based judgment. For instance, a retail firm would possibly want content material for a simple advertising marketing campaign. An agent can deal with that — and execute the marketing campaign.

However wish to revisit a model’s expression and promise throughout a number of markets? An agent wouldn’t be as much as the duty. It will require perception into market developments, model notion, cultural variations throughout markets, and perception into how manufacturers evoke feelings.

Lastly, workflows depending on sometimes “messy” human communication and emotional nuance that require distinctly human parts reminiscent of compassion finest stay with people. Consider customer support points involving irate prospects or healthcare interactions the place a affected person’s emotional or psychological state could also be in danger.

However I’m not describing a binary choice course of: give this to the AI brokers; the whole lot else goes to people. In apply, a hybrid mannequin works finest.

Whereas there must be a transparent delineation between AI and human roles, even when duties must be dealt with by human specialists, AI ought to nonetheless be available to increase their skills and take advantage of their experience.

Usually talking, corporations ought to use agentic AI for transactional, repeatable duties and faucet human experience for high-stakes interactions, emotionally complicated situations, and conditions requiring nuanced judgment. A $50 guarantee declare is perhaps absolutely automated, whereas a $5,000 declare would almost certainly profit from human emotional intelligence and brand-sensitive dealing with.

Faucet into an agentic community

Maybe most necessary, don’t attempt to dive into agentic AI solo. Set up a community of skilled companions. Rising agentic AI platforms can provide the expertise throughout digital and voice channels. A methods integrator and advisor that understands buyer working environments can prepare agentic fashions for particular buyer wants after which combine them into an organization’s operations.

Integrating these fashions into enterprise methods requires deep experience in complicated workflows and industry-specific challenges. It additionally requires an intricate understanding of workflow choice factors and the place human interplay is most wanted – or helpful, in order that agentic AI is a boon to staff and staff productiveness.

Agentic AI affords companies a strong means to enhance effectivity, improve buyer experiences, and drive innovation. However success isn’t about speeding in. It’s about making good, knowledgeable decisions: Beginning in the best place, making use of a hybrid human/AI mannequin, and tapping into the best community.

As a result of with the world of AI altering so shortly, you may’t afford to go it alone.